United States
Environmental Protection
Agency
Health Assessment
Document For
Engine Exhaust

-------
                                    EPA/600/8-90/057F
                                           May 2002
Health Assessment Document
  for Diesel Engine Exhaust
  National Center for Environmental Assessment
     Office of Research and Development
    U.S. Environmental Protection Agency
            Washington, DC

-------
                                      DISCLAIMER

       This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
                                       ABSTRACT

       This assessment examined information regarding the possible health hazards associated
with exposure to diesel engine exhaust (DE), which is a mixture of gases and particles.  The
assessment concludes that long-term (i.e., chronic) inhalation exposure is likely to pose a lung
cancer hazard to humans, as well as damage the lung in other ways depending on exposure.
Short-term (i.e., acute) exposures can cause irritation and inflammatory symptoms of a transient
nature, these being highly variable across the population.  The assessment also indicates that
evidence for exacerbation of existing allergies and asthma symptoms is emerging. The
assessment recognizes that DE emissions, as a mixture of many constituents, also contribute to
ambient concentrations of several criteria air pollutants including nitrogen oxides and fine
particles, as well as other air toxics. The assessment's health hazard conclusions are based on
exposure to exhaust from diesel engines built prior to the mid-1990s. The health hazard
conclusions, in general, are applicable to engines currently in use, which include many older
engines.  As new diesel engines with cleaner exhaust emissions replace existing engines, the
applicability of the conclusions in this Health Assessment Document will need to be reevaluated.
Preferred citation:
U.S. Environmental Protection Agency (EPA). (2002) Health assessment document for diesel engine exhaust.
Prepared by the National Center for Environmental Assessment, Washington, DC, for the Office of Transportation
and Air Quality; EPA/600/8-90/057F. Available from: National Technical Information Service, Springfield, VA;
PB2002-107661, and .
                                             11

-------
                                  CONTENTS

LIST OF TABLES	viii
LIST OF FIGURES	xi
FOREWORD	xiv
PREFACE 	xvi
AUTHORS, CONTRIBUTORS, AND REVIEWERS  	  xvii
ACKNOWLEDGMENTS	  xxii

1.  EXECUTIVE SUMMARY 	1-1
   1.1. INTRODUCTION	1-1
   1.2. COMPOSITION OF DIESEL EXHAUST	1-1
   1.3. DIESEL EXHAUST AS A COMPONENT OF AMBIENT PARTICULATE
      MATTER	1-2
   1.4. ATMOSPHERIC TRANSFORMATION OF DIESEL EXHAUST	1-2
   1.5. EXPOSURE TO DIESEL EXHAUST 	1-3
   1.6. HEALTH EFFECTS OF DIESEL EXHAUST	1-3
      1.6.1.  Acute (Short-Term Exposure) Effects 	1-4
      1.6.2.  Chronic (Long-Term Exposure) Noncancer Respiratory Effects	1-4
      1.6.3.  Chronic (Long-Term Exposure) Carcinogenic Effects	1-4
   1.7. SOURCES OF UNCERTAINTY	1-6

2.  DIESEL EXHAUST EMISSIONS CHARACTERIZATION, ATMOSPHERIC
   TRANSFORMATION, AND EXPOSURES  	2-1
   2.1. INTRODUCTION	2-1
   2.2. PRIMARY DIESEL EXHAUST EMISSIONS  	2-3
      2.2.1.  History of Dieselization  	2-3
      2.2.2.  Diesel Combustion and Formation of Primary Emissions	2-9
      2.2.3.  Diesel Emission Standards and Emission Trends Inventory  	2-15
      2.2.4.  Historical Trends in Diesel Fuel Use and Impact of Fuel
            Properties on Emissions  	2-25
      2.2.5.  Chronological Assessment of Emission Factors	2-29
      2.2.6.  Engine Technology Description and Chronology 	2-43
      2.2.7.  Air Toxic Emissions	2-53
      2.2.8.  Physical and Chemical Composition of Diesel Exhaust Particles  	2-59
   2.3. ATMOSPHERIC TRANSFORMATION OF DIESEL EXHAUST	2-84
      2.3.1.  Gas-Phase Diesel Exhaust 	2-84
      2.3.2.  Particle-Phase Diesel Exhaust 	2-90
      2.3.3.  Diesel Exhaust Aging	2-93
   2.4. AMBIENT DIESEL EXHAUST CONCENTRATIONS AND EXPOSURES  	2-94
      2.4.1. Diesel Exhaust Gases in the Ambient Atmosphere	2-94
      2.4.2. Ambient Concentrations of DPM 	2-95
      2.4.3. Exposures to Diesel Exhaust	2-106
   2.5. SUMMARY AND DISCUSSION	2-118
      2.5.1. History of Diesel Engine Use, Standards, and Technology	2-119
      2.5.2. Physical and Chemical Composition of Diesel Exhaust	2-120
                                       in

-------
                            CONTENTS (continued)

      2.5.3. Atmospheric Transformation of Diesel Exhaust	2-123
      2.5.4. Ambient Concentrations and Exposure to Diesel Exhaust 	2-124
REFERENCES FOR CHAPTER 2	2-126

3. DOSIMETRY OF DIESEL PARTICULATE MATTER	3-1
  3.1. INTRODUCTION	3-1
  3.2. CHARACTERISTICS OF INHALED DIESEL PARTICULATE MATTER	3-2
  3.3. REGIONAL DEPOSITION OF INHALED DIESEL PARTICULATE MATTER ... 3-2
      3.3.1. Deposition Mechanisms	3-3
      3.3.2. Particle Clearance and Translocation Mechanisms 	3-9
      3.3.3. Translocations of Particles to Extra-Alveolar Macrophage Compartment
           Sites	3-22
  3.4. PARTICLE "OVERLOAD"  	3-26
      3.4.1. Introduction	3-26
      3.4.2. Relevance to Humans	3-28
      3.4.3. Potential Mechanisms for an AM  Sequestration Compartment
           for Particles During Particle Overload	3-30
  3.5. BIO AVAILABILITY OF ORGANIC  CONSTITUENTS PRESENT ON
      DIESEL EXHAUST PARTICLES  	3-31
      3.5.1. In Vivo Studies	3-32
      3.5.2. In Vitro Studies	3-34
      3.5.3. Modeling Studies  	3-36
      3.5.4. Summary and Bioavailability	3-37
  3.6. MODELING THE DEPOSITION AND CLEARANCE OF PARTICLES IN
      THE RESPIRATORY TRACT	3-38
      3.6.1. Introduction	3-38
      3.6.2. Dosimetry Models for DPM	3-38
  3.7. SUMMARY AND DISCUSSION	3-54
REFERENCES FOR CHAPTER 3 	3-56

4. MUTAGENICITY  	4-1
  4.1. GENE MUTATIONS  	4-2
  4.2. CHROMOSOME EFFECTS	4-5
  4.3. OTHER GENOTOXIC EFFECTS  	4-7
  4.4. SUMMARY AND DISCUSSION	4-8
REFERENCES FOR CHAPTER 4	4-9

5. NONCANCER HEALTH EFFECTS OF DIESEL EXHAUST	5-1
  5.1. HEALTH EFFECTS OF WHOLE DIESEL EXHAUST 	5-2
      5.1.1. Human Studies 	5-2
      5.1.2. Traffic Studies	5-23
      5.1.3. Laboratory Animal Studies	5-24
  5.2. MODE OF ACTION OF DIESEL EXHAUST-INDUCED NONCANCER
      EFFECTS	5-82
      5.2.1. Comparison of Health Effects of Filtered and Unfiltered Diesel Exhaust .... 5-82

                                     iv

-------
                            CONTENTS (continued)

      5.2.2. Mode of Action for the Noncarcinogenic Effects of DPM	5-89
  5.3. INTERACTIVE EFFECTS OF DIESEL EXHAUST	5-90
  5.4. COMPARATIVE RESPONSIVENESS AMONG SPECIES TO THE
      HISTOPATHOLOGIC EFFECTS OF DIESEL EXHAUST  	5-92
  5.5. DOSE-RATE AND PARTICIPATE CAUSATIVE ISSUES	5-93
  5.6. SUMMARY AND DISCUSSION	5-97
      5.6.1. Effects of Diesel Exhaust on Humans  	5-97
      5.6.2. Effects of Diesel Exhaust on Laboratory Animals	5-99
      5.6.3. Comparison of Filtered and Unfiltered Diesel Exhaust  	5-102
      5.6.4. Interactive Effects of Diesel Exhaust	5-103
      5.6.5. Conclusions	5-103
REFERENCES FOR CHAPTER 5 	5-104

6. ESTIMATING HUMAN NONCANCER HEALTH RISKS OF DIESEL EXHAUST  ... 6-1
  6.1. INTRODUCTION 	6-1
  6.2. THE INHALATION REFERENCE CONCENTRATION APPROACH  	6-3
  6.3. CHRONIC REFERENCE CONCENTRATION FOR DIESEL EXHAUST	6-5
      6.3.1. Principal Studies for Dose-Response Analysis: Chronic,
            Multiple-Dose Level Rat Studies	6-6
      6.3.2. Derivation of Human Continuous Equivalent Concentrations, HECs	6-9
      6.3.3. Dose-Response Analysis—Choice of an Effect Level  	6-11
      6.3.4. Uncertainty Factors (UF) for the RFC—A Composite Factor of 30	6-14
      6.3.5. Derivation of the RfC for Diesel Exhaust	6-16
  6.4. EPIDEMIOLOGICAL EVIDENCE AND NAAQS FOR FINE PM	6-17
      6.4.1. Epidemiological Evidence for Fine PM	6-18
      6.4.2. NAAQS for Fine PM	6-25
      6.4.3. DPM as a Component of Fine PM	6-30
  6.5. CHARACTERIZATION OF THE NONCANCER ASSESSMENT FOR DIESEL
      EXHAUST 	6-30
  6.6. SUMMARY 	6-32
REFERENCES FOR CHAPTER 6	6-33

7. CARCINOGENICITY OF DIESEL EXHAUST	7-1
  7.1. INTRODUCTION 	7-1
      7.1.1. Overview	7-1
      7.1.2. Ambient PM-Lung Cancer Relationships  	7-1
  7.2. EPIDEMIOLOGIC STUDIES OF THE CARCINOGENICITY OF EXPOSURE
      TO DIESEL EXHAUST	7-3
      7.2.1. Cohort Studies	7-6
      7.2.2. Case-Control Studies of Lung Cancer	7-32
      7.2.3. Summaries of Studies and Meta-Analyses of Lung Cancer  	7-61
      7.2.4. Summary and Discussion	7-66

-------
                              CONTENTS (continued)

   7.3. CARCINOGENICITY OF DIESEL EXHAUST IN LABORATORY
       ANIMALS  	7-83
       7.3.1. Inhalation Studies (Whole Diesel Exhaust)	7-84
       7.3.2. Inhalation Studies (Filtered Diesel Exhaust)	7-108
       7.3.3. Inhalation Studies (DE Plus Cocarcinogens)	7-109
       7.3.4. Lung Implantation or Intratracheal Instillation Studies	7-111
       7.3.5. Subcutaneous and Intraperitoneal Injection Studies 	7-117
       7.3.6. Dermal Studies	7-119
       7.3.7. Summary and Conclusions of Laboratory Animal Carcinogeni city Studies . 7-121
   7.4. MODE OF ACTION OF DIESEL EXHAUST-INDUCED
       CARCINOGENESIS 	7-128
       7.4.1. Potential Role of Organic Exhaust Components in Lung Cancer Induction . 7-129
       7.4.2. Role of Inflammatory Cytokines and Proteolytic Enzymes in the
            Induction of Lung Cancer in Rats by Diesel Exhaust	7-132
       7.4.3. Role of Reactive Oxygen Species in Lung Cancer Induction by Diesel
            Exhaust  	7-133
       7.4.4. Relationship of Physical Characteristics of Particles to Cancer Induction .. 7-136
       7.4.5. Integrative Hypothesis for Diesel-Induced Lung Cancer  	7-137
       7.4.6. Summary  	7-139
   7.5. WEIGHT-OF-EVIDENCE EVALUATION FOR POTENTIAL
       HUMAN CARCINOGENICITY	7-140
       7.5.1. Human Evidence 	7-141
       7.5.2. Animal Evidence	7-142
       7.5.3. Other Key Data 	7-143
       7.5.4. Mode of Action 	7-143
       7.5.5. Characterization of Overall Weight of Evidence: EPA's 1986 Guidelines
            for Carcinogen Risk Assessment	7-144
       7.5.6. Weight-of-Evidence Hazard Narrative:  EPA's  Proposed Guidelines
            for Carcinogen Risk Assessment (1996b, 1999)	7-144
   7.6. EVALUATIONS BY OTHER ORGANIZATIONS 	7-146
   7.7. CONCLUSION	7-147
REFERENCES FOR CHAPTER 7	7-148

8.  DOSE-RESPONSE ASSESSMENT:  CARCINOGENIC EFFECTS	8-1
   8.1. INTRODUCTION 	8-1
   8.2. MODE OF ACTION AND DOSE-RESPONSE APPROACH 	8-2
   8.3. USE OF  EPIDEMIOLOGIC STUDIES FOR QUANTITATIVE
       RISK ASSESSMENT	8-4
       8.3.1. Sources of Uncertainty	8-4
       8.3.2. Evaluation of Key Epidemiologic Studies for
            Potential Use  in Quantitative Risk Estimates  	8-5
       8.3.3. Conclusion	8-11
   8.4. PERSPECTIVES ON CANCER RISK  	8-11
   8.5. SUMMARY AND DISCUSSION	8-16
REFERENCES FOR CHAPTER 8	8-17

                                       vi

-------
                           CONTENTS (continued)
9. CHARACTERIZATION OF POTENTIAL HUMAN HEALTH EFFECTS
  OF DIESEL EXHAUST: HAZARD AND DOSE-RESPONSE ASSESSMENTS  	9-1
  9.1. INTRODUCTION  	9-1
  9.2. PHYSICAL AND CHEMICAL COMPOSITION OF DIESEL EXHAUST	9-2
      9.2.1. Diesel Exhaust Components of Possible Health Concern	9-2
      9.2.2. "Fresh" Versus "Aged" Diesel Exhaust	9-4
      9.2.3. Changes of Diesel Exhaust Emissions and Composition Over Time 	9-5
  9.3. AMBIENT CONCENTRATIONS AND EXPOSURE TO DIESEL EXHAUST .... 9-6
  9.4. HAZARD CHARACTERIZATION  	9-8
      9.4.1. Acute and Short-Term Exposures  	9-8
      9.4.2. Chronic Exposure	9-10
  9.5. DOSE-RESPONSE ASSESSMENT  	9-16
      9.5.1. Evaluation of Risk for Noncancer Health Effects	9-17
      9.5.2. Evaluation of Cancer Risks	9-20
  9.6. SUMMARY AND CONCLUSIONS	9-24
REFERENCES FOR CHAPTER 9	9-27

APPENDIX A:  CALCULATION OF HUMAN EQUIVALENT CONTINUOUS
             EXPOSURE CONCENTRATIONS (HECs) 	 A-l

APPENDIX B:  BENCHMARK CONCENTRATION ANALYSIS OF
             DIESEL DATA	B-l

APPENDIX C:  A SUMMARY REVIEW OF CANCER DOSE-RESPONSE
             ANALYSES ON DIESEL EXHAUST	C-l
                                    vn

-------
                                  LIST OF TABLES

2-1.    Vehicle classification and weights for on-road trucks  	2-5
2-2.    Total (gas and diesel) diesel trucks in the fleet in 1992	2-5
2-3.    Typical chemical composition of fine particulate matter	2-14
2-4.    U.S. emission standards: HD highway diesel engines	2-16
2-5.    U.S. emission standards: locomotives (g/bhp-hr) 	2-17
2-6.    U.S. emission standards for nonroad diesel equipment (g/bhp-hr)	2-18
2-7.    Comparison of in-use truck fleet with truck fleet tested on chassis dynamometer,
       percent of total vehicles 	2-31
2-8.    Diesel engine emissions data from engine dynamometer tests	2-34
2-9.    HD diesel emissions results from tunnel tests  	2-40
2-10.   Remote sensing results for HD vehicles	2-41
2-11.   Summary of CDD/CDF emissions from  diesel-fueled vehicles	2-57
2-12.   Baltimore Harbor Tunnel Study: estimated CDD/CDF emission factors for HD
       vehicles	2-60
2-13.   Organic and elemental carbon fractions of diesel and gasoline engine PM exhaust  . 2-70
2-14.   Emission rates of PAH (mg/mi) from LD and HD diesel vehicles	2-73
2-15.   Poly cyclic aromatic hydrocarbons identified in extracts of diesel particles from
       LD diesel engine exhaust  	2-74
2-16.   Emission rates of particle-bound PAH (|lg/mi) from diesel and gasoline engines .. 2-76
2-17.   Concentrations of nitro-PAHs identified inLD diesel particulate extracts 	2-77
2-18.   Average emission rates for polycyclic aromatic hydrocarbons for different fuel
       types  	2-81
2-19.   Classes of compounds in diesel exhaust	2-86
2-20.   Calculated atmospheric lifetimes for gas-phase reactions of selected compounds
       present in automotive emissions with important  reactive species  	2-87
2-21.   Major components  of gas-phase diesel engine emissions, their known atmospheric
       transformation products, and the biological impact of the reactants and products . .  . 2-88
2-22.   Major components  of particle-phase diesel engine emissions, their known
       atmospheric transformation products, and the biological impact of the reactants
       and products 	2-92
2-23.   Ambient DPM concentrations reported from chemical mass balance modeling .... 2-98
2-24.   Ambient diesel particulate  matter concentrations from elemental carbon
       measurements in urban locations  	2-103
2-25.   Ambient diesel particulate  matter concentrations from dispersion modeling  	2-105
2-26.   Nationwide ambient diesel particulate matter concentrations for 1996 from
       the National Air Toxics Assessment National-Scale Assessment
       dispersion modeling  	2-107
2-27.   Occupational exposure to DPM  	2-109
2-28.   Ranges of occupational exposure to DPM by job category with estimates of equivalent
       environmental exposures   	2-111
2-29.   Annual average nationwide DPM exposure estimates (|lg/m3) from on-road sources
       for rural and urban  demographic groups in 1990, 1996, and 2007 using
       HAPEM-MS3 	2-114
                                          Vlll

-------
                              LIST OF TABLES (continued)

2-30.   Draft annual average, 25th, and 75th percentile nationwide DPM exposure
       estimates (|ig/m3) from on-road and nonroad sources for rural and urban counties
       in 1996 using HAPEM4	  2-113
2-31.   Annual average DPM exposures for 1990 and 1996 in the general population and
       among the highest exposed demographic groups in nine urban areas and
       nationwide from on-road sources only using HAPEM-MS3  	  2-114
2-32.   Modeled and estimated concentrations of DPM in microenvironments for
       California for all sources	  2-117
2-33.   Estimated indoor air and total air exposures to DPM in California in 1990	  2-118

3-1.    Predicted doses of inhaled DPM per minute based on total lung volume (M), total
       airway surface area (M^, or surface area in alveolar region (M2)	  3-9
3-2.    Alveolar clearance in laboratory animals exposed to DPM in whole exhaust	  3-17
3-3.    Model comparison for deposition of DPM under equivalent  conditions	  3-50
3-4.    Comparative model estimates of DPM deposition in human  lungs from exposure
       to 5(ig/m3 continuously for one year	  3-53

5-1.    Human studies of exposure to diesel exhaust	  5-18
5-2.    Short-term effects of diesel exhaust on laboratory animals 	  5-26
5-3.    Effects of chronic exposures to diesel exhaust on survival and growth of
       laboratory animals  	  5-31
5-4.    Effects of chronic exposures to diesel exhaust on organ weights and
       organ-to-body-weight ratios	  5-33
5-5.    Effects of diesel exhaust on pulmonary function of laboratory animals  	  5-36
5-6.    Histopathological effects of diesel exhaust in the lungs of laboratory animals 	  5-41
5-7.    Effects of exposure to diesel exhaust on the pulmonary defense mechanisms of laboratory
       animals	  5-51
5-8.    Effects of inhalation of diesel exhaust on the immune system of laboratory animals  . .  5-60
5-9.    Effects of diesel particulate matter on the immune response of laboratory animals  . . .  5-64
5-10.   Effects of exposure to diesel exhaust on the liver of laboratory animals	  5-70
5-11.   Effects of exposure to diesel exhaust on the hematological and cardiovascular
       systems of laboratory animals 	  5-72
5-12.   Effects of chronic exposures to diesel exhaust on serum chemistry of laboratory
       animals	  5-74
5-13.   Effects of chronic exposures to diesel exhaust on microsomal enzymes of
       laboratory animals  	  5-76
5-14.   Effects of chronic exposures to diesel exhaust on behavior and neurophysiology  ....  5-79
5-15.   Effects of chronic exposures to diesel exhaust on reproduction and development
       in laboratory animals	  5-81
5-16.   Composition of exposure atmospheres in studies comparing unfiltered and
       filtered diesel exhaust  	  5-84

6-1.    Histopathological effects of diesel exhaust in the lungs of laboratory animals 	  6-7
6-2.    Human equivalent continuous concentrations: 70-year HECs calculated
       with the model of Yu et al. (1991) from  long-term studies of rats repeatedly
       exposed to DPM	  6-22

                                            ix

-------
                             LIST OF TABLES (continued)
6-3.    Effect estimates per 50 |lg/m3 increase in 24-h PM10 concentrations from U.S. and
       Canadian studies	6-20
6-4.    Effect estimates per variable increments in 24-h concentrations of fine particle
       indicators (PM25,  SO=4, H+) from U.S. and Canadian studies  	6-23
6-5.    Effect estimates per increments in annual average levels of fine particle
       indicators from U.S. and Canadian studies	6-24
6-6.    Decision summary for the quantitative noncancer RfC assessment for continuous
       exposure to diesel paniculate matter (DPM) 	6-33

7-1.    Epidemiologic studies of the health effects of exposure to DE: cohort mortality
       studies	7-27
7-2.    Epidemiologic studies of the health effects of exposure to DE: case-control
       studies of lung cancer	7-55
7-3.    Summary of animal inhalation carcinogenicity studies 	7-85
7-4.    Tumor incidences in rats following intratracheal instillation of DE particles (DPM),
       extracted DPM, carbon black (CB), benzo[a]pyrene (B[a]P), or particles
       plus B[or]P	7-96
7-5.    Tumorigenic effects of dermal application of acetone extracts of DPM  	7-96
7-6.    Tumor incidence and survival time of rats treated by surgical lung implantation
       with fractions from DE condensate (35 rats/group)	7-111
7-7.    Dermal tumorigenic and carcinogenic effects of various emission extracts  	7-120
7-8.    Cumulative (concentration x time) exposure data for rats exposed to whole DE  . . 7-122
7-9.    Evaluations of DE as to human carcinogenic potential 	7-144

8-1.    DPM exposure margins (ratio of occupational + environmental exposures)	8-12

-------
                                  LIST OF FIGURES

2-1.    Diesel truck sales (domestic) for the years 1939-1997	2-6
2-2.    Diesel truck sales as a percentage of total truck sales for the years 1939-1997	2-6
2-3.    Percentage of truck miles attributable to diesel trucks	2-7
2-4.    Model year distribution of in-use HD truck fleet in 1997  	2-8
2-5.    Model year distribution of vehicle miles traveled by the in-use HD truck
       fleet in 1997  	2-8
2-6.    A comparison of IDI (A) and DI (B) combustion systems of high-speed HD diesel
       truck engines	2-10
2-7.    Schematic diagram of diesel engine exhaust particles	2-12
2-8.    Typical chemical composition for diesel particulate matter (PM2 5) from new
       (post-1990) HD diesel vehicle exhaust	2-14
2-9.    Trends in PM10 emissions from on-road and nonroad engines combined and other
       anthropogenic sources of PM10 from 1970 to 1998 (excludes miscellaneous and
       natural sources)	2-19
2-10.   Trends in PM10 emissions from on-road and nonroad diesel engines from 1970
       to 1998 and projections of emissions to 2007 and 2030	2-21
2-11.   Trends in NOX emissions from on-road and nonroad diesel engines combined and
       other anthropogenic sources of NOX from 1970 to 1998 (excludes miscellaneous
       and natural sources)  	2-22
2-12.   Trends in NOX emissions from on-road and nonroad diesel engines from
       1970 to 1998	2-23
2-13.   Trends in SO2 emissions from on-road diesel engines from 1970 to  1998 and
       nonroad diesel engines from 1990 to 1998	2-24
2-14.   Trends in VOC emissions from on-road and nonroad diesel engines from 1970
       to 1998  	2-24
2-15.   Trends in CO emissions from on-road and nonroad diesel engines from
       1970 to 1998	2-25
2-16.   Percentage of total motor fuel use that is on-road diesel fuel since 1949 	2-26
2-17.   On-highway diesel fuel consumption since 1943, values in thousands of gallons . . .  2-26
2-18.   Model year trends in PM, NOX, HC, and CO emissions from HD diesel vehicles
       (g/mile)	2-30
2-19.   Diesel engine certification data for NOX emissions  as a function of model year ....  2-33
2-20.   Diesel engine certification data for PM emissions as a function of model year	2-37
2-21.   Emission factors from HD diesel vehicles from tunnel studies 	2-42
2-22.   Line-haul and switch emissions data  	2-44
2-23.   Effect of turbocharging and aftercooling on NOX ^id PM	2-49
2-24.   Comparison of diesel engine dynamometer PM emissions for four-stroke,
       naturally aspirated, and turbocharged engines	2-50
2-25.   An example of uniflow  scavenging of a two-stroke diesel engine with a positive
       displacement blower	2-52
2-26.   Comparison of two- and four-stroke vehicle diesel  PM emissions from chassis
       dynamometer studies  	2-53
2-27.   Comparison of two- and four-stroke engine diesel PM emissions from engine
       dynamometer studies  	2-54
2-28.   Diesel engine dynamometer SOF emissions from two- and four-stroke engines. . . .  2-54

                                          xi

-------
                            LIST OF FIGURES (continued)

2-29.   Diesel engine aldehyde emissions measured in chassis dynamometer studies  	2-56
2-30.   Diesel engine aldehyde emissions from engine dynamometer studies	2-56
2-31.   Trend in SOF emissions based on chassis dynamometer testing of HD diesel
       vehicles	2-66
2-32.   Trend in SOF emissions for transient engine dynamometer testing of HD diesel
       engines  	2-66
2-33.   Trend in SOF emissions as a percent of total PM based on chassis dynamometer
       testing of FID diesel vehicles  	2-68
2-34.   Trend in SOF emissions as a percentage of total PM from engine dynamometer
       testing  	2-68
2-35.   EC emission rates for diesel vehicles	2-72
2-36.   EC content as percent of fine PM for DPM samples obtained in chassis
       dynamometer studies  	2-72
2-37.   Diesel engine emissions of benzo[a]pyrene and  1-nitropyrene measured in chassis
       dynamometer studies  	2-78
2-38.   Diesel engine dynamometer measurements of benzo[a]pyrene and 1-nitropyrene
       emissions from FID diesel engines	2-79
2-39.   Particle size distribution in DE	2-82

3-1.    Schematic representation of major mechanisms, including diffusion, involved in
       particle deposition	3-4
3-2.    Generalized regional deposition fractions of various sized particles in the human
       respiratory tract	3-6
3-3.    Modeled deposition distribution patterns of inhaled DE particles in the airways of
       different species  	3-8
3-4.    Diagram of known and suspected clearance pathways for poorly soluble particles
       depositing in the alveolar region	3-10
3-5.    Modeled clearance of poorly soluble 4-|im particles deposited in tracheobronchial
       and alveolar regions in humans	3-12
3-6.    Short-term thoracic clearance of inhaled particles as determined by model prediction
       and experimental measurement	3-13
3-7.    Clearance from lungs of rats of 134Cs-FAP fused aluminosilicate tracer particles
       inhaled after 24 months of DE exposure at concentrations of 0 (control),
       0.35 (low), 3.5 (medium), and 7.1 (high) mg DPM/m3  	3-18
3-8.    Lung burdens (in mg DPM soot/g lung) in rats chronically exposed to DE at
       0.35 (low), 3.5 (medium), and 7.1 (high) mg ppm/m3  	3-27
3-9.    Modeled estimates of lung burden in humans after a simulated lifetime exposure
       to DPM using the Yu et al. (1991) and ICRP66 models	3-51

6-1.    Flow diagram of procedure for calculating HECs  	6-10
6-2.    Relative risk (RR) estimates for increased mortality and morbidity endpoints
       associated with 50 |lg/m3 increments in PM10 concentrations as derived from
       studies cited by numbers listed above each given type of health endpoint  	6-22
6-3.    Relative risks of acute mortality in Harvard Six Cities Study, for inhalable
       thoracic particles (PM15/PM10), fine particles (PM2 5), and coarse fraction

                                          xii

-------
                             LIST OF FIGURES (continued)

       particles (PM15-PM2 5)	6-26
6-4.    Adjusted relative risks for mortality are plotted against each of seven long-term
       average particle indices in the Harvard Six Cities Study, from largest range (total
       suspended particles) through sulfate and nonsulfate fine particle concentrations . .  . 6-27
6-5.    Percent of children with <85% normal FVC versus annual-average fine (PM2 x)
       particle concentrations and coarse fraction (PM10_2 x) levels for 22 North
       American cities	6-28

7-1.    Pooled relative risk estimates and heterogeneity-adjusted 95% confidence intervals
       for all studies and subgroups of studies included in the meta-analysis 	7-62
7-2.    Pooled estimates  of relative risk of lung cancer in epidemiologic studies
       involving occupational exposure to DE (random-effects models)	7-64
7-3.    Pathogenesis of lung disease in rats with chronic, high-level exposures to
       particles	7-137
                                           Xlll

-------
                                     FOREWORD
       The diesel engine has been a vital workhorse in the United States, powering many of its
large trucks, buses, and farm, railroad, marine, and construction equipment. Expectations are
that diesel engine use in these areas will increase due to the superior performance characteristics
of the engine. Diesel engine exhaust (DE), however, contains harmful pollutants in a complex
mixture of gases and particulates. Human exposure to this exhaust comes from both highway
uses (on-road) as well as nonroad uses of the diesel engine.
       EPA started evaluating and regulating the gaseous emissions from the heavy-duty
highway use of diesel engines in the 1970s and particle emissions in the 1980s.  The reduction of
harmful exhaust  emissions has taken a large step forward because of standards issued in 2000
which will bring about very large reductions in exhaust emissions for model year 2007 heavy-
duty engines used in trucks, buses, and other on-road uses. A draft of this assessment, along
with the peer review comments of the Clean Air Scientific Advisory Committee, was part of the
scientific basis for EPA's regulation of heavy-duty highway engines completed in December
2000. The information provided by this assessment was  useful in developing EPA's
understanding of the public health implications of exposure to DE and the public health benefits
of taking regulatory action to control exhaust emissions.  EPA anticipates developing similarly
stringent regulations for other diesel engine uses, including those used in nonroad applications.
       Until these regulations take effect, EPA is partnering with state and local agencies to
retrofit older, dirtier, engines to make them run cleaner and to develop model programs to reduce
emissions from idling engines. In addition, EPA and local authorities are working to ensure
early introduction of effective technologies for particulate matter control and the availability of
low- sulfur fuel where possible in advance of the 2007 requirements. Today, at least one engine
manufacturer is producing  new engines with particulate traps that, when coupled with low-sulfur
fuel, meet 2007 particulate emission levels. The Agency expects significant environmental and
public health benefits as the environmental performance  of diesel engines and diesel fuels
improves.
       The health assessment concludes that long-term (i.e., chronic) exposure to DE is likely to
pose a lung cancer hazard as well as damage the lung in other ways depending on exposure. The
health assessment's conclusions are based on exposure to exhaust from diesel engines built prior
to the mid-1990s. Short-term (i.e., acute) exposures can  cause transient irritation and
inflammatory symptoms, although the nature and extent of these symptoms are highly variable
across the population.  The assessment also states that evidence is emerging that diesel exhaust
                                          xiv

-------
exacerbates existing allergies and asthma symptoms. The assessment recognizes that DE
emissions, as a mixture of many constituents, also contribute to ambient concentrations of
several criteria air pollutants including nitrogen oxides, sulfur oxides, and fine particles, as well
as other hazardous air pollutants.
       The particulate fraction of DE and its composition is a key element in EPA's present
understanding of the health issues and formulation of the conclusions in the health assessment.
The amount of exhaust particulate from on-road engines has been decreasing in recent years and
is expected to decrease 90% from today's levels with the engines designed to meet the 2007
regulations. The composition of the exhaust particulates and the gases also will change. While
EPA believes that the assessment's conclusions apply to the general use of diesel engines today,
as cleaner diesel engines replace a substantial number of existing engines, the general
applicability of the conclusions in this health assessment document will need to be reevaluated.
                                          Paul Oilman, Ph.D.
                                          Assistant Administrator
                                          Office of Research and Development
                                          xv

-------
                                      PREFACE

       This document is the U.S. Environmental Protection Agency's science-based Health
Assessment Document for Diesel Engine Exhaust.  The assessment was prepared by the National
Center for Environmental Assessment which is the health risk assessment program in EPA's
Office of Research and Development. The assessment broadly supports activities authorized in
the 1990 Clean Air Act. This assessment was specifically prepared for EPA's Office of
Transportation and Air Quality which requested information regarding the potential health
hazards associated with diesel engine exhaust (DE) exposure.  As DE emissions also contribute
to urban air toxics and ambient particulate matter,  other EPA air programs also have an interest
in this assessment.
       This document was preceded by five earlier drafts: a Workshop Review Draft
(EPA/600/8-90/057A, July 1990), an External Review Draft (EPA/600/8-90/057B, December
1994), an SAB Review Draft (EPA/600/8-90/057C, February 1998), an SAB Review Draft
(EPA/600/8-90/057D, November 1999), and an SAB Review Draft (EPA/600/8-90/057E, July
2000). There was an SAB Environmental Health Committee Review in 1990 of the July 1990
draft. The Science Advisory Board's Clean Air Scientific Advisory Committee (CASAC)
reviewed the 1994 draft in public sessions in May  1995, the 1998 draft in May  1998, the 1999
draft in December 1999, and the July 2000 draft in October 2000. Public comment periods also
were conducted concurrently with the CASAC reviews. In addition many reviewers, both within
and outside the Agency, provided assistance at various review stages. This is the final version of
the assessment which was prepared in response to  CASAC advice and public comments received
on the 2000 draft.
       The scientific literature search for this assessment is generally current through January
2000, although a few later publications have been included.
                                         xvi

-------
                 AUTHORS, CONTRIBUTORS, AND REVIEWERS


      The National Center for Environmental Assessment (NCEA) within EPA's Office of
Research and Development (ORD) was responsible for the preparation of this document.
                       CHAPTER 1. EXECUTIVE SUMMARY
Authors
NCEA Diesel Team
    CHAPTER 2. DIESEL EMISSIONS CHARACTERIZATION, ATMOSPHERIC
                      TRANSFORMATION, AND EXPOSURES

Author
Marion Hoyer, Office of Transportation and Air Quality, U.S. Environmental Protection Agency,
Ann Arbor, MI.

Contributors
Chad Bailey, Office of Transportation and Air Quality, U.S. Environmental Protection Agency,
Ann Arbor, MI.

Tom Baines, Office of Transportation and Air Quality, U.S. Environmental Protection Agency,
Ann Arbor, MI.

David Cleverly, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Washington, DC.

William Ewald, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC.

Robert McCormick, Colorado School of Mines, Golden, CO.

Joseph McDonald, Office of Transportation and Air Quality, U.S. Environmental Protection
Agency, Ann Arbor, MI.

Joseph Somers, Office of Transportation and Air Quality, U.S. Environmental Protection
Agency, Ann Arbor, MI.

Janet Yanowitz, Colorado School of Mines, Golden, CO.

Barbara Zielinska, Desert Research Institute, Reno, NV.
                                        xvn

-------
            AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)

         CHAPTER 3.  DOSIMETRY OF DIESEL PARTICULATE MATTER

Authors
Gary Foureman, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC.

James McGrath, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC.

William Pepelko, National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Washington, DC.


                          CHAPTER 4. MUTAGENICITY

Author
Lawrence Valcovic, National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Washington, DC.


       CHAPTER 5. NONCANCER HEALTH EFFECTS OF DIESEL EXHAUST

Author
James McGrath, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC.

Contributor
Gary Foureman, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC.
          CHAPTER 6. QUANTITATIVE APPROACHES TO ESTIMATING
           HUMAN NONCANCER HEALTH RISKS OF DIESEL EXHAUST

Authors
Gary Foureman, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC.

Lester Grant, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC.

Contributors
Karen Martin, Office of Air Quality Planning and Standards-OAR, U.S. Environmental
Protection Agency, Research Triangle Park, NC.
                                       xvin

-------
            AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)

James McGrath, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC.
             CHAPTER 7. CARCINOGENICITY OF DIESEL EXHAUST

Authors
Aparna Koppikar, National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Washington, DC.

William Pepelko, National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Washington, DC.

Contributors
Drew Levy, University of Washington, Seattle, WA.

Robert Young, Oak Ridge National Laboratory, Oak Ridge, TN.
    CHAPTER 8.  DOSE-RESPONSE ASSESSMENT: CARCINOGENIC EFFECTS

Authors
Chao Chen, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Washington, DC.

William Pepelko, National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Washington, DC.

Contributor
Charles Ris, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Washington, DC.
 CHAPTER 9. CHARACTERIZATION OF POTENTIAL HUMAN HEALTH EFFECTS
     OF DIESEL EXHAUST:  HAZARD AND DOSE-RESPONSE ASSESSMENTS

Author
Charles Ris, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Washington, DC.

Contributors
NCEA Diesel Team
                                       xix

-------
            AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
REVIEWERS
       The Science Advisory Board's Clean Air Scientific Advisory Committee (CASAC)
reviewed the 1994 draft in public sessions in May 1995, the 1998 draft in May 1998, the 1999
draft in December 1999, and the July 2000 draft in October 2000. Public comment periods also
were conducted concurrently with the CASAC reviews. In addition, many reviewers both within
and outside the Agency provided assistance at various review stages. This is the final  version of
the assessment which was prepared in response to the latest CASAC advice and public
comments.
       The authors wish to thank all those who sought to improve the quality of this report with
their comments and are particularly grateful to the CASAC for its advice.
       The following members of the SAB's CASAC participated in the review of the July 2000
draft.

Panel Chair
Dr. Joe Mauderly1, Vice President, Senior Scientist, and Director of National Environmental
       Respiratory Center, Lovelace Respiratory Research Institute, Albuquerque, NM.

CASAC Members2
Mr. John Elston, Administrator, Office of Air Quality Management, State of New Jersey,
       Department of Environmental Protection and Energy, Trenton, NJ.

Dr. Philip K. Hopke3,  R.A. Plane Professor of Chemistry, Clarkson University, Department of
       Chemistry, Potsdam, NY (CASAC Chair).

Dr. Eva J. Pell4,  Steimer Professor of Agriculture Sciences, The Pennsylvania State University,
       University Park, PA.

Dr. Arthur C. Upton,  M.D., Director,  Independent Peer Review, UMDNJ-CRESP,
       Environmental and Occupational Health Sciences  Institute, New Brunswick, NJ.

Dr. Sverre Vedal, M.D., University of British Columbia, Vancouver Hospital, Vancouver, BC,
       Canada.

Dr. Warren White5, Senior Research Associate, Washington University, Chemistry
       Department, St. Louis, MO.

CASAC Consultants6
Dr. David Diaz-Sanchez, Department  of Medicine, UCLA, Los Angeles, CA.

Dr. Eric Garshick, M.D., Staff Physician, Pulmonary and Critical Care Section, WestRoxbury
       Virginia Medical Center, West Roxbury, MA.
                                         xx

-------
Dr. Roger O. McClellan, Advisor, Toxicology and Human Health Risk Analysis, and President
       Emeritus, Chemical Industry Institute of Toxicology (CUT), Albuquerque, NM.

Dr. Gunter Oberdorster, University of Rochester Medical Center, Department of
       Environmental Medicine, Rochester, NY.

Dr. Leslie Stayner7, National Institute for Occupational Safety and Health (NIOSH), Risk
       Evaluation Branch, Taft Laboratories, Cincinnati, OH.

Dr. Ron Wyzga, Electric Power Research Institute (EPRI), Palo Alto, CA.

Science Advisory Board Staff
Mr. Robert Flaak, Designated Federal Official (DFO) and Team Leader, Committee Operations
       Staff, EPA Science Advisory Board (1400A), 1200 Pennsylvania Avenue, NW, U.S.
       Environmental Protection Agency, Washington, DC 20460

Ms. Diana Pozun, Management Assistant, Committee Operations Staff, and Program Specialist,
       Office of the Staff Director, EPA Science Advisory Board (1400A), 1200 Pennsylvania
       Avenue, NW, U.S. Environmental Protection Agency, Washington, DC 20460
        Appointment as Chair of CASAC ended on October 30, 2000. Appointed ex officio Past Chair until
September 3 0,2001.
        CASAC Members are experts appointed by the Administrator to two-year terms to serve on the Clean Air
Scientific Advisory Committee.
        Appointed as Chair of CASAC on October 30, 2000.
       ^Resigned from CASAC on September 28, 2000.
        Appointment as Member of CASAC ended on October 30, 2000.
        CASAC Consultants are experts appointed by the Science Advisory Board Staff Director to a one-year
term to serve on ad hoc Panels formed to address a particular issue; in this case, the CASAC Review of EPA's
Health Assessment Document for Diesel Exhaust.
        Federal Expert.
                                           xxi

-------
                              ACKNOWLEDGMENTS


       The authors would like to acknowledge the contributions of several people who have
made this report and the previous drafts possible.
Document Review
Vanessa Vu                                     Sam Napolitano
National Center for Environmental Assessment      Office of Transportation and Air Quality
U.S. Environmental Protection Agency             U.S. Environmental Protection Agency
Washington, DC                                 Washington, DC

Document Production
Terri Konoza
Judy Theisen
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Kay Marshall
Joanna Taylor
Eric Sorensen
Clara Calderon-Laucho
The COM Group, Inc.
Chevy Chase, MD
                                        xxn

-------
                             1. EXECUTIVE SUMMARY

1.1. INTRODUCTION
       This Health Assessment Document for Diesel Engine Exhaust (DE) represents EPA's
first comprehensive review of the potential health effects from ambient exposure to exhaust from
diesel engines. The assessment was developed to provide information about the potential for DE
to pose environmental health hazards, information that would be useful in evaluating regulatory
needs under provisions of the Clean Air Act. The assessment identifies and characterizes the
potential human health hazards of DE (i.e, hazard assessment) and seeks to estimate the
relationship between exposure and disease response for the key health effects (i.e., dose-response
assessment). A full exposure assessment and risk characterization, the other two components of
a complete risk assessment, are beyond the scope of this document.
       The report has nine chapters and three appendices.  Chapter 2 provides a characterization
of diesel emissions, atmospheric transformation, and human exposures to provide a context for
the hazard evaluation of DE.  Chapters 3, 4, 5, and 7 provide a review of relevant information for
the evaluation of potential health hazards of DE, including dosimetry (Chapter 3), mutagenicity
(Chapter 4), noncancer effects (Chapter 5), and carcinogenic effects (Chapter 7). Chapters 6 and
8 contain dose-response analyses to provide insight about the significance of the key noncancer
and cancer hazards. Chapter 9 summarizes and characterizes the overall nature of the health
hazard potential in the environment and the overall confidence and/or uncertainties associated
with the conclusions.

1.2. COMPOSITION OF DIESEL EXHAUST
       DE is a complex mixture of hundreds of constituents in either a gas  or particle form.
Gaseous components  of DE include carbon dioxide, oxygen, nitrogen, water vapor, carbon
monoxide, nitrogen compounds, sulfur compounds, and numerous low-molecular-weight
hydrocarbons.  Among the gaseous hydrocarbon components of DE that are individually known
to be of toxicologic relevance are the aldehydes (e.g., formaldehyde, acetaldehyde, acrolein),
benzene, 1,3-butadiene, and polycyclic aromatic hydrocarbons (PAHs) and nitro-PAHs.
       The particles present in DE (i.e., diesel particulate matter [DPM]) are composed of a
center core of elemental carbon and adsorbed organic compounds, as well as small amounts of
sulfate, nitrate, metals, and other trace elements. DPM consists  of fine particles (fine particles
have a diameter <2.5  |lm), including a subgroup with a large number of ultrafme particles
(ultrafme particles have a diameter <0.1 |im).  Collectively, these particles have a large surface
area which makes them an excellent medium for adsorbing organics. Also, their small size
makes them highly respirable and able to reach the deep lung. A number of potentially

                                          1-1

-------
lexicologically relevant organic compounds are on the particles. The organics, in general, range
from about 20% to 40 % of the particle weight, though higher and lower percentages are also
reported. Many of the organic compounds present on the particle and in the gases are
individually known to have mutagenic and carcinogenic properties. For example, PAHs, nitro-
PAHs, and oxidized PAH derivatives are present on the diesel particles, with the PAHs and their
derivatives comprising about 1% or less of the DPM mass.
       DE emissions vary significantly in chemical composition and particle sizes between
different engine types (heavy-duty, light-duty), engine operating conditions (idle, accelerate,
decelerate), and fuel formulations (high/low sulfur fuel).  Also, there are emission differences
between on-road and nonroad engines simply because the nonroad engines to date are generally
of older technology. The mass of particles emitted and the organic components on the particles
from on-road diesel engines have been reduced over the  years.  Available data for on-road
engines indicate that lexicologically relevant organic components of DE (e.g., PAHs, nitro-
PAHs) emitted from older vehicle engines are still present in emissions from newer engines,
though relative amounts have decreased.  There is currently insufficient information to
characterize the changes in the composition of DE from  nonroad diesel engines over time.

1.3. DIESEL EXHAUST AS A COMPONENT OF AMBIENT PARTICIPATE MATTER
       DE is emitted from "on-road" diesel engines (vehicle engines) or "nonroad" diesel
engines (e.g., locomotives, marine vessels, heavy-duty equipment, etc.). Nationwide, data in
1998 indicated that DE as measured by DPM made up about 6% of the total ambient PM2 5
inventory (i.e., particles with aerodynamic diameter of 2.5 micrometers or less) and about 23%
of the inventory, if natural and miscellaneous sources of PM25 are excluded. Estimates of the
DPM percentage of the total inventory in urban centers are higher. For example, estimates range
from 10% to 36% in some urban areas in California, Colorado, and Arizona. Available data also
indicate that over the years there have been significant reductions in DPM emissions from the
exhaust of on-road diesel engines, whereas limited data suggest that exhaust emissions from
nonroad engines have increased.

1.4. ATMOSPHERIC TRANSFORMATION OF DIESEL EXHAUST
       After emission from the tailpipe, DE undergoes dilution and chemical and physical
transformations in the atmosphere, as well as dispersion  and transport in the atmosphere. The
atmospheric lifetime for some compounds present in DE ranges from hours to days.  DPM is
directly emitted from diesel-powered engines (primary particulate matter) and can be formed
from the gaseous compounds emitted by diesel engines (secondary particulate matter). Limited
information is available about the physical and chemical transformation of DE in the

                                          1-2

-------
atmosphere. It is not clear what the overall toxicological consequences of DE's transformations
are because some compounds in the DE mixture are altered to more toxic forms while others are
made less toxic.

1.5.  EXPOSURE TO DIESEL EXHAUST
       DPM mass (expressed as |ig DPM/m3) has historically been used as a surrogate measure
of exposure for whole DE. Although uncertainty exists as to whether DPM is the most
appropriate parameter to correlate with human health effects, it is considered a reasonable choice
until more definitive information about the mechanisms of toxicity or mode(s) of action of DE
becomes available. In the ambient environment, human exposure to DE comes from both on-
road and nonroad engine exhaust. A large percentage of the U.S. population also is exposed to
ambient PM25, of which DPM is typically a significant constituent.  Although this document
does not provide an exposure assessment, DE exposure information is included to provide a
context for the health effects information. Exposure estimates for the early to mid-1990s suggest
that national annual average DE exposure from on-road engines alone was in the range of about
0.5 to 0.8 |lg DPM/m3 of inhaled air in many rural and urban areas, respectively. Exposures
could be higher  if there  is a nonroad DE source that adds to the exposure from on-road vehicles.
For example, preliminary estimates show that, on a national average basis, accounting for
nonroad DE emissions adds another twofold to the on-road exposure.  For localized urban areas
where people spend a large portion of their time outdoors, the exposures  are higher and, for
example, may range up  to 4.0 |lg DPM/m3 of inhaled air.

1.6.  HEALTH  EFFECTS OF DIESEL EXHAUST
       Available evidence indicates that there are human health hazards  associated with
exposure to DE. The hazards include acute exposure-related symptoms, chronic exposure-
related noncancer respiratory effects, and lung cancer.  The health hazard conclusions are based
on exhaust emissions from diesel engines built prior to the mid-1990s. With current engine use
including some new and many more older engines (engines typically stay in service for a long
time), the health hazard conclusions, in general, are applicable to engines currently in use.  As
new and cleaner diesel engines, together with different diesel fuels, replace a substantial number
of existing engines, the  general  applicability of the health hazard conclusions will need to be re-
evaluated. With new engine and fuel technology expected to produce significantly cleaner
engine  exhaust by 2007 (e.g., in response to new federal heavy duty engine regulations),
significant reductions in public health hazards are expected for those engine uses affected by the
regulations.
                                          1-3

-------
1.6.1. Acute (Short-Term Exposure) Effects
        Information is limited for characterizing the potential health effects associated with
acute or short-term exposure. However, on the basis of available human and animal evidence, it
is concluded that acute or short-term (e.g., episodic) exposure to DE can cause acute irritation
(e.g., eye, throat, bronchial), neurophysiological symptoms (e.g., lightheadedness, nausea), and
respiratory symptoms (cough, phlegm).  There also is evidence for an immunologic effect-the
exacerbation of allergenic responses to known allergens and asthma-like symptoms. The lack of
adequate exposure-response information in the acute health effect studies precludes the
development of recommendations about levels of exposure that would be presumed safe for
these effects.

1.6.2. Chronic (Long-Term Exposure) Noncancer Respiratory Effects
       Information from the available human studies is inadequate for a definitive evaluation of
possible noncancer health effects from chronic exposure to DE. However, on the basis of
extensive animal evidence, DE is judged to pose a  chronic respiratory hazard to humans.
Chronic-exposure, animal inhalation studies show  a spectrum of dose-dependent inflammation
and histopathological changes in the lung in several animal species including rats, mice,
hamsters, and monkeys.
        This assessment provides an estimate of inhalation exposure of DE (as measured by
DPM) to which humans may be exposed throughout their lifetime without being likely to
experience adverse noncancer respiratory effects.  This exposure level, known as the reference
concentration (RfC) for DE of 5 |ig/m3 of DPM was derived on the basis of dose-response data
on inflammatory and histopathological changes in  the lung from rat inhalation studies. In
recognition of the presence of DPM in ambient PM2 5, it also is appropriate to  consider the
wealth of PM25 human health effects data. In this regard, the  1997 National Ambient Air
Quality Standard for PM2 5 of 15 |ig/m3 (annual average concentration) also would be expected
to provide a measure of protection from DPM, reflecting DPM's current approximate proportion
to PM2 5.

1.6.3. Chronic (Long-Term Exposure) Carcinogenic Effects
       This assessment concludes that DE is "likely to be carcinogenic to humans by inhalation"
and that this hazard applies to environmental exposures. This conclusion is based on the totality
of evidence from human, animal, and other supporting studies. There is considerable evidence
demonstrating an association between DE exposure and increased lung cancer risk among
workers in varied occupations where diesel engines historically have been used. The human
evidence from occupational studies is considered strongly supportive of a finding that DE

                                          1-4

-------
exposure is causally associated with lung cancer, though the evidence is less than that needed to
definitively conclude that DE is carcinogenic to humans. There is some uncertainty about the
degree to which confounders are having an influence on the observed cancer risk in the
occupational studies, and there is uncertainty evolving from the lack of actual DE exposure data
for the workers. In addition to the human evidence, there is supporting evidence of DPM's
carcinogenicity and associated DPM organic compound extracts in rats and mice by
noninhalation routes of exposure. Other supporting evidence includes the demonstrated
mutagenic and chromosomal effects of DE and its organic constituents, and the suggestive
evidence for bioavailability of the DPM organics in humans and animals.  Although high-
exposure chronic rat inhalation studies show a significant lung cancer response, this is not
thought predictive of a human hazard at lower environmental exposures. The rat response is
considered to result from an overload of particles in the lung resulting from the high exposure,
and such an overload is not expected to occur in humans at environmental exposures.
       Although the available human evidence shows a lung  cancer hazard to be present at
occupational exposures that are generally higher than environmental levels, it is  reasonable to
presume that the hazard extends to environmental exposure levels.  While there is an incomplete
understanding of the mode of action for DE-induced lung cancer that may occur in humans, there
is the potential for a nonthreshold mutagenic mode of action stemming from the  organics in the
DE mixture. A case for an environmental hazard also is shown by the simple observation that
the estimated higher environmental exposure levels are close  to, if not overlapping, the lower
range of occupational exposures for which lung cancer increases are reported.  These
considerations taken together support the prudent public health choice of presuming a cancer
hazard for DE at environmental levels of exposure. Overall, the evidence for a potential cancer
hazard to humans resulting from chronic inhalation exposure  to DE is persuasive, even though
assumptions and uncertainties are involved.  While the hazard evidence is persuasive, this does
not lead to similar confidence in understanding the exposure/dose-response relationship.
       Given a carcinogenicity hazard, EPA typically performs a dose-response assessment of
the human or animal data to develop a cancer unit risk estimate that can be used  with exposure
information to characterize the potential cancer disease impact on an exposed population.  The
DE human exposure-response data are considered too uncertain to derive a confident quantitative
estimate of cancer unit risk, and with the chronic rat inhalation studies not being predictive for
environmental levels of exposure, EPA has not developed a quantitative estimate of cancer unit
risk.
       In the absence of a cancer unit risk, simple exploratory analyses were used to provide a
perspective of the range of possible lung cancer risk from environmental exposure to DE.  The
analyses make use of reported lung cancer risk increases in occupational epidemiologic studies,

                                          1-5

-------
and the differences between occupational and environmental exposure. The purpose of having a
risk perspective is to illustrate and have a sense of the possible significance of the lung cancer
hazard from environmental exposure. The risk perspective cannot be viewed  as a definitive
quantitative characterization of cancer risk nor is it suitable for estimation of exposure-specific
population risks.

1.7.  SOURCES OF UNCERTAINTY
       Even though the overall evidence for potential human health effects of DE is persuasive,
many uncertainties exist because of the use of assumptions to bridge data and knowledge gaps
about human exposures to DE and the general lack of understanding about underlying
mechanisms by which DE causes observed toxicities in humans and animals.  A notable
uncertainty of this assessment is whether the health hazards identified from studies using
emissions from older engines can be applied to present-day environmental emissions and related
exposures, as some physical and chemical characteristics of the emissions from certain sources
have changed over time. Available data are not sufficient to provide definitive answers to this
question because changes in DE composition over time cannot be confidently quantified, and the
relationship between the DE components and the mode(s) of action for DE toxicity is/are
unclear. While recognizing the uncertainty, for this assessment a judgment is made that prior-
year toxicologic and epidemiologic findings can be applied to more current exposures, both of
which use DPM mass in air as the measure of DE exposure.
       Other uncertainties include the assumptions that health effects observed at high doses
may be applicable to low doses, and that toxicologic findings in laboratory animals generally are
predictive of human responses. In the absence of a more complete understanding of how DE
may cause adverse health effects in humans and laboratory animals, related assumptions (i.e., the
presence of a biological threshold for chronic respiratory effects based on cumulative dosage and
absence of a threshold for lung cancer stemming from subtle and irreversible effects) are
considered reasonable and prudent.
       Although parts of this assessment, particularly the noncancer RfC estimate, have been
derived with a generic consideration of sensitive subgroups within the population, the actual
spectrum of the population that may have a greater susceptibility to DE is unknown and cannot
be better characterized until more information is available regarding the adverse effects of DPM
in humans. Increased  susceptibility, for example, could result from above-average increases in
DE deposition and retention in the respiratory system or intrinsic differences in respiratory
system tissue sensitivity.  There is no DE-specific information that provides direct insight to the
question of differential human susceptibility.  Given the nature of DE's noncancer effects on the
respiratory system it would be reasonable, for example, to consider possible vulnerable

                                           1-6

-------
subgroups to include infants/children, the elderly, or individuals with preexisting health
conditions, particularly respiratory conditions.
       In developing a perspective on the possible significance of the environmental cancer
hazard of DE, this assessment uses information about the differences in the magnitude of DE
exposures between the occupational and environmental settings. Although an appreciation for
differences in exposure is needed only at an order-of-magnitude level for this assessment, one
should recognize that individual exposure is a function of both the variable concentrations in the
environment and the related breathing and particle retention patterns of the individual. Because
of variations in these factors across the population, different subgroups could receive lower or
higher exposure to DE than those groups mentioned in this assessment.
       Lastly, this assessment considers only potential heath effects from exposures to DE
alone. Effects of DE exposure could be additive to or synergistic with concurrent exposures to
many other air pollutants.  However, in the absence of more definitive data demonstrating
interactive effects (e.g., potentiation of allergenicity effects, potentiation of DPM toxicity by
ambient ozone and oxides of nitrogen) from combined exposures to DE and other pollutants, it is
not possible to address this issue.  Further research is needed to improve the knowledge and data
on DE exposures and potential human health effects, and thereby reduce uncertainties  of future
assessments of the DE health effects data.
                                           1-7

-------
     2. DIESEL EXHAUST EMISSIONS CHARACTERIZATION, ATMOSPHERIC
                       TRANSFORMATION, AND EXPOSURES

2.1.  INTRODUCTION
       This chapter provides background information relating to the diesel engine, the pollutants
it emits, the history of its use in highway vehicles and railroad locomotives, diesel exhaust
composition and emissions trends, and air pollution regulatory standards for diesel engines in the
United States.  The chapter also provides  specific information about the physical and chemical
composition of diesel exhaust, descriptions of its atmospheric transformations, observations of
measured and modeled ambient concentrations (considered alone and as a component of
atmospheric particles in general), some estimates of population exposures as well as a
comparison of DPM with ambient fine particulate matter (PM2 5). In addition, this chapter gives
background information that is used in conjunction with toxicology and epidemiology data to
formulate conclusions about human health hazards that are discussed in later chapters of this
document.  The exposure information does not represent a formal or rigorous exposure
assessment; it is intended only to provide  a context for the health effects data and health hazard
findings.
       For the purposes of this document, carbonaceous matter, diesel exhaust, diesel parti culate
matter, elemental carbon, organic carbon, soluble organic fraction, and soot are defined below.

       Carbonaceous matter: Carbon-containing compounds that are associated with
       particulate matter in diesel exhaust.  In this document, the term carbonaceous matter
       includes all organic and elemental carbon-containing compounds that are found in the
       particle phase. In other documents, this term is sometimes used interchangeably to refer
       to the insoluble fraction of diesel particulate matter or the soot fraction.

       Diesel engine exhaust (DE):  Gaseous and particle-phase emissions resulting from the
       combustion of diesel fuel in an internal-combustion, compression-ignition engine. DE
       includes emissions from a diesel engine or diesel vehicle (inclusive of aftertreatment
       devices), but does not include emissions from brake and tire wear.

       Diesel particulate matter (DPM): The particle-phase compounds emitted in DE. DPM
       can refer to both primary emissions and secondary  particles that are formed by
       atmospheric processes. In this document, DPM refers to primary particles.  Primary
       diesel particles are considered fresh after being emitted and aged after
                                          2-1

-------
       undergoing oxidation, nitration, or other chemical and physical changes in the
       atmosphere. As used in this document, DPM refers to both fresh and aged DPM unless a
       distinction is made.

       Elemental carbon (EC):  Carbon that has undergone pyrolysis (i.e., has been stripped of
       hydrogen).  In pure form, EC contains only carbon atoms, although EC as it exists in
       combustion particulate matter is likely to contain some hydrogen atoms.

       Organic carbon (OC): Carbon- and hydrogen-containing molecules emitted in DE
       largely as the result of unburned diesel fuel and, to a lesser extent, from engine
       lubrication oil.  OC compounds also can contain oxygen, nitrogen, and sulfur, as well as
       other elements in small quantities.

       Soluble organic fraction (SOF):  The organic portion of DPM that can be extracted
       from the particle matrix into solution.  Extraction solutions and procedures vary and are
       described in Section 2.2.8.1.

       Soot:  Agglomerations of EC and OC particles.  Soot also is often characterized as the
       insoluble portion of DPM, and is therefore considered to be mainly EC by some
       investigators.

       This chapter begins with a history of dieselization for on-road vehicles and locomotives,
followed by an introductory discussion of the formation of primary diesel emissions to assist the
reader in understanding the complex factors that influence the formation of particulate matter
(PM) and other DE emissions. The next section is a summary of EPA emission standards for
on-road and locomotive diesel engines and a description of the national trends in emissions from
on-road and nonroad diesel engine sources based on inventory modeling. The chapter continues
with a discussion of diesel fuel use and the impact of fuel properties on emissions.  The
chronological assessment of emissions factors is presented in summaries of chassis and engine
dynamometer testing and tunnel tests. This is followed by a description of engine technologies
and their effect on emissions, and a description of the chemical and physical nature of emissions.
The data describing the important atmospheric transformations of DE are summarized.  The
chapter concludes with  a summary of the available literature regarding atmospheric
concentrations of DPM and  exposures to DE.  EPA has assessed national and urban-area annual
average exposure to DPM using the Hazardous Air Pollutant Exposure Model, and this
assessment is presented in Section 2.4.3. A full exposure assessment would include the

                                          2-2

-------
distribution of ambient DE exposures in different geographic regions and among different
demographic groups, the most highly exposed (90th percentile), exposures in microenvironments
for short and long durations, the maximum exposure range (98th percentile), and the number of
maximum-exposed individuals.  However, such an assessment is not currently available. EPA is
developing tools to provide a more complete exposure assessment.

2.2.  PRIMARY DIESEL EXHAUST EMISSIONS
2.2.1. History of Dieselization
       The diesel engine was patented in!892 by Rudolf Diesel, who conceived it as a prime
mover that would provide much improved fuel efficiency compared with spark-ignition (SI)
engines. To the present day, the diesel engine's excellent fuel economy remains one of its
strongest selling points. In the United States, the diesel engine is used mainly in trucks, buses,
agricultural and other nonroad equipment, locomotives, and ships.
       The chief advantages of the diesel engine over the gasoline engine are its fuel economy
and durability. Diesel engines, however, emit more PM per mile driven compared with gasoline
engines of a similar weight. Over the past decade, modifications of engine components have
substantially reduced particle emissions from both diesel and gasoline engines (Hammerle et al.,
1994; Sawyer and Johnson, 1995).
       The diesel engine compresses air to high pressure and temperature. Fuel, when injected
into this compressed air, autoignites, releasing its chemical energy.  The expanding combustion
gases do work on the piston before being exhausted to the atmosphere. Power output is
controlled by the amount of injected fuel rather than by throttling the air intake.  Compared to its
SI counterpart, the diesel engine's superior efficiency derives from a higher compression ratio
and no part-load throttling. To ensure structural integrity for prolonged reliable operation at the
higher peak  pressures brought about by a higher compression ratio and autoignition, the structure
of a diesel engine generally is more massive than its SI counterpart.
       Diesel engines (also called compression-ignition) may be broadly identified as being
either two- or four-stroke cycle, injected directly or indirectly, and naturally  aspirated or
supercharged.  They also are classified according to service requirements such as light-duty (LD)
or heavy-duty (HD) automotive/truck, small or large industrial, and rail or marine.
       All diesel engines use hydraulic fuel injection in one form or another. The fuel system
must meet four objectives if a diesel engine is to function properly over its entire operating
range. It must: (1) meter the correct quantity of fuel, (2) distribute the fuel to the correct
cylinder, (3) inject the fuel at the correct time, and (4) inject the fuel so that it is atomized and
mixes well with the in-cylinder air. The first two objectives are functions of a well-designed
injection pump, and the last two are mostly functions of the injection nozzle.  Fuel injection

                                          2-3

-------
systems are moving toward the use of electronic components for more flexible control than is
available with purely mechanical systems to obtain lower exhaust emissions without diminishing
fuel efficiency.
       Both the fuel and the lubricants that service diesel engines are highly finished
petroleum-based products combined with chemical additives. Diesel fuel is a mixture of many
different hydrocarbon molecules from about C7 to about C35, with a boiling range from roughly
350 °F to 650 °F. Many of the fuel and oil properties, such as specific energy content (which is
higher than gasoline), ignition quality, and specific gravity, are related to hydrocarbon
composition. Therefore, fuel and lubricant composition affect many aspects of engine
performance, including fuel economy and exhaust emissions.
       Complete and incomplete combustion of fuel in the diesel engine results in the formation
of a complex mixture of gaseous (gas-phase hydrocarbons, CO, CO2, NO, NO2,  SO2) and
particulate exhaust (carbonaceous matter, sulfate, and trace elements). Because of concerns over
health effects associated with DE, EPA began regulating emissions from diesel engines in 1970
(for smoke) and then added regulations for gaseous emissions. EPA first regulated particulate
emissions from FID diesels in  1988.

2.2.1.1. Dieselization of the On-Road Fleet
       Because of their durability and fuel economy, the use of diesel engines, particularly in
long-distance applications, has increased over the years. The Census of Transportation, Truck
Inventory and Use Survey (TIUS) indicates that among Class 3-8 trucks, diesel engine use has
increased more rapidly than gasoline engine use in the past 20 years. Truck classes are defined
by gross vehicle weight as described in Table 2-1. Dieselization first occurred among Class 7
and 8 trucks. The TIUS indicates that 81.5% of diesel trucks on the road in 1963 were Class 7 or
8 trucks (Table 2-2).  Class 7 sales became predominantly (>50%) diesel in the 1970s and Class
8 sales became predominantly diesel in the 1960s. Diesels did not make up a majority of class 5
and 6 sales until the 1990s (Figures 2-1 and 2-2). HD trucks have historically constituted  the
majority of diesel sales  and mileage. However, an increasing number of LD diesel trucks  have
been sold domestically in recent years. In the 1990s,  approximately one in three diesel trucks
sold was a Class 1  or Class 2 vehicle. Diesel trucks have historically been driven more miles per
truck than gasoline trucks.  For example, the TIUS indicates that 59% of diesel trucks were
driven more than 50,000 miles in 1963, compared with  3% of gasoline trucks.
                                          2-4

-------
Table 2-1.  Vehicle classification and weights for on-road trucks
Class
1
2
O
4
5
6
7
8Aa
8Ba
Medium duty (MD)
Light-heavy duty (LHD)
Heavy-heavy duty (HHD)
Gross vehicle weight (Ib)
<6,000
6,001-10,000
10,001-14,000
14,001-16,000
16,001-19,500
19,501-26,000
26,001-33,000
33,001-60,000
>60,000
10,001-19,500 (same as Classes 3-5)
19,501-26,000 (same as Class 6)
>26,001 (same as Class 7-8)
 aClass 8A and Class 8B are often considered together.
Table 2-2. Total (gas and diesel) diesel trucks in the fleet in 1992
Truck class
Class 1 and 2
(Light duty)
Class 3, 4, and 5
(Medium duty)
Class 6
(Light heavy-duty)
Class 7 and 8
(Heavy heavy-duty)
1992 gas and
diesel trucks
55,193,300
1,258,500
732,300
2,016,600
1992 diesel
trucks
1,387,600
326,300
273,800
1,725,300
% Diesels
O
26
37
86
 Source: Census of Transportation, 1995.
                               2-5

-------
Figu
Sour
onn nnn
1 an nnn
ifin nnn
1 AC\ nnn
— -ion nnn
(0
Jr 1 nn nnn
.a
£ «n nnn
Rn nnn
40 onn
2n nnn
0
19


a3 ja
W £J n ^
a
™a a * a OQ
a a3 °o°
a3 a ^°/
o 0^3 oC"^
	 fy*>Q^Y^QrCU^Bi^^^^^^^^^^^^^l^Bli^B^^^^^^^M^£^A^^^^^^^^^^^^I^^^^^^^y^^
30 1940 1950 1960 1970 1980 1990 2000
re 2-1. Diesel truck sales (domestic) for the years 1939-1997.
;e: AAMA, 1927-1974 and 1975-1998.

80 0%
Rn n%
0) OU.U/O

-------
Among combination trucks, consisting of tractor-trailers and single-unit trucks with trailers,
diesel vehicles have driven a majority of the miles since at least 1963, the first year in which
TIUS was conducted (Figure 2-3).
Figure 2-3.  Percentage of truck miles attributable to diesel trucks. VMT= vehicle miles
traveled.
Source: U.S. Bureau of the Census, 1999b.
       The longevity of diesel trucks is an important factor to understand past, current, and
projected exposures to DE because older vehicles are subject to less stringent regulations and
may remain in use for several decades after their manufacture. American Automobile
Manufacturers Association publications (AAMA, 1927-1997) indicate that 53% of trucks from
model years 1947-1956 were still on the road after 14 years.  The proportion of trucks in use
after 14 years was 63% for model years 1974-1983, suggesting that the lifespan of trucks built in
later years is longer. According to the  1997 TIUS, vehicles older than 10 years made up 40% of
Class 7 and 8 trucks and 16% of Class 7-8 vehicle miles traveled (VMT) (Figures 2-4 and 2-5).
Almost all Class 7 and 8 trucks were diesel  vehicles in the period 1982-1997 (93% in 1982 and
99% in 1997).
2.2.1.2. Diesetization of Railroad Locomotive Engines
       Early in the 20th century the political and economic pressure on the railroads to replace
steam locomotives was substantial.  Railroads were losing business to other forms of transport.
The diesel-electric locomotive provided 90% in-service time, compared with only 50% for steam
locomotives, and had three times the thermal efficiency (Klein, 1991; Kirkland, 1983).

                                          2-7

-------
        
-------
Additionally, several cities had passed laws barring steam locomotives within the city limits
because the large quantities of smoke obscured visibility, creating a safety hazard. The first
prototype diesel locomotive was completed in 1917. By 1924 General Electric (GE) was
producing a standard line of switching locomotives on a production basis. Electro-Motive
Corporation was founded the same year to produce diesel locomotives in competition with GE.
This company was purchased in 1929 by General Motors (GM) and became the Electro-Motive
Division. After this acquisition, GM began to develop the two-stroke engine for this application.
Up to this time, all locomotive diesel engines were four-stroke.  Two-strokes offered a much
higher power-to-weight ratio, and GM's strategy was to get a large increase in power by moving
to the two-stroke cycle. The first true high-speed, two-stroke, diesel-electric locomotives were
produced by GM in  1935. However, because of the economic climate of the Great Depression,
few of these were sold until after the Second World War. At the end of the war,  most
locomotives were still  steam-driven but were more than 15 years old, and the railroads were
ready to replace the  entire locomotive fleet. Few, if any, steam locomotives were sold after 1945
because the entire fleet was converted to diesel (Coifman,  1994).
       The locomotive fleet has included significant percentages of both two- and four-stroke
engines.  The four-stroke diesel engines were naturally aspirated in the 1940s and 1950s. It is
unlikely that any of the two-stroke engines used in locomotive applications were strictly
naturally aspirated.  Nearly all two-stroke diesel locomotive engines are uniflow scavenged, with
a positive-displacement blower for scavenging assistance.  In 1975, it was estimated that 75% of
the locomotives in service were two-stroke, of which about one-half used one or more
turbochargers in addition to the existing positive-displacement blower for additional intake boost
pressure.
       Almost all of the four-stroke locomotive engines were naturally aspirated in 1975.
Electronic fuel injection for locomotive engines was first offered in the 1994 model year (U.S.
EPA, 1998b). All locomotive engines manufactured in recent years are turbocharged,
aftercooled or intercooled four-stroke engines.  In part, this is because of the somewhat greater
durability of four-strokes, although impending emissions regulations may have also been a factor
in this shift.  The typical lifespan of a locomotive has been estimated to be more than 40 years
(U.S. EPA, 1998b).  Many of the smaller railroads are still using engines built in the 1940s,
although the engines may have been rebuilt several times since their original manufacture.

2.2.2. Diesel Combustion and Formation of Primary Emissions
       A basic understanding of diesel combustion processes can  assist in understanding the
complex factors that influence the formation of DPM and other DE emissions. Unlike SI
combustion,  diesel combustion is a fairly nonhomogenous process. Fuel is sprayed at high

                                          2-9

-------
pressure into the compressed cylinder contents (primarily air with some residual combustion
products) as the piston nears the top of the compression stroke.  The turbulent mixing of fuel and
air that takes place is enhanced by injection pressure, the orientation of the intake ports
(inducement of intake-swirl tangential to the cylinder wall), piston motion, and piston bowl
shape.  In some cases, fuel and air mixing is induced via injection of the fuel into a turbulence-
generating pre-chamber or swirl chamber located adjacent to the main chamber (primarily in
older, higher speed engines and some LD diesels). Examples of typical direct injection and
indirect injection combustion systems are compared in Figure 2-6. Diesel combustion can be
considered to consist of the following phases (Heywood, 1988; Watson and Janota,  1982):
     Single Hole
   Injection Nozzle
Precombustion
  Chamber
                                                                          Four Hole
                                                                        Injection Nozzle
                       A.           Piston Bowl
  Figure 2-6. A comparison of IDI (A) and DI (B) combustion systems of high-speed
  HD diesel truck engines.  DI engines almost completely replaced IDI engines for
  these applications by the early 1980s. (IDI = indirect injection, DI=direct injection)
                                          2-10

-------
            An ignition delay period, which starts after the initial injection of fuel and continues
            until the initiation of combustion. The delay period is governed by the rate of fuel
            and air mixing, diffusion, turbulence, heat transfer, chemical kinetics, fuel
            vaporization, and fuel composition. Fuel cetane rating is an indication of ignition
            delay.
       •     Rapid, premixed burning of the fuel and air mixture from the ignition delay period.
       •     Diffusion-controlled burning, where the fuel burns as it is injected and diffuses into
            the cylinder.
            A very small amount of rate-controlled burning during the expansion stroke,  after
            the end of injection.

       Engine speed and load are controlled by the quantity of fuel injected. Thus, the overall
fuel-to-air ratio varies greatly as engine speed and load vary. On a macro scale, the cylinder
contents are always fuel-lean. Depending on the time available for combustion and the
proximity of oxygen, the fuel droplets are either completely or partially oxidized.  At
temperatures above 1,300 K, much of the unburned fuel that is not oxidized is pyrolized
(stripped of hydrogen) to form EC (Dec and Espey, 1995). In addition to EC, other
carbonaceous matter is present, largely from unburned fuel. The agglomeration of elemental and
OC forms particles that are frequently referred to as "soot" particles. In this document, the terms
"EC" and "OC" are used to refer to the carbon-containing components of DPM, and collectively,
they are referred to as the carbonaceous fraction of a diesel particle.
       Carbonaceous particle formation occurs primarily during the diffusion-burn phase of
combustion, and is highest during high load and other conditions consistent with high fuel-air
ratios. Most of the carbonaceous matter formed (80% to 98%)  is oxidized during combustion,
most likely by hydroxyl radicals (Kittelson et al., 1986; Foster and Tree, 1994).
       DPM is  defined by the measurement procedures summarized in the Code of Federal
Regulations, Title 40 CFR, Part 86, Subpart N (CFR 40:86.N).  These procedures define DPM
emissions as the mass of material collected on a filter at a temperature of 52 °C or less after
dilution of the exhaust with air.  DPM is formed by a number of physical processes acting in
concert as the exhaust is cooled and diluted. These are nucleation, coagulation, condensation,
and adsorption.  The core DE particles are formed by nucleation and coagulation from primary
spherical particles consisting of solid carbonaceous (EC) material and ash (trace metals and other
elements).  To these, through coagulation, adsorption, and condensation, are added organic and
sulfur compounds (sulfate) combined with other condensed material (Figure 2-7).  Because of
                                          2-11

-------
                                     Solid Carbonaceous/Ash Particle
                                     with adsorbed hydrocarbon/sulfate layer
                                            Sulfuric Acid Particles
                                          Hydrocarbon/Sulfate Particles
              0.2
         Figure 2-7.  Schematic diagram of diesel engine exhaust particles.

         Source: Modified from Kittelson, 1998.
their size, <0.5 mm, these particles have a very large surface area per gram of mass, which
makes them able to adsorb large quantities of ash, organic compounds, and sulfate.  The specific
surface area of the EC core has been measured to be approximately 30-50 m2/g (Frey and Corn,
1967). Pierson and Brachaczek (1976) report that after the extraction of adsorbed organic
material, the surface area of the diesel particle core is approximately 90 m2/g.
       The organic material associated with diesel particles originates from unburned fuel,
engine lubrication oil,  and small quantities of partial combustion and pyrolysis products.  This is
frequently quantified as the SOF, which is discussed in much more detail in Section 2.2.7. The
formation of sulfate in DE depends primarily on fuel sulfur content. During combustion, sulfur
compounds present in  the fuel are oxidized to sulfur dioxide (SO2).  Approximately 1% to 4% of
fuel sulfur is oxidized  to form sulfuric acid (H2SO4) (Wall et al., 1987; Khatri et al., 1978;
Baranescu, 1988; Barry et al., 1985). Upon cooling, sulfuric acid and water condense into an
aerosol that is nonvolatile under ambient conditions. The mass of sulfuric acid DPM is more
than doubled by the mass of water associated with the sulfuric acid under typical DPM
measurement conditions (50% relative humidity, 20-25 °C) (Wall et al., 1987).

                                          2-12

-------
       Emissions from combustion engines produce oxide of nitrogen (NOX) primarily (at least
initially) as of NO.  High combustion temperatures cause reactions between oxygen and nitrogen
to form NO and some NO2. Most NO2 formed during combustion is rapidly decomposed. NO
can also decompose to N2 and O2 but the rate of decomposition is very slow (Heywood, 1988;
Watson and Janota, 1982).  Thus, almost all of the NOX emitted is NO.
       Some organic compounds from unburned fuel and from lubricating oil consumed by the
engine can be trapped in crevices or cool spots within the  cylinder and thus are not sufficiently
available to conditions that would lead to their oxidation or pyrolysis.  These compounds are
emitted from the engine and either contribute to gas-phase organic emissions or to DPM
emissions, depending on their volatility. Within the exhaust system, temperatures are sufficiently
high that these compounds  are entirely present within the gas phase (Johnson and Kittelson,
1996). Upon cooling and mixing with ambient air in the exhaust plume, some of the less volatile
organic compounds can adsorb to the surfaces of the EC agglomerate particles.  Lacking
sufficient EC adsorption sites, the organic compounds may condense on sulfuric acid nuclei to
form a heterogeneously nucleated organic aerosol (Abdul-Khalek et al., 1999).
       Although not unique to DE, the high content of EC associated with typical DPM
emissions has long been used by some investigators to distinguish diesel engine sources of this
particle from other combustion aerosols.  Diesel particles from newer HD engines are typically
composed of-75% EC (EC can range from 33% to 90%), -20% OC (OC can range from 7% to
49%), and small amounts of sulfate, nitrate, trace elements, water, and unidentified components
(Figure 2-8). Metallic compounds from engine component wear, and from compounds in the
fuel and lubricant, contribute to DPM mass. Ash from oil combustion also contributes trace
amounts.
       Ambient PM2 5 measured in the eastern United  States is dominated by sulfate (34%),
whereas ambient PM2 5 in the western United States is dominated by OC (39%) (Table 2-3) (U.S.
EPA,  1999a). Many sources contribute to ambient PM2 5, and these sources and their relative
contribution to ambient PM25 can be identified on the basis of the chemical species present.  The
OC fraction of DPM is increasingly being used to assist investigators in identifying the
contribution of diesel engine emissions to ambient PM2 5.  In particular, hopane and  sterane
compounds (aromatic compounds, >C30) have been used in addition to other polycyclic aromatic
hydrocarbons (PAHs) and long-chain alkanes to distinguish DPM from other mobile source PM
and from ambient PM (Schauer  et al., 1996; Fujita et al., 1998). Although PAH compounds
make up 1% or less of DPM mass, diesel emissions have been observed to have elevated
concentrations of methylated naphthalenes and methylated phenanthrene isomers compared  to
other combustion aerosols (Benner et al., 1989; Lowenthal et al., 1994; Rogge et al., 1993).
Enrichment of benzo[a]anthracene and benzo[a]pyrene (B[a]P) in DPM has also been

                                         2-13

-------
            Diesel  PM2.5 Chemical Composition
         Sulfate, Nitrate
             1%
            (1-4%)
      Organic Carbon
          19%
         (7-49%)
M etals&Elements
      2%
    (1-5%)
                                           Other
                 (1-10%)
                                                  Elemental Carbon
                                                       75%
                                                     (SS-90%)
Figure 2-8. Typical chemical composition for diesel particulate matter (PM2 5)
from new (post-1990) HD diesel vehicle exhaust.
        Table 2-3. Typical chemical composition of fine particulate matter

Elemental carbon
OC
Sulfate, nitrate,
ammonium
Minerals
Unknown
Eastern U.S.
4%
21%
48%
4%
23%
Western U.S.
15%
39%
35%
15%
-
Diesel PM2 5
75%
19%
1%
2%
3%
        Source: U.S. EPA, 1999a.
                                     2-14

-------
observed under some conditions and has been used to assess the relative contribution of DE to
ambient PM.
       Although specific OC species are being used to help distinguish DPM aerosols from
other combustion aerosols, up to 90% of the organic fraction associated with DPM is currently
classified as unresolvable complex material.  Ultrafme DPM (5-50 nm) accounts for the majority
(50% to 90%) of the number of particles but only 1% to 20% of the mass of DPM. A study
conducted by Gertler (1999) in the Tuscarora Mountain tunnel demonstrated an increase in 20
nm diameter particles as the fraction of diesel vehicles in the tunnel increased from 13% to 78%.
The contribution of nuclei-mode particles from a freeway on an ambient aerosol size distribution
was reported by Whitby and Sverdrup (1980).
       In summary, four main characteristics of DPM are (1) the high proportion of EC, (2) the
large surface area associated with the carbonaceous particles in the 0.2 |im size range, (3)
enrichment of certain poly cyclic organic compounds, and (4) 50%-90% of the  number of DPM
particles in diesel engine exhaust are in the nuclei-mode size range, with a mode of 20 nm.

2.2.3. Diesel Emission Standards and Emission Trends Inventory
       EPA set a smoke standard for on-road HD diesel engines beginning with the 1970 model
year and added a carbon monoxide (CO) standard and a combined hydrocarbon (HC)  and NOX
standard for the 1974 model year (Table 2-4). Beginning in the 1979 model year, EPA added a
HC standard while retaining the combined HC and NOX standard.  All of the testing for HC, CO,
and NOX was completed using a steady-state test procedure. Beginning in the 1985 model
year,EPA added a NOX standard (10.7 g/bhp-hr), dropped the combined HC and NOX standard,
and converted from steady-state to transient testing for HC,  CO, and NOX emissions. EPA
introduced a particulate standard for 1988 model year diesel engines using the transient test (0.6
g/bhp-hr). Transient testing involves running an engine on a dynamometer over a range of load
and speed set points.
       Since the 1985 model year, only the NOX and particulate standards have been tightened
for on-road diesel engines. For truck and bus engines, the particulate standard  was reduced to
0.25 g/bhp-hr in 1991, and it was reduced again in 1994 for truck engines to 0.1 g/bhp-hr.  For
urban bus engines, the particulate standard was reduced in 1994 to 0.07 g/bhp-hr and again in
1996 to 0.05 g/bhp-hr. The NOX standard was reduced to 4.0 g/bhp-hr in 1998  for all on-road
diesel engines (bus and truck engines).  The standards for nonmethane hydrocarbon (NMHC)
and NOX combined were further lowered in a 1997 rulemaking, to take effect in 2004. EPA has
recently finalized a regulation that will further reduce NOX NMHC, and PM emissions from
diesel engines starting in 2007.
                                         2-15

-------
Table 2-4. U.S. emission standards: HD highway diesel en
Model
year
1970
1974
1979
1985C
1988
1990
1991
1993
1994
1996
1998
2004
2007
Pollutant (g/bhp-hr)
HC
—
—
1.5
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3

CO
—
40
25
15.5
15.5
15.5
15.5
15.5
15.5
15.5
15.5
15.5
15.5
NOX
—
—
—
10.7
10.7
6.0
5.0
5.0
5.0
5.0
4.0
—
0.2
HC + NOX
—
16b
10b
—
—
—
—
—
—
—
—
2.4 NMHCd
0.14 NMHC
Participate (PM)
t=truck, b=bus,
ub=urban bus
—
—
—
—
0.60
0.60
0.25
0.25 t, 0.10 b
0. lOt, 0.07 ub
0. lOt, 0.05 ub
0. lOt, 0.05 ub
0. lOt, 0.05 ub
0.01
jines
Smoke3
A:40%; L:20%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
A:20%; L:15%; P:50%
       Emissions measured in percent opacity during different operating modes: A=acceleration; L=lug; P=peaks
       dunng either mode.
       bTotal HC.
       °In 1985, test cycle changed from steady-state to transient operation for HC, CO, and NOX measurement and
       in 1988 for PM.
       dOr 2.5 plus a limit of 0.5 nonmethane hydrocarbon (NMHC).
       In December 1997, EPA adopted emission standards for NOX, HC, CO, PM, and smoke
for newly manufactured and remanufactured railroad locomotives and locomotive engines. The
rulemaking, which took effect in the year 2000, applies to locomotives originally manufactured
in 1973 or after, and any time they are manufactured or remanufactured (locomotives originally
manufactured before 1973 are not regulated).  Three sets of emission standards have been
adopted (Tier 0, 1, and 2); they apply to locomotives and locomotive engines originally
manufactured from 1973 through 2001 (Tier 0), from 2002  through 2004 (Tier 1), and in 2005
and later (Tier 2) (Table 2-5; see EPA web page at http://www.epa.gov/omswww/ or
http://www.dieselnet.com/standards/ for current information on mobile source emission
standards). The emissions are measured over two steady-state test cycles that represent two
                                          2-16

-------
Table 2-5. U.S. emission standards: locomotives

Line-haul
Switch
Line-haul
Switch
Line-haul
Switch
Year3
1973 -2001 (Tier 0)
1973 -2001 (Tier 0)
2002-2004 (Tier 1)
2002-2004 (Tier 1)
2005 + (Tier 2)
2005 + (Tier 2)
CO
5.0
8.0
2.2
2.5
1.5
2.4
HC
1.0
2.1
0.55
1.2
0.3
0.6
g/bhp-hr)
NOX
9.5
14.0
7.4
11.0
5.5
8.1
PM
0.6
0.72
0.45
0.54
0.20
0.24
       aDate of engine manufacture.
different types of service, including line-haul (long-distance transport) and switch (involved in
all transfer and switching operations in switchyards) locomotives.
       Emission standards for nonroad equipment are not as stringent as current standards for
on-road equipment and are being phased in within the next decade.  Currently, Federal PM
standards exist for nonroad equipment of several horsepower ratings. For equipment between
175 and 750 horsepower, the PM standard was set at 0.4 g/bhp-hr in 1996 and will decrease to
0.15 g/bhp-hr between 2001 and 2003 depending on the power rating (Table 2-6). This
equipment includes construction, agricultural, and industrial such as bulldozers, graders, cranes,
and tractors. The current PM standard for this equipment is only slightly lower than the 0.6
g/bhp-hr PM standard in place for on-road HD diesel  engines in the late 1980s.
       The EPA emission trends report (U.S. EPA, 2000a) provides emission inventories  for
criteria pollutants (PM10, PM2 5, SO2, NOX, volatile organic compounds [VOC], CO, Pb, and
NH3) from point, area, and mobile sources, which indicate how emissions have changed from
1970 to 1998.  The emission trends are based on the EPA mobile source inventory models
MOBILE, PARTS, and the draft NONROAD model.  PARTS  derives particulate emission rates
for FID diesel vehicles using data generated for new engine certification purposes. PARTS is
currently being modified to account for deterioration, in-use emissions, poor maintenance, and
tampering effects, all of which would  increase emission factors. PM, SO2, NOX, and VOC
emissions  trends from the report are discussed below. Ambient urban/suburban PM samples
rarely reflect the large fraction of natural and miscellaneous sources suggested by the national
inventory, owing to removal of a large portion of these emissions close to their sources as well as
dispersion from these sources to urban/suburban sites. The removal of natural and miscellaneous
PM10 (largely fugitive dust) near their source is a result of the lack of inherent thermal buoyancy,
low release height, and interaction with their surroundings  (impaction and filtration by
vegetation).
                                          2-17

-------
     Table 2-6.  U.S. emission standards for nonroad diesel equipment (g/bhp-hr)
Power rating
IKhp
ll750 hp
Model
year
2000
2005+
2000
2005+
2000
2005+
1998+
2004
2008+
1997+
2003
2007+
1996+
2003
2006+
2001
2006+
2002
2006+
2000+
2006+
Pollutant (g/bhp-hr)
HC
—
—
—
—
—
—
—
—
—
—
—
—
1.0
—
—
—
—
—
—
1.0
—
CO
6.0
6.0
4.9
4.9
4.1
4.1
—
3.7
3.7
—
3.7
3.7
8.5
2.6
2.6
2.6
2.6
2.6
2.6
8.5
2.6
NOX
—
—
—
—
—
—
6.9 (ABT)


6.9 (ABT)
—
—
6.9 (ABT)
—
—
—
—
—
—
6.9 (ABT)
—
NMHC +
NOX
7.8 (ABT)
5.6 (ABT)
7.0 (ABT)
5.6 (ABT)
7.0 (ABT)
5.6 (ABT)
—
5.6 (ABT)
3.5 (ABT)
—
4.9 (ABT)
3.0 (ABT)
—
4.9 (ABT)
3.0 (ABT)
4.8 (ABT)
3.0 (ABT)
4.8 (ABT)
3.0 (ABT)
—
4.8 (ABT)
PM
0.74 (ABT)
0.60 (ABT)
0.60 (ABT)
0.60 (ABT)
0.60 (ABT)
0.44 (ABT)
—
0.30 (ABT)
—
—
0.22 (ABT)
—
0.4
0.15 (ABT)
—
0.15 (ABT)
—
0.15 (ABT)
—
0.4
0.15 (ABT)
Smoke %a






20/15/50


20/15/50


20/15/50






20/15/50

"Emissions measured in percent opacity during different operating modes:  acceleration/lug/peaks during either mode.
ABT=average banking and trading.
Note: The standards for engines less than 50 hp also apply to diesel marine engines.
                                                 2-18

-------
For the summaries presented here, natural and miscellaneous sources are excluded from the
national PM and NOX inventories.
      From 1970 to 1998, PM10 emissions decreased from slightly over 12,200,000 tons to just
over 2,800,000 tons (Figure 2-9). PM10 emissions from on-road and nonroad diesel  engines
increased from 320,000 tons to more than 521,000 tons during this same period, so that in 1970
diesel engine emissions were 3% of the PM10 inventory whereas in 1998, diesel engine emissions
were 18% of the PM10 inventory. Diesel engines also contribute to secondary PM formation
from NOX and SO2 emissions that are converted to nitrate and sulfate. VOCs from diesel engines
also contribute to secondary organic particle formation. The contribution of secondary PM is not
included in the national trends inventories cited here.
      Mobile sources of PM include both gasoline- and diesel-powered on-road vehicles and a
variety of nonroad equipment.  Nonroad diesel engine sources include construction equipment,
agricultural equipment, marine vessels, locomotives, and other sources. The EPA emission
trends report (U.S. EPA, 2000a) indicates that, excluding natural and miscellaneous sources,
mobile sources were responsible for 25% of PM10 emissions in 1998. Diesel engines (on-road
and nonroad combined) were estimated to contribute 72% of mobile-source PM10 emissions.

i zinnn
i 9000 -
§
o i nnnn
o
*3 Qnnn -
o
J2 <^nnn -
c3
w
$ /innn
H
9000 -
n




















1970
• Diesel PM10
D Other Sources of PM10








1980 199
Calendar Year







1
1
0







p=|
\
1998










   Figure 2-9.  Trends in PM10 emissions from on-road and nonroad engines combined
   and other anthropogenic sources of PM10 from 1970 to 1998 (excludes miscellaneous
   and natural sources).

   Source: U.S. EPA, 2000a, National Air Pollutant Emission Trends, 1900-1998.
                                         2-19

-------
Because of the high concentration of fine particles in engine emissions, diesel engines (on-road
and nonroad combined) were estimated to contribute 77% of mobile-source PM25 emissions and
23% of total PM25 in 1998 (excluding natural and miscellaneous emissions). If natural and
miscellaneous PM2 5 sources are included in the inventory, diesel PM2 5 contributes 6% to the
national inventory.
       Gram per mile particulate emissions from diesel vehicles are much greater than those
from gasoline-fueled vehicles, accounting for the large contribution of diesel engine emissions to
the national inventory in spite of the smaller number of diesel engines in use.  Particulate
emissions (PM10) from gasoline-fueled engines decreased dramatically in 1975 with the
widespread introduction of unleaded gasoline. Particulate emissions from diesel highway
vehicles have decreased recently because of EPA emission standards for new model year HD
diesel trucks that were first implemented in 1988 and became increasingly stringent in 1991,
1994, and 2000, as presented in Table 2-4.  A decrease in on-road HD DPM emissions since the
mid-1980s is confirmed by in-use vehicle testing, as described in Section 2.2.5. Because of the
implementation of existing regulations, DPM emissions from on-road sources are expected to
decrease 37% from 1998 to 2007; however, nonroad DPM emissions are expected to increase
15% in the same period (Figure 2-10).
       The EPA emission trends report (U.S. EPA, 2000a) indicates that annual on-road vehicle
PM10 emissions decreased from 397,200 tons to 257,080 tons from 1980 to 1998.l Passenger car
particulate emissions decreased 53% (from 119,000 to 56,000 tons) in this timeframe, while on-
road diesel vehicle PM10 emissions decreased 27% (from 208,000 to 152,000 tons) (Figure 2-10).
Nonroad diesel engine PM10 emissions increased 17% (from 314,000 tons in 1980 to 69,000 tons
in 1998).  Emissions data for PM25 are available only for the period from 1990 to 1998.
Between 1990 and 1998, PM25 emissions from mobile sources decreased by 14%, largely as the
result of decreased on-road emissions.
       From 1970 to 1998, NOX emissions increased from 20,598,000 tons to 24,126,000 tons
(Figure 2-11). NOX emissions from on-road and nonroad diesel engines increased from
1,748,000 tons to 4,753,000 tons during this same period, so that in 1970 diesel engine emissions
were 8% of the NOX inventory while in 1998, diesel engine emissions were 20% of the NOX
       Exhaust emissions constitute the majority of PM emissions from mobile sources, with tire and brake wear
contributing the remainder. To compare trends estimates from past years with future projections (which are provided
for exhaust emissions only), the fraction of brake and tire wear would need to be omitted from these estimates as
reported in the emission trends report (U.S. EPA, 2000a). On average in the late 1990s 39% and 64% of gasoline
vehicle particulate emissions originated from exhaust and 95% and 98% of on-road diesel emissions originated from
exhaust for PM10 and PM2 5, respectively.

                                          2-20

-------
       TOO
       600
       500
      o400
      .n
      
-------
'YYYin
JUJUJ
•ji o^nn
£H Z5LUJ
o
+J
"S Trm
O AJUJU
^
en
<+H
o I»YY>
M iiuuu
T3
Si inmn
2 1UJUJ
o
A
E""1 «Yin
jUJJ
• EteselNOx
y Otter Sources ofNCk










	








19D
P








z




H



R








1 1
1980 1990 1998
Calendar Year
Figure 2-11.  Trends in NOX emissions from on-road and nonroad diesel engines combined
and other anthropogenic sources of NOX from 1970 to 1998 (excludes miscellaneous and
natural sources).

Source: U.S.  EPA, 2000a, National Air Pollutant Emission Trends, 1900-1998.
NOX from LD gasoline vehicles decreased from 1980 to 1998, resulting in an overall decrease in
on-road NOX emissions of 9%, NOX from diesel trucks and buses increased 7% (from 2,463,390
tons in 1980 to 2,630,120 tons in 1998), owing to the illegal use of electronic control devices that
bypassed the trucks' emission control systems, as discussed in Section 2.2.5. NOX emissions
from nonroad diesel engines (including commercial marine and locomotives) have increased
46% (from 3,251,600 tons in 1980 to 4,752,800 tons in 1998) (Figure 2-12).
       About 7% of SO2 came from mobile sources in 1998, with diesels responsible for 74% of
that total. EPA regulations for on-road diesel fuel sulfur content (which started in 1993) have
significantly reduced SO2 emissions from highway diesels.  SO2 emissions from highway diesel
                                         2-22

-------
                                          sNonroad
Figure 2-12. Trends in NOX emissions from on-road and nonroad diesel engines from 1970
to 1998.

Source: U.S. EPA, 2000a, National Air Pollutant Emission Trends, 1900-1998.
engines have decreased 72% (from 303,000 tons in 1980 to 85,000 tons in 1998) (Figure 2-13).
Similar trends are not apparent for nonroad diesels, although in 1998 nonroad diesel engines,
excluding commercial marine vessels, emitted 785,000 tons of SO2, accounting for 56% of
mobile-source SO2 emissions in 1998.
       Diesel engines are not a large source of VOC emissions compared with gasoline engines.
VOC emissions from diesel engines in 1998 were estimated at 2% of the total emissions from all
sources. VOC emissions from diesel mobile sources decreased 9% (from 779,000 tons in 1980
to 721,000 tons in 1998) (Figure 2-14).
       Diesel engines are also not a large source of CO emissions compared with gasoline
engines. In 1998, mobile sources emitted 79% of all CO, and diesel engines accounted for 4% of
the mobile-source CO. CO emissions from on-road diesel vehicles increased 34% between 1980
and 1998, during which time nonroad diesel emissions of CO increased 45% (Figure 2-15).
                                         2-23

-------
                                  E3 O n -ro ad  Q N o n ro ad
                                     Calendar Yea
Figure 2-13. Trends in SO2 emissions from on-road diesel engines from 1970 to 1998 and
nonroad diesel engines from 1990 to 1998.

Source: U.S. EPA, 2000a, National air pollutant emission trends, 1900-1998.
                                  E3 On -Road  H N on ro ad
                                      1

Figure 2-14. Trends in VOC emissions from on-road and nonroad diesel engines from 1970
to 1998.

Source:  U.S. EPA, 2000a, National air pollutant emission trends, 1900-1998.
                                        2-24

-------
Figure 2-15.  Trends in CO emissions from on-road and nonroad diesel engines from 1970
to 1998.

Source:  U.S.  EPA, 2000a, National Air Pollutant Emission Trends, 1900-1998.
2.2.4. Historical Trends in Diesel Fuel Use and Impact of Fuel Properties on Emissions
       Use of diesel fuel increased steadily in the second half of the 20th century.  According to
statistics from the Federal Highway Administration (1995, 1997), in 1949 diesel fuel was
approximately 1% of the total motor fuel used, and in 1995 it was about 18%. Over the same
time, diesel fuel consumption in the United States increased from about 400 million gallons to 26
billion gallons per year, an increase by a factor of more than 60 (Figures 2-16 and 2-17).
       The chemistry and properties of diesel fuel have a direct effect on emissions of regulated
pollutants from diesel engines.  Researchers have studied the NOx and DPM effect of sulfur
content, total aromatic content,  polyaromatic content, fuel density, oxygenate content, cetane
number, and T90 on emissions of regulated pollutants.  T90 is the 90% distillation point
temperature.  An increase in T90 has been observed to cause an increase in DPM emissions
(Cunningham et al., 1990; Sienicki et al.,  1990). Cetane number is a measure of the ignition
quality, or ignition delay time, of a diesel  fuel.  The percent of cetane (less commonly referred to
as hexadecane, C16H34) by volume in a blend with alpha-methylnaphthalene (C10H7CH3) defines
the cetane number that provides the same  ignition delay time as the fuel in use.
                                         2-25

-------
£ zu "
£
' — ' 1 8
3
,_ 1 f.
S lb
+J
0
*5 14
<< 14
13
o 1? -
o in
"53
fa »
13
cS A
Q
« 4 _
1 7
0 ^
53
PH H -







i i






















































          N?   N?
                                       (A
Figure 2-16. Percentage of total motor fuel use that is on-road diesel fuel since 1949.




Source: Federal Highway Administration, 1995.



    30,000,000
    25,000,000
T5  20,000,000

O
^  15,000,000






§  10,000,000

.c

h-

     5,000,000




           0
                 .••Illlllll
              CDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCDCD
              -i^.aiaiaiaiaiO5O5O5O5O5^j^j^j^j^JoooooooooocDCDCD
              oj-i.o3ai^jcD-i-wai^jcD-i-wai^jcD-i-wai^jcD-i-wai



                                           Year


Figure 2-17. On-highway diesel fuel consumption since 1943, values in thousands of


gallons.




Source: Federal Highway Administration, 1995.
                                           2-26

-------
       Before 1993, diesel fuel sulfur levels were not federally regulated in the United States,
although the State of California had such regulations. Industry practices that were in place (e.g.,
the ASTMD 975 specification for No. 2 oils) limited sulfur to 0.5%.  During the years 1960 to
1986, fuel sulfur content showed no chronological increasing or decreasing trends and ranged
from 0.23 to 0.28 wt% (NIPER, 1986). A maximum allowable on-road diesel fuel sulfur content
in the United States was established at 0.05 mass % in 1993, in advance of the  1994 0.10 g/bhp-
hr PM standard for HD on-highway trucks.  Nationally, on-road fuels averaged 0.032% sulfur in
1994 while nonroad fuels averaged 10-fold the sulfur level of on-road fuel, or 0.32% (Dickson
and Sturm, 1994).  The reduction in diesel fuel sulfur reduced total DPM mass emissions
through reduction of sulfate PM (primarily  present as sulfuric acid).
       Considerably higher sulfuric acid PM emissions are possible with DE aftertreatment
systems containing precious metals (oxidation catalysts, lean NOX catalysts, catalyzed DPM
traps). At temperatures over 350 °C to 500 °C (depending on device), SO2 in the exhaust can be
oxidized to sulfuric acid (McClure et al., 1992; McDonald et al., 1995; Wall, 1998).  Sulfur
content remains at unregulated levels for off-highway diesel fuels and fuels used in railroad
locomotives.
       The chemical makeup of diesel fuel has changed over time, in part because of new
regulations and in part because of technological developments in refinery processes.  EPA
currently regulates on-road diesel fuel and requires the cetane index (a surrogate for actual
measurements of cetane number) to be greater than or equal to 40, or the maximum aromatic
content to be 35% or less  (CFR 40:80.29).  EPA recently finalized a regulation that will limit the
sulfur content of on-road diesel fuel to 15 ppm starting in 2006 (U.S. EPA, 2000b). California
has placed additional restrictions on the aromatic content of diesel fuel (California Code of
Regulations, Title 13, Sections 2281-2282)  and requires a minimum cetane number of 50 and an
aromatics cap of 10% by volume, with some exceptions for small refiners and alternative
formulations as long as equivalent emissions are demonstrated. Diesel fuel from larger refiners
is limited to 10% aromatic content, and for  three small refiners (a small fraction of diesel sales)
to 20% aromatic content.  The refiners can also certify a fuel with higher aromatic content as
being emissions-equivalent to the 10% (or 20%) aromatic content fuels by performing a 7-day
engine dynamometer emissions test.  This method is chosen by most, if not all, California
refiners, and so a typical California diesel fuel has an aromatic content above 20%. Emissions
equivalence has been obtained through use  of cetane enhancers, oxygenates, and other
proprietary additives. Nonroad diesel fuel is not regulated, and consequently, cetane index,
aromatic content, and sulfur content vary widely with nominal values for cetane number around
43, 31% aromatics, and sulfur approximately 3,000 ppm.
                                          2-27

-------
       The average cetane number of U.S. diesel fuel declined steadily from 50.0 to 45.1, or
about 0.2% per year, from 1960 to 1986 (NIPER, 1986). The decline in cetane number was
likely accompanied by an increase in aromatic content and density (Lee et al., 1998). A number
of EPA-sponsored studies refer to fuels with nominally 22% aromatics content as "national
average fuel" during the 1970s (Hare, 1977; Springer, 1979), whereas by the 1980s a so-called
national average fuel contained 30% aromatics (Martin, 1981a,b).  Shelton (1979, 1977) has
reported a trend of increasing T90 from 1960 through the late 1970s, which is consistent with
increasing density, aromatic content,  and polyaromatic  content. Unfortunately, aromatic content
was not commonly measured before the 1980s.
       Studies measuring the emissions impact of changes in cetane number and aromatic
content for roughly 1990 model year engine technology find that increasing the aromatic content
from 20% to 30%, with an accompanying decrease in the cetane number from 50 to 44, results in
a 2% to 5% increase in NOX and a 5% to 10% increase in total DPM (McCarthy et al., 1992;
Ullman et al., 1990; Sienicki et al.,  1990; Graboski and McCormick, 1996). These ranges may
be reasonable upper bounds for the effect of changes in fuel quality on NOX and DPM emissions
during the years 1960-1990.
       In the northern United States during wintertime, on-road No. 2 diesel may contain some
percentage of No. 1 diesel to improve cold-flow properties. Discussions with refiners indicate
that a typical wintertime No. 1 diesel blending level is 15 volume %; however, this number must
be taken as  a rough estimate. Blending of No. 1 may lower the aromatic content, resulting in
improved emissions performance. Nationally, on-highway No. 1 fuels averaged 17% aromatic
content in 1994 (Dickson and Sturm, 1994). Thus, there may also be some small but perceptible
seasonal changes in emissions from diesel engines.
       Railroad-grade diesel fuel is currently unregulated. Typically, railroad-grade diesel fuel
is a blend of approximately 10% on-road fuel and 90%  nonroad diesel fuel.  There are no recent
data on the composition of railroad-grade diesel fuel. Somewhat dated diesel fuel oil surveys
(Shelton, 1979) reported that railroad-grade diesels had lower cetane number, higher density,
and higher T90. Also, the cetane index for these fuels can be as much as 9 cetane units higher
than the cetane number, an indication of a high aromatic content in railroad-grade diesels.
       Fuel chemistry is  also important for emission of particle-associated PAHs.  In studies
performed over more than a decade, Williams and Andrews of the University of Leeds have
shown that the solvent-extractable PAHs from diesel particulate originate almost entirely in the
fuel (Williams et al., 1987; Andrews et al., 1998; Hsiao-Hsuan et al., 2000).  The PAH
molecules are relatively refractory,  so a significant fraction survive the combustion process and
condense onto the DPM.  These studies have been confirmed by other research groups (Crebelli
et al., 1995; Tancell et al., 1995). There is a consensus among these researchers that

                                         2-28

-------
pyrosynthesis of PAHs occurs only at the highest temperature operating conditions in a diesel
engine.  Under these conditions, most of the DPM and other pyrolysis products are ultimately
burned before exiting the cylinder. These results indicate that emissions of PAHs are more a
function of the PAH content of the fuel than of engine technology.  For a given refinery and
crude oil, diesel fuel PAH correlates with total aromatic content and T90.  Representative data
on aromatic content for diesel fuels in the United States do not appear to be available before the
mid-1980s. However, the decreasing trend in cetane number, increasing trend in T90,  and the
increasing use of light cycle oil from catalytic cracking beginning in the late 1950s suggest that
diesel PAH content has increased over the past 40 years. Because PAHs have been implicated  as
one potential contributing component to the observed toxicity of DE, changes in PAH  content of
diesel fuel over time, as well as differences between diesel fuels used in different applications
(on-road, nonroad, locomotive), may influence the hazard observed in exposed populations from
different occupations.  However, such a relationship would be difficult to differentiate in an
epidemiologic study because there are several other properties of DE that may be contributing to
the observed toxicity.  Historical trends in PAH-measured emissions are discussed in Section
2.2.8.2.

2.2.5. Chronological Assessment of Emission  Factors
2.2.5.1.  On-Road Vehicles
      Numerous studies have been conducted on emissions from in-use on-road HD diesel
vehicles. HD vehicles are defined as having a rated gross vehicle weight (GVWR) of greater
than 8,500 Ib, and most over-the-road trucks have a GVWR of 80,000 Ib. Emissions of regulated
pollutants from these studies have been reviewed (Yanowitz et al., 2000); the review findings,
which encompass vehicles from model years 1976 to 1998, are summarized below. In addition,
a large amount of engine dynamometer data on HD diesel engines have been published since the
mid-1970s. These data are used below to confirm and expand upon the findings from in-use
vehicle testing.
      Figure 2-18 shows chassis dynamometer data for more than 200 different vehicles
(approximately one-half of which are transit buses), reported in 20 different published studies, as
well as a large amount of additional  data collected by West Virginia University (Yanowitz et al.,
1999; Warner-Selph and Dietzmann, 1984; Dietzmann et al., 1980;  Graboski  et al., 1998a,b;
McCormick et al., 1999; Clark et al., 1995, 1997; Bata et al., 1992;  Brown and Rideout, 1996,
Brown et al., 1997; Dunlap et al.,  1993; Ferguson et al., 1992; Gautam et al.,  1992; Katragadda
                                          2-29

-------
                   10
to
OJ
o
               1
               8
               g
               '&
               LU
                  0.1 -
                                          100
                                       _


                                       1
                                       co
                                       co
                                       C/)
                                           10-
  1975   1980  1985  1990   1995  2000       1975   1980  1985  1990   1995  2000

              Model Year                                Model Year

                                          100-r

               g
               'co
               CO
               O
10-1
 1  -
                  0.1 -
                                                             10-
                                                          CO


                                                         LU
                                                              1 -
                    1975  1980   1985  1990   1995   2000

                                 Model Year
                                            1975   1980  1985  1990   1995  2000

                                                        Model Year
           Figure 2-18. Model year trends in PM, NOX, HC, and CO emissions from HD diesel vehicles (g/mile).


           Source: Yanowitz et al., 2000.

-------
et al., 1993; Rideout et al., 1994; Wang et al., 1993, 1994; Williams et al., 1989; Whitfield and
Harris, 1998; West Virginia University data available on the World Wide Web at
www.afdc.nrel.gov).  The results from vehicles tested more than once using the same test cycle,
and without any additional mileage accumulated between tests, are averaged and reported as one
data point. Buses were tested using the Central Business District (CBD) cycle, while most
trucks were tested using the Urban Dynamometer Driving Schedule (UDDS), also known as the
Schedule Id cycle. Some of the trucks were tested using the West Virginia 5-peak cycle, which
generates considerably lower g/mi emissions than the CBD or UDDS (Yanowitz et al., 1999).
Emissions results from vehicles tested under different test cycles or at different points in the
engine's life cycle have been reported as separate data points. Note that all NOX mass emissions
data are reported as equivalent NO2.  Table 2-7 compares the make-up of the fleet of trucks that
was tested with the in-use truck fleet according to the 1997 Vehicle Inventory and Use Survey
(U.S.  Bureau of the Census, 1999a).  The tested fleet is mostly vehicles in the 33,000-60,000 Ib
range. Analysis of the tested  fleet also shows that the model year distribution is skewed toward
newer vehicles. The 1997 Vehicle Inventory and Use Survey indicates a flat distribution with
roughly the same number of in-use vehicles for each of the model years in the decade preceding
1997. The 1992 Truck Inventory and Use Survey (U.S.  Bureau  of the Census, 1995) shows the
same trend, as  shown in Figure 2-1.  Analysis of odometer mileage for the tested fleet shows that
45% of the vehicles had less than 50,000 miles at the time of testing.  Only 10% of the vehicles
had more than  250,000 miles.  Although the mileage distribution of the in-use fleet is unknown,
it seems unlikely to be as heavily weighted to low-mileage vehicles. Because of the relatively
low mileage of most of the vehicles tested, deterioration of emissions may not be reflected in the
                       Table 2-7. Comparison of in-use truck fleet
                            with truck fleet tested on chassis
                         dynamometer, percent of total vehicles
Class
O
4&5
6&7
8A
8B
In-use trucks,
1995 census
17.7
13.3
25.0
20.9
23.1
Tested
trucks
1
0
17
52
30
                                         2-31

-------
results.  Yanowitz and co-workers (2000) report that average emissions of regulated pollutants
for vehicles of the different classes listed in Table 2-7 are approximately the same.  This is
clearly a reflection of the small number of vehicles in the lighter weight classes for this dataset,
but it also indicates no real difference in emissions for vehicles in Classes 6-8. The data are
mainly for vehicles of 19,500 Ib and greater GVWR (Classes 6 and 7 and heavier),  and
predominantly for vehicles of 33,000 Ib and greater GVWR (Class 8 trucks and buses).
       Figure 2-18 shows emissions trends in g/mi. Least-squares linear regressions and 95%
confidence intervals are plotted on each graph and yield the following equations for predicting
emissions trends (applicable to the years 1976-98):

Log NOX (g/mile) = (Model year * -0.008) + 16.519  R2 = 0.024                        (2-1)
Log PM (g/mile) = (Model year * -0.044) + 88.183  R2= 0.28                         (2-2)
Log HC (g/mile) = (Model year * -0.055) + 109.39  R2= 0.27                         (2-3)
Log CO (g/mile) = (Model Year * -0.041) + 82.876  R2 = 0.22                         (2-4)

       As shown in Figure 2-18, changes in NOX emissions have been relatively small, with an
emission rate averaging about 26 g/mi. The data reported in Figure 2-18 are real-world, in-use
emissions measurements and therefore more accurately reflect emission factors than engine test
data during this period. There are two potential causes for the relative constancy of NOX
emissions as described by Figure 2-18. The first is emissions deterioration due to engine wear.
Weaver and Klausmeier (1988) have shown that diesel engine deterioration results  in lower NOX
emissions and higher DPM emissions, and this finding has recently been confirmed by
McCormick and co-workers (2000). Wear of mechanical devices that limit smoke, fuel pumps,
and fuel injectors alters the effective injection timing to decrease NOX.  Because deterioration is
more a function of maintenance than vehicle age or mileage, deterioration introduces a wide
range in NOX emission factors measured in the chassis dynamometer studies.  The lack of a
decreasing trend in NOX emissions can also be attributed to the use of illegal emissions control
devices that bypassed the trucks' emission control systems under some driving conditions such
as steady-state cruise. EPA has reached a settlement with the diesel engine manufacturers to
discontinue use of these devices.  The illegal devices produced low NOX emissions  on the
transient test (FID FTP) but operated in a high-NOx/high-fuel-economy mode in use under
highway cruise conditions.
       Figure 2-19 shows engine certification data for NOX emissions reported in the many
studies that have employed the transient test over the past 25 years. The engine testing data are
also listed in Table 2-8. The data compiled in Figure 2-19 show a significant decline in NOX
                                          2-32

-------
            .a
            o>
            in
            (fl
            LLI
               14
               12 -
               10 -
                6 -
                4 -
!"
                2
                1975
 1980      1985      1990
          Engine Model Year
1995
2000
Figure 2-19.  Diesel engine certification data for NOX emissions as a function of model year.

Source: Data are from the transient test results provided in Table 2-8.

emissions, and all engines would appear to meet the regulatory standards for their year of
manufacture because of the illegal emissions devices. From 1980 to 1997, the EPA emissions
trends report (U.S. EPA, 1998a) predicted a decline in NOX emissions from HD diesel vehicles
because these data are based on engine test data. The emissions trend includes the growth in
vehicle miles traveled over time as well as changes in emission factors. The more recent trends
inventory (U.S. EPA, 2000a, discussed earlier) includes emission from the illegal emissions
devices and accordingly demonstrates a slight increase in NOX emissions from on-road HD diesel
vehicles from 1990 to 1998.
       DPM, CO, and THC emissions, although widely variable within any model year, have
shown a pronounced declining trend (Figure 2-18).  DPM emissions from chassis dynamometer
tests decreased from an average of 3-4 g/mi in 1977 to an average of about 0.5 g/mi in 1997,
suggesting a decrease in DPM emissions of a factor of about 6. Note that these data are for
vehicles or engines tested on in-use or industry-average fuel at the time they were tested.
Indications are that the observed decline in DPM is caused primarily by changes in engine
                                         2-33

-------
               Table 2-8. Diesel engine emissions data from engine dynamometer tests
Reference


Hare, 1977

Springer, 1979







Perez, 1980


Martin, 198 la

















Martin, 1981b


Ullmanetal., 1984
Martin, 1984
Barry etal., 1985

Engine3


Cat 3208 (NA)
DDC 6V71 (blower)
Mack ETAY(B)673 A (DI,
TC,AC)
Cat 3208 (EGR, NA)
Cat 3406 (DI, TC, AC)
Cat 3406 (DI, TC, AC, EGR)
Cat 3406 (IDI, TC, AC)
DB OM-352A (DI, TC, AC)
DB OM-352A (DI, NA)
Cat (DI, NA)
Cat (DI, EGR)
Cat (DI, TC, AC)
Cat 3208
Cummins NTC350
DDC 6V92T (2S)
Cummins NTCC3 50
DDC8V71N(2S)
DDC 6V92TA (2S)
IH DTI466B
Mack ETAY(B)673A
MackETSX676-01
Cummins VTB-903
Cat 3406
Cat 3406PCTA
Cummins BigCam NTC350
IH DT466
DDC 6V92TA (2S)
DDC 8V71TA (2S)
Cummins NTC290
Cummins NH-250
Cummins VTB-903
DDC 8V71TA (2S)
IH DTI466B
DDAD 6V-71 (2S)
Cummins NTC300
Cat 3406B
DDC 8V-92 TA (2S)
Year


1976
1976
1977

1977
1977
1977
1977
1977
1977
1978
1978
1978
1978
1976
1978
1979
1978
1979
1979
1979
1980
1979
1979
1979
1979
1979
1979
1979
1979
1979
1980
1980
1980
1980
1981
1985
1980
Test"


SS
ss
SS

ss
ss
ss
ss
ss
ss
ss
ss
ss
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
NOX
g/bhp-
hr
7.98
10.24
6.613

3.747
9.79
5.49
5.14
8.93
7.46
8.12
5.16
7.66
7.83
11.41
9.55
6.58
7.15
7.80
7.46
9.01
6.90
8.10
11.28
7.24
9.97
7.91
11.66
9.81
11.10
10.87
5.59
7.91
4.41
6.09
8.13
6.58
8.15
PM
g/bhp-
hr
0.871
1.92
0.61

2.21
0.35
0.93
0.28
0.56
0.99
0.77
1.21
0.33
1.06
0.81
0.72
0.52
0.92
0.65
0.48
0.77
0.85
0.53
0.69
0.49
0.54
0.71
0.73
0.51
0.78
0.97
0.67
0.44
0.62
0.56
0.45
0.48
0.45
CO
g/bhp-
hr
4.04
6.55
1.588

6.200
2.34
4.81
1.26


5.92
5.37
2.20


















2.0
2.28
2.35
3.86
2.70
2.1
2.61
THC
g/bhp-
hr
1.11
0.71
0.476

1.163
0.35
0.17
0.12


0.77
0.57
0.27


















2.23
0.73
0.87
1.42
1.36
0.5
0.53
SOF
g/bhp-
hr
0.103
0.937
0.098


0.063
0.181
0.031
0.190
0.287
0.19
0.079
0.037


















0.228
0.176
0.186
0.298
—
0.061
~
SOF
Methc

c-hexane
c-hexane
Benz/cyc

Benz/cyc
Benz/cyc

Benz/cyc
Benz/cyc
Benz/cyc
DCM
DCM
DCM


















DCM
DCM
DCM
DCM

DCM

Total B[a]P
(PAH) 1-NP (NPAH)
aldehyde, ug/bhp-hrd ug/bhp-hre
mg/bhp-hr
0
0
65 2

161 1
73 0
80 0
80 0
280 0
280 1
1
4
0





















23

70


.76
.24
.23

.72
.15
.08
.11
.87
.07
.08
.34
.34





















—

1

to to

-------
Table 2-8. Diesel engine emissions data from engine dynamometer tests (continued)
Reference






Engaetal., 1985
Baines, 1986
Wachter, 1990
McCarthy et al
., 1992
Perez and Williams,
1989




Needham et al

Kreso et al.,


Bagley etal.,





, 1989

1998


1998
Graboski, 1998b
(and references therein)















Screen etal..
Norbeck et al.,

Sienicki et al.















1995
1998b

1990
Engine3



DDC 8V-71 TAC (2S)
Cummins NTCC-400
Iveco 8460
Navistar DTA466 ES210
Engine 1
Engine 2
Engine 3
Engine 4
Engine 5
Engine 6
Average of 16 engines
Average of 3 engines
Cummins L10-300
Cummins L10-3 10
Cummins Ml 1-330E
Cat 3304 (IDI, NA) non-road
DDC 6V-71N-77 (MUI, 2S)
DDC 6V-92TA-91 (DDECII)
DDC-6V-92TA-87 (2S)
DDC-6V92TA-83 (MUI, 2S)
DDC 6V-92TA -88 (DDECII, 2S)
DDC 6V-92TA-91 (DDECII, 2S)
DDC 6V-71N-77 (MUI, 2S)
DDC 6V-92TA-81/89 (MUI, 2S)
DDC 6V-92TA-91 (DDECII, 2S)
DDC 6V-92TA-89 (DDECII, 2S)
DDC Series 60-91 DDECII
Cummins L-10-87 (MUI)
DDC Series 60-91 (DDECII)
Cummins N-14-87 (MUI)
DDC Series 60-89 (DDECII)
DDC Series 60-91 (DDECII)
Cummins B 5. 9
Navistar DTA466
Cummins L10
DDC Series 60
Navistar DTA466
Year



1984
1985
1991
1993
1982
1982
1982
1982
1982
1982
1988
1991
1988
1991
1995
1983
1977
1991
1987
1983
1988
1991
1977
1981
1991
1989
1991
1987
1991
1987
1989
1991
1995
1994
1991
1994
1991
Test"


SS
T
T
T
T
T
T
T
T
T
T
T
T
SS
SS
SS
SS
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
NOX
g/bhp-
hr
6.64
5.85
4.62

4.93








5.15
4.70
3.82

9.96
4.23
10.77
5.62
8.52
4.4
11.72
10.06
4.84
4.855
4.635
5.64
4.68
6.32
5.128
4.303
4.37
4.779
4.77
4.89
5.25
PM
g/bhp-
hr
0.36
1.26
0.55
0.22
0.082
0.93
0.86
0.59
0.96
1.06
0.88
0.37
0.24
0.103
0.035
0.037
0.56
0.83
0.197
0.59
0.265
0.2
0.276
0.282
0.268
0.227
0.338
0.300
0.309
0.220
0.369
0.252
0.182
0.106
0.090
0.224
0.112
0.22
CO
g/bhp-
hr
1.83
2.99
3.21

1.3












3.59
1.51
0.71
1.19
1.6
1.65
3.18
2.16
1.51
2.499
4.458
2.33
2.26
2.20
4.008
2.004
1.47
0.989
2.26
1.402
—
THC
g/bhp-
hr
0.38
1.48
0.53

0.28








0.26
0.067
0.16

2.01
0.72
-
0.435
0.6
0.42
0.86
0.42
0.44
0.526
0.164
0.89
0.08
0.58
0.154
0.392
0.30
0.181
0.53
0.065
0.23
SOF
g/bhp-
hr
0.0255
~

0.0957
0.0237
0.179
0.145
0.185
0.325
0.076
0.344
0.12
0.10
0.030
0.022
0.016
0.319
0.729
0.0788
~
0.133
0.116
0.07
0.212
0.144
~
~
~
~
0.066
0.100
~
0.061
0.05
0.035
~
0.043
0.05
SOF
Methc




7
SFE
DCM
DCM
DCM
DCM
DCM
DCM
DCM
DCM
DCM
DCM
DCM
Benz/cyc
DCM
?

DCM
Tol/EtOH
Tol/EtOH
DCM
DCM




DCM
?

Tol/EtOH
DCM
DCM

DCM
DCM
Total B[a]P (PAH) 1-NP (NPAH)
aldehyde, ug/bhp-hrd ug/bhp-hre
ms/bhp-hr

..



26 0.83
5.8
4.9 0.89
26 1.2
5.3






1.5(133) 2.2







..
..
..




~
—
0.24((18.5)
26
80 20(1725) 1.95(4.92)
17


-------
        Table 2-8. Diesel engine emissions data from engine dynamometer tests (continued)
Reference


Ullmanetal., 1990
Kadoetal, 1998
Ullman, 1988
Mitchell et al., 1994

Tanakaetal., 1998
Rantanen et al., 1993



Engine3


DDC Series 60
Cat 3406E
Cummins NTCC400
DDC Series 60
Navistar DTA466
Unknown
Scania
Valmet
Volvo
Volvo
Year


1991
1997
1988
1994
1994
1994
1990
1990
1990
1995
Test"


T
T
T
T
T
SS
ss
SS
ss
ss
N0y
g/bhp-
hr
4.552

4.47
4.43
4.86
4.934
9.30
8.67
9.87
4.56
PM
g/bhp-
hr
0.188

0.42
0.111
0.099
0.143
0.157
0.157
0.262
0.135
CO
g/bhp-
hr
2.102

2.22
2.17
1.10
0.807




THC
g/bhp-
hr
0.508

0.53
0.22
0.34
0.352




SOF
g/bhp-
hr
—


0.021
0.046
0.036
0.031



SOF
Methc


DCM

DCM
DCM
DCM
DCM



Total
aldehyde,
ms/bhp-hr



34
56





B[a]P (PAH)
ug/bhp-hr"


0.07(30)

(141)
0.11(242)
.076




1-NP (NPAH)
ug/bhp-hre


0.34

0.04(0.12)
0.3(0.6)





aNA=naturally aspirated.  TC=turbocharged (engines not designated as NA or TC are turbocharged).  AC=aftercooled. DI=direct injection.  IDI=indirect
injection. EGR=exhaust gas recirculation.  2S=two-stroke (engines not designated as 2S are four-stroke). MUI=mechanical unit injector (not electronically
controlled). DDEC=Detroit Diesel Corporation's engine control module (electronic control).
bSS=various single or multimode steady-state tests. T=heavy-duty FTP (transient test).
°SOF extraction method.  SFE=Supercritical fluid extraction.  All others by Soxhlet extraction using the indicated solvents (? for unreported).
DCM=dichloromethane.  Tol/EtOH=toluene/ethanol mixture. Benz/cyc=benzene/cyclohexane mixture.  C-hexane=cyclohexane.
dNumber in parentheses is the total PAH emission obtained by summing emissions of all PAHs reported.
eNumber in parentheses is the total NPAH emission obtained by summing emissions of all NPAHs reported.

-------
technology that often result from emission standards, as well as by the lowering of on-road diesel
fuel sulfur content in 1993.
       As the discussion above indicates, there is a reasonable amount of data upon which to
base emission factor estimates for late 1970s and later HD vehicles.  However, very little
transient test data are available on engines earlier than the mid-1970s.  The limited data available
from six pre-1976 vehicles tested using the transient cycle suggests that PM emission rates
ranged from 1.6 g/mi to 9.0 g/mi, which is a substantially greater range than in post-1976
engines (Fritz et al., 2001).
       Although a substantial decreasing trend in DPM emissions from in-use chassis
dynomometer testing and engine testing (Figure 2-20) is evident, these data reflect a wide range
in emission factors within  any given model year. For example, emission factors for model year
1996 range from less than  0.1 g/mi to more than 1 g/mi (Yanowitz et al., 2000;  Graboski et al.,
1998b). The high variability in DPM emissions measured in the chassis dynamometer tests is
observed because of several factors,  including differences in measurement methods and test
conditions at the various testing facilities, deterioration, and engine-to-engine variation.
Although there  can be excellent agreement between chassis dynamometer testing facilities
(Graboski et al., 1998a), there is no standard HD chassis dynamometer Federal  test procedure,
and no detailed procedures for such testing are described in any authoritative source such as the
Code of Federal Regulations, which  does contain such procedures for engine dynamometer
                 1.£
                 1.25 -

              £  1.00 -
              .Q

              I  0.75 -
              in

              HJ  0.50-
              Q.

                 0.25
                 0.00
i,
      I
                   1975
 1980
1995
2000
                                    1985      1990
                                   Engine Model Year
Figure 2-20. Diesel engine certification data for PM emissions as a function
of model year.

Source:  Data are from the transient test results provided in Table 2-8.

                                       2-37

-------
testing used for EPA emission regulations.  Therefore, each facility has developed its own
approach to HD testing.  Clark et al. (1999) report that the test cycle can have a substantial
effect on DPM emissions, with higher DPM emissions reported from test cycles that incorporate
full-power accelerations.  Test cycles incorporating full-power accelerations reflect urban HD
vehicle driving for several types of vehicles (garbage trucks, buses) operating in urban areas.
Clark et al. (1999) also report that aggressive acceleration produces higher DPM emission rates
than does conservative acceleration, and Clark and co-workers suggest that real in-use driving is
more likely to mimic aggressive acceleration. Although figures are currently unquantified, it is
generally believed that the majority of DPM is generated under transient conditions such as
heavy acceleration.
       Weaver and Klausmeier (1988) have examined potential causes and frequency of DPM
emissions deterioration for in-use HD diesel vehicles. Potential causes include manufacturing
defects and malfunctions such as retarded timing, fuel injector malfunction, smoke-limiting
mechanism problems,  clogged air filter, wrong or worn turbocharger, clogged intercooler, engine
mechanical failure, excess oil consumption, and electronics that have been tampered with or
have failed.  The recent report by McCormick and co-workers (2000) indicates that many of
these malfunctions can have very large effects on DPM emissions, resulting in DPM increases of
typically 50% to 100%. Although Yanowitz and co-workers (1999) found that DPM emissions
were positively correlated with odometer mileage for a fleet of 21 vehicles, it is more likely that
the vehicle state of maintenance will be more important than mileage for determining the degree
of emissions deterioration. In fact, in a similar analysis performed on the chassis dynamometer
results included in the review of Yanowitz et al. (2000), DPM emissions could not be correlated
with odometer mileage. Differences in testing methods between various facilities as well as
varying states of maintenance for vehicles of the same mileage and model year probably account
for this lack of correlation.
       It is difficult, given current information, to quantitatively assess the contribution of high-
emitting or smoking diesel vehicles to ambient DPM. Emission models used to prepare diesel
particulate emission inventories do not account for deterioration.  The relative contribution of
high-emitting diesel vehicles to the total mass and overall chemical composition of diesel
particulates is being quantified.  Some studies report numerous smoking diesel trucks. A study
of the smoke  opacity-based inspection and maintenance program in California found failure rates
of 20% and higher, suggesting that high-emitting vehicles are  not uncommon (CARB/EEAI,
1997). In the Northeast, smoke opacity testing conducted on 781  HD trucks found that 15% of
the vehicles failed the  smoke standard (40% opacity for 1991 and newer HD diesel vehicles and
50% opacity for pre-1991 HD diesel vehicles) (Cooper, 1999). Although the correlation
between smoke and particulate emissions tends to be qualitative or semi quantitative (discussed

                                          2-38

-------
below), there is a good correlation between opacity and EC concentrations, and it is expected
that high-emitting diesel vehicles may be an important part of the DPM emission inventory.
       Others have attempted to determine if the effects of deterioration could be detected for
in-use vehicles. In a study of 21 vehicles (Yanowitz et al., 1999), a linear multivariate regression
analysis found that DPM emissions were positively correlated with odometer mileage (several
other correlation factors were also identified, including model year). A similar analysis
performed on the chassis dynamometer results included in the review of Yanowitz et al. (2000)
found that DPM emissions could not be correlated with odometer mileage, probably because of
differences in testing methods between the various facilities.
       Other approaches for measuring emissions from in-use on-road diesel vehicles include
tunnel tests and remote sensing, the latter of which measures gaseous, but not DPM, emissions.
The literature reports of those studies are summarized in Tables 2-9 and 2-10.  Several tunnel
test studies have reported DPM emission factors (Pierson and Brachaczek, 1976; Japar et al.,
1984; Pierson et al., 1983; Kirchstetter et al., 1999; Gertler et al., 1996, 1999).
       The method for determining emission rates for vehicles traveling through a tunnel is
explained in detail by Pierson et al. (1996). Briefly, the emissions of a species are determined by
measuring the concentration of a pollutant entering and leaving a tunnel along with knowledge
of the cross-section of the tunnel and measurements of the wind flux at the inlet and outlet of the
tunnel. The emission rate is calculated by dividing the mass  of the pollutant by the number of
vehicles that passed through the tunnel and the length of the tunnel.  The diesel and gasoline
vehicle contributions to the total emission of the pollutant are separated by a simple regression
analysis where the intercepts (100% HD and 100% LD) are the diesel and gasoline emission
rates, respectively.
       Emission factors from tunnel studies provide a snapshot of real-world emissions under
driving conditions experienced in the tunnel and reflect emission factors representative of the
mix of in-use vehicles and the atmospheric dilution and short-term transformation processes of
DE.  Emission factors  derived from tunnel studies are often used as one source of information to
study the impact of improved technology and fleet turnover on emissions because they allow
random sampling of large numbers of vehicles, including a range of ages and maintenance
conditions. However,  tunnel studies are limited in that they represent driving conditions on a
single roadway passing through a tunnel and represent mostly steady-state driving conditions,
whereas most DPM is  generated during transient modes of operation; also, tunnel studies do not
include cold-start operations. Both of these factors need to be assessed to understand emission
rates for DPM to which people are exposed (U.S.  EPA, 1992, 1995). DPM emission factors
from in-use fleets derived from tunnel studies in the 1970s and 1980s compared with the 1990s
                                          2-39

-------
               Table 2-9. HD diesel emissions results from tunnel tests (adapted from Yanowitz et al., 1999)
Test
Pierson and
Brachaczeck, 1983







Rogaketal., 1998
Miguel etal., 1998
Weingartneretal.,
1997b
Pierson et al., 1996
Pierson et al., 1996
Pierson et al., 1996
Kirchstetter et al.,
1999
Gertler,1999
Tunnel location,
year of study
Allegheny, 1974
Allegheny, 1975
Allegheny, 1976
Allegheny, 1976
Tuscarora, 1976
Tuscarora, 1976
Allegheny, 1977
Allegheny, 1979
Allegheny, 1979
Cassiar Tunnel,
1995, Vancouver
Caldecott Tunnel,
1996, San Francisco
Gubrist Tunnel,
1993, Zurich
Fort McHenry
Tunnel, downhill,
1992, Baltimore
Fort McHenry
Tunnel, uphill,
1992, Baltimore
Tuscarora Tunnel
1992, Pennsylvania
Caldecott Tunnel,
1997, San Francisco
Tuscarora Tunnel,
1999, Pennsylvania
Fuel
efficiency
(mi/gal)
5.42"








8.03"
5.42C
5.60e
11.46b
5.42b
6.44b
5.42C

NOxa
(g/mi)









19.50
±4.22
23.82±
4.17

9.66
±0.32
22.50
±1.00
19.46
±0.85
23.82±
2.98

NMHC
(g/mi)









-0.16
±0.88


0.92
±0.21
2.55
±1.05
0.68
±0.20


CO
(g/mi)









6.79
±11.78


6.8
±1.5
14.3
±5.5
6.03
±1.61


DPM
(g/mi)
.90-1.80
1.75±0.19
1.5±0.10
1.4 ±0.07
1.3±0.19
1.39 ±26
1.3 ±0.08
1.2 ±0.03
1.4 ±0.04

1.67
± 0.24d
0.62
± 0.02f



1.43
±0.12g
0.29
CO2
(g/mi)









1,280
±40


897
±48
1,897
±168
1,596
±78


NOxa
(g/gal)









157
±34
129
±23

111
±4
122
±5
125
±5
129
±16

NMHC
(g/gal)









-1±7


11
±2
14
±6
4
±1


CO
(g/gal)









55
±95


78
± 17
78
±30
39
± 10


DPM
(g/gal)
4.9-9.8
9.49±1.03
8.1 ±0.54
7.6 ±0.4
7.0 ± 1.0
7.5 ± 1.40
7.0 ± 0.43
6.5 ±0.16
7.6 ±0.19

9.0
±1.3d
3.5
±0.1f



7.7
±0.6g

to
-k
o
       aNOx reported as NO2.
       bCalculated from observed CO2 emissions assuming fuel density 7.1 Ib/gal
       and C is 87% of diesel fuel by weight.
       °Since CO2 emissions not available, fuel efficiency assumed to be the same
       as in slightly uphill tunnel (Fort McHenry).
       dReported as black carbon, assumed that 50% of total PM emissions are
       BC.
eSlope of tunnel unknown, so used average fuel efficiency for the United
States.
fPM3.
8PM25.
hUncertainty reported as ±1.0 standard deviation, except where literature
report did not specify standard deviation; in those cases uncertainty listed as
reported.

-------
                      Table 2-10. Remote sensing results for HD vehicles

NOX


CO


THC

Reference
Jimenez et al., 1998
Cohen etal., 1997
Countess et al., 1999
Bishop etal., 1996
Cohen etal., 1997
Countess et al., 1999
Bishop etal., 1996
Cohen etal., 1997
Year study
conducted
1997
1997
1998
1992
1997
1998
1992
1997
Emissions (g/gal)
150a,b,c
108 a'b'c
187a,b,c
59b
54b
85 b
0.002 HC/CO, mole ratio d
0.00073 HC/CO, mole ratio d
       aRemote sensing measures NO.  The reported value was corrected to a NOX (as NO2) value by
       assuming 90% (mole fraction) of NOX is NO.
       Emissions in g/gal calculated by assuming that fuel density is 7.1 Ib/gal and C is 87% by weight
       of fuel.
       °No humidity correction factor is included.
       dln order to calculate emissions in g/gal, an average molecular weight is needed.
       Source: Yanowitz et al., 1999.

suggest approximately a fivefold decrease in DPM mass emission factors over that time, with the
most recent data from 1999 reporting an emission factor of 0.29 g/mi for the on-highway FID
diesel fleet (Figure 2-21).
       Emission factors vary substantially for the various tunnels, with NOX emissions ranging
from 9.7 to 23.8 g/mi in the 1990s, CO emissions ranging from 6 to 14 g/mi, and THC emissions
ranging from 0.16 to 2.55 g/mi.
       Remote sensing reports emission factors in terms of pollutant emissions per unit of fuel,
not on a per-mile basis.  Agreement between remote sensing and tunnel studies for NOX
emissions is reasonably good for the fleet as a whole, suggesting an average level for the fleet of
about 130 g/gal, comparable to the average emissions factor measured in chassis dynamometer
studies (remote sensing can measure emissions from an individual vehicle, whereas tunnel
studies measure emissions from the fleet as a whole).  Generally, chassis dynamometer tests and
engine dynamometer test results are corrected for ambient humidity, in accordance with the
Federal Test Procedure (CFR 40, Subpart N).  Tunnel tests and remote sensing tests have
typically not included corrections for humidity. Appropriate humidity corrections for NOX and
DPM can be greater than 20% and 10%, respectively (or a total  difference of more than 45% and
20%, respectively, between low- and high-humidity areas), under normally  occurring climatic
conditions. Additionally, the remote sensing literature has not addressed how to determine the
                                          2-41

-------

1.70 -
1.50 -
1 30
— 110
"3)
0 90 -
0 70
0 50 -
0 30 -
0 10





4 Allegheny/Tuscarora - PA
• Fort McHenry - MD
^ ACaldecott - CA
• ^

4

A
" u


1970 1975 1980 1985 1990 1995 2000
Year of Measurement
  Figure 2-21. Emission factors from HD diesel vehicles from tunnel studies.

  Source:  Data from Pierson and Brachaczek, 1976; Japar et al., 1984; Pierson et al., 1996;
  Kirchstetter et al., 1999; Gertler et al., 1995, 1996; Gertler, 1999.

correct value for the NO/NOX ratio, and there is reason to believe that this value may differ
systematically from site to site, although almost all of the NOX is NO as it leaves the vehicle.
       In addition to the humidity correction discussed above, several factors must be taken into
account when comparing DPM measurements from tunnel tests to chassis dynamometer
measurements (Yanowitz et al., 2000):  (1) Chassis testing measures only tailpipe emissions;
tunnel tests can include emissions from other sources (tire wear, etc.), and (2) tunnel tests
typically measure emissions under steady-speed freeway conditions, whereas most chassis
dynamometer tests are measured on cycles that are more representative of stop-and-go urban
driving conditions. This latter limitation also applies to remote sensing readings, which measure
instantaneous emissions versus emissions  over a representative driving cycle.
       Because THC emissions for diesel vehicles are very low in total mass in comparison with
gasoline vehicles, tunnel test results for THC have a high degree of uncertainty.  A regression
analysis to determine the contribution of the limited number of HD vehicles to THC emissions is
unstable; small errors in the total measurements can change estimates substantially.  Similarly,
                                          2-42

-------
CO emissions are comparable to automobile emissions on a per-vehicle-mile basis, but because
there are generally many more automobiles than HD diesels in tunnel tests, CO measurements
from diesels may also have a high degree of uncertainty.

2.2.5.2. Locomotives
       Locomotive engines generally range from 1,000 horsepower up to 6,000 horsepower.
Similar to the much smaller truck diesel engines, the primary pollutants of concern are NOX,
DPM, CO, and HC. Unlike truck engines, most locomotive engines are not mechanically
coupled to the drive wheels. Because of this decoupling, locomotive engines operate in specific
steady-state modes rather than the continuous transient operation normal for trucks. Because the
locomotive engines operate only at certain speeds and torques, the measurement of emissions is
considerably more straightforward for locomotive engines than for truck engines.  Emissions
measurements made during the relatively brief transition periods from one throttle position to
another indicate that transient effects are very short and thus could be neglected for the purposes
of overall emissions estimates.
       Emissions measurements are made at the various possible operating modes with the
engine in the locomotive, and then weighting factors for typical time of operation at each throttle
position are applied to estimate total emissions under one or more reasonable operating
scenarios.  In the studies included in this analysis, two scenarios were considered:  line-haul
(movement between cities or other widely separated points) and switching (the process of
assembling and disassembling trains in a switchyard).
       The Southwest Research Institute made  emissions measurements for three different
engines in locomotives in 1972 (Hare and Springer, 1972) and five more engines in locomotives
using both low- and high-sulfur fuel in 1995 (Fritz, 1995).  Two engine manufacturers (the
Electro-Motive Division of GM, and GE Transportation Systems) tested eight different engine
models and reported the results to EPA (U.S. EPA, 1998b).  All available data on locomotives
are  summarized in the regulatory impact assessment and shown in Figure 2-22.

2.2.6. Engine Technology Description and Chronology
       NOX emissions, DPM emissions, and brake-specific fuel consumption (BSFC) are among
the  parameters that are typically considered during the development of a diesel engine. Many
engine variables that decrease NOX can also increase DPM and BSFC.  One manifestation of the
interplay among NOX, DPM, and BSFC is that an increase in combustion temperatures will tend
to increase NO formation. Higher temperatures will also often improve thermal efficiency, can
improve BSFC, and can increase the rate of DPM oxidation, thus lowering DPM emissions.  One
example of this is the tradeoff of DPM emissions and BSFC versus NOX emissions with fuel
                                         2-43

-------
     20
      15
  O  10
                 Line-Haul Cycle Emissions Data
                           NOx and PM (g/bhp-hr)
                                    o
                                          8
                                 o
                             o
                                              o o
                                                      o
                   0.1
0.2
0.3
0.4
                                 PM
0.5
40
30
^ 20
10
0
0
Switch Cycle Emissions Data
NOx and PM (g/bhp-hr)
a

0 °
- a
Oa ao D
0.2 0.4 0.6 0.8




PM
Figure 2-22. Line-haul and switch emissions data.

Source:  U.S. EPA, 1998a.
                                2-44

-------
injection timing. Many recent advances in reducing the emissions of diesel engines without
aftertreatment are combinations of technologies that provide incremental improvements in the
tradeoffs among these emissions and fuel consumption. The sum total, however, can be
considerable reductions in regulated emissions within acceptable levels of fuel consumption.
       The majority of current HD diesel truck engines certified for use in the United States
utilize:

            A four-stroke cycle
            Direct-injection, high-pressure (1,200 bar to >2,000 bar) fuel injection systems with
            electronic control of injection timing and, in some cases, injection rate
       •     Centrally located multihole injection nozzles
       •     Three or four valves per cylinder
            Turbochargers
       •     In many cases, air-to-air aftercooling
       •     In some cases, the use of an oxidation catalyst.

       These features have phased into use with HD truck engines because they offer a
relatively good combination of fuel consumption, torque-rise, emissions, durability, and the
ability to better "tune" the engines for specific types of applications. Fuel consumption, torque-
rise, and  drivability have been maintained or improved while emissions regulations have become
more stringent.  Many Class 8a and 8b diesel truck engines are now capable of 700,000 to
1,000,000 miles of driving before their first rebuild and can be rebuilt several times because of
their heavy construction and the use of removable cylinder liners.  These engines are expected to
last longer and therefore have a useful life longer than the regulatory estimate of full useful life
for HD engines (-1,000,000 miles) previously used by EPA (for 1980 engines that were driven
less than 300,000 miles between rebuilds and were rebuilt up to three times). Current four-
stroke locomotive engines use engine technology similar to on-highway diesel engines, except
that electronic controls have only recently been introduced.
       It is difficult to separate the components of current high-speed diesel engines for
discussion of their individual effects on emissions. Most of the components interact in numerous
ways that affect emissions, performance, and fuel consumption.

2.2.6.1. Indirect and Direct Injection High-Speed Diesel Engines
       Prior to the 1930s, diesel engine design was limited to relatively low-speed applications
because sufficiently high-pressure fuel injection equipment was not available.  With the advent
of high-speed and higher pressure  pump-line-nozzle systems, introduced by Robert Bosch in the

                                          2-45

-------
1930s, it became possible to inject the fuel directly into the cylinder for the first time, although
indirect injection (IDI) diesel engines continued in use for many years.  As diesels were
introduced into the heavy truck fleet in the 1930s through the 1950s, both IDI and direct
injection (DI) naturally aspirated variants were evident.  A very low-cost rotary injection pump
technology was introduced by Roosa-Master in the 1950s, reducing the cost of DI systems and
allowing their introduction on smaller displacement, higher speed truck engines.  After this time,
only a small fraction of truck engines used an IDI system.
       DI diesel engines have now all but replaced IDI diesel engines for HD on-highway
applications.2 IDI engines typically required much more complicated cylinder head designs but
generally were capable of using less sophisticated, lower pressure injection systems with less
expensive single-hole injection nozzles.  IDI combustion systems are also more tolerant of lower
grades of diesel fuel. Fuel injection systems are likely the single most expensive  component of
many diesel engines. Caterpillar continued producing both turbocharged and naturally aspirated
IDI diesel engines for some on-highway applications into the 1980s. Caterpillar and Deutz still
produce engines of this type, primarily for use in underground mining applications.  IDI
combustion systems are still used in many small-displacement (<0.5 L/cylinder), very high-
speed (>3,000 rpm rated speed) diesel engines for small nonroad equipment (small imported
tractors, skid-steer loaders), auxiliary engines, and small generator  sets, and they were prevalent
in diesel automotive engines in the  1980s; IDI designs continue to be used in automotive diesel
engines.
       IDI engines have practically no premixed burn combustion  and thus are often quieter and
have somewhat lower NOX emissions than DI engines. Electronic controls, high-pressure
injection (e.g., GM 6.5), and four-valve/cylinder designs (e.g., the six-cylinder Daimler LD
engine) can be equally applied to IDI diesel engines as in DI, but they negate advantages in cost
over DI engines. DI diesel engines of the same power output consume 15% to 20% less fuel
than IDI engines (Heywood, 1988).  Considering the sensitivity of the HD truck market to fuel
costs, this factor alone accounts for the demise of IDI diesel engines in these types of
applications. Throttling and convective heat transfer through the chamber-connecting orifice,
and heat rejection from the increased surface area of IDI combustion systems, decrease their
efficiency and can  cause cold-start difficulties when compared to DI designs. Most IDI diesel
engine designs require considerably higher than optimum compression ratios (from an efficiency
standpoint) to aid in cold-starting (19:1 to 21:1 for IDI engines vs. -15:1 to 17:1 for DI engines).
       2The GM Powertrain/AM General 6.5L electronically controlled, turbocharged IDI-swirl chamber engine,
certified as a light HD diesel truck engine, is the last remaining HD on-highway IDI engine sold in the United
States.

                                          2-46

-------
       Because of the early introduction of DI technology into truck fleets, it is likely that by the
end of the 1960s, only a small fraction of the HD diesel engines sold for on-highway use were
IDI engines. It is unlikely that the shift from IDI to DI engine designs through the 1950s and
1960s occurred rapidly and likely that this shift had little significant impact on emissions.
Springer (1979) reports a comparison of nearly identical Caterpillar 3406 engines (turbocharged
and aftercooled) in DI and IDI configurations tested on an engine dynamometer under steady-
state conditions, which limits the usefulness of these data.  There was no significant difference in
emissions of DPM, SOF,  aldehydes, or DPM-associated B[a]P (Table 2-8). Note that IDI
designs continue to be used in automotive diesel engines.

2.2.6.2. Injection Rate
       Decreasing the duration of diffusion combustion and promoting EC oxidation during the
expansion stroke can reduce formation of EC agglomerates (Stone,  1995) and reduce the
particulate carbon fraction at high load (Needham et al., 1989). Both of these effects are
enhanced by increasing the fuel injection rate. The primary means of accomplishing this is by
increasing fuel injection pressure.  In 1977 Robert Bosch introduced a new type of high-pressure
pump capable of producing injection pressures of 1,700 bar at the nozzle (Voss and Vanderpoel,
1977). This increased fuel injection pressure by roughly a factor of 10.  Unit injection, which
combines each fuel injection nozzle with individual cam-driven fuel pumps, can achieve very
high injection pressures (>2,000 bar). The first combination of unit injectors with electronically
controlled solenoids for timing control was offered in the United States by Detroit Diesel
Corporation in the 1988 model year (Hames et al., 1985).  Replacement of the injection cam with
hydraulic pressure, allowing a degree of injection rate control, was made possible with the
hydraulic-electronic unit injection jointly developed by Caterpillar and Navistar, introduced on
the Navistar T444E engine (and variants) in 1993.
       It is widely known that high fuel injection pressures have been used to obtain compliance
with the PM standards that went into effect in 1988 (Zelenka et al.,  1990). Thus, it is likely that
a transition to this technology  began in the 1980s, with the vast majority of new engine sales
employing this technology by  1991, when the 0.25 g/bhp-hr Federal PM standard went into
effect.
       The use of electronic control of injection rate is rapidly increasing on medium HD diesel
engines. Engines are currently under development, perhaps for 2002-2004 introduction, that use
common-rail fuel injection systems with even more flexible control over injection pressure and
timing than previous systems.
       Increased injection rate and pressure can significantly reduce EC emissions, but it can
also increase combustion temperatures and cause an increase in NOX emissions (Springer, 1979;

                                          2-47

-------
Watson and Janota, 1982; Stone, 1995). Low NOX, low DPM, and relatively good BSFC and
brake mean engine pressure (BMEP) are possible when combined with turbocharging,
aftercooling, and injection timing retard.

2.2.6.3. Turbocharging, Charge-Air Cooling, and Electronic Controls
       Use of exhaust-driven turbochargers to increase intake manifold pressure has been
applied to both IDI and DI diesel engines for more than 40 years. Turbocharging can decrease
fuel consumption compared with a naturally aspirated engine of the same power output.
Turbocharging utilizes otherwise wasted exhaust heat and pressure to generate intake boost. The
boosted intake pressure effectively increases air displacement and increases the amount of fuel
that can be injected to achieve a given fuel-air ratio. Turbocharging increases the power density
of an engine.  Boosting intake pressure via turbocharging and reducing fuel-to-air ratio at a
constant power can significantly increase both intake temperatures and NOX emissions.
Increased boost pressure  can significantly reduce ignition delay, which reduces VOC and DPM
SOF emissions (Stone, 1995) and increases the flexibility in selection of injection timing.
Injection timing on turbocharged engines can be retarded further for NOX emission control with
less of an effect on DPM emissions and fuel consumption.  This allows a rough parity in NOX
emissions between turbocharged (non-aftercooled) and naturally aspirated diesel engines
(Watson and Janota,  1982).
       Turbocharging permits the use of higher initial injection rates (higher injection pressure),
           130
           120
           110
           100
             90
             80
             70
                                Naturally Aspirated
Turbocharged/Aftercooled
                60   65   70   75   80  85   90   95  100  105  110  115
                                            NOX  %
       Figure 2-23. Effect of turbocharging and aftercooling on NOX and PM.
       Source: Mori,  1997.
                                         2-48

-------
which can reduce particulate emissions.  Although this may offer advantages for steady-state
operation, hard accelerations can temporarily cause overly fuel-rich conditions because the
turbocharger speed lags behind a rapid change in engine speed (turbo-lag). This can cause
significant increases in DPM emissions during accelerations.  Before the advent of electronic
controls, the effect of acceleration on DPM emissions could be limited by mechanically delaying
demand for maximum fuel rate with a "smoke-puff eliminator." Because this device also limited
engine response, there was considerable incentive for the end-users to remove or otherwise
render the device inactive. Charge-air cooling, for example, using an air-to-air aftercooler (air-
cooled heat exchanger) between the turbocharger compressor and the intake manifold, can
greatly reduce intake air and peak combustion temperatures.  When combined with injection
timing retard, charge-air cooling allows a significant reduction in NOX emissions with acceptable
BSFC and DPM emissions when compared to either non-aftercooled or naturally aspirated diesel
engines (Hardenberg and Fraenkle, 1978; Pischinger and Cartellieri, 1972; Stone, 1995). The
use of charge-air cooling effectively shifts the NOX-DPM tradeoff curve, as shown in
Figure 2-23.
       Electronic control of fuel injection timing allowed engine manufacturers to carefully
tailor the start and length of the fuel injection events much more precisely than through
mechanical means. Because of this, newer on-highway turbocharged truck engines have
virtually no visible smoke on acceleration (although emissions of DPM are substantial during
this driving mode). Electronic controls also allowed fuel injection retard under desirable
conditions for NOX reduction, while still allowing timing optimization for reduced VOC
emissions on start-up, acceptable cold-weather performance, and acceptable performance and
durability at high altitudes.  Previous mechanical unit injected engines (e.g., the 1980s Cummins
L10, the Non-Electronic Control Detroit Diesel 6V92) were capable of reasonably high injection
pressures, but they had fixed injection timing that only varied based on the hydraulic parameters
of the fuel system. Many other engines with mechanical in-line or rotary injection pumps had
only coarse injection timing control or fixed injection timing.
       Precise electronic control of injection timing over differing operating conditions also
allowed HD engine manufacturers to retard injection timing to obtain low NOX emissions during
highly transient urban operation, similar to that found during emissions certification. FID engine
manufacturers also advanced injection timing during less transient operation  (such as freeway
driving) for fuel consumption improvements (-3% to 5%) at the expense of greatly increased
NOX emissions (approximately three to four times regulated levels).  This particular situation
resulted in the recent consent decree settlements between the Federal Government and most FID
engine manufacturers to ensure effective NOX control in all driving conditions, including on-
                                          2-49

-------
highway high-speed steady-state driving.
       Turbocharged engines entered the market very slowly beginning in the 1960s. Data for
DPM emissions from naturally aspirated engines of model years 1976 to 1983 are compared with
DPM emissions from turbocharged engines in Figure 2-24.  There is no consistent difference in
DPM emissions between turbocharged and naturally aspirated engines.  Although not plotted, the
data also show no difference in emissions of NOX, DPM SOF, or DPM-associated B[a]P and 1-
nitropyrene (1-NP).
       Charge-air cooling was introduced during the 1960s and was initially performed in a heat
exchanger using engine coolant.  Cooling of the charge air using ambient air as the coolant was
introduced into heavy trucks by Mack in 1977 with production of the ETAY(B)673A engine
(Heywood, 1988). Use of ambient air allowed cooling of the charge air to much lower
temperatures. Most HD diesel engines sold today employ some form of charge air cooling, with
air-to-air aftercooling being the most common.  Johnson and co-workers (1994) have presented a
comparison of similar engines that differ in that the charge air is cooled by engine coolant (1988
engine) and by ambient air, with  a higher boost pressure for the second (1991 engine). The 1991
engine also used higher pressure  fuel injectors.  The 1991 engine exhibited both lower DPM
emissions  (50% lower than the 1988 engine) and lower NOX emissions.  Higher injection
pressure is likely to have enabled the reduced DPM emissions, whereas the lower charge-air
              o
              'in
              tn
              LU 1
                                                  •  Turbocharged
                                                  O  Naturally Aspirated
                                 •
                                 s
I
I
                 1974       1976       1978       1980
                                    Engine Model Year
                                                          1982
                                                                    1984
Figure 2-24.  Comparison of diesel engine dynamometer PM emissions for four-stroke,
naturally aspirated, and turbocharged engines.

Source: Data are from Table 2-8.
                                         2-50

-------
temperature and the ability to electronically retard the injection timing under some conditions
likely enabled the lower NOX emissions.
       It is apparent on the basis of both the literature and certification data that turbochargers
with aftercoolers can be used in HD engines in conjunction with other changes to produce a
decrease in emissions. On the advent of a NOX standard in 1985, NOX was probably reduced on
the order of 10% to 30% in turbocharged aftercooled engines with retarded injection timing.
This decrease is not evident in the in-use chassis testing data because of deterioration and the use
of illegal emissions devices as described above.  Overall, it is expected that engines in the 1950s
to mid-1970s timeframe would have similar DPM emission rates, whereas post-1970 engines
would have somewhat lower DPM emission rates.

2.2.6.4. Two-Stroke and Four-Stroke High-Speed Diesel Engines
       A detailed discussion of the two- and four-stroke engine cycles can be found in the
literature (Heywood, 1988; Taylor,  1990; Stone, 1995). Nearly all high-speed two-stroke diesel
engines utilize uniflow scavenging assisted by a positive-displacement blower (Figure 2-25).
Uniflow-scavenged two-stroke diesels use poppet exhaust valves similar to those found in four-
stroke engines.  The intake air enters the cylinder through a pressurized port in the cylinder wall.
A crankshaft-driven, positive-displacement blower (usually a roots-type) pressurizes the intake
port to ensure proper scavenging. A turbocharger may be added to the system to provide
additional boost upstream of the blower at higher speeds and to reduce the size and parasitic
losses associated with the positive-displacement blower.
       Two-stroke diesel engines can achieve efficiency comparable to four-stroke counterparts
and have higher BMEP (torque per unit displacement) (Heywood, 1988). It is useful to note that
two-stroke cycle fires each cylinder once every revolution, whereas the four-stroke cycle fires
every other revolution.  Thus, for a given engine size and weight, two-strokes can produce more
power.  However, two-stroke diesel engines are less durable than their four-stroke counterparts.
Lubricating oil is transferred from the piston rings to the intake port, which causes relatively
high oil consumption relative to four-stroke designs.  Durability and low oil consumption are
desirable for on-highway truck applications. This may be why four-stroke engines have been
favored for these applications since the beginning of dieselization in the trucking industry, with
the notable exception of urban bus applications.  Although it is no longer in production, the
Detroit Diesel 6V92 series of two-stroke diesel engines is still the most popular for urban bus
applications, where the high power density allows the engine to be more easily packaged within
limited spaces.  The primary reason that two-stroke engines like the 6V92 are no longer offered
for urban bus applications is excessive DPM emissions.  The lubricating oil control with two-
strokes tends to be lower than for four-stroke engines, and therefore, emissions have higher VOC
                                          2-51

-------
                        Engine
                       .Exhaust
Air Intake
  Ports
                             O
                      rw
                                     \
                                                         Positive
                                                       Displacement
                                                          Blower
             Figure 2-25. An example of uniflow scavenging of a two-
             stroke diesel engine with a positive displacement blower.
             Scavenging is the process of simultaneously emptying the
             cylinder of exhaust and refilling with fresh air.

             Source: Adapted from Taylor, 1990.
and organic DPM emissions relative to four-stroke designs. This was particularly problematic
for urban bus applications because urban bus engines must meet tighter Federal and California
PM emissions standards. The current urban bus PM standard (0.05 g/bhp-hr) is one-half of the
current on-highway HD diesel engine PM standard, although EPA is in the process of proposing
more strict standards for HD diesel truck engines along with further reductions in diesel fuel
sulfur levels.  No two-stroke diesel engine designs have been certified to meet the most recent
urban bus PM emissions standards, and Detroit Diesel Corporation has not certified a two-stroke
diesel engine for on-highway truck use since 1995.
       A comprehensive review of emissions from hundreds of vehicles (1976-98 model years)
that had been tested on chassis dynamometers found that DPM emissions vary substantially
within a given model  year and that within that variation there are no discernible differences in
DPM emissions between two- and four-stroke vehicles (Figure 2-26) (Yanowitz et al., 2000).
DPM emission factors reported for engine tests also indicate that two- and four-stroke engines
have comparable emission factors, as these engines all had to meet the same regulatory standard
(Figure 2-27). In contrast to DPM emissions, evidence suggests that mid-1970s two-stroke
engines exhibited very high SOF levels compared with four-stroke engines, with later model
years showing similar SOF emissions for two- and four-stroke engines (Figure 2-28). For
aldehydes, benzo[a]pyrene, and 1-nitropyrene, data are available for only one two-stroke engine,
                          2-52

-------
            c
            o
            E
            uj
o
4

3


2 -

n
• 4-Stroke Engines
o 2-Stroke Engines

0
o


o
e Q
8 IP
8 8
n *


O
•
p
8JI

Q • •
lll°








•
1.
                1978 1980  1982  1984 1986  1988 1990  1992 1994  1996 1998
                            Engine Emissions Model Year
Figure 2-26. Comparison of two- and four-stroke vehicle diesel PM emissions from chassis
dynamometer studies.

Source: Yanowitz et al., 2000.
but they indicate no significant difference in emissions from comparable model year four-stroke
engines.  Overall, regulated emissions changes attributable to changing proportions of two- and
four-stroke engines in the in-use fleet do not appear to have influenced DPM emission levels, but
the transition to four-stroke engines in the 1970s would have decreased the fraction of SOF
associated with the DPM. It appears that the proportion of two-stroke engines in the in-use fleet
was relatively constant until the late 1980s, when it began to decline.

2.2.7. Air Toxic Emissions
       HD diesel vehicle exhaust contains several  substances that are known, likely, or possible
human or animal carcinogens, or that have serious noncancer health effects. These substances
include, but are not limited to, benzene, formaldehyde, acetaldehyde, 1,3-butadiene, acrolein,
dioxin, PAH, and nitro-PAH (the complete list of chemically characterized compounds present
in DE is provided in Section 2.3.1). Very few historical data are available to examine changes  in
emission rates over time. In this section, trends in aldehyde emissions over time and a summary
                                         2-53

-------
             .Q
             "3)
              (0
              (0
                2
                1 -
                                                  •  4-Stroke
                                                  o  2-Stroke
                     •   §§.••  t   °,
                     8 •     oo      f o 8 •
                 1975      1980      1985      1990

                                   Engine Model Year
1995
2000
Figure 2-27. Comparison of two- and four-stroke engine diesel PM emissions from engine
dynamometer studies.
i.uu -


£ 0.75
Q.
.Q
"3)
.2 0.50
in
E
LLI
LJ_
0
W 0.25 -
n nn -
o
• 4-Stroke
o 2-Stroke
o





• o
8* * •
••• °I° -° 8
:: • • . .! .1:
                 1975      1980      1985       1990      1995      2000

                                    Engine Model Year
Figure 2-28. Diesel engine dynamometer SOF emissions from two- and four-stroke
engines. SOF obtained by dichloromethane extraction in most studies.

Source: Data are from Table 2-8.
                                       2-54

-------
of dioxin emission factors are presented. PAH and nitro-PAH emission factors are discussed in
Section 2.2.8.2.

2.2.7.1. Aldehyde Emissions
       Among the gaseous components emitted by diesel engines, the aldehydes are particularly
important because they constitute an important fraction of the gaseous emissions and they are
probable  carcinogens that also produce noncancer health effects. Formaldehyde makes up the
majority of the aldehyde emissions (65% to 80%), with acetaldehyde being the second most
abundant aldehyde in HD diesel emissions. Total aldehyde emissions reported from chassis
dynamometer testing suggest that aldehyde emissions have declined since 1980; however, only
two tests  reported aldehydes from engines made after 1985 (Figure 2-29). Engine dynamometer
studies also suggest a downward trend in the emissions of aldehydes in the time period from
1976 to 1994 (Figure 2-30). Engine dynamometer studies report aldehyde emission levels of
150-300  mg/bhp-hr for late 1970s engines with no significant effect of turbocharging, or IDI
versus DI. High-pressure fuel injection may have resulted in a marginal increase in aldehyde
emissions (Springer, 1979). By comparison, 1991 model year engines (DI, turbocharged)
exhibited aldehyde emissions in the 30-50 mg/bhp-hr range (Mitchell et al., 1994).

2.2.7.2. Dioxin and Furans
       Ballschmiter et al. (1986) reported detecting polychlorinated dibenzo-p-dioxins (CDDs)
and polychlorinated dibenzofurans (CDFs) in used motor oil and thus provided some of the first
evidence  that CDDs  and CDFs might be emitted by the combustion process in diesel-fueled
engines.  Incomplete combustion and the presence of a chlorine source in the form of additives in
the oil or the fuel were speculated to lead to the formation of CDDs and CDFs. Since 1986,
several studies have been conducted to measure or estimate CDD/CDF concentrations in
emissions from diesel-fueled vehicles. These studies can be characterized as direct
measurements from the engine exhaust and indirect measurements from the sampling of air
within transportation tunnels.
       Table 2-11 is a summary of various CDD/CDF emission characterization studies reported
in the United States and Europe for diesel-fueled cars and trucks.  Hagenmaier et al. (1990)
reported an emission factor for LD diesel vehicles of 24 pg TEQ per liter of diesel fuel
consumed. TEQ, or the toxic equivalency factor, rates each dioxin and furan relative to that of
2,3,7,8-TCDD, which is arbitrarily assigned a TEQ of 1.0 based on animal assays.  Schwind et
al. (1991) and Hutzinger et al. (1992) studied emissions of CDDs/CDFs from German internal
combustion engines running on commercial diesel fuels and reported a range of CDD/CDF
emission  rates across the test conditions (in units of pg TEQ per liter of diesel fuel consumed) of
10-130 pg TEQ/L for diesel car exhaust and 70-81 pg TEQ/L for diesel truck exhaust.

                                         2-55

-------
             1000
          £   800
          E
          g
          '
LU
3  100 H
           CD
           T3
               50 -
                                                       •  4-stroke
                                                       o  2-stroke
                1976 1978 1980 1982 1984  1986  1988  1990  1992 1994 1996
                                    Engine Model Year

Figure 2-30.  Diesel engine aldehyde emissions from engine dynamometer studies.

Source:  Data from Table 2-8.
                                         2-56

-------
                        Table 2-11. Summary of CDD/CDF emissions from diesel-fueled vehicles
to


Study
CARB, 1987; Lew, 1996
Marklundetal, 1990
Hagenmaier et al., 1990
Hagenmaier, 1994
Oehme et al., 1991 (tunnel
study)




Schwind et al., 1991
Hutzinger et al., 1992
Gertler et al., 1996 (tunnel
study)
Gullett and Ryan, 1997


Country
United States
Sweden
Germany
Germany
Norway





Germany

United States

United States


Vehicle tested
Diesel truck
Diesel truck
Diesel car
Diesel bus
—





Diesel car
Diesel truck
Diesel trucks

Diesel truck
Number
of test
vehicles
1
1
1
1
(b)





1
1
(d)

1

Emission factor
(pg TEQ/km driven)
663-1,300
not detected (<18)a
2.4a
not detected (< 1 pg/L)
520C
38C
avg = 280
9,500C
720C
avg = 5,100
5.0-13a
13-15"
mean =172

mean - 29.0


Driving cycle; sampling location
6-hr dynamometer test at 50 km/hr
U.S. Federal mode 13 cycle; before muffler
Comparable to FTP-73 test cycle; in tailpipe
On-the-road testing
Cars moving uphill (3.5% incline) at 60 km/hr
Cars moving downhill (3.5% decline) at 70 km/hr

Trucks moving uphill (3.5% incline) at 60 km/hr
Trucks moving downhill (3.5% decline) at 70 km/hr

Various test conditions (i.e., loads and speeds)
Various test conditions (i.e., loads and speeds)
Mean of seven 12-hour samples

Mean of five sample routes
              "Results reported were in units of pg TEQ/liter of fuel. For purposes of this table, the fuel economy factor used by Marklundetal. (1990), 10 km/L or 24 miles/gal, was used to convert the emission rates
              into units of pg TEQ/km driven for the cars.  For the diesel-fueled truck, the fuel economy factor reported in CARB (1987a)for a 1984 heavy-duty diesel truck, 5.5 km/L (or 13.2 miles/gal), was used.
              "Tests were conducted over portions of 4 days, with traffic rates of 8,000-14,000 vehicles/day. Heavy-duty vehicles (defined as vehicles over 7 meters in length) ranged from 4% to 15% of total.
              "Emission factors are reported in units of pg Nordic TEQ/km driven; the values in units of I-TEQ/km are expected to be about 3% to 6% higher.
              dTests were conducted over 5 days with heavy-duty vehicle rates of 1,800-8,700 vehicles per 12-hour sampling event. Heavy-duty vehicles accounted for 21% to 28% of all vehicles.

-------
       In 1994, Hagenmaier reported CDD/CDF emissions from a diesel-fueled bus and found
no detectable levels in the exhaust (at a detection limit of 1 pg/L of fuel consumed) for
individual congeners. In 1987, the California Air Resources Board (CARB) produced a draft
report of a HD engine tested under steady-state conditions indicating a TEQ emission factor of
7,290 pg/L of fuel burned (or 1,300 pg/km driven) if nondetected values are treated as one-half
the detection limit. Treating nondetected values as zeros yields a TEQ concentration equivalent
to 3,720 pg/L of fuel burned (or 663 pg/km driven) (Lew, 1996). Norbeck et al. (1998c) reported
emission factors for dioxin and furans from a Cummins L10 HD diesel engine running on pre-
1993 fuel of 0.61 pg/L and 0.41 pg/L for the same engine running on reformulated fuel.  The low
emission factors reported by Norbeck et al. (1998c) were attributed to losses of dioxin and furan
compounds to the dilution tunnel walls.
       EPA has directly sampled the exhaust from a HD diesel truck for the presence and
occurrence of CDDs/CDFs (Gullett and Ryan,  1997). The average of five tests (on highway and
city street driving conditions) was 29.0 pg TEQ/km with a standard deviation of 38.3 pg
TEQ/km; this standard deviation reflects the 30-fold variation in the two city driving route tests.
       Tunnel studies are an indirect means of measuring contaminants that may be associated
with emissions from cars and trucks. In these studies, scrapings  of carbonaceous matter from the
interior walls of the transportation tunnel or the tunnel air are sampled and analyzed for the
target contaminants.   Several European studies and one recent U.S. study evaluated CDD/CDF
emissions from vehicles by measuring the presence of CDDs/CDFs in tunnel air. This approach
has the advantage of allowing random sampling of large numbers of vehicles passing through the
tunnel, including a range of ages and maintenance levels. The disadvantage of this approach is
that it relies on indirect measurements (rather than tailpipe measurements), which may introduce
unknown uncertainties into the interpretation of results.
       Oehme et al.  (1991) reported the emission rates  associated with HD diesel trucks as
follows: uphill = 9,500  pg TEQ/km; downhill  = 720 pg TEQ/km; mean = 5,100 pg TEQ/km.
The mean values are the averages of the emission rates  corresponding to the two operating
modes: vehicles moving uphill on a 3.5% incline at an  average speed of 37 mi/hr and vehicles
moving downhill on a 3.5% decline at an average speed of 42 mi/hr.
       Wevers et al. (1992) measured the CDD/CDF content of  air samples taken during the
winter of 1991 inside a tunnel in Antwerp, Belgium.  The results obtained indicated that the
tunnel air had a dioxin TEQ concentration about twice as high as the outside air (80.3 fg TEQ/m3
for tunnel air vs. 35 fg TEQ/m3 for outside air for one set of measurements and 100 fg TEQ/m3
for tunnel air vs. 58 fg TEQ/m3 for outside air for a second set of measurements).
                                         2-58

-------
       During October/November 1995, Gertler et al. (1996, 1998) measured CDDs/CDFs in the
Fort McHenry Tunnel in Baltimore, Maryland. The emission factors calculated, assuming that
all CDDs/CDFs emitted in the tunnel were from FID vehicles, are presented in Table 2-12.  The
average TEQ emission factor was reported to be 172 pg TEQ/km.  The major uncertainties in the
study were tunnel air volume measurement, sampler flow volume control, and analytical
measurement of CDDs/CDFs (Gertler etal., 1996, 1998).
       The relative strengths of the Gertler et al. (1996; 1998) study include: (1) The study is a
recent study conducted in the United  States and thus reflects current U.S. fuels and technology;
(2) virtually no vehicle using the tunnel used leaded gasoline, which is associated with past
emissions of CDDs and CDFs from gasoline-powered vehicles; (3) the tunnel walls and streets
were cleaned 1 week before the start of sampling, and in addition, the study analyzed road dust
and determined that resuspended road dust contributed only about 4% of the estimated emission
factors; and (4) HD vehicles made up, on average 25.7% of vehicles using the tunnel.
       Using the emissions factor from the Gertler et al. studies, the EPA Office of Research and
Development's dioxin source emission inventory estimates that 33.5 g of dioxin TEQ (total
2,3,7,8-TCDD equivalents) were emitted from HD U.S. trucks in 1995.  This is a very small
contribution (1.2%) compared with the national annual emission of 2,800 g CDDs/CDFs.

2.2.8. Physical and Chemical  Composition of Diesel Exhaust Particles
       DPM is defined by the measurement procedures summarized in Title 40 CFR, Part 86,
subpart N. This definition and the basic characteristics of DPM have been summarized in
Section 2.2.2.  As described there, DE particles are aggregates of primary spherical particles that
consist of solid carbonaceous material and ash and contain adsorbed organic and sulfur
compounds (sulfate) combined with other condensed material.  The organic material includes
unburned fuel, engine lubrication oil, and low levels of partial combustion and pyrolysis
products.
       The organic material is absorbed to the EC core and is also found in heterogeneously
nucleated aerosol. This fraction of the DPM is frequently quantified as the SOF (i.e., the fraction
that can be extracted by an organic solvent).  Because of the toxicological significance of the
organic components associated with DPM, it is important to understand, to the extent possible,
the historical changes in the composition of SOF and potential changes in the fraction of SOF
associated with DPM.
       Various researchers have attempted to apportion the SOF to unburned oil and fuel
sources by thermogravimetric analysis and have found that the results vary with test cycle
and engine (Abbass et al., 1991; Wachter, 1990).  Kittelson (1998) estimates that a typical
composition of SOF is about one-fourth unburned fuel and three-fourths unburned
                                         2-59

-------
                 Table 2-12.  Baltimore Harbor Tunnel Study:  estimated CDD/CDF emission factors for HD vehicles



Congener/congener group
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
Total 2,3,7,8-CDD
Total 2,3,7,8-CDF
Total TEQ
Total TCDD
Total PeCDD
Total HxCDD
Total HpCDD
Total OCDD
Total TCDF
Total PeCDF
Total HxCDF
Total HpCDF
Total OCDF
Total CDD/CDF
FID vehicles as % of total
vehicles

Run-specific emission factors
Run no. 2
(pg/km)
24.5
40.2
18.2
37.5
53.6
0
0
0
0
24.5
15.4
0.3
27.7
15.2
12.6
0
0
174
95.7
73.8
245
110
677
0
0
0
124
136
0
0
1,291
21.2

Run no. 3
(pg/km)
61.6
20.6
25.2
28.2
56.5
401
3,361
94.3
48.9
75.7
139
75.1
14.8
82.5
280
58.5
239
3,954
1,108
175
0
21.9
0
802
3361
901
119
319
223
239
5,987
22.0

Run no. 5
(pg/km)
0.0
15.4
46.5
64.3
91.6
729
3,382
67.6
72.6
131
204
73.7
75.6
152
445
60.8
401
4,328
1,684
170
140
83.3
753
1,498
3,382
1,314
1,152
852
814
401
10,390
22.6

Run no. 6
(pg/km)
21.2
5.6
8.3
19.6
48.4
111
1,120
152.8
23.6
46.6
93.8
51.0
0
55.7
154
31.1
175
1,335
784
96
165
35.6
54.5
142
1,120
656
78.4
67.6
144
175
2,638
34.0

Run no. 8
(pg/km)
37.8
38.4
64.5
153
280
2,438
9,730
155.8
53.3
85.0
124
61.3
20.6
93.0
313
25.0
416
12,743
1,347
235
311
174
2,009
5,696
9,730
2,416
1,055
444
513
416
22,766
28.8

Run no. 9
(pg/km)
40.1
0.0
0.0
71.1
126
963
5,829
73.4
0.0
63.9
164
54.4
37.2
86.8
354
2.3
534
7,028
1,371
153
109
0.0
1,666
1,933
5,829
1,007
282
719
354
534
12,434
24.2

Run no. 10
(pg/km)
54.9
83.0
123
186
370
2,080
7,620
61.7
43.3
108
166
95.5
63.5
111
308
34.9
370
10,515
1,362
303
97.3
165
2,971
4,377
7,620
687
626
619
637
370
18,168
27.4

Mean
emission
factors
(pg/km)
34.3
29.0
40.8
80.0
147
960
4,435
86.5
34.5
76.4
129
58.8
34.2
85.2
267
30.4
305
5,725
1,107
172
152
84.2
1,162
2,064
4,435
997
491
451
384
305
10,525
25.7

Notes:
(1) Listed values are based on the difference between the calculated chemical mass entering the tunnel and the mass exiting the tunnel.
(2) All calculated negative emission factors were set equal to zero.
(3) All CDD/CDF emissions were assumed to result from heavy-duty diesel-fueled vehicles. The table presents in the last row the percent of total traffic that was heavy-duty vehicles.
Source: Gertler et al., 1996.

-------
engine oil. Partial combustion and pyrolysis products represented a very small fraction
of the SOF on a mass combustion and pyrolysis products represented a very small fraction
of the SOF on a mass basis (Kittelson,  1998), which is confirmed in numerous other studies.
       A number of investigators have tried to separate the organic fraction into various classes
of compounds.  Schuetzle (1983) analyzed the dichloromethane extract of DPM from a LD
diesel engine and found that approximately 57% of the extracted organic mass is contained in the
nonpolar fraction. About 90% of this fraction consists of aliphatic HCs from approximately C14
to about C40 (Black and High, 1979; Pierson and Brachaczek, 1983).  PAHs and alkyl-substituted
PAHs account for the remainder of the nonpolar mass. The moderately polar fraction (-9% w/w
of extract) consists mainly of oxygenated PAH species, substituted benzaldehydes, and nitrated
PAH. The polar fraction (-32% w/w of extract) is composed mainly of n-alkanoic acids,
carboxylic and dicarboxylic acids of PAH, hydroxy-PAH, hydroxynitro-PAH, and nitrated N-
containing heterocyclic compounds (Schuetzle,  1983; Schuetzle et al., 1985).
       Rogge et al. (1993) reported the composition of the extractable portion of fine DPM
emitted from two HD diesel trucks (1987 model year). The DPM filters were extracted twice
with hexane, then three times with a benzene/2-propanol mixture. The extract was analyzed by
capillary gas chromatography/mass spectrometry (GC/MS) before and after derivatization to
convert organic acids and other compounds having an active H atom to their methoxylated
analogues. Unidentified organic compounds made up 90% of the eluted organic mass and were
shown to be mainly branched and cyclic HCs. From the mass fraction that was resolved as
discrete peaks by GC/MS, -42% were identified as specific organic compounds. Most of the
identified resolved organic mass (-60%) consisted of n-alkanes, followed by n-alkanoic  acids
(-20%). PAH accounted for -3.5% and oxy-PAH (ketones and quinones) for another -3.3%.
       The distribution of the emissions between the gaseous and particulate phases is
determined by the vapor pressure of the individual species, by the amount and type of the DPM
present (adsorption surface available), and by the temperature (Ligocki and Pankow, 1989).
Two-ring and smaller compounds (e.g., naphthalene) exist primarily in the gas phase, whereas
five-ring and larger compounds (e.g., benzo[a]pyrene) are almost completely adsorbed on the
particles. Three- and four-ring compounds are distributed between the two phases. The  vapor
pressures of these intermediate PAHs can  be significantly reduced by their adsorption on various
surfaces. Because of this phenomenon, the amount and type of DPM present play an important
role, together with temperature, in the vapor-particle partitioning of semivolatile organic
compounds (SOCs).
       The measurements of gas/parti culate phase distribution are often accomplished by using a
high-volume filter followed by an adsorbent such as polyurethane foam (PUF), Tenax, or XAD-2
(Cautreels and Van Cauwenberghe, 1978;  Thrane  and Mikalsen,  1981; Yamasaki et al., 1982).

                                         2-61

-------
The pressure drop behind a high-volume filter or cascade impactor can contribute to
volatilization of the three- to five-ring PAHs from the PM proportional to their vapor pressures.
The magnitude of this blow-off artifact depends on a number of factors, including sampling
temperature and the volume of air sampled (Van Vaeck et al., 1984; Coutant et al., 1988).
Despite these problems from volatilization, measurements with the high-volume filters followed
by a solid adsorbent have provided most estimates of vapor-particle partitioning of SOCs in
ambient air, as well as insights into the factors influencing SOC adsorption onto aerosols.
Significant fractions of phenanthrene, anthracene, and their alkylated derivatives, along with
fluoranthene and pyrene, exist in the gas phase. PAHs with molecular weight greater than that of
pyrene are typically not observed on PUF samples. During the collection of particulate organic
compounds, adsorption of semivolatile PAHs can also occur, as well as chemical transformation
of the semivolatile compounds (Schauer et al., 1999; Cantrell et al., 1988; Feilberg et al., 1999;
Cautreels and Van Cauwenberghe, 1978).
       Most of the sulfur in the fuel is oxidized to SO2, but a small amount (1% to 4%) is
oxidized to sulfuric acid in the exhaust.  Sulfate emissions are roughly proportional to sulfur in
the fuel.  Since the reduction of the allowable sulfur content in diesel fuel in 1993, sulfate
emissions have declined from roughly 10% of the DPM mass to around 1%.  Particulate
emissions from numerous vehicles tested using low-sulfur fuel were found to have a sulfate
content of only about 1% (Yanowitz et al., 1999). Water content is on the order of 1.3  times the
amount of sulfate (Wall et al., 1987).
       Metal compounds and other elements in the fuel and engine lubrication oil are exhausted
as ash. Hare (1977) examined 1976 Caterpillar 3208 and Detroit Diesel Corporation 6V-71
engines and found the most abundant elements emitted from the 6V-71 engine were silicon,
copper, calcium, zinc, and phosphorus.  From the Caterpillar engine the most abundant elements
were lead, chlorine, manganese, chromium, zinc, and calcium.  Calcium, phosphorus, and zinc
were present in the engine lubrication oil. The two-stroke 6V-71 engine had higher engine
lubrication oil emissions and therefore emitted higher levels of zinc, calcium, and phosphorus
than the Caterpillar 3208 engine. Other elements may have been products of engine wear or
contaminants from the  exhaust system.  Springer (1979), in his study of 1977 Mack
ETAY(B)673 A and Caterpillar 3208  (EGR) engines, found that calcium was the most abundant
metallic element in DPM samples, with levels  ranging from 0.01 to 0.29 wt% of the DPM.
Phosphorus and silica were the next most abundant elements reported, and sodium, iron, nickel,
barium, chromium, and copper were either present at very low levels or were below detection
limits. Roughly  1 wt% of the total DPM was represented by  the analyzed metals. There was no
consistent difference in metal emissions between the engines tested by Springer or between
modes.  Springer tested both engines on a 13-mode steady-state test.  Dietzmann and co-workers

                                          2-62

-------
(1980) examined metal emission rates from four HD vehicles tested using the UDDS chassis
cycle.  For the single two-stroke engine tested (1977 Detroit Diesel Corporation 8V-71),
calcium, phosphorus, and zinc emission rates were more than 10 times higher than metal levels
observed for three 1979 model year four-stroke engines because of higher engine lubrication oil
emissions. Metals accounted for 0.5% to 5% of total DPM, depending on engine model.  In
addition to these studies, other source profiles for HD diesel engine emissions report levels of
chromium, manganese, mercury compounds, and nickel at levels above the detection limit
(Cooper et al., 1987).
       In more recent studies, Hildemann and co-workers (1991) examined metals in DPM from
the same two 1987 trucks (four-stroke engines) studied by Rogge and co-workers (1993).
Aluminum, silicon, potassium, and titanium were the only metals observed at statistically
significant levels. Taken together these made up less than 0.75 wt% of total DPM mass.
Lowenthal and co-workers (1994) also report metals emission rates for a composite sample of
several diesel vehicles.  The most abundant metals were zinc, iron, calcium, phosphorus, barium,
and lanthanum.  Together these represented less than 0.3% of total DPM mass, with an emissions
rate of 3.3 mg/mi. Norbeck and co-workers (1998b) report engine transient test emissions of
metals for a 1991 Cummins L10 engine. Silicon, iron, zinc, calcium, and phosphorus were
observed and together made up about 0.5% of total DPM, with an emissions rate of 1.2 mg/bhp-
hr.

2.2.8.1. Organic and EC Content of Particles
2.2.8.1.1. Measurement of the organic and EC fraction. Various methods have been used to
quantify the organic fraction of DPM. The most common method has been Soxhlet extraction
with an organic solvent. Following extraction, the solvent can be evaporated and the mass of
extracted material (the SOF) determined, or  alternatively the PM filter is weighed before and
after extraction and the extracted material can be further analyzed to determine concentrations of
individual organic compounds. Vacuum oven sublimation is used to measure a comparable
quantity, the volatile organic fraction (VOF), which can be further speciated by GC with a flame
ionization detector. Other methods have also been employed, including thermal methods,
microwave extraction, sonication with an organic solvent, supercritical fluid extraction,
thermogravimetric analysis, and thermal desorption GC.  Abbass et al. (1991) compared various
methods, including vacuum oven sublimation and 8 hours of Soxhlet extraction, with 4:1
benzene/methanol solvent for determination of SOF and found reasonably good agreement
between the two methods. The VOF value was typically 10% higher; however, this variation
was less than the coefficient of variation between measurements using the same method.
                                         2-63

-------
       Levson (1988) reviewed literature regarding the extraction efficiency of various solvents
and found contradictory results in many cases.  He concluded that there is strong evidence that
the most commonly used solvent, dichloromethane, leads to poor recoveries of higher molecular
weight PAH.  More recently, Lucas et al. (1999) reported the effect of varying
dichloromethane/benzene ratios in the solvent (from 25% to 100% dichloromethane) and
changing extraction times and found that the most effective extraction (i.e., the largest extracted
mass) utilized a 70% dichlorom ethane/30% benzene mixture and extraction times several times
longer than the commonly used 8-hour extraction period. Extractions of 70 hours using pure
dichloromethane were found to result in about twice as much SOF as extractions of only 12
hours.  Between 6 and 24 hours of extraction time (the typical range of extraction times used),
the SOF recovered increased by about one-third.  Using the most effective extraction conditions
(Soxhlet, 70 hours, 70:30 dichloromethane:benzene ratio), Lucas et al. (1999) were able to
extract more than 90% of the total particulate mass.
       Other researchers have investigated the relative quantities of mass removed by sequential
extraction by  polar, moderately polar,  and nonpolar solvents. The extracted nonpolar fraction
(cyclohexane) ranged from 56% to 90% of the SOF, the  moderately polar (dichloromethane)
from 6% to 22%, and the polar fraction (acetonitrile) from 4% to 29% (Dietzmann et al., 1980).
Water and sulfate are not  soluble in cyclohexane or dichloromethane but are soluble in
acetonitrile.
       Although the reports on the extraction efficiencies for PAHs  are in part contradictory, it
appears that Soxhlet extraction and the binary  solvent system composed of aromatic solvent and
alcohol yield  the best recovery of PAHs, as determined by C-B[a]P14(benzo[a]pyrene) spiking
experiments (Schuetzle and Perez, 1983). Limited recovery studies have shown that there is
little degradation or loss of diesel POM on the HPLC column. More than 90% of the mass and
70% to 100% of the Ames S. typhimurium-activQ material injected onto the column has been
recovered (Schuetzle et al., 1985).
       Two thermal methods of organic and EC analysis include thermal optical reflectance
(TOR) and thermal optical transmittance (TOT).  The extractable portion of total carbon,
although commonly used  as a measure of organic compound content, is not equivalent to the OC
fraction as measured by TOR or TOT. In addition, methodological differences between TOR
and TOT also give rise to significant differences in the fraction of total carbon reported as
organic and EC (Birch, 1998; Norris et al., 2000;  Chow et al., 2000).  Although total carbon
reported using TOR or TOT provides results that are comparable (within 10%) (Norris et al.,
2000) the EC content of samples analyzed by TOR is higher than that measured by TOT. This
difference is primarily attributed to the temperature used to evolve carbon from the quartz filter
onto which it is collected. In an analysis of urban PM2 5  samples, Norris et al. (2000) found that

                                         2-64

-------
the EC content of samples analyzed by TOR was a factor of two higher than the EC content of
the same samples analyzed by TOT. Experiments are ongoing to test specific source materials
(including DPM) because some of the difference between methods appears to depend on the type
of OC present on the sample.
       The analytical technique used to measure OC and EC can have a significant effect on the
quantity of reported. In the discussion that follows, every effort has been made to compare only
studies using comparable methods and to state the analysis method employed.

2.2.8.1.2.  Trends in SOF emissions.  SOF emission values are highly dependent on the test
cycle used. Various studies have  shown that SOF generally increases at light engine loads and
high engine speeds because these  conditions lead to low exhaust temperatures, where fuel and oil
are not as effectively oxidized (Scholl et al., 1982; Kittelson, 1998; Springer, 1979; Schuetzle
and Perez, 1983; Martin, 1981b; Shi et al., 2000).  These conditions are more typically observed
in LD diesel vehicle applications,  and thus DPM from these vehicles typically has a higher SOF
component than HD diesel  vehicles (Norbeck et al., 1998c). Acceleration modes normally cause
increased emission of EC and an increase in total DPM emissions, whereas organic components
are more dominant when motoring (Wachter, 1990).  Additionally, cold-start test emissions of
SOF have been shown to be approximately 25% higher than hot-start emissions (Wachter, 1990).
       The quantity of sulfur in diesel fuel has been suggested to have a role in the quantity of
SOF emitted (Sienicki et al., 1990; Tanaka et al., 1998). Sienicki et al. (1990) reported an
approximate 25% increase  in SOF when sulfur concentrations are increased from 0.08% to
0.33%.  The cause is unclear but several explanations have been put forth, including increased
absorption of organic compounds  from the vapor phase onto the DPM by sulfates or sorbed
sulfuric acid. Alternatively, it has been proposed that the measured SOF may include some
sulfate, so that the apparent increase in organic material is due instead to sulfate. Other fuel
effects include an increase  in SOF emissions with a higher T90 (or T95) and with an increase in
aromatic content (Barry et al., 1985; Sienicki et al., 1990; Tanaka et al., 1998; Rantanen et al.,
1993).
       Figures 2-31 and 2-32 show SOF emissions as a function of year for transient  emissions
tests on chassis and engines, respectively. Both  figures suggest a significant decline in SOF
emissions of approximately a factor of 5 since about  1980. The highest SOF emissions are for
two-stroke engines built in the 1970s (up to approximately 1.2 g/mi).  These data indicate that
SOF emission factors for newer model year vehicles are lower than SOF emission factors for
pre-1990 model year vehicles and that this decrease is similar to that observed for emissions of
total DPM by model year.  In a  recent test of six pre-1976 HDDVs, Fritz et al. (2001) reported
                                         2-65

-------





1"
2
u.
O
o
u_
O
(/)






1.4


1.2

1.0

0.8

0.6


0.4


0.2

n n

• Warner-Selph et al.,

1984
o Dietzmann etal., 1980
* T Graboski etal., 1998
• v Rogge etal., 1993
•
• •
•
« c
o • T
•
0 • T
• ' *
0 • .
• . T T
0 • T
T T T ; j
T















                  1975
                           1980
                                     1985       1990

                                     Engine Model Year
                                                        1995
                                                                  2000
Figure 2-31. Trend in SOF emissions based on chassis dynamometer testing of HD diesel
vehicles. Warner-Selph and co-workers: dichloromethane for 8 hours.  Dietzman and co-
workers: hexane followed by dichloromethane, extraction times not reported.  Graboski
and co-workers: VOF by vacuum sublimation at 225° C for 2.5 to 3 hours. Rogge and co-
workers: cyclohexane followed by a benzene/2-propanol mixture that may extract
significantly more organic matter.
u.o -


0.6 -




0.4 -



0.2 -

n n -
_ •
• o
Ullmanetal., 1984
McCarthy etal., 1992


T Perez and Williams, 1989
V
•
n
+
O

A
. *
• o
0 f
• T •
T •
V A

Needham etal., 1989
Graboski, 1998
Spreenetal., 1995
Sienicki etal., 1990
Martin, 1981

Mitchell etal., 1994
Barry etal., 1985


" I o*
O 1!










•

                  1976  1978  1980  1982  1984 1986  1988  1990  1992  1994  1996

                                     Engine Model Year
Figure 2-32. Trend in SOF emissions for transient engine dynamometer testing of HD
diesel engines. Various extraction methods used; see Table 2-8.
                                        2-66

-------
the volatile organic fraction (VOF) ranged from 0.4 g/mi to 4.5 g/mi. These data highlight the wide
range in emission rates for of OC, as have been observed for total PM.
       Steady-state testing conducted on late-1970s engines reported SOF at levels between 0.1
and 0.9 g/bhp-hr, whereas engines from the late 1980s and 1990s all emitted 0.03 g/bhp-hr or
less (Table 2-8).  Hori and Narusawa (1998) measured emissions from engines produced two
decades apart, using identical analytical procedures, and found that SOF emission factors and the
percentage contribution of SOF to DPM were lower in the new engine compared with the old
engine, under all tested engine load and speed conditions and with different fuels.  The authors
reported that the decrease in SOF was due to lower emissions of both lubricating oil and
unburned fuel. To meet the 1991 and 1994 U.S. emission standards, SOF emission rates would
need to be reduced from the levels of the previous decade, although one may expect differences
in SOF fractions of DPM with transient cycles used to determine compliance with emission
standards verus steady-state conditions used in earlier test programs (Kawatani et al., 1993;
Wachter, 1990).  Finally, in the past three decades, for economic reasons engine manufacturers
have made efforts to reduce oil consumption and increase the fuel efficiency of diesel engines,
both of which would be expected to reduce SOF emissions. Problems in achieving SOF
reductions from two-stroke engines were one factor leading to the phaseout of these engines for
on-road use during the 1990s.  No data are available prior to 1976 on SOF emissions from HD
diesel vehicles.  The engine technology changes that occurred between the mid-1950s and mid-
1970s (high-pressure direct injection and turbocharging, primarily) might be expected to increase
the efficiency of combustion and thereby reduce fuel-related SOF. SOF emissions levels in the
mid- to late 1970s may be used as a conservative (low) estimate of SOF emissions during the
preceding two decades.
       The fraction of DPM attributed to SOF from chassis dynamometer studies also shows a
decreasing trend over time, from SOFs that ranged up to approximately 50% in the  1980s to 20%
SOF or less in the 1990s (Figure 2-33).  The recent study by Fritz et al. (2001) reported the
fraction of DPM attributed to VOF from 10% to 60% for HDDVs of model years 1951-1974.
The wide range in SOF as a percent of DPM displayed in Figure 2-33 is suspected to result from
factors such as engine deterioration and test cycle.  The vehicle emissions data reported in Figure
2-33 do not overrepresent buses that are likely to emit DPM with a greater fraction of SOF than
other vehicles. Figure 2-34 presents SOF as a fraction of DPM from the same engine
dynamometer studies reported in Figure 2-32. These data do not reflect a downward trend in
SOF as a fraction of DPM. Because similar extraction methods were used in reports of the SOF
in both the chassis and engine  dynamometer studies, this does not appear to be a source of the
wide variability observed in the fraction of SOF reported. In some of the engine studies,
improved airfuel ratio control was tested in an attempt to lower carbonaceous DPM formation.

                                         2-67

-------
              100
              80 -
              60 -
              40
              20 -
                                     •  Warner-Selph, 1984
                                     O  Dietzmann, et al., 1980
                                     T  Graboski, et al., 1998
                                     V  Rogge, et al., 1993
    1975       1980       1985       1990
                       Engine Model Year
                                                        1995
             2000
Figure 2-33.  Trend in SOF emissions as a percent of total PM based on chassis
dynamometer testing of HD diesel vehicles. Warner-Selph and co-workers:
dichloromethane for 8 hours. Dietzman and co-workers: hexane followed by
dichloromethane, extraction times not reported. Graboski and co-workers: VOF by
vacuum sublimation at 225" C for 2.5 to 3 hours. Rogge and co-workers:  cyclohexane
followed by a benzene/2-propanol mixture that may extract significantly more organic
matter.
               100
                80

            Q_
            3   60 -
O
CO
                40-
                20 -
                           O
                           0
                                T
                                T
g
•
•

f
                                                A

                                                B
                 1976  1978 1980 1982  1984 1986 1988  1990 1992  1994 1996
                                   Engine Model Year
Figure 2-34.  Trend in SOF emissions as a percentage of total PM from engine
dynamometer testing. Data are from Table 2-8.  (See Figure 2-32 for figure key )
                                         2-68

-------
Therefore, substantial differences in SOF as a percent of total DPM could be the result of
different engine technology or test conditions. The engine dynamometer results presented in
Figures 2-32 and 2-34 are from new, or relatively new, engines, that is, engines with no
deterioration, whereas the older engines tested on a chassis dynamometer may have experienced
significant deterioration that would increase SOF emissions as a percent of DPM. One of the
main differences suspected for the lack of a decreasing trend in the percent of SOF in the engine
dynamometer studies is the test cycle used. The engine dynamometer tests typically include test
modes, such as high speed and low load, or low-speed lugging modes, that produce much higher
SOF relative to DPM than the driving cycles used on the chassis tests.
       It appears that as a fraction of total DPM, SOF from new model year FID diesel vehicles
is lower than that from older (pre-1990) HD diesel vehicles.  However, as with total DPM
emissions, a wide range in the fraction of SOF can be observed under different driving
conditions and from vehicles with extensive engine wear.  In general, DPM emissions have  a
lower fraction of organic matter compared to gasoline PM (Table 2-13).  Recent testing of HD
engines at the Desert Research Institute suggests that the OC fraction of DPM  is approximately
19%, whereas earlier studies reported in the U.S. EPA SPECIATE database  suggest a slightly
higher organic fraction of DPM from HD diesel  vehicles, ranging from 21% to 36%.  The
SPECIATE database represents older vehicles that, as discussed above, tend to have higher  SOF
emissions. The OC emissions from LD diesel vehicles recently reported by Norbeck et al.
(1998c) and those reported by the U.S. EPA SPECIATE suggest that LD diesel vehicles emit
DPM with a slightly higher organic content than that from HD diesel vehicles, ranging from 22%
to 43%. Gasoline engine PM emissions have recently been analyzed at the Desert Research
Institute by Fujita et al. (1998) and Watson et al. (1998) for hot stabilized, visibly smoking
vehicles, and cold-starts.  These data all indicate that LD gas vehicles emit PM with a higher
fraction of organic matter than diesel vehicles, with the highest organic content measured from
smoking and high-emitting gasoline vehicles (averaging 76% OC). One new finding from the
data reported by Fujita et al. (1998) is the roughly equivalent emission of organic and EC from
cold-start emissions of gasoline vehicles.  Additional information is needed to  characterize a
range  of OC for DPM from smoking and high-emitting diesel vehicles as well  as cold-start HD
diesel vehicles.

2.2.8.1.3. Trends in EC content.  Because EC is a major component of the chemical source
profile of DE, it is commonly used to determine  the contribution of diesel vehicles to ambient
PM samples (i.e., in source apportionment via chemical mass balance modeling).  EC is not,
strictly speaking, a regulated pollutant,  and so EC emissions are not routinely measured in tests
of diesel vehicles and engines.  The scant data available on measured EC emissions from HD

                                          2-69

-------
       Table 2-13. Organic and elemental carbon fractions of diesel and gasoline engine
       PM exhaust
Engine type
HD diesel engines"
HD diesel engines (SPECIATE)b
LD diesel engines0
LD diesel engines (SPECIATE)b
Gasoline engines (hot stabilized)3
Gasoline engines (smoker and high emitter)3'0
Gasoline engines (cold start)3
%OC
19±8
21-36
30 ±9
22-43
56±11
76 ± 10
46 ±14
%
Elemental
carbon
75 ± 10
52-54
61 ± 16
51-64
25 ± 15
7±6
42 ± 14
       a Fujita et al, 1998, and Watson et al., 1998.
       bU.S. EPA SPECIATE database.
       °Norbecketal., 1998c.
diesel vehicles are plotted in Figure 2-35. Different analytical methods were employed for these
studies, making the comparison of emission rates difficult. Results from the three studies, all
performed on HD trucks, suggest a decline in EC emission rates by model year since the early
1980s. In a study conducted in 1992, four HD vehicles of unknown vintage were tested and a
combined EC emission rate of 0.81 g/mi was reported, which is consistent with the 1990
timeframe in Figure 2-35 (Lowenthal et al., 1994).  EC as a percentage of total DPM in these
studies ranged from 30% to 90%, most likely as a result of different testing cycles and different
engines and different analytical methods.
       Figure 2-36 presents these data as EC fraction of total fine PM. The EC content of DPM
varied widely in the 1980s from approximately 20% to 90%, whereas in more recent years, the
data suggest a smaller range in the EC fraction, from approximately 50%  to 90% (with one data
point at 30%). Recent emission profiles for HD diesel vehicles suggest that 75% ± 10% of the
DPM  is attributable to  EC, whereas approximately  25% of gasoline PM is composed of EC,
except for PM emissions during gasoline vehicle cold-starts, which were found to have an EC
content of approximately 42% (Table 2-13). These data also provide evidence that newer model
year HD engines generally emit DPM that is more rich in EC than older HD engines.

2.2.8.2. PAHs and Nitro-PAH Emissions
       PAHs, nitro-PAHs, and oxidized derivatives of these compounds have attracted
considerable attention because of their known mutagenic and, in some cases, carcinogenic
                                         2-70

-------
                    14

                    12 H
                  1  10 -
I   8H

UJ   6 H
                 00
                 o
                 Q.
                    0 -
                          o o
                          •   Ooo
                              0 Q O O
                                     •  Benzo(a)pyrene
                                     o  1-Nitropyrene
                     1975       1980      1985      1990      1995

                                        Engine Model Year
                                                     2000
                    Figure 2-35. EC emission rates for diesel vehicles.


s
D.
3
o
1—
0
c£
o"
LU



IVJVJ
90
80 -
70 -

60
50 -

40 -
30
20

10 -
n

1
.
m
• *


T
0
• Zielinska, et al., 1998
O Schauer, et al., 1999
T Rogge, et al., 1993
                 82      84     86      88     90      92
                                     Engine Model Year
                                              94
96
Figure 2-36. EC content as percent of fine PM for DPM samples obtained in chassis
dynamometer studies.
                                           2-71

-------
character (National Research Council, 1982). In this section, PAH and nitro-PAH
concentrations and emission rates and trends in emissions over time are presented.

2.2.8.2.1.  PAHs identified in DE.  At least 32 PAHs have been identified in the exhaust of LD
diesel vehicles and HD diesel vehicles (Table 2-14) (Watson et al., 1998; Zielinska et al., 1998).
Table 2-15 lists the PAHs and thioarenes identified in three LD diesel vehicles' DPM extracts,
reported as ng/g of DPM (Tong et al., 1984). SOF fractions accounted for 11% to 15% of the
total DPM mass for the LD diesel vehicles reported by Tong et al. (1984), which is lower than
the LD diesel vehicles organic fraction reported by Norbeck et al. (1998c) in Table 2-13.
Among the PAHs reported by Watson et al. (1998) and Zielinska et al. (1998), the higher
molecular weight compounds (pyrene through coronene) that are expected to partition to the
particle phase have emission rates from HD diesel vehicles ranging from below detection limits
up to 0.071 mg/mi. HD diesel vehicle emission rates for the lower molecular weight PAHs
ranged up to 2.96 mg/mi for dimethylnaphthalenes. In general, among the vehicles tested, PAH
emission rates were higher for LD diesel vehicles compared with HD diesel vehicles. Table 2-16
presents emission rates of four representative particle-phase PAHs from HD diesel vehicles, LD
diesel vehicles, and gasoline (with and without catalytic converter) engines. Emission rates for
benzo[a]pyrene were higher in diesel emissions compared with gasoline emissions, except for
the report  by Rogge et al. (1993), who used extraction methods different from those in other
studies (discussed above).

2.2.8.2.2.  Nitro-PAHs identified in DE. Positive isomer identification for 16 nitro-PAHs has
been made utilizing the GC retention times of authentic standards and low- and high-resolution
mass spectra as identification criteria.  These include 1-nitropyrene; 2-methyl-l-nitronaph-
thalene; 4-nitrobiphenyl; 2-nitrofluorene; 9-nitroanthracene; 9-methyl-10- nitroanthracene; 2-
nitroanthracene; 2-nitrophenanthrene; l-methyl-9-nitroanthracene; l-methyl-3-nitropyrene; 1-
methyl-6-nitropyrene; l-methyl-8-nitropyrene;  1,3-, 1,6-, and 1,8-dinitropyrene; and 6-
nitrobenzo[a]pyrene. In addition, two nitrated heterocyclic compounds were identified,  5- and 8-
nitroquinoline.  Forty-five additional nitro-PAHs were tentatively identified in this diesel
particulate extract (Paputa-Peck et al., 1983). The concentration of nitro-PAHs adsorbed on
diesel particles varies substantially  from sample to sample.  Usually 1-nitropyrene is the
predominant component, and concentrations ranging from 7 to 165 |ig/g of particles are  reported
(Levson, 1988).
       Table 2-17 gives the approximate concentrations of several of the abundant nitro-PAHs
quantified in the early 1980s LD diesel particulate extracts (with the exception of
                                          2-72

-------
Table 2-14. Emission rates of PAH (mg/mi) from LD and HP diesel vehicles
PAH
Naphthalene
2-Menaphthalene
1-Menaphthalene
Dimethylnaphthalenes
Biphenyl
2-Methylbiphenyl
3 -Methylbipheny 1
4-Methylbiphenyl
Trimethylnaphthalenes
Acenaphthylene
Acenaphthene
Phenanthrene
Fluorene
Methylfluorenes
Methylphenanthrenes
Dimethylphenanthrenes
Anthracene
9-Methylanthracene
Fluoranthene
Pyrene
Methyl(pyrenes/fluoranthenes)
Benzonaphthothiophene
B enz [a] anthracene
Chrysene
Benz [b+j +k]fluoranthene
Benzo[e]pyrene
Benzo[a]pyrene
Indeno [ 1 ,2,3 -cd]pyrene
Dibenzo [a] anthracene
Benzo[b]chrysene
Benzo [ghijperlyne
Coronene
Light-duty
diesel
5. 554 ±0.282
3.068 ±0.185
2.313 ±0.134
5.065 ±0.333
0.743 ±0.041
0.203 ±0.015
1.048 ±0.063
0.447 ± 0.028
6.622 ±0.563
0.422 ± 0.024
0.096 ±0.008
1.411 ±0.072
0.442 ±0.038
1.021 ±0.091
1.115 ±0.064
0.637 ±0.047
0.246 ± 0.025
0.013 ±0.002
0.213 ±0.014
0.245 ± 0.020
0.548 ±0.045
0.002 ± 0.002
0.020 ± 0.005
0.029 ±0.005
0.056 ±0.005
0.019 ±0.003
0.013 ±0.004
0.010 ±0.003
0.002 ± 0.003
0.001 ±0.002
0.018 ±0.004
0.006 ± 0.006
Heavy-duty
diesel
2.451 ±0.154
2.234 ±0.152
1.582 ±0.103
2.962 ±0.488
0.505 ±0.037
0.049 ± 0.024
0.401 ±0.036
0.144 ±0.021
1.940 ±0.221
0.059 ±0.087
0.030 ±0.040
0.084 ±0.011
0.066 ± 0.022
0.071 ±0.055
0.124 ±0.069
0.090 ±0.096
0.052 ±0.016
0.434 ±0.082
0.044 ± 0.026
0.071 ±0.017
0.022 ± 0.082
0.001 ±0.027
0.066 ± 0.046
0.009 ±0.021
0.009 ± 0.022
0.010 ±0.014
0.013 ±0.044
0.001 ±0.037
0.000 ±0.053
0.001 ±0.027
0.013 ±0.048
0.001 ±0.095
                                 2-73

-------
Table 2-15. Polycyclic aromatic hydrocarbons identified in
 extracts of diesel particles from LD diesel engine exhaust
Compound
Acenaphthylene
Trimethylnaphthalene
Fluorene
Dimethylbiphenyl
C4 -Naphthalene
Trimethylbiphenyl
Dibenzothiophene
Phenanthrene
Anthracene
Methyldibenzothiophene
Methylphenanthrene
Methylanthracene
Ethylphenanthrene
4H-Cyclopenta[We/]phenanthrene
Ethyldibenzothiophene
2-Phenylnaphthalene
Dimethyl(phenanthrene/anthracene)
Fluoranthene
Benzo [
-------
Table 2-16. Emission rates of particle-bound PAH (ug/mi) from diesel and gasoline engines
PAH


Pyrene
Fluoranthene
Benzo[a]pyrene
Benzo[e]pyrene
Diesel engines
HDD
(a)
71
44
13
10
(b)
17.6
27.2
<0.1
0.24
(c)
36.2
20.8
2.1
4.2
LDD
(a)
245
213
13
19
(d)
66
50
NA
NA
Gasoline engines
Noncatalyst
(c)
49.6
77.3
69.6
73.3
(e)
45
32
3.2
4.8
Catalyst
(a)
248
196
1.0
1.0
(c)
4.0
3.6
3.0
3.6
(a) Watson et al., 1998 included gas-phase PAH .
(b) Westerholmetal., 1991.
(c) Roggeetal., 1993.
(d) Smith, 1989; 1986 Mercedes Benz.
(e) Alsbergetal., 1985.
3-nitrobenzanthrone, reported recently) in |ig/g of particles. Concentrations for some of the
nitro-PAHs identified range from 0.3 |ig/g for 1,3-dinitropyrene to 8.6 |ig/g for 2,7-dinitro-9-
fluorenone and 75 |ig/g for 1-nitropyrene. More recent nitro-PAH and PAH data for HD diesel
engines are reported in units of g/bhp-hr or mass/volume of exhaust, making it impossible to
directly compare them to the older data (Norbeck et al., 1998b; Bagley et al., 1996, 1998;
Baumgard and Johnson, 1992; Opris et al., 1993; Hansen et al., 1994; Harvey et al., 1994;
Kantola et al.,  1992; Kreso et al., 1998; McClure et al., 1992; Pataky et al., 1994).

2.2.8.2.3.  PAH and nitro-PAH emission changes over time.  It is difficult to compare PAH
emissions from different studies because  not all investigators analyze for total PAH or the same
suite of PAH compounds. Most studies have reported emissions of B[a]P or 1-nitropyrene (1-
NP) because of their toxicological activity. The results of chassis dynamometer studies in which
B[a]P or 1-NP were measured are displayed in Figure 2-37. Dietzmann and co-workers (1980)
examined four vehicles equipped with late 1970s turbocharged DI engines. Emissions of B[a]P
from particle extracts ranged from 1.5 to  9 |ig/mi. No relationship between engine technology
(one of the engines was two-stroke) and B[a]P emissions was observed.  Rogge and co-workers
(1993) reported total particle-associated PAH and B[a]P emissions from two  1987 model year
trucks (averaged together, four-stroke and turbocharged engines). The total particle-phase PAH
emission rate was 0.43 mg/mi and the B[a]P emission rate was 2.7 |ig/mi.  Particle-phase PAH in
the Rogge et al. (1993)  study accounted for approximately 0.5% of total DPM mass. Schauer
and co-workers (1999) recently reported a particle-phase PAH emission rate of 1.9 mg/mi
(accounting for about 0.7% of total DPM mass) for a 1995 MD turbocharged and aftercooled
truck. B[a]P emissions were not reported, but emissions of individual species of  similar
                                          2-75

-------
Table 2-17. Concentrations of nitro-PAHs identified in LD diesel
particulate extracts
Nitro-PAHa
4-nitrobiphenyl
2-nitrofluorene
2-nitroanthracene
9-nitroanthracene
9-nitrophenanthrene
3 -nitrophenanthrene
2-methyl-l-nitroanthracene
1 -nitrofluoranthene
7-nitrofluoranthene
3 -nitrofluoranthene
8 -nitrofluoranthene
1-nitropyrene
6-nitrobenzo[a]pyrene
l,3-dinitropyreneb
1 ,6-dinitropyreneb
l,8-dinitropyreneb
2,7-dinitrofluorene°
2,7-dinitro-9-fluorenone°
3 -nitrobenzanthroned
Concentration
(^g/g of
particles)
2.2
-1.8
4.4
1.2
1.0
4.1
8.3
1.8
0.7
4.4
0.8
18.9; 75b
2.5
0.30
0.40
0.53
4.2; 6.0
8.6; 3.0
0.6 to 6. 6
Trom Campbell and Lee (1984) unless noted otherwise. Concentrations recalculated from ug/g of
extract to ug/g of particles using a value of 44% for extractable material (w/w).
bFrom Paputa-Peck et al, 1983.
Trom Schuetzle, 1983.
Trom Enya et al., 1997  (Isuzu Model 6HEL 7127cc).
                                  2-76

-------
   14

   12 H
I
1  10

I   8

LU   6 -

5-   4 -
o
z   2 -

    0 -
                       o o
                       •   0 o o
                           0 Q O O
                                                  •   Benzo(a)pyrene
                                                  o   1-Nitropyrene
                  1975      1980       1985      1990
                                     Engine Model Year
                                          1995
2000
Figure 2-37. Diesel engine emissions of benzo[a]pyrene and 1-nitropyrene measured in
chassis dynamometer studies.

Source: Schuetzle and Perez, 1983; Zielinska et al, 1988; Kado et al., 1996; Dietzmann et al., 1980; Warner-Selph
and Dietzmann, 1984; Rogge et al., 1993; Schauer et al., 1999.
molecular weight were approximately 10 |ig/mi.  Schauer et al. (1999) also reported a gas-phase
PAH emission rate of 6.9 mg/mi for the same truck. Measurements of particle- and gas-phase
PAHs conducted for the Northern Front Range Air Quality Study in Colorado (Zielinska et al.,
1998) showed an average B[a]P emission rate of 13 |ig/mi for 15 vehicles ranging from 1983 to
1993 model years.  The combined gas- and particle-phase PAH emission rate reported for the
NFRAQS study was 13.5 mg/mi. B[a]P emissions from chassis studies are summarized in
Figure 2-37. Zielinska (1999) reports a decreasing trend in particle-associated DE PAH from 11
measurements made on vehicles from model year 1984 to 1993 with a low correlation coefficient
of 0.29.
       B[a]P emissions reported from diesel engine dynamometer studies are summarized in
Figure 2-38. Springer (1979) compared B[a]P emissions from naturally aspirated and
turbocharged engines and found that naturally aspirated engines emitted about 1 jig B[a]P/bhp-
hr, and DI and IDI  engines emitted about 0.15 jig B[a]P/bhp-hr (Table 2-8).  The difference
between 1 and 0.15 jig/bhp-hr could not be attributed to specific technology  changes.  The
majority of engine  test data indicate that B[a]P emissions have generally ranged from
approximately 1 to 4 |ig/bhp-hr over the past 25 years.
                                          2-77

-------
       Emissions reported for 1-NP from diesel engines tested by chassis dynamometer range
from 0.1 to 12 |ig/mi (Figure 2-37), and diesel engine dynamometer studies report 1-NP
emission factors ranging from 1 to 4 |ig/bhp-hr (Figure 2-38). Too few measurements are
available to discern trends in the emission rates of these compounds.
                  30
                  25 -
               O)
               3
               o
               1  20
               m
               01
               si
               g
               £
               01
               c
                U
                  15 H
                  10 -
               S.  5 -
               o
               N
               C
               Ol
               00
                   0 -
•  Transient Test B(a)P
o  Steady-State Test B(a)P
v  Transient Test 1-NP
v  Steady-State Test 1-NP
                   1970     1975     1980    1985    1990
                                     Engine Model Year
                                 1995
2000
Figure 2-38. Diesel engine dynamometer measurements of benzo[a]pyrene and 1-
nitropyrene emissions from HD diesel engines.

Source: Data are from Table 2-8.
       As discussed in Section 2.2.4, Williams et al. (1987) and Andrews et al. (1998) of the
University of Leeds have demonstrated that the solvent-extractable PAH from diesel particulate
originates primarily in the fuel. PAH molecules are relatively refractory, so a significant fraction
survives the combustion process and is exhausted as DPM.  These studies have been confirmed
by other research groups (Crebelli et al., 1995; Tancell et al., 1995) that included the use of
isotopic labeling of fuel PAH. Additionally, engine oil was found to be a reservoir for PAH that
originates in the fuel. Pyrosynthesis of PAH occurs during very high temperature conditions in a
diesel engine, and under these conditions many of the DPM and other pyrolysis products are
ultimately oxidized before exiting the cylinder. Thus, pyrogenic formation of PAH is thought to
contribute a small fraction of the total PAH in diesel engine exhaust. As discussed above, fuel
                                          2-78

-------
PAH content is expected to have slowly increased over a 30-year period until 1993, after which
PAH content of diesel fuel is expected to have remained constant. Increasing use of catalytic
cracking over time may lead to increasing proportions of PAH in distillates; however, fuel
standards limit the aromaticity of fuel to 35% (Section 2.2.4).
       Recently, Norbeck et al. (1998a) reported on the effect of fuel aromatic content on PAH
emissions.  Three diesel fuels were used in a Cummins L10 engine: pre-1993 fuel containing
33% aromatic HC and 8% PAH; low aromatic fuel containing a maximum content of 10%
aromatic HC and maximum of 1.4% PAH; and a reformulated fuel containing 20% to 25%
aromatic HC and 2% to 5% PAH. The investigators found that emission rates for the low-
molecular-weight PAHs (PAHs with three or fewer rings) were significantly lower when the
engine was tested using the low aromatic fuel compared to when the engine was run on the pre-
1993 or reformulated fuel (Table 2-18). Although emission rates reported for several higher
molecular weight (particle-associated) PAHs were lower (ranging from 4% to 28% lower) for
the low aromatic fuel compared with the other two fuels, the differences were not statistically
significant except for coronene.
       On the basis  of these limited data it is difficult to draw a precise, quantitative conclusion
regarding how PAH, B[a]P,  or 1-NP emissions have changed over time and in response to fuel
and engine  changes.  A decrease in the emissions of PAH from post-1990 model year vehicles
and engines compared with pre-1990 vehicles  and engines is suggested by the data; however, the
data also suggest that differences in a vehicle's engine type and make, general engine condition,
fuel composition, and test conditions can influence the emission levels of PAH.

2.2.8.3. Particle Size
       Figure 2-39 shows a generic size distribution for diesel particulate based on mass and
particle number.  Approximately 50%  to 90% of the number of particles in DE are in the
ultrafine size range (nuclei-mode), with the majority of diesel particles ranging in size from
0.005-0.05  (im and the mode at about 0.02 (im. These aerosol particles are formed from exhaust
constituents and consist of sulfuric acid droplets,  ash particles, condensed organic material, and
primary carbon spherules (Abdul-Khalek et al., 1998; Baumgard and Johnson, 1996). Although
it  accounts for the majority of particles, ultrafine DPM accounts for only 1% to 20% of the mass
ofDPM.
       Approximately 80% to 95% of diesel particle mass is in the size range from 0.05 to 1.0
(im, with a mean particle diameter of about 0.2 (im. The EC core has a high specific surface area
of approximately 30 to 50 m2/g (Frey and Corn, 1967), and Pierson and Brachaczek (1976)
report
that after the extraction of adsorbed organic material, the surface area of the diesel particle core

                                         2-79

-------
Table 2-18.  Average emission rates for polycyclic aromatic hydrocarbons for different fuel
types (units are ug/bhp-hr)
PAH
2,3,5-trimethyl naphthalene
Phenanthrene
Anthracene
Methylphenanthrenes/anthracenes
Fluoranthene
Pyrene
Benzo [c]phenanthrene
Benzo [ghijfluoranthene
Cyclopenta[cd]pyrene
Benz [a] anthracene
Chrysene + triphenylene
Benzo [b+j+k]fluoranthene
Benzo [e]pyrene
Benzo [a]pyrene
Perylene
Indeno [ 1 ,2,3 -cd]fluoranthene
Benzo[c]chrysene
Dibenz [a,h] anthracene
Indeno [ 1 ,2,3 -cd]pyrene
Dibenz [a,h+a,c] anthracene
Benzo [b]chrysene
Benzo [ghijperylene
Coronene
Dibenzo [a,l]py rene
Dibenzo [a,e]py rene
Dibenzo [a,i]py rene
Dibenzo [a,h]pyrene
Pre-1993 diesel
fuel
Cetane No. >40
Aromatic 33% v.
PAH 8% wt.
283.68 ±5.27
336.71 ±9.08
38.89 ±1.43
331.32 ±16.07
128.45 ±7.60
193.03 ±16.51
3.03 ±0.24
24.84 ±2.68
21.44±4.11
16.42 ±1.67
17.36 ±1.66
31.05 ±4.17
16.71 ±2.72
20.46 ±3.27
4.32 ±0.88
0.34 ±0.07
0.29 ±0.05
0.93 ±0.05
19.45 ±2.71
1.54 ±0.15
0.40 ±0.01
49.17 ±9.63
9.49 ±3. 13
2.84 ±0.45
1.10 ±0.29
0.91 ±0.21
1.33 ±0.25
Low aromatic
diesel fuel
Cetane No. >48
Aromatic 10% v.
PAH 1.4%wt.
14.77 ±2.42
160.92 ± 15.54
18.54 ±2.13
25. 17 ±1.41
132.36 ± 18.30
211.19±37.35
1.74 ±0.14
18.93 ±2.14
26.15 ±3. 12
10.57 ± 1.15
10.38 ±0.54
23. 17 ±1.98
14.55 ± 1.34
16.48 ±1.56
3.71 ±0.74
0.21 ±0.02
0.18 ±0.05
0.55 ±0.10
14.04 ±1.99
0.87 ±0.12
0.15 ±0.05
39.81 ±7.22
4.93 ±0.47
1.25 ±0.15
0.61 ±0.06
0.27 ±0.09
0.75 ± 0.07
Reformulated diesel
blend
Cetane No. 50-55
Aromatic 20%-25%v.
PAH2%-5%wt.
56.21 ±2.82
220.73 ± 52.68
26.16 ±6.86
111. 98 ±28.74
123.07 ±26.21
206.82 ± 39.04
1.54 ±0.26
16.94 ±2.31
21.25 ±3.46
10.96 ±2.42
12.20 ±2.72
29.18 ±7.93
18.99 ±5. 58
20.59 ±5.75
4.18± 1.16
0.17 ±0.00
0.14 ±0.04
0.67 ±0.09
22.16 ±9.11
1.48 ±0.67
0.27 ±0.05
60.74 ± 26.60
7.48 ± 1.59
2.31 ±0.48
1.13±0.15
0.71 ±0.15
0.84 ±0.20
Source: Norbeck et al., 1998a.
                                         2-80

-------
         a>
         o
         =5
         c
         o
Nanoparticles
Dp < 50 nm
                                                       Fine Particles
                                                       Dp < 2.
                              H
           0.001
                            0.010
                                             Ultrafine Particles
                                             Dp <100 nm
                                              /  Accumulation
                                                   Mode
                                    PM10
                                  Dp< 10 n
                                                                         Coarse
                                                                         Mode
                                             0.100
                                          Diameter (p.m>
                                                              1.000
                                                                               10.000
                                   •Mass Weighting	Number Weighting (
      Figure 2-39. Particle size distribution in DE.

      Source: Kittelson, 1998.
is approximately 90 m2/g.  Because these particles have a very large surface area per gram of
mass, it makes them excellent carriers for adsorbed inorganic and organic compounds;
potentially enhancing penetration of such compounds to lower portions of the respiratory tract
upon inhalation. In addition, ultrafine aerosols can also reach the same areas of the lung.
       Considerable caution is required when reporting particle size measurements from diesel
engine exhaust because dilution conditions during the measurement process significantly affect
size distributions (i.e., the size distribution is largely a function of how it was measured), and
DPM size distributions obtained in dilution tunnel systems may not be relevant to size
distributions resulting from the physical transformation of engine exhaust in the atmosphere.
Measurements made on diluted DE typically show higher numbers of nuclei-mode particles than
do measurements made on raw exhaust because of condensation to form nuclei-mode aerosol
upon cooling of the exhaust.  To understand particle size distributions emitted from diesel
engines, investigators employ various dilution techniques, none of which have been
standardized. Dilution ratio, sampling temperature, humidity, relative concentrations of carbon
and volatile matter, and other sampling factors can therefore have a large impact on the number
and makeup of nuclei-mode particles (Abdul-Khalek et al., 1999; Shi and Harrison, 1999; Liiders
                                           2-81

-------
et al., 1998; Brown et al., 2000). Dilution air temperature and humidity can have a large effect
on particle number and size distribution, especially in the size range below 0.05 (im (also
referred to as nanoparticles).  Shi and Harrison (1999) report that a high dilution ratio and high
relative humidity favor the production of ultrafine particles in diesel engine exhaust.
Abdul-Khalek et al. (1998) report that an increase in the residence time of the exhaust during
dilution resulted in an increase in the number of particles in exhaust.  Khatri et al. (1978) report
increased gas-phase HC condensation to DPM with a decrease in dilution air temperature. Some
studies report no peak in diesel particles in the ultrafine size range (Kleeman et al., 2000).
Kittelson (2000) reports that nanoparticle formation can be prevented by an oxidizing catalyst,
which burns organic components of the exhaust, making them unavailable for nucleation or
condensation to form an aerosol.
       Experiments conducted in a dilution tunnel represent the atmospheric behavior of DE
only under the conditions specific to the dilution tunnel and do not represent the full range of
atmospheric conditions.  Gertler (1999) demonstrated an increase in 0.02 (im particles as the
fraction of diesel vehicles in the  Tuscarora Mountain tunnel increased from 13% to 78%. These
data suggest that the mode at 0.02 jim for ultrafine DPM from DE is evidenced under some
real-world conditions.
       Several groups have shown that decreasing sulfur content decreases the number of
nuclei-mode particles measured in the exhaust, assuming temperature is low enough and
residence time is long enough for nucleation and condensation of sulfate aerosol and water in the
dilution tunnel (Baumgard and Johnson, 1992, 1996; Opris et al., 1993; Abdul-Khalek et al.,
1999). The application of this finding to real-world conditions is difficult to predict, as the
number of nuclei-mode particles formed from sulfate and water in the atmosphere will be
determined by atmospheric conditions, not by dilution tunnel conditions.  With all other factors
held constant, it appears that reducing fuel sulfur content reduces the number of sulfate nuclei-
mode particles. Thus, the reduction in on-road fuel sulfur content that occurred in 1993  reduced
the amount of sulfur dioxide and sulfate available for particle formation. As discussed above,
the contribution of sulfate to total DPM mass ranges from 1% to 5% and is therefore not a
substantial portion of DPM mass.
       More controversial is the suggestion that the DPM emission size distribution from newer
technology engines (1991 and later) may be shifted to a much higher number concentration of
nuclei-mode particles, independent of fuel sulfur content (Kreso et al., 1998; Abdul-Khalek
et al., 1998; Baumgard and Johnson, 1996; Bagley et al.,  1996). For example, Kreso and co-
workers (1998) compared emissions from a 1995 model year engine with measurements made on
1991 and 1988 model year engines in earlier studies (Bagley et al., 1993, 1996). Nuclei-mode
particles made up 40% to 60% of the number fraction of DPM emissions for the 1988 engine and

                                          2-82

-------
97%+ of the DPM from the 1991 and 1995 engines. Number concentrations were roughly two
orders of magnitude higher for the newer engines.  SOF made up 25% to 30% of DPM mass in
the 1988 engine and 40% to 80% of DPM mass for the newer engines.  Total DPM mass was
significantly reduced for the newer engines. It was suggested that increased fuel injection
pressure leads to improved fuel atomization and evaporation, in turn leading to smaller primary
carbonaceous particles. Dilution conditions (relatively low temperature, low primary dilution
ratio, long residence time of more than 3 seconds)  strongly favor the formation of nucleation
products. The 1991 and  1988 engines were tested  with 100 ppm  sulfur fuel whereas the 1995
engine was tested with 310 ppm sulfur fuel, which may confound the results to some extent.
       The results of Kreso and co-workers (1998) and of Bagley and co-workers (1993, 1996)
have been called into question because the high level of SOF emitted by the 1991 engine,
particularly at high-load test modes, was inconsistent with SOF values measured for other
engines using similar types of technology (Last et al.,  1995; Ullman et al., 1995). Kittelson
(1998) notes that there is far less carbonaceous DPM formed in newer engines compared with
older engines.  Accumulation-mode particles may have provided  a high surface area for
adsorption of sulfate and unburned organic compounds. In the absence  of this surface area for
adsorption, higher number concentrations of small particles are formed from nucleation of HCs
and sulfuric acid.
       A study performed at EPA by Pagan (1999) suggested that increased injection pressure
can lead to the formation of more nuclei-mode particles in the exhaust.  Particle size distributions
were measured for diluted exhaust from an engine  in which injection pressure could be varied
from roughly 35 to 110 MPa (about 5,000-16,000  psi), comparable to pressures obtained with
injection technology introduced in the 1980s.  The dilution system and particle size measurement
setup were identical in all experiments, removing some of the uncertainty in earlier studies that
compared engine tests performed years apart.  The results showed a clear increase in the number
of nuclei-mode particles and a decrease in the  number of accumulation-mode particles as
injection pressure was increased. This shift did not occur, however, at high engine speeds and
loads, but only at low to intermediate speeds and loads. The increase in number concentration of
nuclei-mode particles was much lower than the two orders of magnitude increase reported by
Kreso et al. (1998) or Bagley et al.  (1996). One must use caution in applying the results of
Pagan to modern high-injection pressure diesel engines with turbocharging/charge-air cooling
because the engine used by Pagan was a naturally aspirated engine to which high-pressure
common rail injection was applied. This would likely preclude this particular engine from
meeting current on-highway PM or NOX standards. Although some studies have suggested that
increased injection pressure can lead to elevated ultrafme DPM number counts, Kittelson et al.
                                         2-83

-------
(1999) cite a German study that reported a decrease in ultrafme DPM number and mass with
increasing injection pressure.
       Although the majority of particles in DE from modern on-road diesel engines are in the
ultrafme size range, evidence regarding a change in the size distribution over time is unclear. To
understand the size distribution of DPM to which people are exposed will require measurements
under conditions that more closely resemble ambient conditions.

2.3. ATMOSPHERIC TRANSFORMATION OF DIESEL EXHAUST
       Primary diesel emissions are a complex mixture containing hundreds of organic and
inorganic constituents in the gas and particle phases, the most abundant of which are listed in
Table 2-19. The more reactive compounds with short atmospheric lifetimes will undergo rapid
transformation in the presence of the appropriate reactants, whereas more stable pollutants can
be transported over greater distances. A knowledge of the atmospheric transformations of
gaseous and particulate components of diesel emissions and their fate is important in assessing
environmental exposures and risks.  This section describes some of the major atmospheric
transformation processes for gas-phase and particle-phase DE, focusing on the primary and
secondary organic compounds that are of significance for human health.  For a more
comprehensive summary of the atmospheric transport and transformation of diesel emissions,
see Winer and Busby (1995).

2.3.1. Gas-Phase Diesel Exhaust
       Gas-phase DE contains organic and inorganic compounds that undergo various chemical
and physical transformations in the atmosphere, depending on the abundance of reactants and
meteorological factors such as wind speed and direction, solar radiation, humidity, temperature,
and precipitation. Gaseous DE will react primarily with the following species (Atkinson, 1988):

       •   Sunlight, during daylight hours
           Hydroxyl (OH) radical, during daylight hours
           Ozone (O3), during daytime and nighttime
       •   Hydroperoxyl (HO2) radical, typically during afternoon/evening hours
       •   Gaseous nitrate (NO3) radicals or dinitrogen pentoxide (N2O5), during nighttime
           hours
           Gaseous nitric acid (HNO3) and other species such as nitrous acid (HONO) and
           sulfuric acid (H2SO4).
                                         2-84

-------
       Table 2-19.  Classes of compounds in diesel exhaust
      Participate phase
                     Gas phase
      Heterocyclics, hydrocarbons (C14-C35), and
      PAHs and derivatives:
      Acids
      Alcohols
      Alkanoic acids
      n-Alkanes
      Anhydrides
      Aromatic acids
Cycloalkanes
Esters
Halogenated cmpds.
Ketones
Nitrated cmpds.
Sulfonates
Quinones
Heterocyclics, hydrocarbons
derivatives:
Acids
Aldehydes
Alkanoic acids
n-Alkanes
n-Alkenes
Anhydrides
Aromatic acids
                                                  ), and
      Elemental carbon
      Inorganic sulfates and nitrates
      Metals
      Water
                  Cycloalkanes, Cycloakenes
                  Dicarbonyls
                  Ethyne
                  Halogenated cmpds.
                  Ketones
                  Nitrated cmpds.
                  Sulfonates
                  Quinones
Acrolein
Ammonia
Carbon dioxide, carbon monoxide
Benzene
1,3 -Butadiene
Formaldehyde
Formic acid
Hydrogen cyanide, hydrogen sulfide
Methane, methanol
Nitric and nitrous acids
Nitrogen oxides, nitrous oxide
Sulfur dioxide
Toluene
Water _
       Sources: Mauderly, 1992, which summarized the work of Lies et al., 1986; Schuetzle and Frazier,
       1986; Carey, 1987; Zaebst et al., 1988, updated from recent work by Johnson, 1993; McDonald,
       1997;Schaueretal., 1999.

       The major loss process for most of the DE emission constituents is oxidation, which
occurs primarily by daytime reaction with OH radical (Table 2-20). For some pollutants,
photolysis, reaction with O3, and reactions with NO3 radicals during nighttime hours are also
important removal processes. The atmospheric lifetimes do not take into consideration the
potential chemical or biological importance of the products of these various reactions. For
example, the reaction of gas-phase PAHs with NO3 appears to be of minor significance as a PAH
loss process, but it is more important as a route of formation of mutagenic nitro-PAHs. The
reaction products for  some of the major gaseous DE compounds are listed in Table 2-21 and are
discussed briefly below.


 2.3.1.1.  Organic Compounds
       The organic fraction of diesel is a complex mixture of compounds, very few of which
have been characterized.  The atmospheric chemistry of several organic constituents of DE
(which are also produced by other combustion sources) has been studied. A few of these
                                            2-85

-------
Table 2-20. Calculated atmospheric lifetimes for gas-phase reactions of selected
compounds present in automotive emissions with important reactive species
Compound
NO2
NO
HNO3
S02
NH3
Propane
n-Butane
n-Octane
Ethylene
Propylene
Acetylene
Formaldehyde
Acetaldehyde
Benzaldehyde
Acrolein
Formic acid
Benzene
Toluene
m-Xylene
Phenol
Naphthalenef
2-Methylnaphthalenef
1 -Nitronaphthalenef
Acenaphthenef
Acenaphthylenef
Phenanthrenef
Anthracenef
Fluoranthenef
Pyrenef
Atmospheric lifetime resulting from reaction with:
OH"
1.3 days
2.5 days
110 days
16 days
90 days
12 days
5.6 days
1.9 days
1.9 days
7h
19 days
1.9 days
0.6 day
1.2 days
0.6 day
3 1 days
1 1 days
2.5 days
7h
6h
6.8 h
2.8 h
2.3 days
1.5 h
1.3 h
11.2h
8.6 h
-2.9 h
-2.9 h
03b
12 h
1 min
—
>200 years
—
>7,000 years
>4,500 years
—
9 days
1.5 days
6 years
>2 - 104 years
>7 years
—
60 days
—
600 years
300 years
75 years
—
>80 days
>40 days
>28 days
>30 days
-43 min
41 days
—
—
—
NO3C
24 min
1.2 min
—
>1.4xl04years
—
—
3. 6 years
1.2 years
1.2 years
6 days
>5. 6 years
84 days
20 days
24 days
—
—
>6.4 years
3. 6 years
0.8 years
8 min
1.5 years
180 days
1 8 years
1.2 h
6 min
4.6 h
—
-1 year
- 120 days
HO2d
2h
20 min
—
>600 years
—
—
—
—
—
—
—
23 days
—
—
—
—
—
—
—
—
—


—
—
—
—
—
—
hve
2 min
—
—
—
—
—
—
—
—
—
—
4h
60 h
—
—
—
—
—
—
—
—
—
1.7 h
—
—
—
—
—
—
a For 12-h average concentration of OH radical of 1.6* 106 molecule/cm3 (Prinn et al., 1992).
b For 24-h average O3 concentration of 7x 101' molecule/cm3.
0 For 12-h average NO3 concentration of 5><108 molecule/cm3 (Atkinson, 1991).
d For 12-h average HO2 concentration of 108 molecule/cm3.
e For solar zenith angle of 0 °.
f Lifetimes from Arey (1998), for 12-h concentration of OH radical of 1.9xl06 molecule/cm3.

Source: Winer and Busby, 1995, unless noted otherwise.
                                              2-86

-------
   Table 2-21. Major components of gas-phase diesel engine emissions, their
   known atmospheric transformation products, and the biological impact of the
   reactants and products
Gas-phase emission
component
Carbon dioxide
Carbon monoxide
Oxides of nitrogen
Sulfur dioxide
Atmospheric reaction
products
—
—
Nitric acid, ozone
Sulfuric acid
Biological impact
Major contributor to global
warming.
Highly toxic to humans; blocks
oxygen uptake.
Nitrogen dioxide is a respiratory
tract irritant and major ozone
precursor. Nitric acid contributes to
acid rain.
Respiratory tract irritation.
Contributor to acid rain.
Hydrocarbons:
Alkanes (
-------
 2.3.1.1.  Organic Compounds
       The organic fraction of diesel is a complex mixture of compounds, very few of which
have been characterized. The atmospheric chemistry of several organic constituents of DE
(which are also produced by other combustion sources) has been studied. A few of these
reactions and their products are discussed below. For a complete summary  of the atmospheric
chemistry of organic combustion products, see Seinfeld and Pandis (1998).
       Acetaldehyde forms peroxyacetyl nitrate (via formation of peroxyl radicals and reaction
with NO2), which has been shown to be a direct-acting mutagen toward S. typhimurium strain
TA100 (Kleindienst et al., 1985) and is phytotoxic. Benzaldehyde, the simplest aromatic
aldehyde, forms peroxybenzoyl nitrate or nitrophenols following reaction with oxides of nitrogen
(Table 2-21).
       For those PAHs present in the gas phase, reaction with the OH radical is the major
removal route, leading to atmospheric lifetimes of a few hours in daylight.  The gas-phase
reaction of PAHs containing a cyclopenta-fused ring such as acenaphthene,  acenaphthylene, and
acephenanthrylene with the nitrate radical may be an important loss process during nighttime
hours.  Relatively few data are available concerning the products of these gas-phase reactions. It
has been shown that in the presence of NOX, the OH radical reactions with naphthalene,  1- and 2-
methylnaphthalene, acenaphthylene, biphenyl, fluoranthene, pyrene, and acephenanthrylene lead
to the formation of nitroarenes (Arey et al., 1986; Atkinson, 1986; Atkinson et al., 1990;
Zielinska et al., 1988, 1989a; Arey, 1998). In addition, in a two-step process involving OH
radical reaction and NO2 addition, 2-nitrofluoranthene and 2-nitropyrene can be formed and
eventually partition to the particle phase, as will other nitro-PAHs.
       The addition of the NO3 radical to the PAH aromatic ring leads to nitroarene formation
(Sweetman et al., 1986; Atkinson et al., 1987, 1990; Zielinska et al., 1989a). The gas-phase
reactions of NO3 radical with naphthalene, 1- and 2-methylnaphthalene, acenaphthene,
phenanthrene, anthracene, fluoranthene, and  pyrene produce, in general, the same nitro-PAH
isomers as the OH radical reaction, but with different yields (Arey et al., 1989; Sweetman et al.,
1986; Atkinson et al., 1987, 1990; Zielinska et al., 1986, 1989a). For example, the same 2-
nitrofluoranthene is produced from both OH radical and NO3 gas-phase reactions, but the
reaction with NO3 produces a much higher yield. The production of several nitroarene
compounds has been studied in environmental chambers (Arey et al., 1989;  Zielinska et al.,
1990; Atkinson and Arey, 1994; Arey, 1998; Feilberg et al., 1999), and generally the same nitro-
PAH isomers formed from reaction with OH and NO3 radicals are observed in ambient air
samples. Secondary formation of nitroarenes through the  gas-phase reactions of the 2-, 3-, and
4-ring PAHs is the major source for many of the nitroarenes observed in ambient air (Pitts et al.,
1985a-c; Arey et al., 1986; Zielinska et al., 1988). Photolysis is the major removal pathway for

                                            255
                                           -OO

-------
nitroarenes with lifetimes of approximately 2 hours (Feilberg et al., 1999; Nielsen and Ramdahl,
1986).

2.3.1.2. Inorganic Compounds
       SO2 and oxides of nitrogen (primarily NO) are emitted from diesel engines.  SO2 is
readily oxidized by the OH radical in the atmosphere, followed by formation of the HO2 radical
and HSO3, which rapidly reacts with water to form H2SO4 aerosols. Because SO2 is soluble in
water, it is scavenged by fog, cloud water, and raindrops.  In aqueous systems, SO2 is readily
oxidized to sulfate by reaction with hydrogen peroxide (H2O2), O3, or O2 in the presence of a
metal catalyst (Calvert and Stockwell, 1983).  Sulfur emitted from diesel engines is
predominantly (-98%) in the form of SO2, a portion of which will form sulfate aerosols by the
reaction described above. Nonroad equipment, which typically uses fuel containing 3,300 ppm
sulfate, emits more SO2 than on-road diesel engines, which use fuels currently containing an
average of 340 ppm sulfur because of EPA regulations effective in 1993 decreasing diesel fuel
sulfur levels. EPA estimates that mobile sources are responsible for about 7% of nationwide SO2
emissions, with diesel engines contributing 74% of the mobile source total (the majority of the
diesel SO2 emissions originate from nonroad engines) (U.S. EPA, 1998b).
       NO is also oxidized in the atmosphere to form NO2 and particulate nitrate. The fraction
of motor vehicle NOX exhaust converted to particulate nitrate in a 24-hour period has been
calculated using a box model to be approximately  3.5% nationwide, a portion of which can be
attributed to DE (Gray and Kuklin, 1996). EPA estimates that in 1997, mobile sources were
responsible for about 50% of nationwide NOX emissions, with diesel engines being responsible
for approximately one-half of the mobile source total (U.S. EPA, 1998b).

2.3.1.3. Atmospheric Transport of Gas-Phase DE
       Gas-phase DE can be dry deposited, depending on the deposition surface, atmospheric
stability, and the solubility and other chemical properties of the compound.  Dry deposition of
organic species is typically on the order of weeks to months, with dry deposition velocities of
approximately  10"4 cm/sec (Winer and Busby, 1995).  In contrast, inorganic species  such as SO2
and nitric acid have relatively fast deposition rates (0.1-2.5 cm/sec) and will remain in the
atmosphere for shorter time periods compared with the organic exhaust components. Some gas-
phase species will also be scavenged by aqueous aerosols and potentially deposited via
precipitation.  These processes can greatly reduce the atmospheric concentration of some vapor-
phase species.  Atmospheric lifetimes for several gas-phase components of DE are on the order
of hours or days, during which time atmospheric turbulence and advection can disperse these
pollutants widely.

                                         2-89

-------
2.3.2. Particle-Phase Diesel Exhaust
       Particle-associated DE is composed of primarily carbonaceous material (organic and EC)
with a very small fraction composed of inorganic compounds and metals.  The OC fraction
adsorbed on DPM is composed of high-molecular-weight compounds, such as PAHs, which are
generally more resistant to atmospheric reactions than PAHs in the gas phase. The EC
component of DE is inert to atmospheric degradation, whereas the PAH compounds are
degraded by reaction with the following species:

             Sunlight, during daytime hours
             O3, during daytime and nighttime
       •      NO3 and N2O5, during nighttime hours
             OH and HO2
             NO2, during nighttime and daytime hours
             H202
       •      HNO3 and other species such as HONO and H2SO4.

       Because many of the PAH derivatives formed by reaction with some of the reactants
listed above have been found to be highly mutagenic, a brief discussion of PAH photolysis,
nitration, and oxidation follows.  Some of the major degradation products  from particulate DE
and their biological impact are listed in Table 2-22.

2.3.2.1. Particle-Associated PAH Photooxidation
       Laboratory studies  of photolysis of PAHs adsorbed on 18 different fly ashes, carbon
black, silica gel, and alumina (Behymer and Kites, 1985, 1988) and several coal stack ashes
(Yokely et al., 1986; Dunstan et al., 1989) have shown that the extent of photodegradation of
PAHs depends very much on the  nature of the substrate to which they are  adsorbed.  The
dominant factor in the stabilization of PAHs adsorbed on fly ash was the color of the fly ash,
which is related to the amount of carbon black present.  It appears that PAHs were stabilized if
the carbon black content of the fly ash was greater than approximately 5%. On black substrates,
half-lives of PAHs studied were on the order of several days (Behymer and Kites, 1988). The
environmental chamber studies of Kamens et  al. (1988) on the daytime decay of PAHs present
on residential wood smoke particles and on gasoline internal combustion emission
particles showed PAH half-lives of approximately 1 hour at moderate humidities and
temperatures.  At very low angle  sunlight, very low water vapor concentration, or very low
temperatures, PAH daytime half-lives increased to a period of days.  The presence and
                                         2-90

-------
       Table 2-22. Major components of particle-phase diesel engine emissions,
       their known atmospheric transformation products, and the biological impact
       of the reactants and products
Particle-phase emission
component
Elemental carbon
Inorganic sulfate and
nitrate
Hydrocarbons (C14-C35)
PAHs (> 4 rings) (e.g.,
pyrene, benzo[a]pyrene)
Nitro-PAHs(>3rings)
(e.g., nitropyrenes)
Atmospheric reaction
products
—
—
Little information;
possibly aldehydes, ketones,
and alkyl nitrates
Nitro-PAHs(>4rings)a
Nitro-PAH lactones
Hydroxylated-nitro derivatives
Biological impact
Nuclei adsorb organic compounds;
size permits transport deep into the
lungs (alveoli)
Respiratory tract irritation
Unknown
Larger PAHs are major contributors
of carcinogens in combustion
emissions. Many nitro-PAHs are
potent mutagens and carcinogens.
Many nitro-PAHs are potent
mutagens and carcinogens. Some
reaction products are mutagenic in
bacteria (Ames assay).
     aNitro-PAHs with more than two rings will partition into the particle phase.
     Source: Health Effects Institute, 1995.
composition of an organic layer on the aerosol seems to influence the rate of PAH photolysis
(Jang and McDow, 1995; McDow et al., 1994; Odum et al., 1994).
       Because of limited understanding of the mechanisms of these complex heterogeneous
reactions, it is currently impossible to draw any firm conclusion concerning the photostability of
particle-bound PAHs in the atmosphere. Because DPM contains a relatively high quantity of
EC, it is reasonable to speculate that PAHs adsorbed onto these particles might be relatively
stable under standard atmospheric conditions, leading to an anticipated half-life of 1 or more
days.

2.3.2.2. Particle-Associated PAH Nitration
       Since 1978, when Pitts et al. (1978) first demonstrated that B[a]P deposited on glass-
fiber filters exposed to air containing 0.25  ppm NO2 with traces of HNO3 formed nitro-B[a]P,
numerous studies of the heterogeneous nitration reactions of PAHs adsorbed on a variety of
substrates in different simulated atmospheres have been carried out (Finlayson-Pitts and Pitts,
1986).  PAHs deposited on glass-fiber and Teflon-impregnated glass-fiber filters react with
gaseous N2O5, yielding their nitro derivatives (Pitts et al., 1985b,c).  The most abundant isomers
formed were 1-NP from pyrene, 6-nitro-B[a]P from B[a]P, and 3-nitroperylene from perylene.
                                          2-91

-------
       The formation of nitro-PAHs during sampling may be an important consideration for
DPM collection because of the presence of NO2 and HNO3 (Feilberg et al., 1999). However,
Schuetzle (1983) concluded that the artifact formation of 1-NP was less than 10% to 20% of the
1-NP present in the diesel particles if the sampling time was less than 23 min (one FTP cycle)
and if the sampling temperature was not higher than 43 °C. The formation of nitroarenes during
ambient high-volume sampling conditions has been reported to be minimal, at least for the most
abundant nitropyrene and nitrofluoranthene isomers (Arey et al., 1988).
       DPM contains a variety of nitroarenes, with 1-NP being the most abundant among
identified nitro-PAHs. The concentration of 1-NP was measured in the extract of particulate
samples collected at the Allegheny Mountain Tunnel on the Pennsylvania Turnpike as 2.1 ppm
and ~5 ppm by mass of the extractable material from diesel and SI vehicle PM, respectively.
These values are much lower than would be predicted on the basis of laboratory measurements
for either diesel or SI engines (Gorse et al., 1983). Several nitroarene measurements have been
conducted in airsheds heavily affected by motor vehicle emissions (Arey et al., 1987; Atkinson
et al., 1988; Zielinska et al., 1989a,b; Ciccioli et al., 1989, 1993). Ambient PM samples were
collected at three sites in the Los Angeles Basin during two summertime periods and one
wintertime period.  Concentrations of 1-NP ranged from 3 pg/m3, to 60 pg/m3, and 3-
nitrofluoranthene was also present in DPM at concentrations ranging from not detectable to  70
pg/m3.

2.3.2.3. Particle-Associated PAH Ozonolysis
       Numerous laboratory studies have shown that PAHs deposited on combustion-generated
fine particles and on model substrates undergo reaction with O3 (Katz et al., 1979; Pitts et al.,
1980, 1986; Van Vaeck and Van Cauwenberghe, 1984; Finlayson-Pitts and Pitts, 1986).  The
dark reaction toward O3 of several PAHs deposited on model substrates has been shown to be
relatively fast under simulated atmospheric conditions (Katz et al.,  1979; Pitts et al., 1980, 1986).
Half-lives on the order of 1 to several hours were reported for the more reactive PAHs, such as
B[a]P,  anthracene, and benz[a]anthracene (Katz et al., 1979).
       The reaction of PAHs deposited on diesel particles with 1.5 ppm O3 under high-volume
sampling conditions has been shown to be relatively fast, and half-lives on the order of 0.5 to 1
hour have been reported for most PAHs studied (Van Vaeck and Van Cauwenberghe, 1984).
The most reactive PAHs include B[a]P, perylene,  benz[a]anthracene, cyclopenta[cd]pyrene, and
benzo[ghi]perylene. The benzofluoranthene isomers are the least reactive of the PAHs studied,
and benzo[e]perylene is less reactive than its isomer B[a]P.  The implications of this study for
the high-volume sampling ambient POM are important: reaction of PAHs with O3 could
possibly occur under high-volume sampling conditions during severe photochemical smog

                                         2-92

-------
episodes, when the ambient level of O3 is high. However, the magnitude of this artifact is
difficult to assess from available data.

2.3.2.4. Atmospheric Transport ofDE Particulate Matter
       Ultrafine particles emitted by diesel engines undergo nucleation, coagulation, and
condensation to form fine particles.  DPM can be removed from the atmosphere by dry and wet
deposition.  Particles of small diameter (<1 jim), such as DPM, are removed less efficiently than
larger particles by wet and dry deposition and thus have longer atmospheric residence times.
Dry deposition rates vary depending on the particle size.  Because of their small size, DE
particles have residence times of several days (dry deposition velocities of approximately 0.01
cm/sec) (Winer and Busby, 1995). Diesel particulates may be removed by wet deposition if they
serve as condensation nuclei for water vapor deposition or are scavenged by precipitation in- or
below-cloud.
       In a study designed to assess the atmospheric concentrations and transport of DE
particles, Horvath et al.  (1988) doped the sole source of diesel fuel in Vienna with an
organometallic compound of the heavy earth element dysprosium. The authors found that in
some of the more remote sampling areas, DPM composed more than 30% of the paniculate
mass, indicating that DPM can be dispersed widely.

2.3.3. Diesel Exhaust Aging
       Primary DE is considered "fresh," whereas "aged" DE is considered to have undergone
chemical and physical transformation and dispersion over a period of a day or two.  Laboratory
dilution tunnel measurements represent a homogeneous environment compared to the complex
and dynamic system into which real-world DE is emitted. The physical and chemical
transformation of DE will vary depending on the environment into which it is emitted. In an
urban or industrial environment, DE may enter an atmosphere with high concentrations of
oxidizing and nitrating radicals, as well as nondiesel organic and inorganic compounds that may
influence the toxicity, chemical stability, and atmospheric residence time.
       In general, secondary pollutants formed in an aged aerosol mass are more oxidized, and
therefore have increased polarity and water solubility (Finlayson-Pitts and Pitts, 1986). Kamens
et al. (1988) reported that photooxidation of particle-bound PAH is enhanced as relative
humidity is increased. Weingartner et al. (1997a) and Dua et al. (1999) have reported that unlike
many other types of particles, diesel particles do not appear to undergo hygroscopic growth once
emitted to the atmosphere and may even shrink in size to some extent under increasing relative
humidity conditions. Weingartner et al. (1997a) evaluated the hygroscopic growth of diesel
particles and found that freshly emitted diesel particles demonstrated minimal hygroscopic

                                         2-93

-------
growth (2.5%), whereas aged particles subjected to UV radiation and ozonolysis exhibited
somewhat greater but still minimal hygroscopic growth. An increase in the sulfur content of
diesel fuel has also been observed to result in somewhat greater water condensation onto diesel
particles. To the extent that DE components are oxidized or nitrated in the atmosphere, they may
be removed at rates different from their precursor compounds and may exhibit different
biological reactivities. Data suggesting that minimal hygroscopic growth of DPM occurs also
has implications for the dosimetry of these particles in the lung because the smaller particles will
reach the lowest airways of the lung, whereas growth of the particle would result in deposition in
the upper airways. The dosimetry of DPM is discussed in Chapter 3.
       In a recent experiment, the biological activity of DPM exposed to 0.1 ppm ozone for 48
hours was compared with that of DPM not exposed to ozone (Ohio et al., 2000). Instillations of
the ozonated DPM in rat lung resulted in an increase in biological activity (neutrophil influx,
increased protein, and lactate dehydrogenase activity) compared with DPM that had not been
treated with ozone. These data suggest that ambient levels of ozone can alter DPM constituents
causing an increase in toxicity compared with nonozonated DPM.
       In addition to changes in particle composition with aging, particle size distributions may
vary depending on aggregation and coagulation phenomena in the aging process. People in
vehicles, near roadways (e.g., cyclists, pedestrians, people in nearby buildings), and on
motorcycles will be exposed to more fresh exhaust than the general population. In some settings
where emissions are entrained for long periods  through meteorological or other factors,
exposures would be expected to include both fresh and aged  DE.  The complexities of transport
and dispersion of emission arising from motor vehicles have been the subject of extensive
modeling and experimental studies over the past decades and have been summarized by
Sampson (1988); exposures to DPM are discussed in the next section of this chapter.
       The major organic constituents of DE and their potential  degradation pathways described
above provide evidence for (1) direct emission  of PAHs, (2)  secondary formation of nitroarenes,
and (3) secondary sulfate and nitrate formation. Because nitro-PAH products are often more
mutagenic than their precursors, the formation,  transport, and concentrations of these compounds
in an aged aerosol mass are of significant interest.

2.4. AMBIENT DIESEL EXHAUST CONCENTRATIONS AND EXPOSURES
2.4.1. Diesel Exhaust Gases in the Ambient Atmosphere
       Although emissions of several DE components have been measured, few studies have
attempted to elucidate the contribution of diesel-powered engines to atmospheric concentrations
of these components. The emission profile of gaseous organic compounds is different for diesel
and SI vehicles; the low-molecular-weight aromatic HCs and alkanes (
-------
characteristic of SI engine emissions, whereas the heavier alkanes (>C10) and aromatic HCs
(such as naphthalene, methyl- and dimethyl- naphthalenes, methyl- and dimethyl-indans) are
more characteristic of diesel engine emissions. These differences were the basis for
apportionment of gasoline- and diesel-powered vehicle emissions to ambient nonmethane
hydrocarbon (NMHC) concentrations in the Boston and Los Angeles (South Coast Air Basin)
urban areas.
       The chemical mass balance receptor model (described below) was applied to ambient
samples collected in these areas, along with appropriate fuel, stationary, and area source profiles
(Fujita et al., 1997).  The average of the sum of NMHC attributed to DE, gasoline-vehicle
exhaust, liquid gasoline, and gasoline vapor was 73% and 76% for Boston and the South Coast
Air Basin (SoCAB), respectively. The average source contributions of DE to NMHC
concentrations were 22% and 13% for Boston and the SoCAB, respectively. Diesel vehicles
emit lower levels of NMHC in the exhaust compared with gasoline vehicles. The relative
contribution of DE clearly depends on several factors, including fleet composition, sampling
location (e.g., near a bus station vs. near a highway or other sources), and the contribution from
point and area sources. The contribution of DE to ambient NMHC showed large variations
among sampling sites in the Boston area.  The source apportionment in the Fujita et al. (1997)
study indicates that mobile vehicle-related emissions account for the majority of ambient NMHC
in the two urban areas studied, and the results can likely be extrapolated to other urban areas
with similar source compositions. Other source apportionment methods such as those used by
Henry et al. (1994) have been applied to speciated HC data to separate the mobile source direct
emission from gasoline evaporative emissions.  This method uses a combination of graphical
analysis (Graphical Ratio Analysis for Composition Estimates, GRACE) and multivariate
receptor modeling methods (Source Apportionment by Factors with Explicit Restrictions,
SAFER) and was not used to identify the diesel engine contribution to the HCs measured.

2.4.2. Ambient Concentrations of DPM
       Because DPM is chemically complex, an assessment of ambient DPM concentrations
relies primarily on (1) studies that collect ambient samples and adequately characterize their
chemical composition, or (2) modeling studies that attempt to recreate emissions and
atmospheric conditions. Ambient concentrations of DPM also have been reported from studies
using surrogate species. The results of these studies are summarized below. Studies conducted
in Europe and Japan were reviewed, but for the most part were not included because of questions
surrounding the applicability  of measurements in locations that use different diesel technology
and control measures from those in the United States.
                                         2-95

-------
2.4.2.1. Source Apportionment Studies
       Receptor models are used to infer the types and relative contributions of sources to
pollutant measurements made at a receptor site.  Receptor models assume that the mass is
conserved between the source and receptor site and that the measured mass of each pollutant is a
sum of the contributions from each source. Receptor models are referred to as "top-down" in
contrast to "bottom-up" methods, which use emission inventory data, activity patterns, and
dispersion modeling from the source to predict concentrations at a receptor site.
       The most commonly used receptor model for quantifying concentrations of DPM at a
receptor site is the chemical mass balance (CMB) model. Input to the CMB model includes
measurements of PM mass and chemistry made at the receptor site as well as measurements
made of each of the source types suspected to impact the site.  Because of problems involving
the elemental similarity between diesel and gasoline emission profiles and their co-emission in
time and space, chemical molecular species that provide markers for separation of these sources
have been identified (Lowenthal et al., 1992). Recent advances in chemical analytical
techniques have facilitated the development of sophisticated molecular source profiles, including
detailed speciation of PM-associated organic compounds that allow the apportionment of PM to
gasoline and diesel sources with increased confidence.  CMB analysis that uses speciation of
organic compounds in the source profiles is typically referred to as extended species CMB.
Older studies that made use of only EC, total OC, trace elements, and major ions in the source
profiles (conventional CMB) have been published and are summarized here, but they are subject
to more uncertainty. It should be noted that because receptor modeling is based on the
application of source profiles to ambient measurements, estimates of DPM  concentration
generated by this method include the contribution from on-road and nonroad sources to the
extent the source profiles are similar (which would include military sources depending on the
sampling locations and fleet composition).  In addition, this method identifies sources of primary
emissions of DPM only, and the contribution of secondary aerosols is not attributed to sources.
       The CMB model has been used to assess concentrations of DPM in  areas of California,
Phoenix, Denver, and Manhattan (Table 2-23). DPM concentrations reported by Schauer et al.
(1996) for samples collected in California in 1982 ranged from 4.4 |ig/m3 in west Los Angeles to
11.6 |ig/m3 in downtown Los Angeles.  The average contribution of DPM to total PM2 5 mass
ranged from 13% in Rubidoux to 36% in downtown Los Angeles.  As mentioned above, this
model  accounts for primary emissions of DPM only; the contribution of secondary aerosol
formation (both acid and organic aerosols) is not included. In sites downwind from urban areas,
such as Rubidoux in this study,  secondary nitrate formation can account for a substantial fraction
of the mass (25% of the fine mass measured in Rubidoux was attributed to secondary nitrate), a
portion of which comes from DE (Gray and Kuklin, 1996).

                                         2-96

-------
                  Table 2-23.  Ambient DPM concentrations reported from chemical mass balance modeling
Reference
Schaueretal., 1996
Chowetal., 1991
California EPA, 1998a
Wittorffetal., 1994
Maricopa Association of
Governments, 1999
Fujitaetal., 1998
Location
West LA, CA
Pasadena, CA
Rubidoux, CA
Los Angeles, CA
West Phoenix, AZ
Central Phoenix, AZ
South Scottsdale, AZ
Estrella Park, AZ
Gunnery Park, AZ
Pinnacle Peak, AZ
California, 6 air basins
California, 9 air basins
Manhattan, NY
Phoenix, AZ
Welby, CO
Brighton, CO
Year of sampling
1982, annual average (~60
samples at each site)
1989-90, winter
1 1 days at each site
1988-92, annual
1993, spring 3 days
1994-95, winter 12 days
1996-97, winter 60 days
Location type
Urban
Urban
Urban
Urban
Urban
Urban
Urban
Nonurban
Nonurban
Nonurban
Urban0
Nonurban0
Urban
Urban
Urban
Suburban
Diesel PM2 s
Mg/m3 mean,
(range)
4.4
5.3
5.4
11.6
13 (max. 22)
13 (max. 16)
10 (max. 12)
5
3
2
1.8-3.6a
0.2-2.6a
29.2(13.2-46.7)a
2.4 (0-5.3)
1.7 (0-7.3)
1.2 (0-3.4)
Average
DPM %of
total PM
(range)
18
19
13
36
18
20
17
9
10
12


53(31-68)
15 (0-27)
10 (0-26)
10 (0-38)
Source profile
used
EC, OCS,
elements
EC, OCT, MI,
elements
EC, OCT, MI,
elements
EC, OCT, MI,
elements
EC, OCT, MI,
elements
EC, OCS, MI,
elements
EC, OCS, MI,
elements
to
           TM10.
           bNot available.
           °Urban air basins are qualitatively defined as those areas that are moderately or largely urbanized, and nonurban air basins are those areas that are largely
           nonurban, but may have one or more densely populated areas.
           Abbreviations: EC: Elemental carbon; OCT: OC total; OCS: OC species; MI: Major ions including nitrate, sulfate, chloride and, in some cases, ammonium,
           sodium, potassium.

-------
       The California Environmental Protection Agency (Cal EPA) reported ambient DPM
concentrations for 15 air basins in California based on ambient measurements taken statewide
from 1988 to 1992 (Cal EPA, 1998a).  Cal EPA used CMB analysis of ambient measurements
from the San Joaquin Valley (1988-89), South Coast (1986), and San Jose (winters for 1991-92
and 1992-93) to determine mobile source contributions and then applied the California 1990
PM10 emissions inventory to determine the fraction of mobile source PM10 attributable to diesel
emissions.  The results of this analysis indicate that annual average basin-wide levels of direct
DPM may be as low as 0.2 |ig/m3 and may range up to 2.6 |ig/m3 for basins that are largely
nonurban but may have one or more densely populated areas (such as Palm Springs in  the Salton
Sea basin). DPM concentrations for air basins that are moderately or largely urbanized ranged
from 1.8 |ig/m3 to 3.6 |ig/m3.
       Two studies using CMB analysis that report DPM concentrations have been conducted in
the Phoenix area. A wintertime study in 1989-90 reported DPM concentrations for nonurban
areas ranging from 2 (ig/m3 to 5 (ig/m3 and DPM concentrations for central and south Phoenix
urban areas ranging from 10 (ig/m3 to 13 |ig/m3 (Chow et al., 1991). Chow et al. (1991) reported
that DPM levels on single days can range up to 22 |ig/m3 at the central Phoenix site. A more
recent study conducted from November 1994 through March 1995 reported DPM concentrations
for Phoenix averaging 2.4 |ig/m3and reaching 5.3 |ig/m3 (Maricopa Association of Governments,
1999).  The extended species CMB was used for this study, providing a more confident
identification of DPM separate from gasoline PM emissions than the earlier Phoenix study.
DPM accounted for an average 15% of ambient PM2 5, and gasoline PM accounted for an
average of 52% of ambient PM25 in the 1994-95 Phoenix study.
       In a recently published study designed to investigate the ability of a new type of factor
analysis, positive matrix factorization, to separate sources contributing to the urban aerosol in
Phoenix, Ramadan et al. (2000) report their  success in separating the DE PM from other motor
vehicle PM. Fine PM samples were collected by two different types of samplers in Phoenix, one
set collected from March 1995 through June 1998 and a  second set from June 1996 through June
1998. Elemental and OC were analyzed using TOT.  Particles of DE origin were identified by
their high EC content in addition to specific trace elements, including manganese, sulfur, and
iron. DPM concentrations exceeding 5 |ig/m3 were reported for winter months during  the study
period.  The investigators concluded that motor vehicles, vegetative burning, and FID DE were
the three major sources contributing to ambient fine PM in Phoenix, with higher contributions in
the winter than in summer.
       During the winter of 1997, a study assessed DPM concentrations at two urban sites in the
Denver area (Fujita et al.,  1998). The Northern Front Range Air Quality Study (NFRAQS),
initiated to assess the sources of the "brown cloud" observed along Colorado's Front Range,

                                         2-98

-------
conducted air quality sampling during the winter of 1996, summer of 1996, and winter of 1997.
For a 60-day period from December 1996 through January 1997, ambient samples collected at
two urban Denver sites were analyzed for OC species for use in the extended-species CMB. The
average DPM concentrations reported for the urban site at Welby, CO, and the suburban site at
Brighton, CO, were 1.7 |ig/m3 and 1.2 |ig/m3, respectively.  During the study period, DPM
concentrations exceeded 5 |ig/m3 on two occasions in Welby, with reported DPM concentrations
of 5.7 |ig/m3 and 7.3 |ig/m3.  DPM accounted for an average of 10% of ambient PM2 5, and
gasoline PM accounted for an  average of 27% of ambient PM25.
       One of the major claims from the NFRAQS was a substantial contribution of EC from
gasoline-powered vehicles, mainly from cold-start and high-emitting vehicles. At the Welby
site, the contribution of diesel  and gasoline emissions to EC measurements was 52% and 42%,
respectively. At the Brighton  site, the contribution of diesel and gasoline emissions to EC
measurements was 71% and 26%, respectively.  The findings from the NFRAQS are compelling
and suggest the need for further investigations to quantify the contribution from cold-start and
high-emitting vehicle emissions for both gasoline and diesel vehicles. Geographical, temporal,
and other site-specific parameters that influence PM concentrations, such as altitude, must be
considered when extrapolating the NFRAQS findings to other locations.
       In addition to the need for urban and rural average DPM concentrations, an assessment of
potential health effects resulting from DPM exposure includes an assessment of people in
environments with potentially elevated levels of DPM. Limited data are available to allow a
characterization of DPM concentrations in "hotspots" such as near heavily traveled roadways,
bus stations, train stations, and marinas. Only one CMB study has attempted to apportion PM
measured in an urban hotspot.  Wittorff et al. (1994) reported results of conventional CMB
performed on PM samples collected in the spring of 1993 over a 3-day period at a site adjacent
to a major bus stop on Madison Avenue in midtown Manhattan. Buses in this area idle for as
long as 10 minutes, and PM emissions are augmented by the elevated levels of DPM emitted
during acceleration away from the bus stop (discussed in  Section 2.2.5).  DPM concentrations
reported from this study ranged from 13.0 |ig/m3 to 46.7 |ig/m3.  This study attributed, on
average, 53% of the PM10 to DE.  The DPM concentrations resulting from the source
apportionment method used in this study require some caution because the CMB model
overpredicted PM10 concentrations by an average 30%, which suggests that additional sources of
the mass were not accounted for in the model. The relevance of the Manhattan bus stop
concentrations and potential exposure for large urban populations provide strong motivation for
further studies in the vicinity of such hotspots.
       In summary, source apportionment studies of ambient samples collected before 1990
suggest that seasonal and annual average DPM concentrations for nonurban areas ranged from 2

                                          2-99

-------
(ig/m3 to 5 (ig/m3.  DPM concentrations reported from CMB studies for urban areas in the pre-
1990 timeframe ranged from 4.4 (ig/m3 to 13 |ig/m3, with concentrations on individual days
ranging up to 22 (ig/m3.  Source apportionment applied to ambient measurements taken in 1990
or later suggest that seasonal or annual average DPM levels in suburban/nonurban locations can
range from 0.2 (ig/m3 to 2.6 |ig/m3, with maximum reported values ranging up to 3.4 |ig/m3.
DPM concentrations reported from CMB studies in urban areas during 1990 or later range from
1.7 (ig/m3 to 3.6 |ig/m3, with maximum concentrations up to 7.3 |ig/m3.  The highest DPM
concentrations reported from CMB analysis of ambient measurements were those in the vicinity
of a bus stop in midtown Manhattan, which ranged from 13.2 (ig/m3 to 46.7 |ig/m3.

2.4.2.2. EC Surrogate for DPM
      EC is a major component of DE, contributing approximately 50% to 85% of diesel
particulate mass, depending on engine technology, fuel type, duty cycle, engine lubrication oil
consumption, and  state of engine maintenance (Graboski et al., 1998b; Zaebst et al., 1991;
Pierson and Brachaczek, 1983; Warner-Selph and Dietzmann, 1984). In urban ambient
environments, DE is one of the major contributors to EC, with other potential sources including
spark-engine exhaust; combustion of coal, oil, or wood; charbroiling; cigarette smoke; and road
dust.  Although cold-start emissions from gasoline combustion vehicles were reported to be an
important source of EC in wintertime samples collected in two cities in the Denver area (Fujita et
al., 1998), it is  currently unclear to what extent these results are transferable to other locations.  It
is noteworthy that the EC content of the cold-start emissions from gasoline combustion vehicles
was lower than that from diesel  combustion engines in the same study by almost a factor of 2.
      Fowler (1985) evaluated several components of DE and concluded that EC is the most
reliable overall measure of ambient DE exposure. Because of the large portion of EC in DPM,
and the fact that DE is one of the major contributors to EC in many ambient environments, DPM
concentrations can be bounded using EC measurements.  Surrogate calculations of DPM have
been based on the  fraction of ambient EC measured in a sample that is attributable to diesel
engine exhaust and the fraction  of the diesel particle mass accounted for by EC. In the  recent
Multiple Air Toxics Exposure Study in the South  Coast Air Basin (MATES-II, SCAQMD,
2000), EC measurements were used to estimate DPM concentrations by the following
relationship: approximately  67% of fine EC in the ambient air in the Los Angeles area originates
from diesel engine exhaust (Gray, 1986), and the average EC fraction of diesel particles
measured was 64%.  Therefore, in the MATES-II study, the South Coast Air Quality
Management District calculated DPM concentrations from EC measurements by multiplying a
measured EC concentration by 67% and dividing by the fraction of DPM mass accounted for by
EC of 64%, for example, DPM concentration = (EC * 0.67)70.64, or DPM = EC *  1.04  (not

                                         2-100

-------
appreciably different from EODPM). This calculation, used in the MATES-II study, relies on
data collected in the 1982 timeframe and may not accurately represent the current day
contributions of diesel engines to the ambient EC inventory.  Using a 1998 emissions inventory
for the South Coast Air Basin, it is now estimated that a more appropriate conversion from EC to
DPM is to multiply EC by 1.24 (MATES-II, SCAQMD, 2000).
       An alternative calculation can be  derived using data from recent studies in Colorado and
Arizona (Fujita et al., 1998; Maricopa Association of Governments, 1999). The fraction of EC
attributable to DE can be estimated from detailed source profiles applied to a CMB model as
discussed above.  The contribution of diesel engines to EC averaged 68% ± 20% for Brighton,
CO, and 49% ± 26% at Welby, CO, as part of the winter 1996-1997 NFRAQS. In Phoenix,
diesel engine exhaust was estimated to account for approximately 46% ± 22% of the ambient
EC. For some environments, such as certain occupational settings in which diesel engines are in
proximity to workers, all the EC may realistically be attributed to DE as a reasonable upper
bound estimate of DPM concentrations.
       As discussed in Section 2.2, the EC content of DPM can vary widely depending on
engine type, load conditions, and the test cycle. However, typical profiles for HD and LD diesel
engines have been determined and the typical EC fraction of DPM ranges from approximately
52% to 75%.
       Ambient EC attributed to DE in the studies described  above ranges from 46% to 68%.  A
lower-bound estimate of DPM from ambient EC measurements in areas with similar source
contributions to those in the Phoenix and Denver areas can be derived using the equation:

                    DPM = (EC * 0.46)70.75 or DPM = EC * 0.62

An upper-bound estimate uses the equation:

                    DPM = (EC * 0.68)70.52 or DPM = EC * 1.31

Using the average of the ranges provides the equation:

                                 DPM = EC* 0.89.

       Clearly the choice of a point estimate can provide a surrogate calculation of DPM that
can vary by at least a factor of two. Although a recommended surrogate DPM calculation
method is not provided here, the surrogate DPM calculation is used to illustrate the usefulness of
                                         2-101

-------
this approach for estimating DPM in the absence of a more sophisticated receptor modeling
analysis for locations where fine PM EC concentrations are available.
       One source of variability in EC concentrations reported for ambient studies is the
measurement method used to quantify EC. As discussed in Section 2.2.8.1, EC and OC are
operationally defined. Ambient samples are typically analyzed for EC using thermal  optical
reflectance or thermal optical transmittance.  The measurement technique used in the  NFRAQS
and Phoenix studies was TOR, which, as discussed in Section 2.2.8.2, often results in higher EC
levels compared to TOT analyses.
       Table 2-24 provides  a lower- and upper-bound DPM estimate from annual average EC
concentrations for three urban areas, in addition to DPM concentrations reported from EC
measurements for the MATES- II (SCAQMD, 2000). Under an EPA research grant with the
Northeastern States for Coordinated Air Use Management (NESCAUM), PM2 5 samples were
collected every 6 days for 1  year (1995) in Boston (Kenmore Square), MA, and Rochester, NY,
and were analyzed for EC using TOT (Salmon et al., 1997). DPM concentrations were estimated
to be in the range from 0.8 |ig/m3 to 1.7 |ig/m3 in Boston, and from 0.4 |ig/m3 to 0.8 |ig/m3 in
Rochester (Table 2-24).
       Table 2-24.  Ambient diesel particulate matter concentrations from elemental
       carbon measurements in urban locations
Reference
Salmon etal., 1997

Sisler, 1996
Year of
sampling
1995, annual

1992-1995,
annual

South Coast
Air Quality
Management
District, 1999
1995-6, annual
Location
Boston, MA
Rochester, NY
Washington, DC
MATES II c
Anaheim, CA
Burbank, CA
Los Angeles, CA
Fontanta, CA
Huntington Park, CA
Long Beach, CA
Pico Rivera, CA
Rubidoux, CA
DPM25
/ug/m3 lower-upper bound
range (point estimate)3
0.8-1.7(1.1)
0.4-0.8 (0.5)
0.9-2.2(1.5)
Diesel PM2 5
Mg/m3
avg± std dev.
2.4 ± 1.8
3. 3 ±1.9
3.5 ± 1.9
3.4 ±2.3
4.5 ±2.4
2.5 ±1.7
4.4 ±2.2
3.4 ±2.0
DPM %
of total
PM
6-12
3-6
4-12

b
b
b
b
b
b
b
b
a Lower-bound range: DPM=EC*0.62; upper-bound range: DPM=EC*1.31; point estimate: DPM=EC*0.89.
b Not available.
The Multiple Air Toxics Exposure Study in the South Coast Air Basin reported DPM calculated from EC
concentrations as DPM=EC* 1.04.  Standard deviations are reported.
                                         2-102

-------
       The Interagency Monitoring of Protected Visual Environments (IMPROVE) project
being conducted by the National Park Service includes an extensive aerosol monitoring network
mainly in rural or remote areas of the country (national parks, national monuments, wilderness
areas, national wildlife refuges, and national seashores), and also in Washington, DC (Sisler,
1996). PM25 samples, collected from March 1992 through February 1995 twice weekly for 24-
hour duration at 43 sites (some co-located in the same rural park area), were analyzed for a suite
of chemical constituents, including EC (using TOR). EC concentrations in these rural locations
may have EC source contributions quite different from those in the urban areas in which the
fraction of EC attributable to DE has been reported. The lack of information regarding EC
sources in these rural locations makes the application of the EC surrogate highly uncertain.  It is
noteworthy that annual average EC concentrations in the rural and remote regions reported as
part of the IMPROVE network range from 0.1 |ig/m3 for Denali National Park, AK, to 0.9 |ig/m3
for the Lake Tahoe, CA, area.  In Washington, DC, the annual average EC concentration of 1.7
|ig/m3 is estimated as an annual average DPM concentration of 1.4  |ig/m3.
       The annual average EC measurements in Washington, DC, suggest that the DPM
concentrations are in the range from 1.0 |ig/m3 to 2.2 |ig/m3, accounting for 5% to 12% of
ambient PM2 5.  Seasonally averaged data for the Washington, DC, site indicate that EC
concentrations and, by extension, DPM concentrations peak in the autumn and winter (2.0 |ig/m3
and 0.9 |ig/m3 EC, respectively).
       DPM concentrations reported recently as part of the MATES-II study at eight locations
ranged from 2.4 |ig/m3 to 4.5 |ig/m3.  DPM concentrations at Huntington Park  and Pico Rivera,
CA, were higher than other DPM concentrations in the South Coast Air Basin, perhaps because
of higher diesel  truck traffic, proximity to nonroad diesel sources, or nondiesel sources of EC,
including gasoline vehicle traffic.
       In a recent study of the trends in fine particle and EC concentrations in Southern
California, Christoforou et al.  (2000) report that EC concentrations measured in 1993 were 29%-
40% of EC concentrations measured in 1982 at four urban Los Angeles sites.  The authors credit
lower PM emission rates from on-road diesel engines as well as cleaner-burning diesel fuel for
the  observed EC decrease.  The extent to which nonroad diesel equipment impacts a given site
will influence the trend in ambient EC concentrations because fewer regulations have been
promulgated to control the PM emissions from these engines.

2.4.2.3. Dispersion Modeling Results
       Dispersion models estimate ambient levels of PM at a receptor site on the basis of
emission factors for the relevant sources and parameters that simulate atmospheric processes
such as the advection, mixing, deposition, and chemical transformation of compounds as they are

                                         2-103

-------
transported from the source to the receptor site(s). Cass and Gray (1995), Gray and Cass (1998),
and Kleeman and Cass (1998) have applied dispersion models to the South Coast Air Basin to
estimate DPM concentrations.  The models used by these investigators applied emission factors
from 1982 and consequently  are representative of concentrations prior to the implementation of
DPM emission controls.  In addition to offering another approach for estimating ambient DPM
concentrations, dispersion models can provide the ability to distinguish on-highway from
nonroad diesel source contributions and have presented an approach for quantifying the
concentrations of secondary aerosols from DE.
       Cass and Gray (1995) used a Lagrangian particle-in-cell model to estimate the source
contributions to  atmospheric  fine carbon particle concentrations in the Los Angeles area,
including diesel  emission factors from on-highway and off-highway sources.  Their dispersion
model indicates  that for 1982, the annual average ambient concentrations of DPM ranged from
1.9 |ig/m3 in Azusa, CA, to 5.6 |ig/m3 in downtown Los Angeles (Table 2-25). The contribution
of on-highway sources to DPM ranged from 63.3% in downtown Los Angeles to 89% in west
Los Angeles.  Of the on-highway diesel contribution, the model predicted that for southern
California, HD trucks made up the majority (85%) of the DPM inventory, and overall they
contributed 66% of the DPM in the ambient air. Nonroad sources of DE include pumping
stations, construction sites, shipping docks, railroad yards, and heavy equipment repair facilities.
Cass and Gray (1995) also report that wintertime peaks in DPM concentrations can reach 10
|ig/m3.
Table 2-25. Ambient diesel particulate matter concentrations from dispersion modeling
Reference
Cass and Gray,
1995
Kleeman and Cass,
1998
Kleeman et al.,
2000
Location
Azusa, CA
Lennox, CA
Anaheim, CA
Pasadena, CA
Long Beach, CA
Downtown LA, CA
West LA, CA
Claremont, CA
Long Beach, CA
Fullerton, CA
Riverside, CA
Year of sampling
1982, annual
1982, annual
1982, annual
1982, annual
1982, annual
1982, annual
1982, annual
18-19 Aug 1987
24 Sept 1996
24 Sept 1996
25 Sept 1996
Location type
Nonurban
Nonurban
Urban
Urban
Urban
Urban
Urban
Nonurban
Urban
Nonurban
Suburban
DPM25
g/m3 (mean)
1.4a
3.8a
2.T
2.0a
3.5a
3.5a
3.8a
2.4 (4.0)a'b
1.9(2.6)b
2.4(3. 9)b
4.4(13.3)b
DPM % of
total PM
5
13
12
7
13
11
16
8(6)b
8(7)b
9(8)b
12(13)b
a On-road diesel vehicles only; all other values are for on-road plus nonroad diesel emissions.
b Value in parentheses includes secondary DPM (nitrate, ammonium, sulfate and hydrocarbons) attributable to atmospheric reactions of primary
diesel emissions of NOX, SO2 and hydrocarbons. For the fraction of ambient PM attributable to DPM, the value in parenthesis reports total DPM
(primary plus secondary) as a fraction of total ambient PM (primary plus secondary).
                                           2-104

-------
       Kleeman and Cass (1998) developed a Lagrangian model that examines the size and
chemical evolution of aerosols, including gas-to-particle conversion processes during transport.
This model was applied to one well-characterized episode in Claremont, CA, on August 27-28,
1987. The model provided reasonable predictions of PM10 (overpredicting PM10 by 13%), EC,
and OC, and it adequately reconstructed the size distribution of the aerosols. The model
indicated that on August 27-28, 1987, the PM25 concentration was 76.7 |ig/m3, 13.2% (10.1
|ig/m3) of which was attributable to diesel engine emissions. This estimate includes secondary
aerosol formation for sulfate, ammonium, nitrate, and organic compounds, which accounted for
4.9 |ig/m3 of the total estimated DPM mass. The secondary organic aerosol was estimated to be
1.1 |ig/m3, or 31% of the total secondary aerosol mass, with the remainder composed  of nitrate,
ammonium, and sulfate aerosols.
       Dispersion modeling estimates of diesel PM concentrations from on-highway  and
nonroad sources have recently been developed as part of the EPA National Air Toxics
Assessment (NATA) National Scale Assessment. This assessment uses the Assessment System
for Population Exposure Nationwide  (ASPEN) dispersion model to estimate ambient
concentrations for the year 1996. The NATA national scale assessment reports concentrations of
DPM and 32 additional urban air toxic compounds at the  county, State, and national level
(NATA, 2001).
       ASPEN makes a number of simplifying assumptions in order to model  concentrations on
a nationwide scale. For instance, concentration estimates at the census tract level were  estimated
using modeling assumptions to allocate emissions from the county level, and the model is very
sensitive to the assumptions used.  In addition, dispersion of emissions from nonpoint sources
(e.g., on-highway and nonroad vehicles) was treated simplistically. For resident tracts that have
radii greater than 0.3 km, non-point-source ambient concentrations are estimated on the basis of
five pseudo-point sources.  The average concentration for the census tract is determined by
spatially averaging the ambient concentrations associated with the receptors defined for the five
pseudosources that fall within the bounds of the tract.  Other limitations include the following:
terrain impacts on dispersion were  not included; the study relied on long-term climate summary
data, and no long-range transport was included for DPM (medium-range transport for DPM,
within 300 km, was included). Because of the limitations, the results are most meaningfully
interpreted when viewed over large geographic areas.  The 1996 results from ASPEN compare
well  (generally within a factor of 1.5) with estimated concentrations from EC measurements and
receptor modeling, as well as data from other dispersion modeling studies. The complete results
of the assessment are available at http://www.epa.gov/ttn/uatw/nata.
                                         2-105

-------
       Table 2-26 presents 25th percentile, average, and 75th percentile nationwide
concentrations from the 1996 National-Scale Assessment as well as the contribution of on-road
and nonroad DE the sources to the nationwide average. The national average DPM
concentration reported in the National-Scale Assessment is 2.1 |ig/m3, of which nonroad sources
are estimated to contribute 67% and on-road sources contribute the remainder. Less than 2% of
the nationwide DE inventory is attributed to point sources, and these were not included in the
modeling as part of National-Scale Assessment. A wide range in average State-specific ambient
DPM concentrations was reported by the National-Scale Assessment with the lowest values for
mainly rural States with few DE sources, such as Wyoming (annual average of 0.2 |ig/m3), and
the highest values for States with large urban centers such as New York (annual average of 5.4
|ig/m3).
Table 2-26. Nationwide ambient diesel particulate matter concentrations for 1996 from the
National Air Toxics Assessment National-Scale Assessment dispersion modeling
Location
Nationwide
All urban
counties
All rural
counties
25th
percentile,
DPM10
mg/m3
0.9
1.2
0.4
Average,
DPM10
mg/m3
2.1
2.4
0.7
75th
percentile,
DPM10
mg/m3
2.5
2.7
1.0
Contribution to
average from on-
road sources, DPM10
mg/m3
0.6
0.7
0.3
Contribution to
average from
nonroad sources,
DPM10
mg/m3
1.4
1.7
0.5
Source: NATA, 2001. Data available at http://www.epa.gov/ttn/uatw/nata.
2.4.3. Exposures to Diesel Exhaust
       Ultimately, it is personal exposure that determines health impacts. In the following
sections, modeled average exposures and some information reflecting potential exposures for
those who spend a large portion of their time outdoors are presented. Occupational exposures to
DPM are summarized for the variety of workplaces in which diesel engines are used.  These
occupational exposures are placed into context with equivalent environmental exposures to
understand the potential for overlap in average occupational and average ambient exposures.
Because DE is a mixture of particles and gases, one must choose a measure of exposure (i.e.,
dosimeter); |ig/m3 of DPM has historically been used in many  studies as the dosimeter for the
entire DE mixture.
                                         2-106

-------
2.4.3.1. Occupational Exposure to DE
       The National Institute for Occupational Safety and Health (NIOSH, 1988) estimates that
approximately 1.35 million workers are occupationally exposed to DE emissions. Such workers
emissions include mine workers, railroad workers, bus and truck drivers, truck and bus
maintenance garage workers, loading dock workers, firefighters, heavy equipment operators, and
farm workers.
       Measurements of DPM exposure in occupational environments have included respirable
particulate (<3.5 jim), smoking-corrected respirable particulate, combustible respirable
particulate, and EC, among other methods. The measurement method used in each of the studies
discussed below is listed in Table 2-27. Occupational exposures to DPM as well as breathing
zone concentrations of DPM have been described in some detail by Watts (1995), Groves and
Cain (2000), Hammond (1998), the World Health Organization (1996), and Birch and Cary
(1996) and are briefly, but not comprehensively, summarized here.
       The highest occupational exposures to DPM are for workers in coal mines and noncoal
mines using diesel-powered equipment. These exposures, reported by several investigators,
range from approximately 10 |ig/m3 to 1,280 |ig/m3 (Table 2-27). Rogers and Whelan (1999)
report exposures to specific DPM-associated PAHs (including  naphthalene, fluorene,
phenanthrene, pyrene, and benz[a]anthracene) for mine workers using diesel fuels containing
low and high levels of sulfur, aliphatic, and aromatic compounds. Results of this study indicate
that the composition of DPM to which workers were exposed varies considerably based on
engine condition, fuel, and other operating parameters. Mine worker exposures to PAH
compounds were highest for naphthalene, ranging from 1,312 |ig/g to 3,228 |ig/g of organics,
and exposures were lowest for benz[a]anthracene, ranging from less than 3 jig/g up to 18 jig/g of
organics.
       Other investigators have reported DPM-associated PAH concentrations that do not
necessarily represent personal exposures but are a snapshot of  short periods of elevated
concentration that make up a portion of a worker's daily exposure.  Bagley et al. (1991, 1992)
reported levels of B[a]P ranging from below the detection limit of 0.05 ng/m3 to 61 ng/m3
collected only during periods of mining activity. Watts (1995) reported DPM concentrations in
four mines collected during significant  diesel activity, ranging  from 850  |ig/m3 to 3,260 |ig/m3.
Heino (1978) reports DPM concentrations for locomotive engineers reaching 2,000 |ig/m3.
       In a study of four railroads, Woskie et al. (1988) reported concentrations of respirable
dust (corrected for cigarette smoke paniculate) that ranged from 39 |ig/m3 for engineers/firers to
134 |ig/m3 for locomotive shop workers and  191 |ig/m3 for hostlers. Woskie et al. (1988) also
reported smoking-corrected respirable dust for railroad clerks (17 jig/m3), who are considered to
be not exposed to DE. Although these exposures may have included  nondiesel PM (background

                                         2-107

-------
        Table 2-27. Occupational exposure to PPM
Author
Gangal and Dainty,
1993a
Saverin, 1999
Rogers and Whelan,
1999
Haney, 1990a
Arabs, 1991aa
Woskieetal., 1988
Groves and Cain,
2000
Froines et al., 1987
NIOSH, 1992a
Birch and Gary,
1996
Birch and Gary,
1996
NIOSH, 1990
Zaebstetal., 1991
Groves and Cain,
2000
Kittelsonetal.,2000
Year of
sample
NA
1992
1990-99
1980s
NA
3-years in mid-1980s
NA
1985
NA
NA
NA
NA
NA
1990
1990
NA
1999-2000
Location/job type,
typical work schedule
of 8 hours
Noncoal mine workers
Noncoal mine workers
Coal mine workers
Coal mine workers
(five mines)
Coal mine workers
(four mines)
Railroad engineer/frier
Railroad braker/conductor
Railroad shop workers
Railway repair
Firefighters (two stations)
Firefighters
(three stations)
Firefighters
Fire station employees (four
stations)
Airport ground crew
Public transit workers
Diesel forklift dockworkers
Dockworkers
Mechanics
Long- and short-haul truckers
Bus garage/repair
Forklift trucks
Bus drivers
Parking ramp attendants
n
-200
255b
>1,300
NA
NA
128
158
176
64
238
18
NA
NA
NA
NA
24
75
80
128
53
27
39
12
Sample
type
RCD
RTC
DPSMM
SJI
PDEAS
ARP
ARP
ARP
EC(U)
TSP
EC(T)
EC(U)
EC(U)
EC(U)
EC(U)
EC(T)
EC(T)
EC(T)
EC(T)
EC(U)
EC(U)
EC(T)
EC(T)
Range in
DPM,
Hg/ni3
100-900
38-1,280
10-640
180-1,000
750-780
39-73
52-191
114-134
7-50
63-748
6-70
20-79
4-52
7-15
15-98
12-61
9-20
5-28
2-7
7-217
7-403
1-3
2 ±0.4
3 Cited in Watts (1995).  NA: not available.
b Personal exposure and area samples were not reported separately for this study.
RCD: respirable combustible dust; RTC: respirable total carbon SPM: submicrometer PM; DPSMM: diesel
paniculate submicron mass (two-stage impaction sampler used to separate PM by size); EC(T): elemental carbon
analyzed by TOT; EC(R) elemental carbon analyzed by TOR; EC(U) elemental carbon analyzed by colouremetric
method or method not reported;  SJI: single-jet impactor agreed within 10% with simultaneous PDEAS
measurements;  PDEAS: personal DE aerosol sampler collects DPM <0.8 ,wm, SPM: paniculate matter; ARP:
respirable paniculate adjusted to remove the influence of cigarette smoke; TSP: total suspended paniculate matter.
                                              2-108

-------
respirable dust levels have been estimated to have contributed approximately 10 |ig/m3 to 33
|ig/m3 for this study), the majority of the respirable PM is believed to have originated from diesel
locomotive emissions.  Groves and Cain (2000) reported EC exposures among railway repair
workers averaging 21 |ig/m3 with a range from 7-50 |ig/m3.  DPM exposures reported for
firefighters operating diesel engine vehicles range from 4 |ig/m3 to 748 |ig/m3, which also
encompasses the range of DPM exposures reported for airport ground crew and public
transportation system personnel (7 |ig/m3 to 98 |ig/m3).
       Studies reporting DE exposure among fire station employees typically report particulate
levels below 100 |ig/m3 (ranging from 4 |ig/m3 to 79 |ig/m3) (NIOSH, 1992; Birch and  Cary,
1996). In a study by Froines et al. (1987), DPM exposures for firefighters in two stations ranged
from 39 |ig/m3 to 73 |ig/m3.  Birch and Cary (1996) also reported DPM exposures for airport
ground crew and public transit workers, ranging from 7 |ig/m3 to 15 |ig/m3 for airport ground
crews and 15 |ig/m3 to 98 |ig/m3 for public transit workers. Dock workers using diesel-powered
forklifts have been reported to have DPM exposures ranging from 6 |ig/m3 to 403 |ig/m3
(NIOSH, 1990; Zaebst et al., 1991; Groves and Cain, 2000). In studies by NIOSH (1990) and
Fowler (1985), the organic material measured accounted for about  one-half to almost all of the
carbonaceous DPM exposures, providing evidence that some pieces of nonroad equipment
(forklifts and construction equipment) emitted DPM with a significant OC fraction in the 1980s
and early 1990s.
       Zaebst et al. (1991) also reported DPM exposures for mechanics, road drivers, and local
drivers for 8-hour shifts at each of six large hub truck terminals. Residential background and
highway background samples at fixed sites were also collected during warm-weather and cold-
weather periods, and the geometric mean for DPM concentrations ranged from 1 |ig/m3 to 5
|ig/m3. DPM exposures for road and local truckers in warm- and cold-weather periods ranged
from 2 |ig/m3 to 7 |ig/m3, whereas exposure levels for mechanics were reported between 5 |ig/m3
and 28 |ig/m3 (geometric means).
       Kittelson et al. (2000) are measuring DPM exposures for bus drivers, parking garage
attendants, and mechanics using TOT to quantify EC.  Personal exposures for bus drivers on four
different routes range from 1 |ig/m3 to 3 |ig/m3 and exposure among parking ramp attendants
averaged 2  |ig/m3.  These results are preliminary, and data for  the mechanics have not yet been
analyzed. This study will also characterize PAH compounds to which these workers are
exposed.
       Bus garage workers have also been assessed for exposure to DE using urinary excretion
of 8-oxo-2'-deoxyguanosine  (Loft et al., 1999).  Other biomarkers of DE exposure in
occupational workers have included measurements of urinary  1-hydroxypyrene,  adducts of DNA
                                         2-109

-------
and hemoglobin, and 8-hydroxyguanosine in lung tissue (Nielsen et al., 1996; Tokiwa et al.,
1999; Zwirner-Baier and Neumann, 1999; Kara et al., 1997).
       To estimate an environmental exposure that is equivalent to an occupational lifetime
exposure, the fraction of lifetime worker inhalation exposure (calculated as the amount of air
breathed on the job multiplied by the typical amount of time spent on the job) is calculated
relative to 70-year lifetime inhalation exposure: (10 m3/shift/20 m3/day) * (5 days/7 days) * (48
weeks/52 weeks) * (45-year career/70-year lifetime) = 0.21.  Using this calculation, 21% of an
annual average occupational lifetime exposure is roughly equivalent to a 70-year annual  average
lifetime environmental exposure. The equivalent environmental exposures for the occupational
exposures presented in Table 2-28 range from 0.6 |ig/m3 to 14 |ig/m3 for truckers, dock workers,
and mechanics, and from 2 |ig/m3 to 269 |ig/m3 for miners.  The low end of the range of
environmental equivalent exposures for several of the occupational settings overlaps with
average modeled exposures and with ambient concentrations of DPM in urban areas in the
1990-1996 timeframe.  The overlap between some occupational exposures and environmental
exposures, as well as the small difference between occupational environmental equivalent
exposures and environmental exposures, is a significant concern and suggests the potential for
significant risk in the general population.  The possible magnitude of the cancer risk in the
general population is discussed in Chapter 8, Section 8.3.
Table 2-28. Ranges of occupational exposure to DPM by job category with estimates of
equivalent environmental exposures
Year of sampling
1980s and 1990s
1980s
1985 and later
NA
1990
1990
Occupations
Miners
Railroad workers
Firefighters
Airport crew, public transit workers
Dockworkers, mechanics
Long- and short-haul truckers
Occupational
DPM, yUg/m3
10-1,280
39-191
4-748
7-98
5-61
2-7
Environmental
equivalent3
exposure, MS/ni3
2-269
8-40
1-157
2-21
1-13
0.4-2
"Environmental equivalent exposure is calculated as the occupational exposure * (10 mVshift / 20 mVday)*
(5 days / 7days) * (48 weeks / 52 weeks) * (45 year career / 70 year lifetime), or occupational exposure * 0.21
(discussed in section 2.4.3.1.
2.4.3.2. Ambient Exposure to DE
       Modeled estimates of population exposures to DPM integrate exposure in various indoor
and outdoor environments and also account for the demographic distribution, time-activity
                                         2-110

-------
patterns, and DPM concentrations in various environments, including job-related exposures.
Two modeling efforts have been developed to determine DPM exposures in the general
population: the Hazardous Air Pollutant Exposure Model for Mobile Sources, version 3
(HAPEM-MS3) and the California Population Indoor Exposure Model (CPIEM).  EPA has also
developed version 4 of the HAPEM, which provides State-specific average exposures for DPM
and 32 other urban air toxic compounds. The draft exposure assessment using HAPEM version
4 (HAPEM4) has been conducted as part of the National Air Toxics Assessment National-Scale
Analysis described in Section 2.4.2.3 above and results are provided here.

2.4.3.2.1.  The Hazardous Air Pollutant Exposure Model. To estimate population exposures to
DPM, EPA has used HAPEM-MS3 (U.S. EPA,  1999b). This model provides national and
urban-area-specific exposures to DPM from on-road sources only. HAPEM-MS3 is based on the
CO probabilistic National Ambient Air Quality Standards (NAAQS) exposure model
(pNEM/CO), which is used to estimate the frequency distribution of population exposure to CO
and the resulting carboxyhemaglobin levels (Law et al., 1997). HAPEM simulates the CO
exposure scenario of individuals in 22  demographic groups for 37 microenvironments. CO
concentrations are based on ambient measurements made in 1990 and are related to exposures of
individuals in a 10-km radius around the sampling site.  DPM exposures are calculated as in
Equation 2-5, using a ratiometric approach to CO.

                      DPM  ,  3=(CC> , JCOglmi)xDPMglmi                      (2-5)
                           /ilg/m  ^    fjg/m     glmi /       glmi

       Data provided to the model include CO monitoring data for 1990; time-activity data
collected in Denver, Washington, DC,  and Cincinnati from 1982 to 1985; microenvironmental
data; and 1990 census population data. Motor vehicle DPM and CO emission rates reported by
EPA (1999c) are used to calculate mobile-source DPM exposures, and exposures in future years
are projected based on the increase in vehicle miles traveled.  EPA's PARTS model is used to
estimate DPM emission rates (g/mi)  for the fleet as a whole in any given calendar year. PARTS
is currently being modified to account  for deterioration, actual in-use emissions, poor
maintenance, and tampering effects,  all of which increase emission factors. As a result,
HAPEM-MS3 exposure estimates based on PARTS emission factors may underestimate true
exposures from on-road sources. A comparison of PARTS HD diesel vehicle emission factors
with those presented earlier in this chapter suggests that PARTS may underestimate HD diesel
vehicle emissions by up to 50%.
       HAPEM-MS3 assumes that the highway fleet (gasoline plus diesel) emissions ratio of
CO to DPM can be used as an adjustment factor to convert estimated CO personal exposure to

                                        2-111

-------
DPM exposure estimates.  This assumption is supported by the observation that even though
gasoline vehicles emit the large majority of CO, gasoline and diesel highway vehicles travel on
the same roadways. DPM and CO are both relatively long-lived atmospheric species (1-3 days)
except under certain conditions (Seinfeld and Pandis, 1998); therefore, the model does not
account for chemical and physical differences between the DPM and CO, and the model assumes
that for the average person in a modeled air district, CO and DPM are well mixed. Exposure in
microscale environments in which these assumptions may not be valid were not modeled.
      A validation study conducted for the pNEM/CO model on which HAPEM-MS3 is based
indicates that CO exposures for the population in the 5"1 percentile were overestimated by
approximately 33%, whereas those with exposures in the 98th percentile were underestimated by
about 30%. This validation study is considered applicable to the HAPEM-MS3 model. To
address the underestimate of exposures  for the most highly exposed, Brodowicz (1999) used CO
concentrations relevant to the most highly exposed populations to determine DPM exposures for
different demographic groups within this population; the results are discussed below.
      Annual average DPM exposures from on-road vehicles and nonroad sources nationwide
for the general population, rural and urban population, outdoor workers, and urban children are
reported in Tables 2-29 and 2-30. The modeled annual average DPM exposure nationwide
(urban and rural areas) in 1996 from on-road sources only was 0.8 |ig/m3. The modeled annual
average exposure in urban areas for the  same year was 0.8 (ig/m3, and the modeled exposure for
rural areas was 0.4 |ig/m3. Among the demographic groups modeled, urban outdoor workers in
general were found to have the highest average exposure to DPM, averaging 1.0 |ig/m3 from on-
road sources in 1996.  DPM exposures attributable to on-road sources are projected to decrease
until approximately 2007 because of fleet turnover and the full implementation of Federal
regulations that are currently in place. Full implementation of the recently finalized Heavy-Duty
Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements would
significantly lower DPM exposures from on-road sources in the post-2007 timeframe (U.S. EPA,
2000b).
      Because diesel vehicle traffic, and therefore exposure to DPM, varies for different urban
areas, HAPEM-MS3 was used to estimate annual average population exposures for 10 urban
areas. Modeled 1996 DPM exposures in the cities ranged from 0.6 |ig/m3 in Chicago and St.
Louis to 1.3 |ig/m3 in Phoenix (Table 2-31).  In 1996, estimated average DPM exposure from on-
road sources was higher than the national average in five cities: Atlanta, Minneapolis, New
York, Phoenix, and Spokane. Nationally in  1996, 97% of DPM exposure from on-road vehicles
was attributable to HD diesel vehicles, and the rest was  generated mainly by LD diesel trucks.
                                         2-112

-------
       Table 2-29. Annual average nationwide DPM exposure estimates (ug/m3)
       from on-road sources for rural and urban demographic groups in 1990,1996,
       and 2007 using HAPEM-MS3
Demographic group
50-State population
Rural population
Urban population
Urban outdoor workers
Urban children (0-17)
1990
0.8
0.5
0.9
1.1
0.9
1996
0.8
0.4
0.8
1.0
0.8
2007
0.4
0.2
0.4
0.5
0.4
        Source: U.S. EPA, 1999b, adjusted to reflect HDDV VMT described in U.S. EPA, 2000b.
Table 2-30.  Draft annual average, 25th, and 75th percentile nationwide DPM exposure
estimates (ug/m3) from on-road and nonroad sources for rural and urban counties in 1996
using HAPEM4
Demographic
group



Nationwide
Rural population
Urban population
25th Percentile,
DPM10
mg/m3


0.6
0.3
0.8
Average,
DPM10
mg/m3


1.4
0.6
1.6
75th Percentile,
DPM10
mg/m3


1.8
0.7
2.0
Contribution to
average from on-
road sources, DPM10
mg/m3

0.5
0.2
0.5
Contribution to
average from
nonroad sources,
DPM10
mg/m3
0.9
0.3
1.1
Source: NAT A, 2001. Data available at http://www.epa.gov/ttn/uatw/nata.
       Because HAPEM-MS3 is suspected to underestimate exposures in highly exposed
populations, 1990 CO concentrations relevant to the most highly exposed populations were used
to determine 1990 DPM exposures for different demographic groups in this population.  The
highest DPM exposures ranged from 0.8 |ig/m3 for outdoor workers in St. Louis to 2.0 |ig/m3 for
outdoor workers in Spokane and up to 4.0 |ig/m3 for outdoor children in New York (Table 2-31).
The highest exposed demographic groups were those who spend a large portion of their time
outdoors. It is important to note that these exposure estimates are lower than the total exposure
to DPM because they reflect only DPM from on-road sources and not exposure to nonroad DPM
emissions.
                                        2-113

-------
Table 2-31.  Annual average DPM exposures for 1990 and 1996 in the general
population and among the highest exposed demographic groups in nine urban areas
and nationwide from on-road sources only using HAPEM-MS3
Urban area
Nationwide
Atlanta, GA
Chicago, IL
Denver, CO
Houston, TX
Minneapolis, MN
New York, NY
Philadelphia, PA
Phoenix, AZ
Spokane, WA
St. Louis, MO
1990
Population average
exposure, jig/m3
0.8
0.8
0.8
0.7
0.6
1.0
1.6
0.7
1.4
1.3
0.6
1996
Population average
exposure, jig/m3
0.8
0.9
0.6
0.8
0.9
1.0
1.2
0.7
1.3
1.1
0.6
Highest DPM exposure in 1990,
jig/m3 (demographic group
experiencing this exposure)
NA
NA
1.3 (outdoor workers)
1.2 (outdoor workers)
0.8 (outdoor workers)
1.5 (outdoor workers)
4.0 (outdoor children)
1.2 (outdoor children)
2.4 (nonworking men 18-44)
2.0 (outdoor workers)
0.8 (outdoor workers)
NA - Not available.
Source: U.S. EPA, 1999b,
adjusted to reflect HDDV VMT described in U.S. EPA, 2000b.
       The HAPEM4 modeling approach provides exposure estimates from on-road and
nonroad sources as well as point and area sources for pollutants other than DPM. In addition,
HAPEM4 incorporates technical advancements over previous Agency exposure assessments.
Instead of using a surrogate pollutant such as CO to estimate exposure, HAPEM4 uses census
tract DPM concentrations provided by the ASPEN dispersion model described in Section 2.4.2.3
to estimate DPM exposure for individuals in each census tract in the United States.  The
exposure modeling results are aggregated to provide county, State, and nationwide exposure
estimates. HAPEM4 also incorporates the latest data regarding time-activity patterns from the
Consolidated Human Activity Database and the latest data available regarding penetration of PM
to indoor environments. The results of this modeling approach are currently undergoing peer
review and  are therefore considered a draft  and subject to change.
       Nationwide exposure estimates from HAPEM4 are provided in Table 2-30.  The draft
National-Scale Assessment 1996 national average estimate of DPM exposure attributable to on-
road and nonroad sources is  1.4 |ig/m3.  On-road sources are estimated to account for 0.5 |ig/m3
and nonroad sources 0.9 |ig/m3. The HAPEM-MS3 1996 exposure value of 0.8 |ig/m3 and the
                                        2-114

-------
most recent draft National-Scale Assessment value of 0.5 |ig/m3 differ slightly as a result of the
different modeling approaches. Both the HAPEM-MS3 and HAPEM4 exposure results support
the risk perspective provided in Chapter 8, Section 8.3.

2.4.3.2.2. Personal exposures: microenvironments/hotspots.  Personal monitoring for DPM
exposure has focused on occupationally exposed groups, including railroad workers, mine
workers, mechanics, and truck drivers. Although some studies have measured personal
exposures to ambient PM, none have conducted detailed chemical analysis to quantify the
portion of PM attributable to DE (e.g., using extended species CMB, discussed above). EC
concentrations have been reported for some microenvironments and are discussed in this section.
Microenvironmental exposures of significant concern include in-vehicle exposures such as
school buses and passenger cars as well as near highways and in urban canyons. Because DPM
from mobile sources is emitted into the breathing zone of humans, this source has  a greater
potential for human exposure (per kg of emissions) compared to combustion particulates emitted
from point sources.
       Recent EC measurements reported for enclosed vehicles driving on Sacramento
roadways ranged from below detection limits up to 10 |ig/m3 and from 3 |ig/m3 to 40 |ig/m3 on
Los Angeles roadways.  Elevated levels of PM25 and EC were observed when the  vehicle being
followed was powered by a HD diesel truck or bus (Cal EPA, 1998b).  EC is also present in the
exhaust of gasoline vehicles, so these measurements are likely to include some EC from gasoline
vehicles. The SHEDS (Stochastic Human Exposure and Dose Simulation) model  for PM
predicts that although the typical person spends only about 5% of his or her time in a vehicle,
this microenvironment can contribute on average 20% and as much as 40% of a person's total
PM exposure (Burke et al., 2000).
        The California Air Resources Board also collected EC near the Long Beach Freeway for
4 days in May 1993 and 3 days in December 1993 (Cal EPA, 1998a). Using emission estimates
from their EMFAC7G model  and EC and OC composition profiles for diesel and gasoline
exhaust, tire wear, and road dust,  CARB estimated the contribution of the freeway to DPM
concentrations.  For the 2 days of sampling in December 1993, DE from vehicles on the nearby
freeway was estimated to contribute from 0.7 |ig/m3 to 4.0 |ig/m3 excess DPM above background
concentrations, with a maximum of 7.5 |ig/m3.
       In 1986, EC concentrations were measured in Glendora, CA, during a carbonaceous
aerosol intercomparison study (Cadle and Mulawa, 1990; Hansen and Novakov, 1990). One
technique used during the study reported EC concentrations in 1-minute intervals, reflecting the
impact from diesel vehicles 50 m from the study site.  The diesel vehicles were estimated to
contribute up to 5 |ig/m3 EC above the background concentration.
                                        2-115

-------
       In a study designed to investigate relationships between DE exposure and respiratory
health of children in the Netherlands, EC measurements were collected in 23 schools located
from 47 m to 377 m from a freeway and in 8 schools located at a distance greater than 400 m
from a freeway (Brunekreef, 1999).  EC concentrations in schools near freeways ranged from 1.1
|ig/m3 to 6.3 |ig/m3, with a mean of 3.4 |ig/m3, and EC concentrations in schools more than 400
m from freeways ranged from 0.8 |ig/m3 to 2.1 |ig/m3, with a mean of 1.4 |ig/m3.  Brunekreef et
al. (2000), using a reflectance method to report "soot" or carbonaceous particulate concentrations
as a surrogate for EC, found a statistically significant increase in carbonaceous particle
concentrations inside and outside of the schools with increasing truck traffic (predominantly
diesel), with decreasing distance between the school and the highway, and with an increase in
the percent of time the school was downwind of the highway. In additional studies in elderly
subjects in Helsinki and Amsterdam, Janssen et al. (2000) reported that outdoor measurements of
EC were highly correlated with indoor and personal exposure measurements of EC, supporting
the position that  short-term increases in outdoor EC concentrations are reflected in increased
personal exposures even for those who spend much of their time indoors.
       Although there is little quantitative information regarding personal exposure to DPM,
certain exposure situations are expected to result in higher than average exposures. Those in the
more highly exposed categories would generally include people living in urban areas in which
diesel delivery trucks, buses, and garbage trucks frequent the roadways, but also included would
be people living  near freeways, bus stations, construction sites, train  stations, marinas frequented
by diesel-powered vessels, and distribution hubs using diesel truck transport.  One study using
the 1-hydroxypyrene biomarker of DE exposure reported exposure among most (76%) of the 26
adolescents sampled in Harlem (Northridge et al., 1999). In a follow-on study, Kinney  et al.
(2000) reported EC concentrations from personal monitors worn by study staff on sidewalks at
four Harlem intersections that ranged from 1.5 |ig/m3 to 6 |ig/m3.  The EC concentrations were
found to be associated with diesel bus and truck counts such that spatial variations in sidewalk
concentrations of EC were attributed to local diesel sources in Harlem.
       In any situation in which diesel engines operate and a majority of time is spent outdoors,
personal exposures to DE are expected to exceed average exposures. Because a large but
currently undefined portion of DPM is emitted during acceleration, those living and working in
the vicinity of sources operating in this transient mode could experience highly  elevated levels of
DPM.  DPM enriched in soluble organic material (as opposed to EC) is emitted from LD
vehicles, some nonroad equipment, on-road diesel engines during cold-start and motoring
conditions, and poorly maintained vehicles.  The potential health effects of acute exposures to
elevated DPM levels as well as health effects resulting from chronic  exposures are discussed in
subsequent chapters in this document.
                                         2-116

-------
2.4.3.2.3.  The California Population Indoor Exposure Model  CPIEM, developed under
contract to the CARB, estimates Californians' exposure to DPM using distributions of input data
and a Monte Carlo approach (Cal EPA, 1998a). This model uses population-weighted outdoor
DPM concentrations in a mass balance model to estimate DPM concentrations in four indoor
environments: residences, office buildings, schools, and stores/retail buildings.  The model takes
into account air exchange rates, penetration factors, and a net loss factor for deposition/removal.
In four additional environments (industrial plants, restaurants/lounges, other indoor places, and
enclosed vehicles), assumptions were made about the similarity of each of these spaces to
environments for which DPM exposures had been calculated.  Industrial plants and enclosed
vehicles were assumed to have DPM exposures similar to those in the outdoor environment;
restaurants/lounges were assumed to have DPM concentrations similar to stores; and other
indoor places were assumed to have DPM concentrations similar to offices. The estimated DPM
concentrations in the indoor and outdoor environments range from 1.6 |ig/m3 to 3.0 |ig/m3 (Table
2-32).
                 Table 2-32. Modeled and estimated concentrations of
                 DPM in microenvironments for California for all sources
Microenvironment
Residences
Offices
Schools
Stores/public/retail bldgs
Outdoor places
Industrial plants3
Restaurants/lounges3
Other indoor places3
Enclosed vehicles3
Estimated mean DPM
(stdev), jig/m3
1.9(0.9)
1.6(0.7)
1.9(0.8)
2.1 (0.9)
3.0(1.1)
3.0(1.1)
2.1(0.9)
1.6(0.7)
3.0(1.1)
                 "Concentrations assumed based on similarity with modeled environments.
                 Source: California EPA, 1998a.
                                         2-117

-------
    The DPM concentrations reported in Table 2-32 were used as input to CPIEM, and time-
activity patterns for children and adults were used to estimate total indoor and total air exposures
to DPM.  Overall, total indoor exposures were estimated to be 2.0 ± 0.7 ng/m3, and total air
exposures (indoor and outdoor exposures) were 2.1 ± 0.7 |ig/m3 (Table 2-33). The South Coast
Air Basin and the San Francisco Bay Area were also modeled using CPIEM, where total air
exposures to DPM were estimated to be 2.5 ± 0.9 |ig/m3 and 1.7 ± 0.9 |ig/m3,  respectively.
          Table 2-33. Estimated indoor air and total air exposures to DPM in
          California in 1990
Exposed population
All Californians
South Coast Air Basin
San Francisco Bay Area
Total indoor
exposure (stdev),
ug/m3
2.0 (0.7)
2.4 (0.9)
1.7(0.9)
Total air exposure,
(stdev), ug/m3
2.1 (0.8)
2.5 (0.9)
1.7(0.9)
           Source:  California EPA, 1998a.
      Exposure estimates were also made by Cal EPA (1998a) for 1995, 2000, and 2010 using a
ratiometric approach to 1990 exposures. Total air exposures reported for 1995 and projected for
2000 and 2010 were 1.5 |ig/m3, 1.3 |ig/m3, and 1.2 |ig/m3, respectively.

2.5. SUMMARY AND DISCUSSION
       This chapter summarizes information regarding the history of the use of diesel engines,
technological developments and their impact on emissions over time, Federal standards on DE,
the chemical and physical character of DE, atmospheric transformations of DE, and ambient DE
concentrations and exposures. The aspects of each of these topics that are most relevant to the
discussion of health effects in later chapters of this document are summarized here. Because the
majority of information regarding the chemical composition and historical changes in DE
pertains to on-road diesel engines, these data are discussed in greater detail than diesel emissions
from nonroad equipment.  Where possible, nonroad emissions were discussed in Chapter 2 and
are briefly summarized here.
                                         2-118

-------
 2.5.1. History of Diesel Engine Use, Standards, and Technology
       The use of diesel engines in the trucking industry began in the 1940s, and diesel engines
slowly displaced gasoline engines among HD trucks, accounting for 36% of new HD truck sales
in 1960, 85% of sales in 1970, and almost 100% of sales in 1997. It is estimated that in 2000,
HD diesel vehicles will travel more than 224 billion miles (U.S. EPA, 2000b).  In 1997, on-
highway HD diesel engines contributed 66% of the PM25 emitted by on-highway vehicles.
       To understand changes in emissions over time, it is important to note the difference
between model year emission trends and calendar year emission trends.  Emission trends by
model year refer to the year in which an engine was made; the emission rate is  specific to the
technology and regulations in effect for that year. Emissions in a specific calendar year refer to
aggregate emissions due to the mix of model year engines on the road. Because of the time
required for fleet turnover, emission rates for the on-road fleet in any calendar year are not as
low as the most recent model year emission rate. In 1997, 40% of the HD vehicles on the road
were at least 10 years old and traveled approximately 17% of total HD vehicle  miles.
       EPA set a  smoke standard for on-road HD diesel engines beginning with the  1970 model
year.  In the ensuing years, standards for PM from diesel engines for on-road applications
decreased from 0.6 g/bhp-hr in 1988 to  0.1 g/bhp-hr for trucks in 1994-1995 and 0.05 g/bhp-hr
for buses in 1996-1997.  Calendar year emission contributions of PM from diesel engines to
national PM10 inventories reflect decreases expected to result from Federal regulations, because
the emission factor models (MOBILES  and PARTS) used  to provide emission estimates for
mobile sources largely use engine test data required for certification. The U.S. EPA Trends
Report estimates that PM10 emissions attributable to on-road diesel vehicles decreased 27%
between 1980 and 1998.  DPM emission factors (g/mi by model year) measured from in-use
vehicles decreased on average by a factor of six from the mid-1970s to the mid-1990s.
       It is important to note that in spite of the decreasing trend in  DPM emission factors by
model year, a wide range in emission factors from in-use testing is reported, even for newer
model year HD vehicles (from less than 0.1 g/mi to more than  1 g/mi for model year 1996
vehicles). The high variability in DPM emissions within one model year has been attributed to
deterioration3 and differences in measurement methods and test conditions at the various testing
facilities.  Studies in which consistent testing methods were used suggest that deterioration (even
for newer model year engines) causes some of the variability in emission factors, whereas other
       Deterioration includes increases in emission rates (g/bhp hr) due to normal wear as well as manufacturing
defects and malfunctions such as retarded timing, fuel injector malfunction, smoke limiting mechanism problems,
clogged air filter, wrong or worn turbocharger, clogged intercooler, engine mechanical failure, excess oil
consumption, and electronics that have been tampered with or have failed.
                                          2-119

-------
studies clearly demonstrate the important influence of test conditions and driving protocols (e.g.,
aggressive driving) on DPM emission factors.
       Even though significant reductions in DPM from diesel vehicle emissions for on-road
applications have been realized, diesel engines (nonroad and on-road combined) are still
significant contributors to 1998 inventories of particulate matter, contributing approximately
23% of PM25 emissions (not including the contribution from natural and miscellaneous sources).
       Technology innovations that impact diesel engine emissions have occurred in the years
since 1960, in particular the advent of turbocharging with charge air cooling and direct-injection
engines.  The use of these new technologies tends to lower emissions from on-road diesel
engines; until the late 1970s, however, engines were optimized for performance rather than
emissions, so the effect on emissions prior to this time was small. The limited amount of data
available indicates that on-road engines in the 1950 to 1975 timeframe had DPM emissions
similar to, and in some cases higher than, those of the mid-1970 engines that were not yet
controlled for particulates.
       Few data are available to assess the changes in emission rates from locomotive, marine,
or other nonroad diesel engine sources over time.  It is expected that because  the typical lifespan
of a locomotive engine is at least 40 years and PM regulations for these engines do not take
effect until  2000, PM emission rates by model year from locomotives are not likely to have
changed substantially since the introduction of the diesel engine into the railroad industry in the
early 1950s.
       Particulate matter regulations for nonroad diesel equipment are not as stringent as PM
regulations for on-road diesel  engines. Although PM emissions have declined for on-road
trucks, it is estimated that PM10 emissions from nonroad diesel engines increased 17% between
1980 and 1998. DPM emissions from nonroad diesel engines are expected to continue to
increase from current levels in the absence of new regulations. No information is available
regarding changes in the chemical composition of nonroad engine emissions over time.

2.5.2. Physical and Chemical Composition of Diesel Exhaust
       Complete and incomplete combustion of fuel in the diesel engine results in the formation
of a complex  mixture of hundreds of organic and inorganic compounds in the gas and particle
phases. Among the gaseous components of DE, the aldehydes are particularly important because
of their health effects and because they are an important fraction of the gaseous emissions.
Formaldehyde makes up a majority of the aldehyde emissions (65%-80%) from diesel engines,
with the next  most abundant aldehydes being acetaldehyde and acrolein. Other gaseous
components of DE that are notable for their health effects include benzene, 1,3-butadiene, PAH,
and nitro-PAH. Dioxin compounds have also been detected in trace quantities in DE and
                                         2-120

-------
currently account for 1.2% of the national inventory. Dioxin compounds are known to
accumulate in certain foods, such as beef, poultry, and dairy products.  It is unknown whether
deposition of DE emissions has an impact on food chains in local areas.
       DPM contains EC, OC, and small amounts of sulfate, nitrate, metals, trace elements,
water, and unidentified compounds. DPM is typically composed of more than 50% to
approximately 75% EC depending on the age of the engine, deterioration, HD versus LD, fuel
characteristics, and driving conditions.  The OC portion of DPM originates from unburned fuel,
engine lubrication oil, and low levels of partial combustion and pyrolysis products and typically
ranges from approximately 19% to 43%, although the range can be broader depending on many
of the same factors that influence the EC content of DPM. Polyaromatic hydrocarbons generally
constitute less than 1% of the DPM mass. Metal compounds and other elements in the fuel and
engine lubrication oil are exhausted as ash and typically make up l%-5% of the DPM mass.
Elements and metals detected in DE include barium, calcium, chlorine, chromium, copper, iron,
lead, manganese, mercury,  nickel, phosphorus, sodium, silicon, and zinc. The composition of
DPM contrasts strongly with the typical chemical composition of ambient DPM25 that is
dominated by sulfate for aerosols measured in the eastern United States and by nitrate,
ammonium, and OC in the western United States.
       Approximately 1% to 20% of the mass of DPM in DE is in the ultrafine size range
(nuclei-mode), with the majority of particles ranging in size from 0.005 to 0.05 microns and
having a mean diameter of about 0.02 microns. These particles account for 50%-90% of the
number of particles. These ultrafine particles are largely composed of sulfate and/or sulfate with
condensed OC.
       Evidence regarding an increase in the number of ultrafine particles from new HD engines
is inconclusive.  The dilution conditions used to measure the size distribution of DE have a large
impact on the number of ultrafine particles quantified.  To understand the size distribution of
DPM to which people are exposed will  require measurements under conditions that more closely
resemble ambient conditions.
       Approximately 80%-95% of the mass of particles in DE is in the size range from 0.05-1.0
microns, with a mean particle diameter  of about 0.2 microns, and therefore in the fine PM size
range.  Diesel particles in the 0.05-1.0 micron range are aggregates of primary spherical particles
consisting of an EC core, adsorbed organic compounds, sulfate,  nitrate, and trace elements.
These particles have a very large surface area per gram of mass, which makes them an excellent
carrier for adsorbed inorganic and organic compounds and, due to their small size, they can
effectively reach the lower portions of the respiratory tract.  The EC core has a high specific
surface area of approximately 30-90 m2/g.
                                         2-121

-------
       Because of the potential toxicological significance of the organic components associated
with DPM, it is important to understand, to the extent possible, the historical changes in the
amount and composition of the DPM-associated organic fraction.  The organic component of
DPM has typically been characterized by extraction with organic solvents, although other
techniques such as thermogravimetric methods have also been used. Results from studies using
similar extraction methods were compared to characterize historical changes in the SOF
emission rates, the percentage of DPM comprised by SOF, and the composition of SOF.  Data
from both engine and chassis dynamometer tests  suggest that SOF emission rates have decreased
by model year from 1975 to 1995. When expressed as a percentage of total DPM, the
contribution of SOF to total DPM demonstrates a wide range of variability that may be attributed
to different test cycles, different engine types, and different deterioration rates among the
vehicles tested.  Currently, LD diesel engines emit DPM with a higher fraction of SOF than do
FID engines.
       Chassis dynamometer tests demonstrate an overall decrease in the mass percentage
contribution of SOF to DPM, ranging from 10% to 60% in the 1980s and -5% to 20% in the
1990s.  In contrast, engine dynamometer tests demonstrate that typically 10%-50% of DPM mass
is soluble organic matter for engines in model years 1980-1995. The higher SOF fraction of
DPM from 1990s model year engine dynamometer tests is attributed primarily to the differences
in the engine and chassis dynamometer driving cycles.  The engine dynamometer testing
includes high- speed and low-load or low-speed lugging test modes in the engine Federal Test
Procedure that produce DPM with a high SOF fraction.
       The chassis dynamometer data are considered to reflect real-world trends in emissions
from heavy HD vehicles by model year because vehicles from different model years, with
different mileage and different levels of deterioration, are represented.  Thus,  it is expected that
the percentage of SOF from new (1990 or later) model year heavy HD diesel vehicles is lower
than that from older vehicles. This expectation is supported by  data demonstrating an overall
increase in the fraction of EC in the carbonaceous component of DPM.  The important
observation from the engine test data is that some driving modes occurring in real-world
applications even with new (post-1990) engines may produce DPM with a high SOF component
(up to 50%).
       PAH and nitro-PAH are present in DPM from both new and older engine exhaust. There
is no information to suggest that the overall PAH composition profile for DPM has changed.
There are too few data to speculate on the changes in emissions of total PAH, nitro-PAH, or
PAH and nitro-PAH components such as BaP and 1-NP. The data suggest that differences in a
vehicle's engine type and make, general engine condition, fuel composition, and test conditions
can influence the emissions levels of PAH.  Some studies suggest that fuel composition is the
most important determinant of PAH emissions. There is limited evidence that gas-phase PAH
                                        2-122

-------
emission rates increase with higher fuel PAH content and that some particle-phase PAH
emission rates increase with higher fuel PAH content. These data suggest that during the period
from 1960 to  1986, when the aromatic content of fuel increased, PAH emissions may have
increased until the aromatic content of diesel fuel was capped in 1993.  The aromatic content of
nonroad diesel fuel is not federally regulated and is typically greater than 30%. PAH emissions
from nonroad equipment would also be expected to vary with the PAH content of the fuel.
       Currently, information regarding emission rates, chemical composition, and relative
contribution of DPM from high-emitting HD diesel vehicles is not available and may
significantly change the understanding of DPM composition to which people are exposed. Some
studies have reported a substantial number of smoking diesel trucks in the in-use fleet.  Although
the correlation between smoke and paniculate concentration varies with the driving cycle and
measurement  method, the results of smoke opacity tests suggest that high-emitting HD diesel
vehicles may be important contributors to ambient DE and DPM concentrations.
       The chemical composition of DPM to which people are currently exposed is determined
by a combination of older and newer technology on-road and nonroad engines. Consequently,
the decrease in the SOF of DPM by model year does not directly translate into a proportional
decrease in DPM-associated organic material to which people are currently exposed.  In
addition, the impact from high-emitting and/or smoking diesel engines is not quantified at this
time.  Because of these uncertainties, the changes in DPM composition over time cannot
presently be quantified. The data clearly indicate that lexicologically significant organic
components of DE (e.g., PAHs, PAH derivatives, nitro-PAHs) were present in DPM  and DE in
the 1970s and are still present in DPM and DE as a whole.
       Although a significant fraction  of ambient DPM (over 50% is possible) is also emitted by
nonroad equipment, there are no data available to characterize changes in the chemical
composition of DPM from nonroad equipment over time.
       Some analysts project that diesel engines will increase substantially in the LD fleet in
coming years. Although LD engines currently emit DPM with higher SOF than  HD engines of
the same model year, recently promulgated Tier 2 standards will require control  measures in the
2004-2007 timeframe that will reduce PM emissions from these vehicles. These control
measures provide some assurance that  even if LD diesel use increases, DPM emitted  from these
vehicles will likely have a smaller SOF component than such engines currently emit.

2.5.3. Atmospheric Transformation  of Diesel Exhaust
       An understanding of the physical as well as chemical transformations of DE in the
atmosphere is necessary to fully understand the impact of this complex chemical mixture  on
human health. In the past two decades, data acquired from laboratory and ambient experiments
                                         2-123

-------
have provided information regarding the atmospheric loss processes and transformation of DE,
but knowledge concerning the products of these chemical transformations is still limited.  A
recent study has suggested that DPM exposed to ambient levels of ozone is sufficiently altered to
increase the rat lung inflammatory effect compared with DPM not exposed to ozone.
       Studies investigating the chemical and physical changes of DE emissions suggest that
there is little or no hygroscopic growth of primary diesel particles.  This observation suggests
that the small size of DPM particles might be maintained upon inhalation, particularly near the
emission source, allowing these particles to reach the lower portions of the respiratory tract.
Increased solubility can increase the removal efficiency of secondary  diesel particles compared
with their precursor compounds. Secondary aerosols from DE may also exhibit different
biological reactivities from the primary particles.  For example, there  is  evidence for nitration  of
some PAH  compounds resulting in the formation of nitroarenes that are often more mutagenic
than their precursors.

2.5.4. Ambient Concentrations and Exposure to Diesel Exhaust
      Because of changes in engine technology and DPM emissions over time, ambient
concentrations reported from studies before 1990 are compared here to those reported  after 1990.
There are no studies in which direct comparisons can be made because of different analytical and
modeling tools used to assess DPM ambient levels.
      DPM concentrations reported from CMB and dispersion modeling studies in the 1980s
suggest that in urban and suburban areas (Phoenix, AZ, and Southern California), annual average
DPM concentrations ranged from 2 to 13 |ig/m3, with possible maximum daily values in Phoenix
of 22 |ig/m3. In these studies, the average contribution of DPM in urban areas to total  ambient
PM ranged  from 7% in Pasadena, CA, to 36% in Los Angeles.
      In the 1990 timeframe, annual or seasonal average DPM concentrations reported in CMB
studies and from EC measurements for urban and suburban areas range from 1.2 to 4.5 |ig/m3.
The contribution of DPM to ambient PM at these sites averaged 10%-15% on a seasonal or
annual basis, with contributions up to 38% on individual days (Brighton, CO). Dispersion
modeling on individual  days in Southern California in the 1990s predicts DPM concentrations
ranging from 1.9 to 4.4  |ig/m3 (8%-12% of ambient PM).  On individual days at a major bus stop
in New York City, DPM concentrations were reported to reach 46.7 |ig/m3 and averaged 53%  of
ambient PM, highlighting the important influence of diesel bus traffic in an urban street canyon.
      In nonurban and rural areas in the 1980s, DPM concentrations reported range from 1.4  to
5 |ig/m3and on average comprised 5%-12% of the ambient aerosol. In the 1990s, nonurban air
basins in California were reported to have DPM concentrations ranging from 0.2-2.6 |ig/m3.
                                         2-124

-------
       Although estimates from emissions models suggest that DPM emissions from on-road
sources decreased during the 1990s, the atmospheric data available do not provide a clear
indication of trends in DPM concentrations but are likely to be more a reflection of the choice in
sampling sites, source apportionment methods, and modeling techniques.  In general, from the
limited number of studies available it appears that DPM concentrations averaged over at least a
season in the 1990s typically ranged from 1-4 |ig/m3. These data can be used in model-monitor
comparisons and to provide  an indication of long-term average exposures in some urban areas.
Additional work is needed to assess ambient DPM and DE concentrations in several urban
environments, to assess microenvironments, and to evaluate the relative impact of nonroad and
on-road sources on concentrations.
       A comprehensive exposure assessment cannot currently be conducted because of the lack
of data. Information regarding DPM in occupational environments suggests that exposure
ranges up to approximately 1,280 |ig/m3 for miners, with lower exposure measured for railroad
workers (39-191 |ig/m3), firefighters (4-748 |ig/m3), public transit personnel who work with
diesel equipment (7-98 |ig/m3),  mechanics  and dockworkers (5-65 |ig/m3), truck drivers (2-7
|ig/m3), and bus drivers (1-3 |ig/m3).  Work area concentrations at fixed sites are often higher
than measured exposures, especially for mining operations or other enclosed spaces. For several
occupations involving DE exposure, an increased risk of lung cancer has been reported by
epidemiologic  studies (discussed in Chapter 7).  An estimate of the 70-year lifetime
environmental  exposure equivalent to these occupational exposures provides one means of
comparing the  potential overlap between occupational exposures and exposures modeled for the
general public.  The estimated 70-year lifetime exposures equivalent to those for the
occupational groups discussed above range from 0.4-2 |ig/m3 on the low end to 2-269 |ig/m3 on
the high end.
       The EPA has performed a national-scale exposure assessment for DPM from on-road
sources.  Current national exposure modeling using the HAPEM-MS3 model suggests that in
1996, annual average DPM exposure from  on-road DE  sources in urban areas was 0.8 |ig/m3,
whereas in rural areas, exposures were 0.4  |ig/m3.  Among 10 urban areas in which DPM
exposures were modeled,  1996 annual average exposure from on-road DE sources ranged from
0.6 |ig/m3 to 1.2 |ig/m3. Outdoor workers and children who spent a large amount of time
outdoors were  estimated to have elevated DPM exposures in 1990, ranging up to 4.0 |ig/m3from
on-road sources only. Based on the national inventory, nonroad emission  sources could
contribute at least twofold more DPM than that emitted by on-road sources.  Results of the draft
National-Scale Assessment for  1996 indicate that national average exposure to DPM, including
nonroad sources, is 1.4 |ig/m3, with 0.9 |ig/m3 of that average attributed to emissions from
nonroad sources.
                                         2-125

-------
       Low-end exposures for many of the occupational groups overlap 1990 and 1996

exposures from on-road sources modeled for the general population (0.8 |ig/m3) and for the more

highly exposed groups. This potential overlap, or small difference between occupational and

ambient exposures, presents a concern that health effects observed in occupational groups may

also be evidenced in the general population.  The potential magnitude of this  risk is discussed in

Chapter 8.

       In different exposure environments, the types of diesel vehicles, their mode of operation,

maintenance, atmospheric transformation, and many additional factors influence the chemical

nature and quantity of DPM to which people are exposed.  The potential health consequences of

both short- and long-term exposures to DE are discussed in the following chapters of this

document.


                              REFERENCES FOR CHAPTER 2

Abbass, MK; Andrews, GE; Ishaq, RB; et al. (1991) A comparison of the paniculate composition between
turbocharged and naturally aspirated DI diesel engines. SAE Technical Paper Ser. No. 910733.

Abdul-Khalek, IS; Kittelson, DB; Graskow, BR; et al. (1998) Diesel exhaust particle size: measurement issues and
trends. SAE Technical Paper Ser.  No. 980525.

Abdul-Khalek, IS; Kittelson, DB; Brear, F. (1999)  The influence of dilution conditions on diesel exhaust particle
size distribution measurements. SAE Technical Paper Ser. No.  1999-01-1142.

American Automobile Manufacturers Association (AAMA). (1927-1974) Motor truck facts. Motor Truck
Committee.  Washington, DC.

AAMA. (1975-1998) Motor vehicle facts and figures.  Washington, DC.

Andrews, GE; Ishaq, RB; Farrar-Khan, JR; et al. (1998) The influence of speciated diesel fuel composition on
speciated paniculate SOF emissions.  SAE Technical Paper Ser. No. 980527 (and references therein).

Arey, J. (1998) Atmospheric reactions of PAHs including formation of nitroarenes. In: The handbook of
environmental chemistry, Vol. 3, Part I. PAHs and related compounds. Nielsen, AH, ed. Berlin/Heidelberg,
Germany: Springer-Verlag, pp. 347-385.

Arey, J; Zielinska, B; Atkinson, R; et al. (1986) The formation of nitro-PAHs from the gas-phase reactions of
fluoranthene and pyrene with the OH radical in the presence of NOX. Atmos Environ 20:2339-2345.

Arey, J; Zielinska, B; Atkinson, R; et al. (1987) Polycyclic aromatic hydrocarbon and nitroarene concentrations in
ambient air during a wintertime high-NOx episode in the Los Angeles Basin. Atmos Environ 21:1437-1444.

Arey, J; Zielinska, B; Atkinson, R; et al. (1988) Formation of nitroarenes during ambient high-volume sampling.
Environ Sci Technol 22:457-462.

Arey, J; Zielinska, B; Atkinson, R; et al. (1989) Nitroarene products from the gas-phase reactions of volatile
poly cyclic aromatic hydrocarbons with OH radical and N2O5. Int J Chem Kinet 21:775-799.

Atkinson, R. (1986) Kinetics and mechanisms of the gas-phase reactions of the hydroxyl radical with organic
compounds under atmospheric conditions. Chem Rev 86:69-201.


                                             2-126

-------
Atkinson, R. (1991) Kinetics and mechanisms of the gas-phase reactions of the NO3 radical with organic
compounds. J Phys ChemRef Data 20:459-507.

Atkinson, R; Arey, J. (1994) Atmospheric chemistry of gas-phase polycyclic aromatic hydrocarbons: formation of
atmospheric mutagen. Environ Health Perspect 102(Suppl. 4): 117-126.

Atkinson, R; Arey, J; Zielinska, B; et al. (1987) Kinetics and products of the gas-phase reactions of OH radicals and
N2O5 with naphthalene andbiphenyl. Environ Sci Technol 21:1014-1022.

Atkinson, R; Arey, J; Winer, AM; et al. (1988) A survey of ambient concentrations of selected polycyclic aromatic
hydrocarbons (PAH) at various locations in California. Final report prepared under contract no. A5-185-32 for
California Air Resources Board, Sacramento, CA.

Atkinson, R; Arey, J; Zielinska, B; et al. (1990) Kinetics and nitro-products of the gas-phase OH and NO3 radical-
initiated reactions of naphthalene-d8 fluoranthene-d10 and pyrene. Int J Chern Kinet 22:999-1014.

Bagley, ST; Baumgard, KJ; Gratz, LD. (1991) Comparison of in-min and laboratory-generated diesel paniculate
matter, biological activity, and polynuclear aromatic hydrocarbon levels. In: Proceedings of the Third Symposium
of Respirable Dust in the Mineral Industries.  Franz, RL; Ramini, RV, eds. Littleton, CO: Society for Mining,
Metallurgy, and Exploration, pp. 61-70.

Bagley, ST; Baumgard, KJ; Gratz, LD. (1992) Polynuclear aromatic hydrocarbons and biological activity associated
with diesel paniculate matter collected in underground coal mines.  IC9324. Pittsburgh, PA: Bureau of Mines, pp.
40-48.

Bagley, ST; Gratz, LD; Leddy, DG; et al. (1993) Characterization of particle and vapor phase organic fraction
emissions from a HD diesel engine equipped with a particle trap and regeneration controls. Health Effects Institute,
Research Report No. 56.  Cambridge, MA: HEI.

Bagley, ST; Baumgard, KJ; Gratz, LD; et al. (1996) Characterization of fuel and aftertreatment effects on diesel
emissions.  Health Effects Institute, Research Report No. 76. Cambridge, MA: HEI.

Bagley, ST; Gratz, LD; Johnson, JH; et al. (1998) Effects of an oxidation catalytic converter and a biodiesel fuel on
the chemical, mutagenic, and particle size characteristics of emissions from a diesel engine.  Environ Sci Technol
32:1183-1191.

Ballschmiter, K.; Buchert, H.; Niemczyk, R.; et al. (1986) Automobile exhausts versus municipal waste
incineration as sources of the polychloro-dibenzodioxins (PCDD) and -furans (PCDF) found in the environment.
Chemosphere 15(7):901-915.

Baranescu,  RA. (1988) Influence of fuel sulfur on diesel paniculate emissions. SAE Technical Paper Ser. No.
881174.

Barry, EG; McCabe, LJ;  Gerke, DH; et al. (1985) Heavy-duty diesel engine/fuels combustion performance and
emissions~a cooperative research program.  SAE Technical Paper Ser. No. 852078.

Bata, R; Wang, W; Gautam, M; et al. (1992) Fleet-site measurements of exhaust gas emissions from urban buses.
ASME, ICE. New Dev Off-Highway Eng 18:185-196.

Baumgard,  KJ; Johnson,  JH. (1992) The effect of low sulfur fuel and a ceramic particle filter on diesel exhaust
particle size distributions. SAE Technical Paper Ser. No. 920566.

Baumgard,  KJ; Johnson,  JH. (1996) The effect of fuel and engine design on diesel exhaust particle  size
distributions.  SAE Technical Paper Ser. No. 960131.
                                                2-127

-------
Behymer, TD; Hites, RA. (1985) Photolysis of polycyclic aromatic hydrocarbons adsorbed on simulated
atmospheric participates. Environ Sci Technol 19:1004-1006.

Behymer, TD; Hites, RA. (1988) Photolysis of polycyclic aromatic hydrocarbons adsorbed on fly ash. Environ Sci
Technol 22:1311-1319.

Benner, BA; Gordon, GE; Wise, SA. (1989) Mobile sources of atmospheric PAH: a roadway tunnel study. Environ
Sci Technol 23:1269-1278.

Birch, ME. (1998) Analysis of carbonaceous aerosols: interlaboratory comparison. Analyst 123:851-857.

Birch, ME; Gary, RA. (1996) Elemental carbon-based method for monitoring occupational exposures to paniculate
diesel exhaust.  Aerosol Sci Technol 25:221-241.

Bishop, GA; McLaren, SE; Stedman, DH; et al. (1996) Method comparisons of vehicle emissions measurements in
the Fort McHenry and Tuscarora Mountain Tunnels. Atmos Environ 30:2307.

Black, F; High, L. (1979) Methodology for determining paniculate and gaseous diesel hydrocarbon emissions. SAE
Technical Paper Ser. No. 790422.

Brodowicz, P. (1999) Determination of the demographic groups with the highest annual average modeled diesel
PM exposure. Memorandum. U.S. Environmental Protection Agency, National Fuel and Vehicle Emissions
Laboratory.

Brown, DF; Rideout, GR. (1996) Urban driving cycle test results of retrofitted diesel oxidation catalysts on heavy-
duty vehicles. SAE Technical Paper Ser. No.  960134.

Brown, KF; Rideout, GR; Turner, JE. (1997) Urban driving cycle results of retrofitted diesel oxidation catalysts on
heavy duty vehicles: one year later. SAE Technical Paper Ser. No. 970186.

Brown, JE; Clayton, JM; Harris, DB; et al. (2000) Comparison of the particle size distribution of heavy-duty DE
using a dilution tailpipe  sampler and an in-plume sampler during on-road operation. J Air Waste Manage Assoc
50:1407-1416.

Brunekreef, B.  (1999) Environmental diesel exhaust exposure and respiratory health of children in the Netherlands.
Presented at the Health Effects Institute Diesel Workshop: Building a Research Strategy to Improve Risk
Assessment. Stone Mountain, GA, March 7-9, 1999.

Brunekreef, B;  Janssen,  NAH; van Vliet, PHN; et al. (2000) Paniculate matter concentrations in relation to degree
of urbanization and proximity to highways in the Netherlands. Presented at: PM2000: Paniculate Matter and
Health - The Scientific Basis for Regulatory Decision-Making. Specialty Conference and Exhibition. January 24-
28, 2000, Charleston, SC.

Burke, JB; Aufall, M; Ozkaynak, H; et al. (2000) Predicting population exposures to PM: the importance of
microenvironmental concentrations and human activities. Presented at: PM2000: Paniculate Matter and Health -
The Scientific Basis for Regulatory Decision-Making. Specialty Conference and Exhibition. January 24-28, 2000,
Charleston, SC.

Cadle, SH; Mulawa, PA. (1990) Atmospheric carbonaceous species measurement methods comparison study:
General Motors results.  Aerosol Sci Technol  12(1): 128-141.

California Air Resources Board (CARB).  (1987)  Determination of PCDD and PCDF emissions from motor
vehicles.  Draft report. October 1987.  Test report no. C-86-029.
                                                2-128

-------
California Air Resources Board and Energy and Environmental Analysis, Inc. (CARB/EEAI). (1997) Regulatory
Amendments to California's heavy-duty inspection program (HDVIP) and periodic smoke inspection program
(PSIP).  August 1997.

California Environmental Protection Agency (Cal-EPA). (1998a) Report to the California Air Resources Board on
the proposed identification of diesel exhaust as a toxic air contaminant. Appendix III, part A: Exposure assessment.
April 1998. http://www.arb.ca.gov/regact/diesltac/diesltac.htm.

Cal-EPA. (1998b) Measuring concentrations of selected air pollutants inside California vehicles. Final report.

Calvert, JG; Stockwell, WR. (1983) Acid generation in the troposphere by gas-phase chemistry. Environ Sci Technol
17:428A-443A.

Campbell, RM; Lee, ML. (1984) Capillary column gas chromatographic determination of nitro poly cyclic aromatic
compounds in paniculate extracts. Anal Chem 56:1026-1030.

Cantrell, BK; Salas, LJ; Johnson, WB; et al. (1988) Phase distributions of low volatility organics in ambient air.
Prepared for the U.S. Environmental Protection Agency, Washington, DC, under EPA contract no. 68-02-3748.

Carey, PM. (1987) Air toxics emissions from motor vehicles. U.S. Environmental Protection Agency,
Washington, DC, Office of Mobile Sources, Report No. EPA/AA-TSS-PA-86-5.

Cass, GR; Gray, HA. (1995) Regional emissions and atmospheric concentrations of diesel engine paniculate matter:
Los Angeles as a case study. In: Diesel exhaust: a critical analysis of emissions, exposure, and health effects. A
special report of the Institute's Diesel Working Group. Cambridge, MA: Health Effects Institute, pp.  125-137.

Cautreels, W; Van Cauwenberghe, K. (1978) Experiments on the distribution of organic pollutants between
airborne paniculate matter and the corresponding gas phase. Atmos Environ 12:1133-1141.

Chow, JC; Watson, JG; Richards, LW; et al. (1991) The 1989-1990 Phoenix PM10 study. Volume II: source
apportionment. Final report.  DRI Document No. 8931.6F1. Prepared for Arizona Department of Environmental Air
Quality, Phoenix, AZ, by Desert Research Institute, Reno, NV.

Chow, JC; Watson, JG; Crow, D; et al. (2000) Comparison of IMPROVE and NIOSH carbon measurements.
Accepted for publication in Aerosol Sci Technol.

Christoforou, CS; Salmon, LG; Hannigan, MP; et al. (2000) Trends in fine particle concentration and chemical
composition in Southern California. J Air Waste Manage Assoc 50:43-53.

Ciccioli, P; Cecinato, A;  Brancaleoni, E; et al. (1989) Evaluation of nitrated PAH in anthropogenic emission and air
samples: a possible means of detecting reactions of carbonaceous particles in the atmosphere. Aerosol Sci Technol
10:296-310.

Ciccioli, P; Cecinato, A;  Cabella, R; et al. (1993) The contribution of gas-phase reactions to the nitroarene fraction
of molecular weight 246 present in carbon particles sampled in an urban area of northern Italy. Atmos Environ
27A:1261.

Clark, N; Messer, JT; McKain, DL; et al. (1995) Use of the West Virginia University truck test cycle to evaluate
emissions from Class 8 trucks. SAE Technical Paper Ser. No. 951016.

Clark, N; Gautam, M; Lyons, D; et al. (1997) Natural gas and diesel transit bus emissions: review and recent data.
SAE Technical Paper Ser. No.  973203.

Clark, NN; Gautam, M; Rapp, BL; et al. (1999) Diesel and CNG transit bus emissions characterization by two
chassis dynamometer laboratories: results and issues. SAE Technical Paper Ser. No. 1999-01-1469.
                                                2-129

-------
Cohen, LH; Countess, RJ; Countess, SJ. (1997) Advanced remote sensing technology demonstration. Mobile
Sources Air Pollution Review Committee.

Coifman, B. (1994) Evolution of the diesel locomotive in the United States.
http://www.cyberus.ca/~yardlimit/guide/locopaper.html

Cooper, C. (1999) Heavy-duty diesel smoke enforcement program in the Northeast. Presented at the Mobile
Sources Technical Review Subcommittee, July 1999.

Cooper, JA. et al.; and NEA, Inc. (1987) PM10 source composition library for the South Coast Air Basin, volumes I
and II. Prepared for the South Coast Air Quality Management District, El Monte, CA. July 15,  1987.

Countess, R; Countess, S; Cohen, R. (1999) Remote sensing of heavy duty diesel trucks revisited. 9th Coordinating
Research Council On-Road Vehicle Emissions Workshop, San Diego, CA.

Coutant, RW; Grown, L; Chuang, JC; et al. (1988) Phase distribution and artifact formation in ambient air sampling
for polynuclear aromatic hydrocarbons. Atmos Environ 22:403-409.

Crebelli, R; Conti, L; Crochi, B; et al. (1995) The effect of fuel composition on the mutagenicity of diesel engine
exhaust.  Mutat Res 346:167-172.

Cunningham, LF; Henly, TJ; Kulinowski, AM. (1990) The effects of diesel ignition improver in low-sulfur fuels on
heavy-duty diesel emissions. SAE Technical Paper Ser. No. 902173.

Dec, JE; Espey,  C. (1995) Ignition and early soot formation in a diesel engine using multiple 2-D imaging
diagnostics.  SAE Technical Paper Ser. No. 950456.

Dickson, CL; Sturm, GP. (1994) Diesel fuel oils,  1994. National Institute for Petroleum and Energy Research.
NIPER-187 PPS 94/5.

Dietzmann, HE; Parness, MA; Bradow, RL. (1980) Emissions from trucks by chassis version of 1983  transient
procedure. SAE Technical Paper Ser. No. 801371.

Dua, SK; Hopke, PK; Raunemaa, T.  (1999) Hygroscopicity of diesel aerosols. Water Air Soil Pollut 112:247-257.

Dunlap, LS; Pellegrin, V; Ikeda, R; et al.  (1993) Chassis dynamometer emissions testing results  for diesel and
alternative-fueled transit buses. SAE Technical Paper Ser. No. 931783.

Dunstan, TDJ; Mauldin, RF; Jinxian, Z; et al. (1989) Adsorption and photodegradation of pyrene on magnetic,
carbonaceous, and mineral subtractions of coal stack ash. Environ Sci Technol 23:303-308.

Enya, T; Suzuki, H; Watanabe, T; et al. (1997) 3-Nitrobenzanthrone, a powerful bacterial mutagen and suspected
human carcinogen found in diesel exhaust and airborne particles. Environ Sci Technol 31:2772-2776.

Federal Highway Administration (FHA).  (1995) Motor fuel use, 1919-1995. U.S. Department of Transportation:
www.fhwa.dot.gov/ohim/Summary95/sectionl.html; Vol. MF221.

FHA. (1997) Highway statistics summary to 1995. U.S. Department of Transportation. FHWA-PL-97-009.

Feilberg, A; Kamens, RM; Strommen, MR; et al.  (1999) Modeling the formation, decay, and partitioning of
semivolatile nitro-polycyclic aromatic hydrocarbons (nitronaphthalenes) in the atmosphere. Atmos Environ
33:1231-1243.

Ferguson, DH; Gautam, M; Wang, WG; et al. (1992) Exhaust emissions from in-use heavy duty vehicles tested on a
transportable transient chassis dynamometer. SAE Technical Paper Ser. No. 922436.
                                                2-130

-------
Finlayson-Pitts, BJ; Pitts, JN, Jr. (1986) Atmospheric chemistry: fundamentals and experimental techniques. New
York: John Wiley & Sons.

Foster, DE; Tree, DR. (1994) Optical measurements of soot particle size, number density and temperature in a
direct injection diesel engine as a function of speed and load. SAE Technical Paper Ser. No. 940270.

Fowler, DP. (1985) Industrial hygiene/environmental sampling program to develop qualitative and quantitative data
for diesel exhaust emission exposure. Final report. Coordinating Research Council-APRAC Project No. CAPM-24-
78.

Frey, JW; Corn, M. (1967)  Diesel exhaust particulates. Nature 216:615-616.

Friones, JR; Hinds, WC; Duffy, RM; Lafuente, EJ; Liu, WV. (1987) Exposure of firefighters to diesel emissions in
fire stations. Am Ind Hyg Assoc J 48:202-207.

Fritz, SG. (1995) Emission measurements-locomotives. Prepared by Southwest Research Institute under U.S. EPA
Contract no. 68-C2-0144.

Fritz, SG; Bailey, CR; Scarbro, CA; et al. (2001) Heavy-duty diesel truck in-use emission test program for model
years 1950 through 1975.  SAE Technical Paper Ser. No. 2001-01-1327.

Fujita, EM; Lu, Z; Sheetz, L; et al. (1997) Determination of mobile source emission source fraction using ambient
field measurements. Final report.  Prepared for Coordinating Research Council, Atlanta, GA, by Desert Research
Institute, Reno, NV, under CRC Project No. E-5-1.

Fujita, E; Watson, JG; Chow, JC;  et al. (1998) Northern Front Range Air Quality Study, volume C: source
apportionment and simulation methods and evaluation.  Prepared for Colorado State University, Cooperative
Institute for Research in the Atmosphere, by Desert Research Institute, Reno, NV.

Gautam, M; Ferguson, D; Wang, WG; et al. (1992) In-use emissions and performance monitoring of heavy duty
vehicles using a transportable transient chassis test facility. SAE Technical Paper Ser. No. 921751.

Gertler, AW. (1999) Real-world measurements of diesel paniculate matter. Presentation to CASAC-Mobile
Sources Technical Review Subcommittee, October 13, 1999.

Gertler, AW; Sagebiel, JC;  Dippel, WA; et al.  (1996) A study to quantify on-road emissions of dioxins and furans
from mobile sources:  phase 2. Reno, NV: Desert Research Institute.

Gertler, AW; Sagebiel, JC;  Dippel, WA; et al. (1998) Measurements of dioxin and furan emission factors  from HD
diesel vehicles. J Air Waste Manage Assoc 48:276-278.

Ohio, AJ; Richards, JH; Dailey, LA; et al. (2000) The effects of oxidants on modifying the toxicity of ambient PM.
Society of Toxicology Annual Meeting.  San Francisco, CA.

Gorse, RA, Jr.; Riley, TL; Ferris, FC; et al. (1983) 1-Nitropyrene concentration and bacterial mutagenicity in on-
road vehicle paniculate emissions. Environ Sci Technol 17:198-202.

Graboski, MS; McCormick, RL. (1996) Effect of diesel fuel chemistry on emissions at high altitude.  SAE
Technical Paper Ser. No. 961947.

Graboski, MS; McCormick, RL; Alleman, T. (1998a) Testing of natural gas and diesel buses for comparison with
WVU mobile dynamometer. Colorado Institute for Fuels and High Altitude Engine Research, Colorado School of
Mines, Golden, CO.
                                                2-131

-------
Graboski, MS; McCormick, RL; Yanowitz, J; et al. (1998b) HD diesel testing for the Northern Front Range Air
Quality Study. Colorado Institute for Fuels and Engine Research, Colorado School of Mines, Golden, CO.
http://nfraqs.cira.colostate.edu/index2.html.

Gray, HA. (1986) Control of atmospheric fine primary carbon particle concentrations. Ph.D. thesis, California
Institute of Technology.

Gray, HA; Kuklin, A. (1996) Benefits of mobile source NOX related paticulate matter reductions. Final report.
Prepared by Systems Applications International for U.S. Environmental Protection Agency.  SYSAPP-96/61.

Gray, HA; Cass, GR. (1998) Source contributions to atmospheric fine carbon particle concentrations. Atmos
Environ 32(22):3805-3825.

Groves, J; Cain, JR. (2000) A survey of exposure to diesel engine exhaust emissions in the workplace. Ann Occup
Hyg44(6):435-447.

Gullett, BK; Ryan, JV. (1997) On-road sampling of diesel engine emissions of polychlorinated dibenzo-p-dioxin
and polychlorinated dibenzofuran.  Organo Comp 32:451-456.

Hagenmaier, H; Dawidowsky, V; Weber, UB; et al. (1990) Emission of polyhalogenated dibenzodioxins and
dibenzofurans from combustion-engines. Organo Comp 2:329-334.

Hames, RJ; Straub, RD; Amann, RW. (1985) DDEC Detroit Diesel electronic control.  SAE Technical Paper Ser.
No. 850542.

Hammerle, RH; Schuetzle, D; Adams, W. (1994) A perspective on the potential development of environmentally
acceptable light-duty diesel engines. Environ Health Perspect (Suppl.) 102:25-30.

Hammond, SK. (1998) Occupational exposure to diesel exhaust.  Presented at the Scientific Review Panel Meeting,
California Air Resources Board, March 11, 1998.

Hansen, ADA; Novakov, T. (1990) Real-time measurement of aerosol black carbon during the carbonaceous
species methods comparison study. Aerosol Sci Technol 12(1): 194-199.

Hansen, KF; Bak, F; Andersen, EM; et al.  (1994) The influence of an oxidation catalytic converter on the chemical
and biological characteristics of diesel exhaust emissions. SAE Technical Paper Ser. No. 940241.

Kara, K; Hanaoka, T; Yamano, Y; et al. (1997) Urinary 1-hydroxypyrene levels of garbage collectors with low-
level exposure to polycyclic aromatic hydrocarbons. Sci Total Environ 199:159-164.

Hardenberg, H; Fraenkle, G. (1978) The effect of charge air cooling on exhaust emissions and power output of
turbocharged engines.  Institute of Mechanical Engineers Paper No. C71/78.

Hare, CT.  (1977) Characterization of diesel gaseous and paniculate emissions. Final report. Prepared by Southwest
Research Institute under contract no. 68-02-1777.

Hare, CT;  Springer, KJ. (1972) Exhaust emissions from uncontrolled vehicles and related equipment using internal
combustion engines. Part I, locomotive diesel engines and marine counterparts. Southwest Research Institute under
contract no. EHS 70-108.

Harvey, GD; Baumgard, KJ; Johnson, JH; et al. (1994) Effects of a ceramic particle trap and copper fuel additive on
HD diesel emissions. SAE Technical Paper Ser. No. 942068.

Health Effects Institute. (1995) Diesel exhaust: a critical analysis of emissions, exposure, and health effects.
Cambridge, MA.
                                                2-132

-------
Henry, RC; Lewis, CW; Collins, JF. (1994) Vehicle-related hydrocarbon source compositions from ambient data:
the GRACE/SAFER method.  Environ Sci Technol 28:823-832.

Hey wood, JB. (1988) Internal combustion engine fundamentals. New York: McGraw-Hill, Inc.

Hildemann, LM; Markowski, GR; Cass, GR. (1991) Chemical composition of emissions from urban sources of fine
organic aerosol. Environ Sci Technol 25:744-759.

Hori, S; Narusawa, K. (1998) Fuel composition effects on SOF and PAH exhaust emissions from a DI diesel
engine. SAE Technical Paper Ser. No. 980507.

Horvath, H; Kreiner, I; Norek, C; et al. (1988) Diesel emissions in Vienna. Atmos Environ 22:1255-1269.

Hsiao-Hsuan, M; Wen-Jhy, L; Chung-Ban, C; et al. (2000) Effect of fuel aromatic content on PAH emission from a
heavy-duty diesel engine. Chemosphere 41:1783-1790.

Hutzinger, O; Essers, U.; Hagenmaier, H.  (1992) Untersuchungen zur emission halogenierter dibenzodioxine und
dibenzofurane aus verbrennungsmortoren beim betrieb mit handelsublichen betriebsstoffen. Universities of
Bayreuth, Stuttgart and Tubingen, Germany.  GSF-Forschungszentrum, Munich, Germany, ISSN 0937-9932.

Jang, M; McDow, SR.  (1995) Benz[a]anthracene photodegradation in the presence of known organic constituents
of atmospheric aerosols. Environ Sci Technol 29:2654.

Janssen, NAH; de Hartog, JJ; Hoek, G; et al., (2000) Peroanl exposure to fine paniculate matter in elderly subjects:
relation between personal, indoor, and outdoor concentrations. J Air Waste Manage Assoc 50:1133-1143.

Japar, SM; Szkarlat, AC; Gorse, RA; et al. (1984) Comparison of solvent extraction and thermal-optical carbon
analysis methods: application to diesel vehicle exhaust aerosol. Environ Sci Technol  18:231-234.

Jimenez, J; McRae, G; Nelson, D; et al. (1998) Remote sensing of HD diesel truck NOX emissions using TILDAS.
8th Coordinating Research Council On-Road Vehicle Emissions Workshop: San Diego, CA, pp. 7-45-7-61.

Johnson, JE. (1993) Hydrocarbon oxidation in a diesel oxidation  catalytic converter.  MS thesis, University of
Minnesota.

Johnson, JE; Kittelson, DB. (1996) Deposition, diffusion and adsorption in the diesel oxidation catalyst.  Appl Catal
B: Environ 10:117-137.

Johnson, JH; Bagley, ST; Gratz, LD; et al. (1994) A review of diesel paniculate control technology and emissions
effects~1992 Horning Memorial Award Lecture. SAE Technical Paper Ser. No. 940233.

Kado, NY; Okamoto, RA; Kuzmicky, PA. (1996) Chemical andbioassay analyses of diesel andbiodiesel
paniculate matter: pilot study. Final report to Montana State Department of Environmental Quality and the U.S.
Department of Energy by Department of Environmental Toxicology, University of California, Davis.

Kado, NY; Kuzmicky, PA; Kiefer, KL; et al. (1998) Emissions from biodiesel fuel combustion: bioassay and
chemical analyses of the particle and semi-volatile emissions from hydrogenated biodiesel fuels.  Final report to
U.S. Department of Energy by Department of Environmental Toxicology, University  of California, Davis.

Kamens, RM; Guo, Z; Fulcher, JN; et al. (1988) Influence of humidity, sunlight, and temperature on the daytime
decay of polyaromatic hydrocarbons on atmospheric soot particles. Environ Sci Technol 22:103-108.

Kantola, TC; Bagley, ST; Gratz, LD; et al. (1992) The influence of a low sulfur fuel and a ceramic particle trap on
the physical, chemical, and biological character of HD diesel emissions. SAE Technical Paper Ser. No. 920565.
                                                2-133

-------
Katragadda, S; Bata, R; Wang, WG; et al. (1993) A correlation study between two HD vehicle chassis dynamometer
emissions testing facilities. SAE Technical Paper Ser. No. 931788.

Katz, M; Chan, C; Tosine, H; et al. (1979) Relative rates of photochemical and biological oxidation (in vitro) of
poly cyclic aromatic hydrocarbons. In: Polynuclear aromatic hydrocarbons: Third International Symposium on
Chemistry and Biology-Carcinogenesis and Mutagenesis, October 1978, Columbus, OH. Jones, PW; Leber, P, eds.
Ann Arbor, MI: Ann Arbor Science Publishers, Inc., pp. 171-189.

Kawatani, T; Mori, K; Fukano, I; et al. (1993) Technology for meeting the 1994 USA exhaust emission regulations
on heavy-duty diesel engine.  SAE Technical Paper Ser. No. 932654.

Khatri, NJ; Johnson, JH; Leddy, DG. (1978) The characterization of the hydrocarbon and sulfate fractions of diesel
paniculate matter. SAE Technical Paper Ser. No. 780111.

Kinney, PL; Agarwal, M; Northridge, ME; et al. (2000) Airborne concentrations of PM2.5 and DE particles on
Harlem sidewalks: a community-based pilot study. Environ Health Perspect 108:213-218.

Kirchstetter, TW; Harley, RA; Kreisberg, NM; et al. (1999) On-road measurement of fine particle and nitrogen
oxide emissions from light- and HD motor vehicles. Atmos Environ 33:2955-2968.

Kirkland, JF. (1983) Dawn of the diesel age. Glendale, CA: Interurban Press.

Kittelson, DB. (1998) Engines and nanoparticles:  a review. J Aerosol Sci 29:575-588.

Kittelson, DB. (2000) What do we know about the nature of the nanoparticles emitted by diesel engines?
Coordinating Research Council, January 2000.

Kittelson, DB; Pipho, ML; Ambs, JL; et al.  (1986) Particle concentrations in a diesel cylinder: comparison of theory
and experiment. SAE Technical Paper Ser. No. 861569.

Kittelson, DB; Winthrop, W; Baltensperger, R; et al. (1999) Diesel aerosol sampling methodology. University of
Minnesota Center for Diesel Research. http://www.me.umn.edu/centers/cdr/Proj_EPA.html.

Kittelson, D; Watts,  W; Ramachandran, G. (2000) Exposures to diesel exhaust aerosol: relationship between 6-hour
time weighted average personal exposures to EC/OC and real-time variations in aerosol properties. Health Effects
Institute Annual Conference 2000: exploring exposure and risk issues. April 9-11, 2000, Atlanta, GA.

Kleeman, MJ; Cass,  GR. (1998) Source contributions to the size and composition distribution of urban paniculate
air pollution. Atmos  Environ 32(16):2803-2816.

Kleeman, MJ; Schauer, JJ; Cass, GR. (2000) Size and composition distribution of fine paniculate matter emitted
from mo tor vehicles. Environ Sci Technol 34:1132-1142.

Klein, M. (1991) The diesel revolution.  Am Heritage Invention Technol 6.

Kleindienst, TE;  Shepson, PB; Edney, EO; et al. (1985) Peroxyacetyl nitrate:  measurement of its mutagenic activity
using the Salmonella/mammalian microsome reversion assay. MutatRes 157:123-128.

Kreso,  AM; Johnson, JH; Gratz, LD; et al. (1998) A study of the effects of exhaust gas recirculation onHD diesel
engine  emissions. SAE Technical Paper Ser. No. 981422.

Last, RJ; Krueger, M; Duernholz, M. (1995) Emissions and performance characteristics of a 4-stroke, direct injected
diesel engine fueled  with blends of biodiesel and low sulfur diesel fuel. SAE Technical Paper Ser. No. 950054.

Law, PL; Lioy, PJ; Zelenka, MP; et al. (1997) Evaluation of a probabilistic exposure model applied to carbon
monoxide (pNEM/CO) using Denver personal exposure monitoring  data.  J Air Waste Manage Assoc 47:491-500.


                                                2-134

-------
Lee, R; Hobbs, CH; Pedley, JF. (1998) Fuel quality impact on heavy duty diesel emissions: literature review. SAE
Technical Paper Ser. No. 982649.

Levson, K. (1988) The analysis of diesel paniculate. Fresenius Z Anal Chem 331:467-478.

Lew, G.  (1996) Letter to G. Schweer (Versar, Inc.) dated January 11, 1996.  Sacramento, CA:  State of California
Air Resources Board, Engineering and Laboratory Branch. Available for inspection at: U.S. Environmental
Protection Agency, Office of Research and Development, National Center for Environmental Assessment,
Washington, DC.

Lies, KH; Hartung, A; Postulka, A; et al. (1986)  Composition of diesel exhaust with particular reference to particle
bound organics including formation of artifacts. In: Carcinogenic and mutagenic effects of diesel engine exhaust.
London: Elsevier Science Limited, pp. 65-82.

Ligocki, MP; Pankow, JF. (1989) Measurements of the gas/particle distributions of atmospheric organic
compounds. Environ Sci Technol 23:75-83.

Loft, S; Poulsen, HE; Vistisen, K; et al. (1999) Increased urinary excretion of 8-oxo-2'-deoxyguanosine, a
biomarker of oxidative DNA damage, in urban bus drivers. MutatRes 441:11-19.

Lowenthal, DH; Chow, JC; Watson, JG; et al. (1992) The effects of collinearity on the ability to determine aerosol
contributions from diesel- and gasoline-powered vehicles using the chemical mass balance model. Atmos Environ
26A(13):2341-2351.

Lowenthal, DH, Zielinska, B., Chow, JC, et al. (1994) Characterization of heavy-duty diesel vehicle emissions.
Atmos. Environ. 28:731-743.

Lucas, A; Duran, A; Carmona, M; et al. (1999) SAE Technical Paper Ser. No. 1999-01-3532.

Ltiders, H; Kruger, M; Strommel, P; et al.  (1998) The role of sampling conditions in particle size distribution
measurements. SAE Technical Paper Ser. No. 981374.

Maricopa Association of Governments. (1999) The 1999 Brown Cloud Project for the Maricopa Association of
Governments Area, Final Report, December 1999.

Marklund, S.; Andersson, R.; Tysklind, M.; et al. (1990)  Emissions of PCDDs and PCDFs in gasoline and diesel
fueled cars. Chemosphere 20(5):553-561.

Martin, S. (1981a) Emissions from heavy-duty engines using the  1984 transient test procedure. Volume II-diesel.
EPA/460/3-81/031.

Martin, S. (1981b) Diesel paniculate by 1986 HD transient Federal test procedure.  Final report. Prepared by
Southwest Research Institute under contract no. 68-03-2603.

Mauderly, J. (1992) Diesel exhaust. In: Environmental toxicants: human exposures and their health effects.
Lippmann, M, ed. New York: Van Nostrand Reinhold, pp. 119-162.

McCarthy, CI; Slodowske, WJ; Sienicki, EJ; et al. (1992) Diesel fuel property effects on exhaust emissions from a
heavy duty diesel engine that meets 1994 emissions requirements. SAE Technical Paper Ser. No. 922261.

McClure, BT; Bagley ST; Gratz, LD. (1992) The influence of an oxidation catalytic converter and the fuel
composition on the chemical and biological characteristics of diesel exhaust emissions. SAE Technical Paper Ser.
No. 920854.

McCormick, RL; Graboski, MS; Alleman, TL; et al. (1999) In-use emissions from natural gas fueled HD vehicles.
SAE Technical Paper Ser. No.  1999-01-1507.


                                                2-135

-------
McCormick, RL; Alleman, TL; Graboski, MS; et al. (2000) Quantifying the emissions benefit of opacity testing and
repair of heavy-duty diesel vehicles. Proceedings of the 10th Coordinating Research Council On-Road Vehicle
Emissions Workshop, March 27-29, 2000, San Diego, CA.

McDonald, JF. (1997) The emissions and combustion characteristics of a soy-methyl-ester biodiesel fuel in a
naturally aspirated indirect injection diesel engine. MS thesis, University  of Minnesota.

McDonald, J; Purcell, DL; McClure, BT. (1995) Methyl ester oxygenated fuels for diesel mining applications.  In:
Proceedings of the 7th U.S. Mine Ventilation Symposium, Society for Mining, Metallurgy, and Exploration, Inc.

McDow, SR; Sun, Q; Vartiainen, M; et al. (1994) Effect of composition and state of organic components on
polycyclic aromatic hydrocarbon decay in atmospheric aerosols. Environ Sci Technol 28:2147.

Miguel, AH; Kirchstetter, TW; Harley, RA; et al. (1998) On-road emissions of paniculate polycyclic aromatic
hydrocarbons and black carbon from gasoline and diesel vehicles. Environ Sci Technol 32:450-455.

Mitchell, K; Steere, DE; Taylor, JA; et al. (1994) Impact of diesel fuel aromatics on paniculate, PAH, and nitro-
PAH emissions. SAE Technical Paper Ser. No. 942053.

Mori, K. (1997) Worldwide trends in HD diesel engine exhaust emission legislation and  compliance technologies.
SAE Technical Paper Ser. No. 970753.

National Air Toxics Assessment (NATA). (2001) DPM and additional urban air toxics concentration and exposure
data are available at http://www.epa.gov/ttn/uatw/nata.

National Institute for Occupational Safety and Health (NIOSH). (1988) Carcinogenic effects of exposure to diesel
exhaust. Current Intelligence Bulletin 50. No. 88-116. U.S. Department of Health and Human Services, Public
Health Service, Centers for Disease Control, National Institute for Occupational Safety and Health, Cincinnati, OH.

NIOSH. (1990) Health hazard evaluation report. Prepared by Yellow Freight System, Inc. HETA 90-088-2110.

NIOSH. (1992) Health hazard evaluation report: City of Lancaster, Division of Fire, Lancaster, OH. NIOSH Report
No. HHE HETA 92-0160-2360, NTIA No. PB-94140142. U.S. Department of Health and Human Services, Public
Health Service, Centers for Disease Control, National Institute for Occupational Safety and Health, Cincinnati, OH.

National Institute for Petroleum and Energy Research (NIPER). (1986 and earlier years)  Diesel fuel oils, annual
survey.

National Research Council. (1982) Diesel cars: benefits, risks and public policy. Final report of the Diesel Impacts
Study Committee. Washington, DC: National Academy Press.

Needham, JR; Doyle, DM; Faulkner, SA; et al. (1989) Technology for 1994. SAE Technical Paper Ser. No.
891949.

Nielsen, T; Ramdahl, T. (1986) Discussion on determination of 2-nitrofluoranthene and 2-nitropyrene in ambient
paniculate matter: evidence for atmospheric reactions. Atmos Environ 20:1507.

Nielsen, PS; Andreassen, A; Farmer, PB; et al. (1996) Biomonitoring of diesel exhaust-exposed workers. DNA and
hemoglobin adducts and urinary 1-hydroxypyrene as markers of exposure. Toxicol Lett  86:27-37.

Norbeck, JM; Truex, TJ; Smith, MR; et al. (1998a) Evaluation of factors that affect diesel exhaust toxicity. Final
report, prepared under Contract No. 94-312, for California Air Resources Board, by the Center for Environmental
Research and Technology, College of Engineering, University of California, Riverside, CA.

Norbeck, JM; Durbin, TD; Truex, TJ; et al. (1998b) Characterizing paniculate emissions from medium- and light
heavy-duty diesel-fueled vehicles. Final report. Prepared for South Coast Air Quality Management District by for


                                                 2-136

-------
California Air Resources Board by the Center for Environmental Research and Technology, College of
Engineering, University of California, Riverside, CA, under contract no. 97031.

Norbeck, JM; Durbin, TD; Truex, TJ. (1998c) Measurement of primary paniculate matter emissions from light-duty
motor vehicles. Prepared for Coordinating Research Council, Inc., and South Coast Air Quality Management
District, by Center for Environmental Research and Technology, College of Engineering, University of California,
Riverside, CA, 1998.

Norris, GA; Birch, ME; Tolocka, MP; et al. (2000) Comparison of paniculate organic and elemental carbon
measurements made with the IMPROVE and NIOSH method 5040 protocols. Aerosol Sci Technol, in press.

Northridge, ME; Yankura, J; Kinney, PL; et al. (1999) Diesel exhaust exposure among adolescents in Harlem: a
community-driven study. Am J Public Health 89:998-1002.

Odum, JR; McDow, SR; Kamens, RM. (1994) Mechanistic and kinetic studies of the photodegradation of
benz[a]anthracene in the presence of methoxyphenols. Environ Sci Technol 28:1285.

Oehme, M; Larssen, S; Brevik, EM. (1991) Emission factors of PCDD/CDF for road vehicles obtained by a tunnel
experiment. Chemosphere 23:1699-1708.

Opris, CN; Gratz, LD; Bagley, ST; et al. (1993) The effects of fuel sulfur concentration on regulated and
unregulated HD diesel emissions. SAE Technical Paper Ser. No. 930730.

Pagan, J. (1999) Study of particle size distributions emitted by  a diesel engine. SAE Technical Paper Ser. No.
1999-01-1141.

Paputa-Peck, MC; Marano, RS; Schuetzle, D; et al. (1983) Determination of nitrated polynuclear aromatic
hydrocarbons in paniculate extracts by capillary column gas chromatography with nitrogen selective detection. Anal
Chem 55:1946-1954.

Pataky, GM; Baumgard, KJ; Gratz, LD; et al. (1994) Effects of an oxidation catalytic converter on regulated and
unregulated diesel emissions. SAE Technical Paper Ser. No. 940243.

Perez, JM. (1980) Measurement of unregulated emissions-some heavy-duty diesel engine results. Health Effects of
Diesel Engine Emissions, EPA/600/9-80/057a.

Perez, JM; Williams, RL. (1989) A study of paniculate extracts from 1980s heavy duty diesel engines run on
steady-state and transient cycles.  SAE Technical Paper Ser. No. 892491.

Pierson, WR; Brachaczek, WW. (1976) Paniculate matter associated with vehicles on the road. SAE Trans 85:209-
227.

Pierson, WR; Brachaczek, WW. (1983) Paniculate matter associated with vehicles on the road. Aerosol Sci Tech
2:1-40.

Pierson, WR; Gertler, AW; Robinson, NF; et al. (1996) Real-world automotive emissions-summary of studies in
the Fort McHenry and Tuscarora mountain tunnels. Atmos Environ 30:2233-2256.

Pischinger, R; Cartellieri, W. (1972) Combustion system parameters and their effect on diesel engine exhaust
emissions.  SAE Technical Paper Ser. No. 720756.

Pitts, JN, Jr; Van Cauwenberghe, KA; Grosjean, D; et al. (1978) Atmospheric reactions of polycyclic aromatic
hydrocarbons: facile formation of mutagenic nitro derivatives.  Science 202:515-519.

Pitts, JN, Jr; Lokensgard, DM; Ripley, PS; et al. (1980) Atmospheric epoxidation of benzo[a]pyrene by ozone:
formation of the metabolite benzo[a]pyrene-4,5-oxide. Science 210:1347-1349.


                                                2-137

-------
Pitts, JN, Jr; Sweetman, JA; Zielinska, B; et al. (1985a) Determination of 2-nitrofluoranthene and 2-nitropyrene in
ambient paniculate matter: evidence for atmospheric reaction. Atmos Environ 19:1601-1608.

Pitts, JN, Jr; Zielinska, B; Sweetman, JA; et al. (1985b) Reaction of adsorbed pyrene and perylene with gaseous
N2O5 under simulated atmospheric conditions. Atmos Environ 19:911-915.

Pitts, JN, Jr; Sweetman, JA; Zielinska, B; et al. (1985c) Formation of nitroarenes from the reaction of poly cyclic
aromatic hydrocarbons with dinitrogen pentoxide. Environ Sci Technol 19:1115-1121.

Pitts, JN, Jr; Paur, HR; Zielinska, B; et al. (1986) Factors influencing the reactivity of poly cyclic aromatic
hydrocarbons adsorbed on filters and ambient POM with ozone. Chemosphere 15:675-685.

Prinn, R; Cunnold, D; Simmonds, P; et al. (1992) Global average concentration and trend for hydroxyl radicals
deduced from ALE/GAGE trichloroethane (methyl chloroform data for 1978-1990). J Geophys Res 97:2445-2461.

Ramadan, Z; Xin-Hua, S; Hopke, PK. (2000) Identification of sources of Phonix aerosol by positive matrix
factorization. J. Air & Waste Manage. Assoc. 50:1308-1320.

Rantanen, L; Mikkonen, S; Nylund, L; et al. (1993) Effect of fuel on the regulated, unregulated and mutagenic
emissions of DI diesel engines. SAE Technical Paper Ser. No. 932686.

Rideout, G; Kirshenblatt, M; Prakash, C. (1994) Emissions from methanol, ethanol, and diesel powered urban
transit buses. SAE Technical Paper Ser. No. 942261.

Rogak, SN; Pott, U; Darin, T; et al. (1998) Gaseous emissions from vehicles in a traffic tunnel in Vancouver, British
Columbia. J Air Waste Manage Assoc 48:604-615.

Rogers, A; Whelan, B. (1999) Exposures in Australian mines. Health Effects Institute Number 7. March 7-9, 1999
Stone Mountain,  GA.

Rogge, WF; Hildemann, LM; Mazurek, MA; et al.  (1993) Sources of fine organic aerosol. 2. Noncatalyst and
catalyst-equipped automobiles and HD diesel trucks. Environ Sci Technol  27:636-651.

Salmon, LG; Cass, GR; Pedersen, DU; et al. (1997) Determination of fine particle concentration and chemical
composition in the northeastern United States, 1995.  Progress Report to Northeast States for Coordinated Air Use
Management (NESCAUM), October 1997.

Sampson, PJ.  (1988) Atmospheric transport and dispersion of air pollutants associated with vehicular emissions. In:
Air pollution, the automobile  and public health. Watson, AY; Bates, RR; Kennedy, D, eds. Washington, DC:
National Academy Press, pp.  77-97.

Saverin, R.  (1999) German potash miners: cancer mortality. Health Effects Institute Number 7. March 7-9, 1999
Stone Mountain,  GA. pp. 220-229.

Sawyer, RF; Johnson, JH. (1995) Diesel emissions and control technology. In: Diesel exhaust: a critical analysis of
emissions, exposure, and health effects. Cambridge, MA: Health Effects Institute, pp. 65-81.

Schauer, JJ; Rogge, WF; Hildemann, LM; et al. (1996) Source apportionment of airborne paniculate matter using
organic compounds as tracers. Atmos Environ 30(22):3837-3855.

Schauer, JJ; Kleeman, MJ; Cass, GR; et al.  (1999) Measurement of emissions from air pollution sources. 2. Q
through C30 organic compounds from medium duty diesel trucks. Environ Sci Technol 33:1578-1587.

Scholl, JP; Bagley, ST; Leddy, DG; et al. (1982) Study of aftertreatment and fuel injection variables for paniculate
control in heavy-duty diesel engines.  EPA-460/3-83-002.
                                                2-138

-------
Schuetzle, D. (1983) Sampling of vehicle emissions of chemical analysis and biological testing. Environ Health
Perspect 47:65-80.

Schuetzle, D; Frazier, JA. (1986) Factors influencing the emission of vapor and paniculate phase components from
diesel engines. In: Carcinogenic and mutagenic effects of diesel engine exhaust. London: Elsevier Science Limited,
pp. 41-63.

Schuetzle, D; Perez, JM. (1983) Factors influencing the emissions of nitrated-polynuclear aromatic hydrocarbons
(nitro-PAH) from diesel engines. J Air Pollut Control Assoc 33:751-755.

Schuetzle, D; Jensen, TE; Ball, JC. (1985) Polar polynuclear aromatic hydrocarbon derivatives in extracts of
particulates: biological characterization and techniques for chemical analysis. Environ Int 11:169-181.

Schwind, K-H; Thoma, H; Hutzinger, O; et al.  (1991) Emission halogenierter dibenzodioxine (PXDD) und
dibenzofurane (PXDF) aus verbrennungsmotoren. UWSF-Z Umweltchem Oekotox 3: 291-298. [English
translation]

Seinfeld,  JH; Pandis, SN. (1998) Atmospheric  chemistry and physics: from air pollution to climate change. New
York: John Wiley & Sons, Inc.

Shelton, EM. (1977) Diesel fuel oils. BERC/PPS-77/5.

Shelton, EM. (1979) Diesel fuel oils. BETC/PPS-79/5.

Shi, JP; Harrison, RM. (1999) Investigation of ultrafine particle formation during diesel exhaust dilution. Environ
Sci Technol 33:3730-3736.

Shi, JP; Mark, D; Harrison, RM.  (2000) Characterization of particles from a current technology heavy-duty diesel
engine. Environ Sci Technol 34:748-755.

Sienicki,  EJ; Jass, RE;  Slodowske, WJ; et al. (1990) Diesel fuel aromatic and cetane number effects on combustion
and emissions from a prototype 1991 diesel engine. SAE Technical Paper Ser. No. 902172.

Sisler, JF. (1996) Spatial and seasonal patterns and long term variability of the composition of the haze in the
United States: an analysis of data from the IMPROVE network. Cooperative Institute for Research in the
Atmosphere. Colorado State University. ISSN: 0737-5352-32.

South Coast Air Quality Management District (SCAQMD). (2000) Multiple air toxics exposure study (MATES-II).
Final Report and Appendices, March 2000.

Spreen, KB; Ullman, TL; Mason, RL. (1995) Effects of fuel oxygenates, cetane number, and aromatic content on
emissions from 1994 and 1998 prototype heavy-duty diesel engines. Final report. Prepared by Southwest Research
Institute under Contract No. VE-10.

Springer, KJ. (1979) Characterization of sulfates, odor, smoke, POM and particulates from light and heavy duty
engines-part IX. Prepared by Southwest Research Institute. EPA/460/3-79/007.

Stone, R. (1995) Introduction to internal combustion engines. Warrendale, PA: Society of Automotive Engineers.

Sweetman, JA; Zielinska, B; Atkinson, R; et al. (1986) A possible formation pathway for the 2-nitrofluoranthene
observed in ambient paniculate organic matter. Atmos Environ 20:235-238.

Tanaka, S; Takizawa, H; Shimizu, T; et al. (1998) Effect of fuel compositions on PAH in paniculate matter from DI
diesel engines. SAE Technical Paper Ser. No. 982648.
                                                2-139

-------
Tancell, PJ; Rhead, MM; Trier, CJ; et al. (1995) The sources of benzo[a]pyrene in diesel exhaust emissions. Sci
Total Environ 162:179-186.

Taylor, CF. (1990) The internal combustion engine in theory and practice - volume 1.  Cambridge, MA: MIT Press.

Thrane, KE; Mikalsen, A. (1981) High-volume sampling of airborne polycyclic aromatic hydrocarbons using glass
fibre filters and polyurethane foam. Atmos Environ 15:909-918.

Tokiwa, H; Sera, N; Nakanishi, Y; et al. (1999) 8-Hydroyguanosine formed in human lung tissues and the
association with diesel exhaust particles. Free Rad Biol Med 27:1251-1258.

Tong, HY; Sweetman, JA; Karasek, FW; et al. (1984) Quantitative analysis of polycyclic aromatic compounds in
Diesel exhaust paniculate extracts by combined chromatographic techniques. J Chromatogr 312:183-202.

Ullman, TL; Mason, RL; Montalvo, DA. (1990) Effects of fuel aromatics, cetane number, and cetane improver on
emissions from a 1991 prototype heavy-duty diesel engine.  SAE Technical Paper Ser. No. 902171.

Ullman, TL; Spreen, KB; Mason, RL. (1995) Effects of cetane number on emissions from a prototype 1998 HD
diesel engine.  SAE Technical Paper Ser. No. 950251.

U.S. Bureau of the Census. (1995) Truck inventory and use survey. 1992 Census of Transportation, TC92-T-52.

U.S. Bureau of the Census. (1999a) 1997 Economic census, vehicle inventory and use survey. EC97TV-US.

U.S. Bureau of the Census. (1999b) Truck inventory and use survey. 1997 Census of Transportation, EC97-TV-
US.

U.S. Environmental Protection Agency (EPA). (1992) Highway vehicle emission estimates. Office of Mobile
Sources.

U.S. EPA. (1995) Highway vehicle emissions estimates.  Office of Mobile Sources.

U.S. EPA. (1998a) National air pollutant emission trends update, 1970-1997. EPA/454/E-98-007.

U.S. EPA. (1998b) Locomotive emission standards. Office of Mobile Sources Regulatory Support Document.

U.S. EPA. (1999a) Air quality criteria for paniculate matter. External review draft, October 1999.

U.S. EPA. (1999b) Analysis of the Impacts of control programs on motor vehicle toxic emissions and exposure in
urban areas and nationwide: volume I and II.  Prepared for EPA by Sierra Research, Inc. and Radian International
Corporation/Eastern Research Group, November 30, 1999. Report Nos. EPA/420/R-99/029 and
EPA/420/R-99/030.

U.S. EPA. (2000a) National air pollutant emission trends, 1900-1998. EPA-454/R-00-002, March 2000.

U.S. EPA. (2000b) Control of Air Pollution From New Motor Vehicles: Heavy-Duty Engine and Vehicle Standards
and Highway Diesel Fuel  Sulfur Control Requirements. Final Rule. Regulatory impact analysis. Office of
Transportation and Air Quality. EPA420-R-00-026. This document can also be found in the Code of Federal
Regulations 40 CFR Parts 69,  80, and 86.

Unnasch, S; Lowell, D; Lonyai, F; et al. (1993) Performance and emissions of clean fuels in transit buses with
Cummins L10 engines. SAE Technical Paper Ser. No. 931782.

Van Vaeck, L; Van Cauwenberghe, K. (1984) Conversion of polycyclic aromatic hydrocarbons on diesel paniculate
matter upon exposure to ppm levels of ozone. Atmos Environ 18:323-328.
                                               2-140

-------
Van Vaeck, L; Van Cauwenberghe, K; Janssens, J. (1984) The gas-particle distribution of organic aerosol
constituents: measurements of the volatilisation artifact in hi-vol cascade impactor sampling. Atmos Environ
18:417-430.

Voss, JR; Vanderpoel, RE. (1977) The shuttle distributor for a diesel fuel injection pump. SAE Technical Paper Ser.
No. 770083.

Wachter, WF. (1990) Analysis of transient emission data of a model year 1991 heavy duty diesel engine. SAE
Technical Paper Ser. No. 900443.

Wall, JC. (1998) Diesel fuel composition for future emissions regulations. Cummins Engine Co. panel discussion,
SAE International Fall Fuels and Lubricants Meeting and Exposition, October 21,  1998.

Wall, JC; Shimpi,  SA; Yu, ML. (1987) Fuel sulfur reduction for control of diesel paniculate emissions. SAE
Technical Paper Ser. No. 872139.

Wang, W; Gautam, M; Sun, X; et al. (1993) Emissions comparisons of twenty-six  heavy duty vehicles operated on
conventional and alternative fuels. SAE Technical Paper Ser. No. 932952.

Wang, WG; Lyons, D; Bata, R; et al. (1994) In-use emissions tests of alternatively fueled heavy duty vehicles by a
chassis dynamometer testing facility.  SAE Technical Paper Ser. No.   945124.

Warner-Selph, MA; Dietzmann, HE. (1984) Characterization of heavy duty motor vehicle emissions under transient
driving conditions. Prepared by Southwest Research Institute.  EPA/600/3-84/104.

Watson, N; Janota, MS. (1982)  Turbocharging the internal combustion engine.  New York: John Wiley and Sons.

Watson, JG; Fujita, EM; Chow, JC; et al. (1998) Northern Front Range Air Quality Study final report. Prepared by
Desert Research Institute for Colorado State University, Cooperative Institute for Research in the Atmosphere,
1998.

Watts, FW, Jr. (1995) Assessment of occupational exposure to diesel emissions. In: Diesel exhaust: a critical
analysis of emissions, exposure, and health effects. Cambridge, MA:  Health Effects Institute, pp. 107-123.

Weaver, CS; Klausmeier, RF. (1988) Heavy-duty diesel vehicle inspection and maintenance study.  Final Report
Vol. 1: Summary Report to California Air Resources Board, Contract No. A4-151-32.

Weingartner, E; Burtscher, H; Baltensperger, U. (1997a) Hygroscopic properties of carbon and diesel soot particles.
Atmos Environ 31:2311-2327.

Weingartner, E; Keller,  C; Stahel, WA; et al. (1997b) Aerosol emissions in a road tunnel. Atmos Environ
31(3):451-462.

Wevers, M; De Fre, R; Rymen, T. (1992) Dioxins and dibenzofurans in tunnel air. In: Sources of exposure.
Extended Abstracts, vol. 9. Organo Comp 9:321-324.

Whitby, KT; Sverdrup, GM. (1980) California aerosols: their physical and chemical characteristics. In: Hidy, GM;
Mueleeer, PD; Grosjean, D; et al., eds. The character and origins of smog aerosols: a digest of results from the
California Aerosol Characterization Experiment (ACHEX).  New York: John Wiley & Sons, Inc., pp. 477-517.
(Advances in environmental science and technology: v.9).

Whitfield, JK; Harris, DB. (1998) Comparison of heavy duty diesel emissions from engine and chassis
dynamometers and on-road testing. 8th Coordinating Research Council On-Road Vehicle Emissions Workshop: San
Diego, CA.
                                                 2-141

-------
Williams, PT; Andrews, GE; Battle, KD. (1987) The role of lubricating oil in diesel paniculate and paniculate PAH
emissions.  SAE Technical Paper Ser. No. 872084.

Williams, DJ; Milne, JW; Quigley, SM; et al. (1989) Paniculate emissions from in-use motor vehicles. II. Diesel
vehicles. Atmos Environ 23(12):2647.

Winer, AM; Busby, WF, Jr. (1995) Atmospheric transport and transformation of diesel emissions.  In: Diesel
exhaust: a critical analysis of emissions, exposure, and health effects. Cambridge, MA: Health Effects Institute, pp.
84-105.

Wittorff, DN; Gertler, AW; Chow, JC; et al. (1994) The impact of diesel paniculate emissions on ambient
paniculate loadings.  Air & Waste Management Association 87th Annual Meeting, Cincinnati, OH, June 19-24,
1994.

World Health Organization (WHO). (1996) Diesel fuel and emissions, environmental health criteria #171, WHO,
Geneva, Switzerland.

Woskie, SR; Smith, TJ; Hammond, SK; et al. (1988) Estimation of the diesel exhaust exposures of railroad workers.
I.  Current exposures. Am JIndMed 13:381-394.

Yamasaki, H; Kuwata, K; Miyamoto, H. (1982) Effects of ambient temperature on aspects of airborne polycyclic
aromatic hydrocarbons. Environ Sci Technol 16:189-194.

Yanowitz, J; Graboski, MS; Ryan, LBA; et al. (1999) Chassis dynamometer study of emissions from 21 in-use
heavy duty diesel vehicles. Environ Sci Technol 33:209-216.

Yanowitz, J; McCormick, RL; Graboski, MS. (2000) Critical review: in-use emissions from heavy duty diesel
vehicles. Environ Sci Technol 34:729-740.

Yokely, RA; Garrison, AA; Wehry, EL; et al. (1986) Photochemical transformation of pyrene andbenzo[a]pyrene
vapor-deposited on eight coal stack ashes. Environ Sci Technol 20:86-90.

Zaebst, DD; Blad, LM; Morris, JA; et al. (1988) Elemental carbon as a surrogate index of diesel exhaust exposure.
In: Proceedings of the American Industrial Hygiene  Conference, May 15-20, 1988, San Francisco, CA.

Zaebst, DD; Clapp, DE; Blake, LM; et al.  (1991) Quantitative determination of trucking industry workers'
exposures to diesel exhaust particles. Am Ind Hyg Assoc J 52:529-541.

Zelenka, P; Kriegler, W; Herzog, PL; et al. (1990) Ways toward the clean heavy duty diesel. SAE Technical Paper
Ser. No. 900602.

Zielinska, B. (1999)  Changes in diesel engine emissions over the last two decades. Diesel workshop: building a
research strategy to improve risk assessment. Health Effects Institute Number 7. March 7-9, 1999, Stone Mountain,
GA.

Zielinska, B; Arey, J; Atkinson, R; et al. (1986) Reaction of dinitrogen pentoxide with fluoranthene. J Am Chem
Soc 108:4126-4132.

Zielinska, B; Arey, J; Atkinson, R; et al. (1988) Nitration of acephenanthrylene under simulated atmospheric
conditions and in solution and the presence of nitroacephenanthrylene(s) in ambient air. Environ Sci Technol
22:1044-1048.

Zielinska, B; Arey, J; Atkinson, R; et al. (1989a) Formation of methylnitronaphthalenes from the gas-phase
reactions of 1- and 2-methylnaphthalene with OH radicals and N2O5 and their occurrence in ambient air. Environ
Sci Technol 23:723-729.
                                                2-142

-------
Zielinska, B; Arey, J; Atkinson, R; et al. (1989b) The nitroarenes of molecular weight 247 in ambient paniculate
samples collected in southern California. Atmos Environ 23:223-229.

Zielinska, B; Arey, J; Atkinson, R. (1990) The atmospheric formation of nitroarenes and their occurrence in ambient
air. In: Proceedings of the Fourth International Conference on N-Substituted Aryl Compounds: Occurrence,
Metabolism and Biological Impact of Nitroarenes. Cleveland, OH, July 1989.

Zielinska, B; McDonald, J; Hayes, T; et al. (1998) Northern Front Range Air Quality Study, volume B: source
measurements. Desert Research Institute.

Zwirner-Baier, I; Neumann, H-G.  (1999) Poly cyclic nitroaarenes (nitro-PAHs) as biomarkers of exposure to diesel
exhaust. MutatRes 441:135-144.
                                                 2-143

-------
                3.  DOSIMETRY OF DIESEL PARTICIPATE MATTER

3.1.  INTRODUCTION
       Animals and humans receive different internal doses when breathing the same external
concentrations of airborne materials such as diesel paniculate matter (DPM) (Brain and Mensah,
1983; Schlesinger, 1985). The dose received in different species differs from the aspects of the
total amount deposited within the respiratory tract, the relative distribution of the dose to specific
regions in the respiratory tract, and the residence time of these materials within the respiratory
tract, i.e., clearance. Using an external concentration breathed by laboratory animals as a basis
for any guidance for human exposure to DPM would then be an inadequate approximation of the
total and regional dose that humans may receive.
       The reason for the existence of this chapter and for consideration about interspecies
dosimetry is the lack of human health effect data on DPM and the concomitant need to be able to
evaluate existing animal data from the aspect of an equivalent human dose. The objective of this
chapter is to evaluate and address this issue of interspecies dosimetric differences through:

       •     A general overview of what is known about how particles like DPM are deposited,
            transported to,  and  cleared from the respiratory tract. Information on both
            laboratory animals  (mainly rodents) and humans will be considered  and interspecies
            similarities and differences highlighted.
       •     An overview of what is known about the bioavailability of the organic compounds
            adsorbed onto DPM from information in humans, animals, and in vitro studies, and
            from model predictions.
       •     An evaluation of the suitability of available dosimetric models and procedures for
            DPM to estimate interspecies extrapolations whereby an exposure scenario,
            conditions, and outcome in laboratory animals are adjusted to an equivalent
            outcome in humans via calculation of an internal dose.

       The focus in this chapter will be on the particulate fraction of diesel emissions, i.e, DPM.
Although diesel engine exhaust consists of a complex mixture of typical combustion gases,
vapors, low-molecular-weight hydrocarbons, and particles, it is the particle phase that is
considered to be of major health concern. The major constituents of diesel engine exhaust (DE)
and their atmospheric reaction products are described in Chapter 2.
       As will be deduced in Chapter 5, pulmonary toxicity and carcinogenicity are the major
focal points of diesel toxicity and of DPM deposition. Therefore, dosimetric considerations are
limited to the lung although DPM deposition would occur throughout the respiratory tract, from

                                          3-1

-------
the nares to the alveoli. Aspects of respiratory tract dosimetry to be considered in this chapter
include the characteristics of DPM, deposition of DPM throughout the respiratory tract, the
conducting airways and alveolar regions, normal DPM clearance mechanisms and rates of
clearance in both these regions, clearance rates during lung overload (in rats), elution of organics
from DPM, transport of DPM to extra-alveolar sites, and the interrelationships of these factors.
       The overall goal in this chapter follows from the objective—to judge the feasibility and
suitability of procedures allowing for derivation of an internal dose estimate of DPM for humans,
i.e., of a human equivalent concentration to exposure concentrations and conditions used in
animal studies. This goal is of significance especially in the quantitative dose-response analysis
of DPM effects in laboratory animals proposed in Chapter 6.

3.2. CHARACTERISTICS OF INHALED DIESEL PARTICIPATE MATTER
       The formation, transport, and characteristics of DPM are among the subjects considered
in detail in Chapter 2.  DPM consists of aggregates of spherical carbonaceous particles (typically
about 0.2 |_im mass median aerodynamic diameter [MMAD] or, more appropriately, mass median
thermodynamic diameter [MMTD]) to which significant amounts of higher-molecular-weight
organic compounds are adsorbed.  DPM has an extremely large surface area that allows for the
adsorption of organic compounds (see Chapter 2, Section 2.2.2). The organic carbon portion of
DPM can range from at least 19% to 43% from highway diesel engines; no data are available to
characterize the organic content of DPM from nonroad engines.  The lexicologically relevant
organic chemicals include high-molecular-weight hydrocarbons  such as the polycyclic aromatic
hydrocarbons (PAHs) and their derivatives (Chapter 2, Section 2.2.8).

3.3. REGIONAL DEPOSITION OF INHALED DIESEL PARTICIPATE MATTER
       This section discusses the major factors controlling the disposition of inhaled particles.
Note that disposition is defined as encompassing the processes of deposition, absorption,
distribution, metabolism, and elimination. The regional deposition of particulate matter in the
respiratory tract is dependent on the interaction of a number of factors, including respiratory tract
anatomy (airway dimensions and branching configurations), ventilatory characteristics (breathing
mode and rate, ventilatory volumes and capacities), physical processes (diffusion, sedimentation,
impaction, and interception), and the physicochemical characteristics (particle size, shape,
density, and electrostatic attraction) of the inhaled particles. Regional deposition of parti culate
material is usually expressed as deposition fraction of the total particles or mass inhaled and may
be represented by the ratio of the particles or mass deposited in a specific region to the number or
mass of particles inspired.  The factors affecting deposition in these various regions and their
importance in understanding the fate of inhaled DPM are discussed in the following sections.
                                           O  O
                                           3-2

-------
       It is beyond the scope of this document to present a comprehensive account of the
complexities of respiratory mechanics, physiology, and toxicology, and only a brief review will
be presented here.  The reader is referred to publications that provide a more in-depth treatment
of these topics (Weibel, 1963; Brain and Mensah, 1983; Raabe et al., 1988; Stober et al., 1993;
U.S. EPA, 1996).
       The respiratory tract in both humans and experimental mammals can be divided into three
general regions on the basis of structure, size, and function: the extrathoracic (ET), the
tracheobronchial (TB), and the alveolar (A). In humans, inhalation can occur through the nose or
mouth or both (oronasal breathing).  Animal models used in respiratory toxicology studies,
particularly the rat, however, are obligate nose breathers.

3.3.1.  Deposition Mechanisms
       This section provides an overview of the basic mechanisms by which inhaled particles
deposit within the respiratory tract. Details concerning the aerosol physics that explain both how
and why particle deposition occurs as well as data on total human respiratory tract deposition are
presented in detail in the earlier PM Criteria Document (U.S. EPA, 1996) and will only be briefly
summarized here. For more extensive discussions of deposition processes, refer to reviews by
Morrow (1966), Raabe (1982), U.S. EPA (1982), Phalen and Oldham (1983), Lippmann and
Schlesinger (1984), Raabe et al. (1988), and Stober et al. (1993).
       As pictorially represented in Figure 3-1, particles may deposit by five major mechanisms
(inertial impaction, gravitational settling, Brownian diffusion, electrostatic attraction, and
interception).  The relative contribution of each deposition mechanism to the fraction of inhaled
particles deposited varies for each region of the respiratory tract.
       It is important to appreciate that these processes are not necessarily independent but may,
in some instances, interact with one another such that total deposition in the respiratory tract may
be less than the calculated probabilities for deposition by the individual processes (Raabe, 1982).
Depending on the particle size and mass, varying degrees of deposition may occur in the ET (or
nasopharyngeal), TB, and A regions of the respiratory tract.
       Upon inhalation of particulate matter such as that found in DE, particle deposition will
occur throughout the respiratory tract.  Because of high airflow velocities and abrupt directional
changes in the ET and TB regions, inertial impaction is a primary deposition mechanism,
especially for particles >2.5 pm dae (aerodynamic equivalent diameter). Although inertial
impaction is a prominent process for deposition of larger particles in the tracheobronchial region,
it is of considerably less significance as a determinant of regional deposition patterns for
                                            5-3

-------
         Directional
           Change
             Very
           Abrupt
   Air
Velocity
             Less
           Abrupt
             Mild
              Electrostatic
              Precipitation
Figure 3-1. Schematic representation of major mechanisms, including diffusion, involved
in particle deposition. Airflow is signified by the arrows and particle trajectories by the
dashed line.
                                      3-4

-------
DPM, which have a dae <0.2 |_im and may be considered a rather polydisperse distribution with
sigma g values of 2.4 and greater.
       All aerosol particles are continuously influenced by gravity, but particles with a
dae > 0.5 |_im are affected to the greatest extent. A spherical compact particle will acquire a
terminal settling velocity when a balance is achieved between the acceleration of gravity acting
on the particle and the viscous resistance of the air; it is this velocity that brings the particle into
contact with airway surfaces. Both sedimentation and inertial impaction cause the deposition of
many particles within the same size range. These deposition processes act together in the ET and
TB regions, with inertial impaction dominating in the upper airways and sedimentation becoming
increasingly dominant in the lower conducting airways, especially for the largest particles that
can penetrate into the smaller bronchial airways.
       As particle diameters become <1 [am, the particles are increasingly subjected to diffusive
deposition because of random bombardment by air molecules, which results in contact with
airway surfaces. A dae of 0.5 \im is often considered a boundary between diffusion and
aerodynamic (sedimentation and impaction) mechanisms of deposition.  Thus, instead of having
a dae, diffusive particles  of different shapes can be related to the diffusivity of a thermodynamic
equivalent size based on spherical particles (Heyder et al., 1986).  Diffusive deposition of
particles is favored in the A region of the respiratory tract as particles of this size are likely to
penetrate past the ET and TB regions.
       Electrostatic precipitation is deposition related to particle charge. The electrical charge
on some particles may result in an enhanced deposition over what would be expected from size
alone.  This is due to image charges induced on the surface of the airway by these particles,  or to
space-charge effects whereby repulsion of particles containing like charges results in increased
migration toward the airway wall. The effect of charge on deposition is inversely proportional to
particle size and airflow rate. A recent study employing hollow airway casts of the human
tracheobronchial tree that assessed deposition of ultrafme (0.02 |_im) and fine (0.125 |_im)
particles found that deposition of singly charged particles was 5-6 times that of particles  having
no charge, and 2-3 times that of particles at Boltzmann equilibrium (Cohen et al., 1998).  This
suggests that within the  TB region of humans, electrostatic precipitation may be a significant
deposition mechanism for ultrafme and some fine particles, the latter of which are inclusive of
DPM.  Thus, although electrostatic precipitation is generally a minor contributor to overall
particle deposition, it may be important for DPM.
       Interception is deposition by physical contact with airway surfaces and is most important
for fiber deposition (U.S. EPA,  1996).
       Figure 3-2 shows the regional (ET, TB, A) deposition in the human respiratory tract as
influenced by particle size.  Keeping in mind that DPM is a polydisperse distribution with 0.2 |_im
                                           3-5

-------
         0.001
0.01
0.1
1
                                Particle Diameter (|jm)
  Figure 3-2. Generalized regional deposition fractions of various sized particles in the
  human respiratory tract.  (Adapted from the International Commission on
  Radiological Protection (ICRP) Publication 66 (1994) model. For unit density,
  spherical particles inhaled through the nose by an adult male with a tidal volume of
  1250 mL, respiratory frequency of 20 min"1, and functional residual capacity (FRC) of
  3300 mL.) ET, extrathoracic; TB, tracheobronchial; A, alveolar.


being only the median diameter, it can be seen that principal fraction particles sized from < 0.2
down to around 0.002 |_im would, as predicted based on their size and the expected mechanism of
diffusion, deposit in the alveolar region. Particles below this size range (and above around 4 |_im)
tend to deposit in the ET region.  Specific modeling results for deposition of DPM particles
inclusive of their distribution (i.e., og) are presented in Section 3.6.
3.3.1.1. Biological Factors Modifying Deposition
       The available experimental deposition data in humans are commonly derived using
healthy adult Caucasian males. Various factors can act to alter deposition patterns from those
obtained in this group. The effects of different biological factors, including gender, age, and
respiratory tract disease, on particle deposition have been reviewed previously (U.S. EPA, 1996,
Section 10.4.1.6).  In general, there appears to be an inverse relationship between airway
resistance and total deposition.

-------
       Differences in patterns of deposition between humans and animals have been summarized
(U.S. EPA, 1996; Schlesinger, 1985) and show clearly that when exposed to the same aerosol or
gas, humans and animals receive doses that may differ in both total and regional (i.e., ET, TB, or
A) deposition from a number of variables including particle size, especially for larger sized
particles, i.e. dae > 1 \im.  Such interspecies differences are important because the adverse toxic
effect is likely more related to the quantitative pattern of deposition within the respiratory tract
than to the exposure concentration; this pattern determines not only the initial respiratory tract
tissue dose but also the specific pathways by which the inhaled material is cleared and
redistributed (Schlesinger, 1985).  Such differences in initial deposition must be considered when
relating biological responses obtained in laboratory animal studies to effects in humans.
       The deposition patterns of inhaled diesel  particles in the respiratory tract of humans and
mammalian species has been reviewed (Health Effects Institute, 1995). Schlesinger (1985)
showed that physiological differences in the breathing mode for humans (nasal or oronasal
breathers) and laboratory rats (obligatory nose breathers), combined with different airway
geometries, resulted in  significant differences in  lower respiratory tract deposition patterns for
larger sized particles (>1 [am dae) in that a much lower fraction of inhaled larger particles is
deposited in the alveolar region of the rat compared with humans. However, alveolar deposition
of the much smaller DPM (around 0.2  |_im dae) was not affected as much by the differences
among species, as was demonstrated in model calculations by Xu and Yu (1987).  These
investigators modeled the deposition efficiency of inhaled DPM in rats, hamsters, and humans on
the basis of calculations of the models of Schum and Yeh (1980) and Weibel (1963). These
simulations (Figure 3-3) indicate relative deposition patterns in the lower respiratory tract
(trachea = generation 1; alveoli = generation 23) and are similar among hamsters,  rats, and
humans. Variations in  alveolar deposition of DPM over one breathing cycle in these different
species were predicted to be within 30% of one another (Xu and Yu, 1987). Xu and Yu (1987)
note that this similarity is concordant with the premise that deposition of the submicron diesel
particles is dominated by diffusion rather than sedimentation or impaction. Although these data
assumed nose-breathing by humans, the results would not be very different for mouth-breathing
because of the low filtering capacity of the nose for particles in the 0.1 to 0.5 [am range (see
Figure 3-2).
       The preceding discussion addresses deposition patterns and deposition efficiencies of
DPM in the respiratory tract of various species including humans. The alveolar region was
focused upon primarily because, as shown in Chapter 5, this region is where adverse effects from
long-term DPM exposure are typically observed.  For dosimetric calculations and modeling,
however, it would be of much greater importance to consider the actual deposited dose. Table
3-1 presents the analysis of Xu and Yu (1987) on prediction of the deposited doses of DPM
                                           3-7

-------
                o
                15
                   10
                      -2
                   Af\-
                o 10
                Q.
                (U
                Q
                   10-
                                                  Fischer rat
Human
                               4       8     12     16     20     24
                                  Generation Number
                Figure 3-3. Modeled deposition distribution patterns of
                inhaled DE particles in the airways of different species.
                Generation 1-18 are TB; >18 are A.
inhaled in 1 min in the lungs of humans, rats, and hamsters on three different bases: the total
lung volume (M), the surface area of all lung airways (Mj), or the surface area of the epithelium
of the alveolar region only (M2).  According to this analysis, the deposited dose is lower in
humans than in the two rodent species regardless of how the deposited dose is expressed. These
results are most certainly due predominately to the greater respiratory  exchange rate in rodents
and smaller size of the rodent lung.  Table 3-1 also indicates that the differences (between
humans to animals) are less on a surface area basis («3-fold) than on a lung volume basis (~ 14-
fold).  This is due to larger alveolar diameters and concomitant lower  surface area per unit of
lung volume in humans.  Such differences in the deposited dose in relevant target areas such as
the alveolar region are important and have to be considered when extrapolating the results from
DPM  exposure studies in animals to humans.  As will be discussed elsewhere in this document,
procedures for dose extrapolation from animals to humans includes considering the process of
clearance, with clearance measurements being in relation to surface area rather than to volume.
Thus predicted doses of particulates would be based on surface areas,  such as Ml andM2 in Table
3-1, rather than on volume, M.

-------
Table 3-1. Predicted doses of inhaled DPM per minute based on total lung volume (M),
total airway surface area (M,), or surface area in alveolar region (M2)
Species
Hamster
Fischer rat
Human
iyr - mass DPM deposited
M
(10 3 ng/min/cm3)
3.548
3.434
0.249
in luns per minute
Mt
(10 6 |ig/min/cm2)
3.088
3.463
1.237

M2
(10 6 ng/min/cm2)
2.382
2.608
0.775

            total lung volume
M = mass DPM deposited in lung per minute
           total airway surface area
M2 = mass DPM deposited on the unciliated airways per minute
           surface area of the unciliated airways

Based on the following conditions:  (1) mass median aerodynamic diameter (MMAD) = 0.2 \im; geometric standard
deviation (og) = 1.9; packing density (cj>) = 0.3; and particle mass density (p) = 1.5 g/cm2; (2) particle concentration =
1 mg/m3; and (3) nose-breathing. For humans, total lung volume = 3200 cm2, total airway surface area = 633,000
cm2, surface area of the unciliated airways = 627,000 cm2. Corresponding values for Fisher rats are 418cm2, 412cm2,
and 409cm2; for hamsters, 282cm2, 262cm2, and 261cm2. Tidal volumes (in cm2) and respiratory frequency (per min)
used for humans were 500 and 14; for Fisher rats, 1.6 and 98; for hamsters, 67 and 1.0.

Source: XuandYu, 1987.

     Particle deposition will initiate particle redistribution processes (e.g., clearance
mechanisms, phagocytosis) that transfer the particles to various subcompartments, including the
alveolar macrophage pool, pulmonary interstitium, and lymph nodes.  Over time, therefore, only
small amounts of the original particle intake would be associated with the alveolar surface areas.

3.3.2. Particle Clearance and Translocation Mechanisms
       This section provides an overview of the mechanisms and pathways by which particles
are cleared from the respiratory tract. The mechanisms of particle clearance as well as clearance
routes from the various regions of the respiratory tract have been considered in the PM Criteria
Document (U.S. EPA, 1996) and reviewed by Schlesinger et al. (1997).
       Particles that deposit upon airway surfaces may be cleared from the respiratory tract
completely, or be translocated to other sites within this system, by various regionally distinct
processes. These clearance mechanisms can be categorized as either absorptive (i.e., dissolution)
or nonabsorptive (i.e., transport of intact particles) and may occur simultaneously or with

                                             3-9

-------
temporal variations.  Particle solubility in terms of clearance refers to solubility within the
respiratory tract fluids and cells. Thus, a poorly soluble particle is one whose rate of clearance by
dissolution is insignificant compared to its rate of clearance as an intact particle (as is the case
with DPM).  The same clearance mechanisms act on different particles to different degrees, with
their ultimate fate being a function of deposition site, physicochemical properties (including any
toxicity), and sometimes deposited mass or number concentration. However, the duration of
clearance for poorly soluble particles such as DPM as it exists between species, months for rats
vs. years or even decades for humans, can make dissolution of DPM a significant contributor for
humans (Kreyling, 1992).
       Figure 3-4 outlines many of the known and suspected clearance pathways for poorly
soluble particles, such as DPM, that deposit in the alveolar region. Included are the
representations of the translocation pathways from the alveolar epithelium through the
insterstitium and on through the lymph nodes; this latter path will be referred to frequently later
in this chapter.
       Deposited Particle
        Phagocytosis by
     Alveolar Macrophages
                                               Endocytosis by
                                          -^  Type I Alveolar
                                           i ypc i /-VIVGUICM     i       i
                                           Epithelial Cells     I Blood k-
                                                             ^r
       Movement within
       Alveolar Lumen

Bronchiolar/ Bronchial
       Lumen       4
  Passage Through
 Alveolar Epithelium
         ir~
I—  Interstitium
        t
      Mucociliary Blanket
              I
                          Lymphatic Channels
                                   I
                               Lymph
                                                              Passage through
                                                            Pulmonary Capillary
                                                                Endothelium
                                                                t
                                                               Phagocytosis by
                                                                  Interstitial
                                                                Macrophages
           Gl Tract
Figure 3-4. Diagram of known and suspected clearance pathways for poorly soluble
particles depositing in the alveolar region. (Modified from Schlesinger, 1995).
                                         3-10

-------
3.3.2.1. Extrathoracic Region
       The clearance of poorly soluble particles deposited in the nasal passages occurs via
mucociliary transport, and the general flow of mucus is backwards, i.e., towards the nasopharynx.
Mucus flow in the most anterior portion of the nasal passages is forward, clearing deposited
particles to the vestibular region where removal is by sneezing, wiping, or blowing.
       Soluble material deposited on the nasal  epithelium is accessible to underlying cells via
diffusion through the mucus. Dissolved substances may be subsequently translocated into the
bloodstream.  The nasal passages have a rich vasculature, and uptake into the blood from this
region may occur rapidly.
       Clearance of poorly soluble particles deposited in the oral passages is by  expectoration or
by swallowing into the gastrointestinal tract.

3.3.2.2. Tracheobronchial Region
       The dynamic relationship between deposition and clearance is responsible for
determining lung burden at any point in time. Clearance of poorly soluble particles from the TB
region is mediated primarily by mucociliary transport, a more rapid process than those operating
in alveolar regions.  Mucociliary transport (often referred to as the mucociliary escalator) is
accomplished by the rhythmic beating of cilia that line the respiratory tract from the trachea
through the terminal bronchioles.  This movement propels the mucous layer containing deposited
particles (or particles within alveolar macrophages [AMs]) toward the larynx. Clearance rate by
this system is determined primarily by the flow velocity of the mucus, which is greater in the
proximal airways and decreases distally.  These rates also exhibit interspecies and individual
variability. Considerable species-dependent variability in tracheobronchial  clearance has been
reported, with dogs  generally having faster clearance rates than guinea pigs, rats, or rabbits
(Felicetti et al., 1981).  The half-time (t1/2) values for tracheobronchial clearance of relatively
insoluble particles are usually on the order of hours, as compared to alveolar clearance, which is
on the order of hundreds of days in humans and dogs.  The clearance of particulate matter from
the tracheobronchial region is generally recognized  as being biphasic or multiphasic (Raabe,
1982). Some studies have shown that particles are cleared from large, intermediate, and small
airways with t1/2  of 0.5, 2.5, and 5 h, respectively. However, reports have indicated that clearance
from airways is biphasic and that the long-term component for humans may take much longer for
a significant fraction of particles deposited in this region, and may not be complete within 24 h as
generally believed (Stahlhofen et al., 1990; ICRP, 1994).
       Although most of the parti culate matter will be cleared from the tracheobronchial region
towards the larynx and ultimately swallowed, the contribution  of this fraction relative to
carcinogenic potential is unclear. With the exception of conditions of impaired bronchial
                                          3-11

-------
clearance, the desorption t1/2 for particle-associated organics is generally longer than the
tracheobronchial clearance times, thereby making uncertain the importance of this fraction
relative to toxicity in the respiratory tract (Pepelko, 1987).  However, Gerde et al. (1991a)
showed that for low-dose exposures, particle-associated PAHs were released rapidly at the site of
deposition indicating that they would be available for involvement in postulated carcinogenic
processes. The relationship between the early clearance of poorly soluble particles of 4 \im
aerodynamic diameter from the tracheobronchial regions and their longer-term clearance from
the alveolar region is illustrated in Figure 3-5, clearly showing the rapid depuration from the TB
region compared with the A region.  This relationship, although demonstrated with 4 |_im
particles, is probably relevant and applicable to DPM-sized particles (i.e., 0.2 |_im) as clearance
mechanisms are believed not to be particularly particle-sized dependent (Morrow et al., 1967a,b;
Snipes etal., 1983).
       Cuddihy and Yeh (1986) reviewed respiratory tract clearance of particles inhaled by
humans.  Depending on the type of particle (ferric oxide, Teflon discs, or albumin microspheres),
the technique employed, and the anatomic region (midtrachea, trachea, or main bronchi), particle
                             Tracheobronchial
                             Deposition
                                    Alveolar Deposition
                         20
80
                                    40            60
                             Hours After Inhalation
Figure 3-5. Modeled clearance of poorly soluble 4-|im particles deposited in
tracheobronchial and alveolar regions in humans.
                                      3-12

-------
velocity (moved by mucociliary transport) ranged from 2.4 to 21.5 mm/min. The highest
velocities were recorded for midtracheal transport, and the lowest were for main bronchi.
      Cuddihy and Yeh (1986) described salient points to be considered when estimating
particle clearance velocities from tracheobronchial regions: these include respiratory tract airway
dimensions, calculated inhaled particle deposition fractions for individual airways, and thoracic
(A + TB) clearance measurements. Predicted clearance velocities for the trachea and main
bronchi were found to be similar to those experimentally determined for inhaled radiolabeled
particles, but not those for intratracheally instilled particles.  The velocities observed for
inhalation studies were generally lower than those of instillation studies. Figure 3-6 illustrates a
comparison of the short-term clearance of inhaled particles by human subjects and the model
predictions for this clearance. However, tracheobronchial clearance via the mucociliary escalator
is of limited importance for long-term clearance.
      1.0
0.8
  0)
  §.
  o>
 Q
  o
      0.6
 .2   0.4
      0.2
  o
  (0
                                       Range of Three
                                       Measurements
                                              Model Projection
                                              Same as Lower
                                              Limit of Range
                     20        40       60        80
                             Hours After Inhalation
                                                         100
120
 Figure 3-6. Short-term thoracic clearance of inhaled particles as determined by model
 prediction and experimental measurement.

 Source: Cuddihy and Yeh, 1986 (from Stahlhofen et al., 1980).
                                      3-13

-------
       Exposure of F344 rats to whole exhaust containing DPM at concentrations of 0.35, 3.5, or
7.1 mg/m3 for up to 24 mo did not significantly alter tracheal mucociliary clearance as assessed
by clearance of 99mTc-macroaggregated albumin instilled into the trachea (Wolff et al., 1987).
The authors stated that measuring retention would yield estimates of clearance efficiency
comparable to measuring the velocity for transport of the markers in the trachea. The results of
this study were in agreement with similar findings of unaltered tracheal mucociliary clearance in
rats exposed to DPM (0.21, 1.0, or 4.4 mg/m3) for up to 4 mo (Wolff and Gray, 1980).  However,
the 1980 study by Wolff and Gray, as well as an earlier study by Battigelli et al. (1966), showed
that acute exposure to high concentrations of DE soot (1.0 and 4.4 mg/m3 in the study by Wolff
and Gray [1980] and 8 to 17 mg/m3 in the study by Battigelli et al. [1966]) produced transient
reductions in tracheal mucociliary clearance. Battigelli et al. (1966) also noted that the
compromised tracheal clearance was not observed following cessation of exhaust exposure.
       That tracheal  clearance does not appear to be significantly impaired or is impaired only
transiently following exposure to high concentrations of DPM is consistent with the absence of
pathological effects in the tracheobronchial region of the respiratory tract in experimental
animals exposed to DPM. The apparent retention of a fraction of the deposited dose in
the airways could be  cause for some concern regarding possible effects in this region, especially
in light of the results from simulation studies by Gerde et al. (1991b) suggesting that release of
PAHs from particles  may occur within minutes and therefore at the site of initial deposition.
However, the absence of effects in the TB areas in long-term DPM studies and experimental
evidence that particle-associated PAHs are released at the site of particle deposition together
suggest that these PAHs and other organics may be of lesser importance in tumorigenic responses
of rats  than originally suspected.  On the other hand, the data of Nikula et al. (1997a,b) could be
interpreted to suggest that a larger fraction of particles are translocated to the interstitium of the
respiratory tract in primates that are heavily  exposed (and therefore presumably in humans) than
in rats that are heavily exposed, including the interstitium of the respiratory bronchioles, an
anatomical site absent in rats (Section 3.6).  Moreover, eluted PAHs in the TB region are retained
longer  than those in the alveoli (Gerde et al., 1999), allowing time for activation.  Also, the
results  of Kreyling (1992) indicate that appreciable dissolution of even poorly soluble particles
may occur as a consequence of long absolute duration of clearance, such as years or decades, in
humans.  Thus PAHs may have a role in human response to DE that cannot be evaluated with the
rat model.
       Also,  impairment of mucociliary clearance function as a result of exposure to
occupational  or environmental respiratory tract toxicants or to cigarette smoke may significantly
enhance the retention of particles in the TB region.  For example, Vastag et al. (1986)
demonstrated that not only smokers with clinical symptoms of bronchitis but also symptom-free
                                          3-14

-------
smokers have significantly reduced mucociliary clearance rates.  Although impaired
tracheobronchial clearance could conceivably have an impact on the effects of deposited DPM in
the conducting airways, it does not appear to be relevant to the epigenetic mechanism likely
responsible for DE-induced rat pulmonary tumors as the tumors observed in these studies were
all or nearly all of A vice TB origin.
       Poorly soluble particles such as DPM that are deposited within the TB region are cleared
predominantly by mucociliary transport towards the oropharynx, followed by swallowing.
Poorly soluble particles may also be cleared by traversing the epithelium by endocytotic
processes, and enter the peribronchial region. Clearance may occur following phagocytosis by
airway macrophages, located on or beneath the mucous lining throughout the bronchial tree, or
via macrophages that enter the airway lumen from the bronchial or bronchiolar mucosa
(Robertson, 1980).

3.3.2.3. A Region
       A number of investigators have reported on the alveolar clearance kinetics of human
subjects. Bohning et al.  (1980) examined alveolar clearance in eight humans who had inhaled
<0.4 mg of 85Sr-labeled polystyrene particles (3.6 ± 1.6  |_im diam.). A double-exponential model
best described the clearance of the particles and provided t1/2 values of 29 ± 19 days and 298 ±
114 days for short-term and long-term phases, respectively. It was noted that of the particles
deposited in the alveolar region, 75% ± 13% were cleared via the long-term phase.  Alveolar
retention t1/2 values of 330 and 420 days were reported for humans who had inhaled
aluminosilicate particles of MMAD 1.9 and 6.1  |_im (Bailey et al., 1982).  In a comprehensive
study Bailey et al. (1985) followed the long-term retention of inhaled particles in  a human
respiratory tract.  The retention of 1 and 4 |_im fused aluminosilicate particles labeled with
strontium-85 and yttrium-88, respectively, was followed in male volunteers for about 533 days.
Approximately 7% of the initial  lung deposit of 1 [am particles and 40% of the 4 |_im particles
were associated with a rapid clearance phase corresponding to the calculated tracheobronchial
deposits. Retention of the remaining material followed a two-component exponential function,
with phases having half-times of the order of tens of days  and several hundred days, respectively.
       Quantitative data on clearance rates in humans having large lung burdens  of paniculate
matter are lacking.  Bohning et al. (1982) and Cohen et  al. (1979), however, did provide evidence
for slower clearance in smokers, and Freedman  and Robinson (1988) reported slower clearance
rates in coal miners who had mild pneumoconiosis with presumably high lung burdens of coal
dust. Although information on particle burden and particle overload relationships in humans is
much more limited than in experimental animal models, inhibition of clearance does seem to
occur.  Stober et al. (1967) estimated a clearance t1/2 of 4.9 years in coal miners with nil or slight
                                          3-15

-------
silicosis, based on postmortem lung burdens.  The lung burdens and estimated exposure histories
ranged from 2 to 50 mg/g of lung or more, well above the value at which clearance impairment is
observed in the rat. Furthermore, impaired clearance resulting from smoking or exposure to
other respiratory toxicants may increase the possibility of an enhanced particle accumulation
effect resulting from exposure to other particle sources such as DPM.
       Normal alveolar clearance rates in laboratory animals exposed to DPM have been
reported by a number of investigators (Table 3-2). Because the rat is, historically, the species for
which experimentally induced lung cancer data are available and for which most clearance data
exist, it is the species most often used for assessing human  risk, and reviews of alveolar clearance
studies have been generally limited to this species.
       Chan et al. (1981) subjected 24 male F344 rats to nose-only inhalation of diluted DE
generated from a diesel engine (6 mg/m3) labeled with 131Ba or 14C for 40 to 45  min and assessed
total lung deposition, retention, and elimination.  Based on  radiolabel inventory, the deposition
efficiency in the respiratory tract was  15% to 17%. Measurement of 131Ba label in the feces
during the first 4 days following exposure indicated that 40% of the deposited DPM was
eliminated via mucociliary clearance.  Clearance of the particles from the lower respiratory tract
followed a two-phase elimination process consisting of a rapid (ti/2 of 1  day) elimination by
mucociliary transport and a slower (ti/2 of 62 days) macrophage-mediated alveolar clearance.
This study provided data for normal alveolar clearance rates of DPM not affected by prolonged
exposure or particle overloading.
       Several studies have investigated the effects of exposure concentration on the alveolar
clearance of DPM by laboratory animals. Wolff et al. (1986, 1987) provided clearance data (ti/2)
and lung burden values for F344 rats exposed to DE for 7 h/day, 5 days/week for 24 mo.
Exposure concentrations of 0.35,  3.5,  and 7.1  mg of DPM/m3 were employed in this whole body-
inhalation exposure experiment. Intermediate (hours-days) clearance of 67Ga2O3 particles (30
min, nose-only inhalation) was assessed after 6, 12, 18, and 24 mo of exposure  at all of the DPM
concentrations.  A two-component function described the clearance  of the administered
radiolabel:

             F(t) = ^exp(-0.693  t/ij +  Sexp(-0.693t/T2),             (3-1)

where F(t) was the percentage retained throughout the respiratory tract, A and B were the
magnitudes of the two components (component^ included nasal,  lung,  and gastrointestinal
clearance, while component B represented intermediate lung clearance) and TJ and i2 were the
                                          3-16

-------
      Table 3-2.  Alveolar clearance in laboratory animals exposed to DPM in whole exhaust
Species/sex
Rats, F-344, M



Rats, F-344



Rats


Rats, F-344, MF


Rats, F-344;


Guinea pigs,
Hartley
Rats, F-344




Exposure
technique
Nose only;
Radiolabeled DPM


Whole body;
assessed effect
on clearance of
67Ga2O3 particles
Whole body


Whole body


Nose-only;
Radiolabeled 14C








Exposure
duration
40-45 min



7h/day
5 days/week
24 mo

19 h/day
5 days/week
2.5 years
7 h/day
5 days/week
18 mo
45 min
140 min


45 min
20 h/day
7 days/week
7-1 12 days


Particles
mg/m3
6



0.35
3.5
7.1

4


0.15
0.94
4.1
7
2


7
0.25
6



Observed effects
Four days after exposure, 40% of DPM eliminated by
mucociliary clearance. Clearance from lower RT was in
2 phases. Rapid mucociliary (t1/2 = 1 day); slower
macrophage-mediated (t1/2 = 62 days).
T! significantly higher with exposure to 7. 1 mg/m3 for
24 mo; T2 significantly longer after exposure to 7.1 mg/m3
for 6 mo and to 3.5 mg/m3 for 18 mo.

Estimated alveolar deposition = 60 mg; particle burden
caused lung overload. Estimated 6-15 mg particle-bound
organics deposited.
Long-term clearance was 87 ± 28 and 99 ± 8 days for
0.15 and 0.94 mg/m3 groups, respectively; t1/2 = 165 days
for 4. 1 mg/m3 group.
Rats demonstrated 3 phases of clearance witht1/2 = 1, 6,
and 80 days, representing tracheobronchial, respiratory
bronchioles, and alveolar clearance, respectively. Guinea
pigs demonstrated negligible alveolar clearance from
day 10 to 432.
Monitored rats for a year. Proposed two clearance models.
Clearance depends on initial particle burden; t1/2 increases
with higher exposure. Increases in t1/2 indicate increasing
impairment of AM mobility and transition into overload
condition.
Reference
Chan etal. (1981)



Wolff etal. (1986,
1987)


Heinrich et al.
(1986)

Griffis etal. (1983)


Lee etal. (1983)




Chan etal. (1984)




RT = respiratory tract.
AM = alveolar macrophage.
T! = clearance from primary, ciliated airways.
T2 = clearance from nonciliated passages.

-------
half-times for the A and B components, respectively. The early clearance half-times (TJ), were
similar for rats in all exposure groups at all time points except in the high-exposure (7.1 mg/m3)
group following 24 mo of exposure, which was faster than the controls. Significantly longer B
component retention half-times, representing intermediate clearance probably from nonciliated
structures such as alveolar ducts and alveoli, were noted after as little as 6 mo exposure to DPM
at 7.1 mg/m3 and 18 mo exposure to 3.5 mg/m3.
       Nose-only exposures to 134Cs fused aluminosilicate particles (FAP) were used to assess
long-term (weeks-months) clearance. Following 24-mo exposure to DPM, long-term clearance
of 134Cs-FAP was significantly (p<0.01) altered in the 3.5 (cumulative exposure [C x T] of
11,760 mg-h/m3) and 7.1 mg/m3, C x T = 23,520 mg-h/m3) exposure groups (ti/2 of 264 and 240
days, respectively) relative to the 0.35 mg/m3 and control groups (ti/2 of 81 and 79 days,
respectively). Long-term clearance represents the slow component of particle removal from the
alveoli.  The decreased clearance correlated with the greater particle burden in the lungs of the
3.5 and 7.1 mg/m3 exposure groups. Based on these findings, the cumulative exposure of
> 11,760 mg-h/m3 (or 3.5 mg/m3 for a lifetime exposure) represented a particle overload
condition resulting in compromised alveolar clearance mechanisms; the clearance rate at the
lowest concentration (0.35 mg/m3; cumulative exposure of 118 mg-h/m3) was not different from
control rates (Figure 3-7).
          100-1
      O
      m
      O)
          10
                                                                        High
                                                                Control
                   20
 i
40
60
 I
80
100
 I   i    I
120   140
160
180
200
                                         Time (Days)
      Figure 3-7.  Clearance from lungs of rats of 134Cs-FAP fused aluminosilicate
      tracer particles inhaled after 24 months of DE exposure at concentrations of
      0 (control), 0.35 (low), 3.5 (medium), and 7.1 (high) mg DPM/m3.
                                          3-18

-------
       Heinrich et al. (1986) exposed rats 19 h/day, 5 days/week for 2.5 years to DPM at a
particle concentration of about 4 mg/m3, equal to a "C x T" of 53,200 mg-h/m3.  The deposition
in the alveolar region was estimated to equal 60 mg. The lung particle burden was apparently
sufficient to result in a "particle overload" condition (Section 3.4). With respect to the organic
matter adsorbed onto the particles, the authors estimated that over the 2.5-year period, 6-15 mg of
particle-bound organic matter had been deposited and was potentially available for biological
effects. This estimation was based on the analysis of the DE used in the experiments, values for
rat ventilatory functions, and estimates of deposition and clearance.
       Accumulated burden of DPM in the lungs following an 18-mo, 7 h/day, 5 days/week
exposure to whole DE was reported by Griffis et al. (1983).  Male and female F344 rats exposed
to 0.15, 0.94, or 4.1  mg DPM/m3 were sacrificed at 1 day and 1, 5, 15, 33, and 52 weeks after
exposure, and DPM was extracted from lung tissue dissolved in tetramethylammonium
hydroxide.  Following centrifugation and washing of the supernatant, DPM content of the tissue
was quantitated using spectrophotometric techniques. The analytical procedure was verified by
comparing  results to recovery studies using known amounts of DPM with lungs of unexposed
rats.  Lung  burdens were 0.035, 0.220, and 1.890 mg/g lung tissue, respectively, in rats exposed
to diluted whole exhaust at 0.15, 0.94, and 4.1 mg DPM/m3. Long-term retention for the 0.15
and 0.94 mg/m3 groups had estimated half-times of 87 ± 28 and 99 ± 8 days,  respectively. The
retention t1/2 for the 4.1-mg/m3 exposure group was 165 ± 8 days, which was  significantly
(/X0.0001) greater than those of the lower exposure groups. The 18-mo exposures to 0.15 or
0.96 mg/m3 levels of DPM [C x  T] equivalent of 378 and 2,368 mg-h/m3, respectively) did not
affect clearance rates, whereas the exposure to the 4.1 mg/m3 concentration C x T = 10,332
mg-h/m3) resulted in impaired clearance.
       Lee et al. (1983) described the clearance of DPM (7 mg/m3 for 45 min or 2 mg/m3 for 140
min) by F344 rats (24 per group) and Hartley guinea pigs exposed by nose-only inhalation to
diluted whole exhaust with no apparent particle overload in the lungs as being in three  distinct
phases. The exposure protocols  provided comparable total doses based on a  14C radiolabel.
14CO2 resulting from combustion of 14C-labeled diesel fuel was removed by a diffusion scrubber
to avoid erroneous assessment of 14C intake by the animals.  Retention of the radiolabeled
particles was determined up to 335 days after exposure and resulted in a three-phase clearance
with retention t1/2 values of 1,  6,  and 80 days. The three clearance phases are taken to represent
removal of tracheobronchial deposits by the mucociliary escalator, removal of particles deposited
in the respiratory bronchioles, and alveolar clearance, respectively.  Species variability in
clearance of DPM was also  demonstrated because the Hartley guinea pigs exhibited negligible
alveolar clearance from day 10 to day 432 following a 45-min exposure to a DPM concentration
                                          3-19

-------
of 7 mg/m3.  Initial deposition efficiency (20% ± 2%) and short-term clearance were, however,
similar to those for rats.
       Lung clearance in male F344 rats preexposed to diluted whole DE containing DPM at
0.25 or 6 mg/m3 20 h/day, 7 days/week for periods lasting from 7 to 112 days was studied by
Chan et al. (1984). Following this preexposure protocol, rats were subjected to 45-min nose-
only exposure to 14C-DE, and alveolar clearance of radiolabel was monitored for up to 1 year.
Two models were proposed:  a normal biphasic clearance model and a modified lung retention
model that included a slow-clearing residual component to account for sequestered aggregates of
macrophages.  The first model described a first-order clearance for two compartments:  R(t) =
Ae"ult + Be"u2t.  This yielded clearance t1/2 values  of 166 and 562 days for rats preexposed to 6.0
mg/m3 for 7 and 62 days, respectively.  These values were significantly (p<0.05) greater than the
retention t1/2 of 77 ± 17 days for control rats. The same retention values for rats of the 0.25
mg/m3 groups were 90 ± 14 and 92 ± 15 days, respectively, for 52- and 112-day exposures and
were not significantly different from controls. The two-compartment model represents overall
clearance of the tracer particles, even if some of the particles were sequestered in particle-laden
macrophages with substantially slower clearance rates. For the second model, which excluded
transport of the residual fractions in sequestered macrophage aggregates, slower clearance was
observed in the group with a lung burden of 6.5  mg (exposed to 6.0 mg/m3 for 62 days), and no
clearance was observed in the 11.8 mg group (exposed to 6.0 mg/m3 for 112 days).  Clearance
was shown to be dependent on the initial burden of particles, and therefore the clearance t1/2
would increase in higher exposure scenarios.  This study emphasizes the importance of particle
overloading of the lung and the ramifications on clearance of particles; the significant increases
in half-times indicate an increasing impairment of the alveolar macrophage mobility and
subsequent transition into an overload condition as is discussed further in Section 3.4.
       Long-term alveolar clearance rates of particles in various laboratory animals and humans
have been reviewed by Pepelko (1987). Although retention t1/2 varies both among and within
species and is also dependent on the physicochemical properties of the inhaled particles, the
retention t1/2 for humans is much longer (>8 mo) than the average retention t1/2 of 60 days for rats.
       Clearance from the A region occurs via a number of mechanisms and pathways, but the
relative importance of each is not always certain and may vary between species.  Particle removal
by macrophages comprises the main nonabsorptive clearance process in this region. Alveolar
macrophages reside on the epithelium, where they phagocytize and transport deposited material,
which they contact by random motion or via directed migration under the influence of local
chemotactic factors (Warheit et al., 1988).
       Particle-laden macrophages may be cleared from the A region along a number of
pathways (U.S. EPA, 1996). Uningested particles or macrophages in the interstitium may
                                          3-20

-------
traverse the alveolar-capillary endothelium, directly entering the blood (Raabe, 1982; Holt,
1981); endocytosis by endothelial cells followed by exocytosis into the vessel lumen seems,
however, to be restricted to particles <0.1 pm diameter, and may increase with increasing lung
burden (Lee et al., 1985; Oberdorster, 1988).  Once in the systemic circulation, transmigrated
macrophages, as well as uningested particles, can travel to extrapulmonary organs.
       Alveolar macrophages constitute an important first-line cellular defense mechanism
against inhaled particles that deposit in the alveolar region of the lung. It is well established that
a host of diverse materials, including DPM, are phagocytized by AMs shortly after deposition
(White and Garg, 1981; Lehnert and Morrow, 1985) and that such cell-contained particles are
generally rapidly sequestered from both the extracellular fluid lining in the alveolar region and
the potentially sensitive alveolar epithelial cells. In addition to this role in compartmentalizing
particles from other lung constituents, AMs are prominently involved in mediating the clearance
of relatively insoluble particles from the air spaces (Lehnert and Morrow, 1985). Although the
details of the actual process have not been delineated, AMs with their particle burdens gain
access and become coupled to the mucociliary escalator and are subsequently transported from
the lung via the conducting airways. Although circumstantial, numerous lines of evidence
indicate that such AM-mediated particle clearance is the predominant mechanism by which
relatively insoluble particles are removed from the alveolar region of the lungs (Gibb and
Morrow, 1962; Ferin,  1982; Harmsen et al., 1985; Lehnert and Morrow,  1985; Powdrill et al.,
1989).
       The removal characteristics for particles deposited in the alveolar region of the lung have
been descriptively represented by numerous investigators as a multicompartment or
multicomponent process in which each component follows simple first-order kinetics (Snipes
and Clem, 1981; Snipes et al., 1988; Lee et al., 1983). Although the various compartments can
be described mathematically, the actual physiological mechanisms determining these differing
clearance rates have not been well characterized.
       Lehnert et al. (1988, 1989) performed studies using laboratory rats to  examine
particle-AM relationships over the course of alveolar clearance of low to high lung burdens of
noncytotoxic microspheres (2.13  [am diam.) to obtain information on potential AM-related
mechanisms that form the underlying bases for kinetic patterns of alveolar clearance as a function
of particle lung burdens. The intratracheally instilled lung burdens varied from 1.6 x 107
particles (about 85  |_ig) for the low lung burden to 2.0 x 108 particles (about 1.06 mg) for the mid-
dose and 6.8 x io8 particles (about 3.6 mg) for the highest lung burden. The lungs were lavaged
at various times postexposure and the numbers of spheres in each macrophage counted.
Although such experiments provide information regarding the response of the lung to particulate
matter, intratracheal instillation is not likely to result in the same depositional characteristics as
                                          3-21

-------
inhalation of particles.  Therefore, it is unlikely that the response of alveolar macrophages to
these different depositional characteristics will be quantitatively similar.
       The t1/2 values of both the early and later components of clearance were virtually identical
following deposition of the low and medium lung burdens. For the highest lung burden,
significant prolongations were found in both the early, more rapid, as well as the slower
component of alveolar clearance. The percentages of the particle burden associated with the
earlier and later components, however, were similar to those  of the lesser lung burdens. On the
basis of the data, the authors concluded that translocation of AMs from alveolar spaces by way of
the conducting airways is fundamentally influenced by the particle burden of the cells so
translocated.  In the case of particle overload that occurred at the highest lung burden, the
translocation of AMs with the heaviest cellular burdens of particles (i.e., greater than about
100 microspheres per AM) was definitely compromised.
       On the other hand, analysis of the disappearance of AMs with various numbers of
particles indicates that the particles may not exclusively reflect the translocation of AMs from the
lung. The observations are also consistent with a gradual redistribution of retained particles
among the AMs in the lung concurrent with the removal of particle-containing AMs via the
conducting airways. Experimental support suggestive of potential processes for such particle
redistribution comes from a variety of investigations involving AMs and other endocytic cells
(Heppleston and Young, 1973; Evans et al., 1986; Aronson, 1963;  Sandusky et al., 1977;
Heppleston, 1961; Riley and Dean, 1978).

3.3.3.  Translocations of Particles to Extra-Alveolar Macrophage Compartment Sites
       Although the phagocytosis of particles by cells free within the lung and the mucociliary
clearance of the cells with their particulate matter burdens represent the most prominent
mechanisms that govern the fate of particles deposited in the alveolar region, other mechanisms
exist that can affect both the retention characteristics  of relatively insoluble particles in the lung
and the lung clearance pathways for the particles.  One mechanism is endocytosis of particles by
alveolar lining (Type I) cells (Sorokin and Brain, 1975; Adamson and Bowden, 1978, 1981) that
normally provide >90% of the cell surface of the alveoli in the lungs of a variety of mammalian
species (Crapo et al., 1983).  This process may be related to the size of the particles that deposit
in the lungs and the numbers of particles that are deposited.  Adamson and Bowden (1981) found
that with increasing loads of carbon particles (0.03 |_im diam.) instilled in the lungs of mice, more
free particles were observed in the alveoli within a few  days; it should be noted, however, that
this phenomenon was demonstrated with very high doses given as a bolus such that the
mechanism and relevance of this  phenomenon at lower concentrations may be different or even
unrelated to what may happen at much lower concentrations.  The relative abundance of particles
                                          3-22

-------
endocytosed by Type I cells also increased with increasing lung burdens of the particles, but
instillation of large particles (1.0 |_im) rarely resulted in their undergoing endocytosis.  A 4 mg
burden of 0.1 |_im diameter latex particles is equivalent to 8 x  1012 particles, whereas a 4 mg
burden of 1.0 |_im particles is composed of 8 x io9 particles. Regardless, DPM with volume
median diameters between 0.05 and 0.3 [am (Frey and Corn, 1967; Kittleson et al., 1978) would
be expected to be within the size range for engulfment by Type I cells should suitable encounters
occur.  Indeed, it has been demonstrated that DPM is endocytosed by Type I cells in vivo (White
andGarg, 1981).
       Unfortunately, information on the kinetics of particle engulfment (endocytosis) by Type I
cells relative to that by AMs is scanty. Even when relatively low burdens of particulate matter
are deposited in the lungs, some fraction of the particles usually appears in the regional lymph
nodes (Ferin and Feldstein, 1978; Lehnert, 1989). As will be  discussed, endocytosis of particles
by Type I cells is an initial, early step in the passage of particles to the lymph nodes.  Assuming
particle phagocytosis is not sufficiently rapid or perfectly efficient, increasing numbers of
particles would be expected to gain entry into the Type I epithelial cell compartment during
chronic aerosol exposures. Additionally, if particles are released on a continual basis by AMs
that initially sequestered them after lung deposition, some fraction of the "free" particles so
released could also undergo passage from the alveolar space into Type I cells.
       The endocytosis of particles by Type I cells represents only the initial  stage of a process
that can lead to the accumulation of particles in the lung's interstitial compartment and the
subsequent translocation of particles to the regional lymph nodes.  As suggested by the results of
Adamson and Bowden (1981), a vesicular transport mechanism in the Type I cell can transfer
particles administered at high concentrations by instillation from the air surface of the alveolar
epithelium into the lung's interstitium, where particles may be phagocytized by interstitial
macrophages or remain in a "free" state for a poorly defined period that may be dependent on the
physicochemical characteristics of the particle.  The lung's interstitial compartment accordingly
represents an anatomical site for the retention of particles in the lung, although the kinetics on
movement into and out of this site remain obscure for both humans and test species. Whether or
not AMs, and perhaps polymorphonuclear neutrophils (PMNs) that have gained access to the
alveolar space compartment and phagocytize particles there, also contribute to the particle
translocation process into the lung's interstitium also remains a controversial issue.
       Translocation of parti culate matter to the various interstitial spaces within the lung is a
prominent phenomenon occurring at least at high (occupational) exposures that has been
examined extensively for both DPM and coal dust in a species comparison between rats and
primates (Nikula et al., 1997a,b). Detailed pulmonary morphometry conducted on F344 rats and
cynomolgus monkeys that had been exposed for 24 months to occupational levels of DPM (1.95
                                          3-23

-------
mg/m3; see Lewis et al., 1989) showed major differences in the pulmonary sites of particulate
deposition. In rats, about 73% of DPM was present in the alveolar ducts/alveoli and 27% in
interstitial compartments; for monkeys the corresponding figures were markedly different at 43%
and 57%.  The corresponding pulmonary histopathology confirmed that both species were
affected, although rats are more sensitive, as incidence and severity scores for alveolar effects
ranged from 15 of 15 with severity scores from 1-4 (minimal to moderate), whereas for monkeys
the corresponding values were only 4 of 15 at a range of 0-2 (not observed to minimal).
Similarly, both species exhibited histopathology at the interstitial sites of deposition but with
effects in monkeys being  slightly more severe (1 of 15 graded as slight, 14 of 15 graded as
minimal) than those in rats (14 of 15 graded as slight, 1 of 15 graded as minimal). The basis for
this interspecies difference may be due to any number of clear contrasts that exist between rat
and primate lungs, including anatomical (primates and humans have respiratory bronchioles
whereas rats do not), kinetic (primates and human clearance processes allow more residence time
of particles in the lung than do those in rats or rats may have faster interstitial to lymph node
clearance rates than do humans and primates),  or morphological (primates and humans have
more interstitial tissue, more and thicker pleura, and wider interstitial  spaces than do rats).
Aspects of the study itself that may obscure its interpretation include the relative lifespan the
exposure represented between the tested species (lifetime for rat vs. about 10% lifetime of
primate), that there was only the single time point at which the relative burdens were determined,
and that rat lymph node burdens were not included in the analysis. The analysis of Kuempel
(2000) using human occupational data clearly showed that models require an interstitialization
process to provide adequate fits to the empirical human (miners') lung deposition data discussed
in that study.  Hypotheses about possible mechanisms for the interstitialization process are scant,
although Harmsen et al. (1985) provided some evidence in dogs that migration of AMs may
contribute to the passage of particles to the interstitial compartment and also may be involved in
the subsequent translocation of particles to draining lymph nodes.  Translocation to the
extrapulmonary regional lymph nodes apparently can involve the passage of free particles as well
as particle-containing cells via lymphatic channels in the lungs (Harmsen et al., 1985; Ferin and
Feldstein, 1978; Lee et al., 1985). Further, it has been noted that particles accumulate both more
rapidly and more abundantly in lymph nodes that receive lymphatic drainage from the lung (Ferin
and Feldstein, 1978; Lee et al.,  1985). It should be stressed that further investigation is required
to confirm the character and even existence of the interstitialization process in the lungs of
humans with exposures to particles at lower environmental concentrations, or to submicrometer
particles such as DPM, or to examine the kinetics and time course of the interstitialization
process.
                                          3-24

-------
3.3.3.1. Clearance Kinetics
       The clearance kinetics of PM have been reviewed in the PM CD (U.S. EPA, 1996) and by
Schlesinger et al. (1997), the results of which indicate that clearance kinetics may be profoundly
influenced by several factors. The influence of time, for example, is definitively showed by the
work of Bailey et al.  (1985; discussed above), who showed that the rate of clearance from the
pulmonary region to  the GI tract decreased nearly fourfold from initial values to those noted at
200 days and beyond after particle inhalation.

3.3.3.2. Interspecies Patterns of Clearance
       The inability  to study the retention of certain materials in humans for direct risk
assessment requires the use of laboratory animals.  Adequate toxicological assessment
necessitates that interspecies comparisons consider aspects of dosimetry including knowledge of
clearance rates and routes.  The basic mechanisms and overall patterns  of clearance from the
respiratory tract are similar in humans  and most other mammals.  Regional clearance rates,
however, can show substantial variation between species, even for similar particles deposited
under comparable exposure conditions (U.S. EPA, 1996; Schlesinger et al., 1997; Snipes et al.,
1989).
       In general, there are species-dependent rate constants for various clearance pathways.
Differences in regional and total clearance rates between some species are a reflection of
differences in mechanical clearance processes.  For consideration in assessing particle dosimetry,
the end result of interspecies differences in clearance is that the retained doses in the lower
respiratory tract can differ between species, which may result in differences in response to similar
particulate exposures.

3.3.3.3. Clearance Modifying Factors and Susceptible Populations
       A number of host and environmental factors may modify clearance kinetics and may
consequently make individuals exhibiting or afflicted with these factors particularly susceptible
to the effects resulting from exposure to DPM.  These include age, gender, physical activity,
respiratory tract disease, and inhalation of irritants (U.S. EPA, 1996, Section 10.4.2.5).
Respiratory tract clearance appears to be prolonged in a number of pathophysiological conditions
in humans, including chronic sinusitis, chronic bronchitis, asthma, chronic obstructive lung
disease, and various acute respiratory infections.

3.3.3.4. Respiratory Tract Disease
       Earlier studies reviewed in the PM CD (U.S. EPA, 1996) noted that various respiratory
tract diseases are associated with alterations in  overall clearance and clearance rates. Prolonged
                                           3-25

-------
nasal mucociliary clearance in humans is associated with chronic sinusitis or rhinitis, and cystic
fibrosis. Bronchial mucus transport may be impaired in people with bronchial carcinoma,
chronic bronchitis, asthma, and various acute infections.  In certain of these cases, coughing may
enhance mucus clearance, but it generally is effective only if excess secretions are present.
       The rates of A region particle clearance are reduced in humans with chronic obstructive
lung disease and in laboratory animals with viral infections, whereas the viability and functional
activity of macrophages are impaired in human asthmatics and in animals with viral-induced lung
infections (U.S. EPA, 1996). However, any modification of functional properties of
macrophages appears to be injury specific, reflecting the nature and anatomic pattern of disease.

3.4. PARTICLE "OVERLOAD"
3.4.1. Introduction
       Some experimental studies using laboratory rodents employed high exposure
concentrations of relatively nontoxic, poorly soluble particles.  These particle loads interfered
with normal clearance mechanisms, producing clearance rates different from those that would
occur at lower exposure levels. Prolonged exposure to high particle concentrations is associated
with what is termed particle overload. This is defined as the overwhelming of macrophage-
mediated clearance by the deposition of particles at a rate exceeding the capacity of that
clearance pathway. Aspects and occurrence of this  phenomenon have already been alluded to in
earlier portions of this chapter on alveolar clearance (Section 3.3.2.3). The relevance of this
phenomenon for human risk assessment has long been the object of scientific inquiry. A
monograph on this matter and many others relevant to DPM has appeared (ILSI, 2000), and the
results, opinions, and judgments put forth therein are used extensively in this chapter and in this
assessment.
       Wolff et al. (1987) used 134Cs-labeled fused aluminosilicate particles to measure alveolar
clearance in rats following 24-mo exposure to low,  medium, and high concentrations of DE
(targeted concentrations of DPM of 0.35, 3.5 and 7.1 mg/m3).  The short-term component of the
multicomponent clearance curves was similar for all groups, but long-term clearance was
retarded in the medium- and high-exposure groups (Figure 3-7). The half times  of the long-term
clearance curves were 79, 81, 264, and 240 days, respectively,  for the control, low-, medium-,
and high-exposure groups. Clearance was overloaded at the high and medium but not at the low
exposure level. Lung burdens of DPM were measured after 6,  12, 18, and 24 mo of exposure.
The results (Figure 3-8) indicate that the lung burden of deposited particles was appreciably
increased or "overloaded" compared with the low level of exposure in the two highest exposures
post 6 months. Figure 3-8 also compares these observational results of lung burden with
simulated results where no overload would occur (McClellan, 2000). Comparison
                                          3-26

-------
     16-.

     14-
 D)
  -  12H
 0)
 C
 O
 E
 CO
 0)
 Q.
"o
 O
C/}
10-

 8-

 6-

 4-

 2-

 0
mg/m3
•7
i
3C
.0
n "^
u.oo
Data



Model



                          i
        o
                               10
15
20
25
                       Diesel  Soot Inhalation, Months
Figure 3-8. Lung burdens (in mg DPM soot/g lung) in rats chronically exposed to DE at
 0.35 (low) (•), 3.5 (medium) (A), and 7.1 (high) mg / m3 (•). The solid figures
represent actual data with means and standard errors from animals sacrificed at 6,12,
and 18 months after initiation of exposures.  Lines are simulated model results from
these same exposure levels, assuming no effect of exposure concentration on deposition
or clearance of particles (from Wolff et al., 1987; McClellan, 2000).

of the observed and  simulated results clearly shows that the two highest exposure levels resulted
in lung burdens that were ever-increasing and not at all concordant with the simulated results,
whereas the burdens at the low-exposure level were closely approximated by the simulation.
Thus, at the two highest exposure levels, deposition processes were outpacing clearance
mechanisms. Results from the low-exposure level indicate that clearance processes were not
inhibited, the lung burden remaining the same throughout all time periods examined.
       Morrow (1988) has proposed that the condition of particle overloading in the lungs is
caused by a loss in the mobility of particle-engorged AMs and that such an impediment is related
to the cumulative volumetric load of particles in the AMs.  Morrow (1988) has further estimated
that the clearance function of an AM may be completely impaired when the particle burden in the
AM is of a volumetric size  equivalent to about 60% of the normal volume of the AM.  Morrow's
                                        3-27

-------
hypothesis was the initial basis for the physiology-oriented multicompartmental kinetic (POCK)
model derived by Stober et al. (1989) for estimating alveolar clearance and retention of relatively
insoluble, respirable particles in rats.
       A revised version of this model refines the characterization of the macrophage pool by
including both the mobile and immobilized macrophages (Stober et al., 1994). Application of
the revised version of the model to experimental data suggested that lung overload does not cause
a dramatic increase in the total burden of the macrophage pool but results in a great increase in
the particle burden of the interstitial space, a compartment that is not available for macrophage-
mediated clearance. The revised version of the POCK model is discussed in greater detail in the
context of other dosimetry models below.
       Oberdorster and co-workers (1992) assessed the alveolar clearance of smaller (3.3 pm
diam.) and larger (10.3 |_im diam.) polystyrene particles, the latter of which are volumetrically
equivalent to about 60% of the average normal volume of a rat AM, after intratracheal instillation
into the lungs of rats. Even though both sizes of particles were found to be phagocytized by AMs
within a day after deposition, and the smaller particles were cleared at a normal rate, only
minimal  lung clearance of the larger particles was observed over an approximately 200-day
postinstillation period, thus supporting the volumetric AM overload hypothesis.
       It has been hypothesized that when the retained lung burden approaches 1 mg particles/g
lung tissue, overloading will  begin in the rat (Morrow, 1988); at 10 mg particles/g lung tissue
macrophage-mediated clearance of particles would effectively cease. Overloading appears to be
a nonspecific effect noted in  experimental studies, generally in rats, using many different kinds of
poorly soluble particles (including TiO2, volcanic ash, DPM, carbon black, and fly ash) and
results in A region clearance slowing or stasis, with an associated inflammation and aggregation
of macrophages in the lungs  and increased translocation of particles into the interstitium (Muhle
et al., 1990a,b; Lehnert, 1990; Morrow,  1994). Following overloading, the subsequent
retardation of lung clearance, accumulation of particles, chronic inflammation, and the
interaction of inflammatory mediators with cell proliferative processes and DNA may lead to the
development of fibrosis, epithelial cell mutations, and fibrosis in rats (Mauderly, 1996). The
phenomenon of overload has been discussed in greater detail in the previous PM CD (U.S. EPA,
1996).

3.4.2. Relevance to Humans
       The relevance of "lung overload" to humans, and even to species other than laboratory
species (rats and mice and hamsters; Muhle et al., 1990a,b), is not clear. Although likely to be  of
little relevance for most "real world" ambient exposures  of humans, this phenomenon is of
concern in interpreting some long-term experimental exposure data and perhaps for human
                                           3-28

-------
occupational exposure. In addition, relevance to humans is clouded by the fact that
macrophage-mediated clearance is slower and perhaps less important in humans than in rats
(Morrow, 1994).
       Particle overload appears to be an important factor in the pulmonary carcinogenicity
observed in rats exposed to DPM.  A study by Griffis et al. (1983) demonstrated that exposure (7
h/day, 5 days/week) of rats to diluted whole DE containing DPM at concentrations of 0.15, 0.94,
or 4.1 mg/m3 for 18 mo resulted in lung burdens of 0.035, 0.220,  and 1.89 mg/g of lung tissue,
respectively. The alveolar clearance of those rats with the highest lung burden (1.89 mg/g of
lung) was impaired, as determined by a significantly greater (/XO.OOOl) retention t1/2 for DPM.
Impaired clearance was reflected in the greater lung burden/exposure concentration ratio at the
highest exposure level. Similarly, in the study by Chan et al. (1984), rats exposed for 20 h/day, 7
days/week to diluted whole DE containing DPM (6 mg/m3) for 112 days had an extraordinarily
high lung particle burden of 11.8 mg, with no alveolar particle clearance being detected over 1
year.
       Muhle  et al. (1990a,b) indicated that overloading of rat lungs occurred when lung particle
burdens reached 0.5 to 1.5 mg/g of lung tissue and that clearance mechanisms were totally
compromised at lung particle burdens > 10 mg/g for particles with a specific density close to 1,
observations that are concordant with  those of Morrow (1988).
       Pritchard (1989), utilizing data from a number of DE exposure studies, examined alveolar
clearance in rats as a function of cumulative exposure. The resulting analysis noted a significant
increase in retention t1/2 values at exposures above 10 mg/m3-h/day and also showed that normal
lung clearance mechanisms appeared to be compromised as the lung DPM burden approached
0.5 mg/g of lung.
       Animal studies have revealed that impairment of alveolar clearance can occur following
chronic exposure to DPM (Griffis et al., 1983; Wolff et al., 1987; Vostal et al., 1982; Lee et al.,
1983) or a variety of other diverse poorly soluble particles of low toxicity (Lee et al., 1986, 1988;
Ferin and Feldstein, 1978; Muhle et al., 1990). Because high lung burdens of relatively
insoluble, biochemically inert particles result in diminution of normal lung clearance kinetics or
in what is now called particle overloading, this effect appears to be more related to the mass
and/or volume of particles in the lung than to the nature of the particles per se. Particle overload
relates only to  poorly soluble particles of low toxicity. It must be noted, however, that some
types of particles may be cytotoxic and impair clearance at lower lung burdens (e.g., crystalline
silica may impair clearance at much lower lung burdens than DPM).  Regardless, as pointed out
by Morrow (1988), particle overloading in the lung modifies the dosimetry for particles in the
lung and thereby can alter toxicologic responses.
                                          3-29

-------
       Although quantitative data are limited regarding lung overload associated with impaired
alveolar clearance in humans, impairment of clearance mechanisms appears to occur, and at a
lung burden generally in the range reported to impair clearance in rats, i.e., approximately 1 mg/g
lung tissue. Stober et al. (1967), in their study of coal miners,  reported lung particle burdens of 2
to 50 mg/g lung tissue, for which estimated clearance t1/2 values were very long (4.9 years).
Freedman and Robinson (1988) also reported slower alveolar clearance rates in coal miners,
some of whom had a mild degree of pneumoconiosis. It must  be noted, however, as has been
reported even in some studies with rats exposed lifetime to overload conditions (50 mg/m3 TiO2;
Lee et al., 1986) that no lung cancer was reported among those miners with apparent particle
overload.
       Consideration of the above information further clarifies the human relevance of
noncancer effects that may be elicited from overload-type conditions in rats studies.  Under
conditions that would be most likely to elicit overload conditions in humans, such as the
excessive dust burdens in the lungs of miners, cancer is not observed although noncancer
responses such as fibrosis and macrophage responses are documented (Freedman and Robinson,
1988; Haschek and Witschi, 1991; Oberdorster, 1994).  In deliberation on the matter of whether
the rat lung nonneoplastic responses to poorly soluble particles (such as DPM) are predictive of a
similar hazard in humans, an expert panel (ILSI, 2000) opined that such responses would indeed
be a useful predictor for similar responses in humans.

3.4.3.  Potential Mechanisms for an AM Sequestration Compartment for Particles During
       Particle Overload
       Several factors may be involved in the particle-load-dependent retardations in the rate of
particle removal from the lung and the corresponding functional appearance of an abnormally
slow clearing or particle sequestration compartment. As previously mentioned, one potential  site
for particle sequestration is the  containment of particles in the  Type I cells.  Information on the
retention kinetics for particles in the Type I cells is not currently available. Also, no
morphometric analyses have been performed to date to estimate what fraction of a retained lung
burden may be contained in the Type I cell population of the lung during lung overloading.
       Another anatomical region in the lung that may be a slow clearing site is the interstitial
compartment (Kuempel, 2000).  Little is known about the kinetics of removal of free particles or
particle-containing macrophages from the interstitial spaces, or what fraction of a retained burden
of particles is contained in the lung's interstitium during particle overload. The gradual
accumulation of particles in the regional lymph nodes and the  appearance of particles and cells
with associated particles in lymphatic channels and in the peribronchial and perivascular
                                          3-30

-------
lymphoid tissue (Lee et al., 1985; White and Garg, 1981) suggest that the mobilization of
particles from interstitial sites via local lymphatics is a continual process.
       Indeed, it is clear from histologic observations of the lungs of rodents chronically exposed
to DPM that Type I cells, the interstitium, the lymphatic channels, and pulmonary lymphoid
tissues could collectively comprise subcompartments of a more generalized slow clearing
compartment.
       Although these sites must be considered potential contributors to the increased retention
of particles during particle overload, a disturbance in particle-associated AM-mediated clearance
is undoubtedly the predominant cause, inasmuch as, at least in rodents, the AMs are the primary
reservoirs of deposited particles.  The factors responsible for a failure of AMs to translocate from
the alveolar space compartment in lungs with high particulate matter burdens remain uncertain,
although a hypothesis concerning the process involving volumetric AM burden has been offered
(Morrow, 1988).
       Other processes also may be involved in preventing particle-laden AMs from leaving the
alveolar compartment under conditions of particle overload in the lung.  Clusters or aggregates of
particle-laden AMs in the alveoli are typically found in the lungs of laboratory animals that have
received large lung burdens of a variety of types of particles (Lee et al., 1985), including DPM
(White and Garg, 1981; McClellan et al., 1982). The aggregation of AMs may explain, in part,
the reduced clearance of particle-laden AM during particle overload.  The definitive
mechanism(s) responsible for this clustering of AMs has not been elucidated to date. Whatever
the underlying mechanism(s) for the AM aggregation response, it is noteworthy that AMs
lavaged from the lungs of DE-exposed animals continue to demonstrate a propensity to aggregate
(Strom, 1984). This observation could result either from the surface characteristics of AMs
being fundamentally altered or from macrophage activation by phagocytized particles that then
release chemotactic factors (Bellmann et al., 1990) in a manner that promotes their adherence to
one another in the alveolar region. AM aggregation may not simply be directly caused by their
abundant accumulation as a result of immobilization by large particle loads. Furthermore, even
though overloaded macrophages may redistribute particle burden to other AMs, clearance may
remain inhibited (Lehnert, 1988). This may, in part, be because attractants from the overloaded
AMs cause aggregation of those that are not carrying a particle burden.

3.5.  BIOAVAILABILITY OF ORGANIC CONSTITUENTS PRESENT ON DIESEL
     EXHAUST PARTICLES
       Because it has been shown that DPM extract is not only mutagenic but also contains
known carcinogens, the organic fraction was originally considered to be the primary source of
carcinogenicity in animal studies. Since then, evidence has been presented that carbon black,
                                          3-31

-------
lacking an organic component, is capable of inducing lung cancer at exposure concentrations
sufficient to induce lung particle overload.  This suggested that the relatively insoluble carbon
core of the particle may be of greater importance for the pathogenic and carcinogenic processes
observed in the rat inhalation studies conducted at high exposure concentrations. (See Chapter 7
for a discussion of this issue.) However, lung cancer reported in epidemiologic studies was
associated with diesel exposure levels far below those inducing particle overload in lifetime
studies in rats. It is therefore suggested that compounds in the organic fraction of DPM may
have some role in the etiology of human lung cancers. This leads to an interest in characterizing
the bioavailability of organics.
       The bioavailability of toxic organic compounds adsorbed to DPM can be influenced by a
variety of factors. Although the agent may be active while present on the particle, most particles
are taken up by AMs, a cell type not generally considered to be a target site. In order to reach the
target site, elution from the particle surface is necessary followed by diffusion and uptake by the
target cell. Metabolism to an active form by either the phagocytes or the target cells is also
required for  activity of many of the compounds present.
       This  section describes only the various manner and mechanisms by which organics
adsorbed onto DPM may become bioavailable.  In vivo and in vitro results involving various
biological extraction media as well as modeled scenarios of bioavailability are presented. Actual
estimates of the amount of organics from DPM to which respiratory tract tissues may be exposed
are discussed and presented in Section 3.6.2.7.

3.5.1. In Vivo Studies
3.5.1.1. Laboratory Investigations
       Several studies reported on the retention  of particle-adsorbed organics following
administration to various rodent species. In studies reported by Sun et al. (1982, 1984) and Bond
et al. (1986), labeled organics were deposited on DPM following heating to vaporize away the
organics originally present. Sun et al. (1982) compared the disposition of either pure or diesel
particle-adsorbed benzo[a]pyrene (B[a]P) following nose-only inhalation by F344 rats. About
50% of particle-adsorbed B[a]P was cleared with a half-time of 1 h, predominantly by
mucociliary  clearance.  The long-term retention of particle-adsorbed 3H-B[a]P at 18 days was
approximately 230-fold greater than that for pure 3H-B[a]P (Sun et al., 1982). At the end of
exposure, about  15% of the 3H label was found in blood, liver, and kidney. Similar results were
reported in a companion study by Bond et al. (1986), and by Sun et al. (1984) with another PAH,
1-nitropyrene, except the retention half-time was 36 days.
       Ball and King (1985) studied the disposition and metabolism of intratracheally instilled
14C-labeled 1-NP (>99.9% purity) coated onto DPM. About 50% of the 14C was excreted within
                                          3-32

-------
the first 24 h; 20% to 30% of this appeared in the urine, and 40% to 60% was excreted in the
feces. Traces of radiolabel were detected in the trachea and esophagus.  Five percent to 12% of
the radiolabel in the lung co-purified with the protein fraction, indicating some protein binding.
The corresponding DNA fraction contained no 14C above background levels.
       Bevan and Ruggio (1991) assessed the bioavailability of B[a]P adsorbed to DPM from a
5.7-L Oldsmobile diesel engine. In this study, exhaust particles containing 1.03 |_ig B[a]P/g
particles were supplemented with exogenous 3H-B[a]P to provide 2.62 |_ig B[a]P/g of exhaust
particles.  In vitro analysis indicated that the supplemented B[a]P eluted from the particles at the
same rate  as the original B[a]P.  Twenty-four hours after intratracheal instillation in Sprague-
Dawley rats, 68.5%  of the radiolabel remained in the lungs. This is approximately a 3.5-fold
greater proportion than that reported by Sun et al. (1984), possibly because smaller amounts of
B[a]P adsorbed on the particles resulted in  stronger binding or possibly because of differences
between inhalation exposure and intratracheal exposure. At 3 days following administration,
more than 50% of the radioactivity remained in  the lungs, nearly 30% had been excreted into the
feces, and the remainder was distributed throughout the body. Experiments using rats with
cannulated bile ducts showed that approximately 10% of the administered radioactivity appeared
in the bile over a 10-h period and that less than 5% of the radioactivity entered the feces via
mucociliary transport. Results of these studies showed that when organics are adsorbed to DPM
the retention of organics in the lungs is increased considerably.  Because retention time is very
short following exposure to pure compounds not bound to particles, it can be concluded that the
increased retention time is primarily the result of continued binding to DPM.  The detection of
labeled compounds in blood, systemic organs, urine, and bile as well as the trachea, however,
provides evidence that at least some of the organics are eluted from the particles following
deposition in the lungs and would not be available as a carcinogenic dose to the lung. As
discussed  above, the results of Gerde (1999a,b)  indicate that most of the organics eluted from
particles deposited in the alveolar region, especially PAHs, are  predicted to rapidly enter the
bloodstream and thus not to contribute to potential induction of lung cancer.

3.5.1.2. Studies in Occupationally Exposed Humans
       DNA adducts in the lungs of experimental animals exposed to DE  have been measured in
a number of animal experiments (World Health Organization, 1996). Such studies, however,
provide limited information regarding bioavailability of organics, as positive results may well
have been related to factors associated with lung particle overload, a circumstance reported by
Bond et al. (1990), who found carbon black, a substance virtually devoid of organics, to induce
DNA adducts in rats at lung overload doses. These authors showed that levels of DNA adducts
present in  pulmonary type n cells from the lungs of rats (n=15) exposed to equivalent conditions
                                          3-33

-------
of either carbon black or DE (each at 6.2 mg/m3) were nearly the same and 4- to 5-fold more than
air-exposed controls. This similarity was noted despite a difference of nearly three orders of
magnitude in solvent-extractable organic content between DE (30%) and carbon black (0.04%).
None of the DE or carbon black adducts comigrated with B[a]P diol epoxide.
       On the other hand, DNA adduct formation and/or mutations in blood cells following
exposure to DPM, especially at levels insufficient to induce lung overload, can be presumed to be
the result of organics diffusing into the blood.  Hemminki et al. (1994) reported increased levels
of DNA adducts in lymphocytes of bus maintenance and truck terminal workers.  Osterholm
et al. (1995) studied mutations at the hprt-locus of T-lymphocytes in bus maintenance workers.
Although they were unable to identify clear-cut exposure-related differences in types of
mutations, adduct formation was significantly increased in the exposed workers.  Nielsen et al.
(1996) reported significantly increased levels of lymphocyte DNA adducts, hydroxyvaline
adducts in hemoglobin, and 1-hydroxypyrene in urine of garage workers exposed to DE.

3.5.2. In Vitro Studies
3.5.2.1. Extraction of Diesel Particle-Associated Organics by Biological Fluids
       In vitro extraction of mutagenic organics by biological fluids can be estimated by
measurement of mutagenic activity in the particular fluid. Using this approach, Brooks et al.
(1981) reported extraction efficiencies of only  3% to 10% that of dichloromethane following
DPM incubation in lavage fluid, serum, saline, albumin, or dipalmitoyl lecithin. Moreover,
extraction efficiency did not increase with incubation time up to 120 h. Similar findings were
reported by King et al. (1981), who also reported that lung lavage fluid and lung cytosol fluid
extracts of DPM were not mutagenic. Serum extracts of DPM did exhibit some mutagenic
activity, but considerably less than that of organic solvent extracts. Furthermore, the mutagenic
activity of the solvent extract was significantly reduced when combined with serum or lung
cytosol fluid, suggesting protein binding or biotransformation of the mutagenic components.
Siak et al. (1980) assessed the mutagenicity of material extracted from DPM by bovine serum
albumin in solution, simulated lung surfactant, fetal calf serum (PCS), and physiological saline.
Only PCS was found to extract some mutagenic activity from the DPM. Keane et al. (1991),
however, reported positive effects for mutagenicity in salmonella and sister chromatid exchange
in V79 cells exclusively in the supernatant fraction of DPM dispersed in aqueous mixtures of
dipalmitoyl phosphatidyl choline, a major component of pulmonary surfactant, indicating that
pulmonary surfactant components can extract active components of DPM and result in
bioavailability.
       The ability of biological fluids to extract organics in vitro and their effectiveness in vivo
remains equivocal because of the character of the particular fluid.  For example, extracellular
                                          3-34

-------
lung fluid is a complex mixture of constituents that undoubtedly have a broad range of
hydrophobicity (George and Hook, 1984; Wright and Clements, 1987), which is fundamentally
different from  serum in terms of chemical composition (Gurley et al., 1988). Moreover,
assessments of the ability of lavage fluids, which actually represent substantially diluted
extracellular lung fluid, to extract mutagenic activity from DPM clearly do not reflect the in vivo
condition. Finally, except under very high exposure concentrations, few particles escape
phagocytosis and possible intracellular extraction. In this respect,  Hiura et al. (1999) have
shown that whole exhaust containing DPM, but not carbon black or diesel particles devoid of
organics, induces apoptosis, apparently through generation of oxygen radicals. This study
implicates organic compounds present on DPM. It also indicates the bioavailability of organics
for generation  of radicals from reaction with particle-associated organics or following elution
from DPM.

3.5.2.2.  Extraction of DPM-Associated Organics by Lung Cells and Cellular Components
      A more likely means by which organics may be extracted from DPM and metabolized in
the lung is either through particle dissolution or extraction of organics from the particle surface
within the phagolysosomes of AMs and other lung cells. This mechanism presupposes that the
particles are internalized. Specific details about the physicochemical conditions of the
intraphagolysosomal environment, where particle dissolution in AMs presumably occurs in vivo,
have not been well characterized. It is known that phagolysosomes constitute an acidic (pH 4 to
5) compartment in macrophages (Nilsen et al., 1988; Ohkuma  and Poole, 1978). The relatively
low pH in the phagolysosomes has been associated with the dissolution of some types of
inorganic particles (some metals) by macrophages (Marafante  et al., 1987; Lundborg et al.,
1984), but few studies provide quantitative information concerning how organics from DPM may
be extracted in the phagolysosomes (Bond et al., 1983). Whatever the mechanism, assuming
elution occurs, the end result is a prolonged exposure of the respiratory epithelium to DPM
organics, which include low concentrations of carcinogenic agents such as PAH.
      Early studies by King et al. (1981) found that when pulmonary alveolar macrophages
were incubated with DPM, amounts of organic compounds and mutagenic activity decreased
measurably from the amount originally associated with the particles, suggesting that organics
were removed  from the phagocytized particles.  Leung et al. (1988) studied  the ability of rat lung
and liver microsomes to facilitate transfer and metabolism of B[a]P from diesel particles.  14C-
B[a]P coated diesel particles, previously extracted to remove the original organics, were
incubated directly with liver or lung microsomes.  About 3% of the particle-adsorbed B[a]P was
transferred to the lung microsomes within 2 h.  Of this amount about 1.5% was metabolized, for
a total of about 0.05% of the B[a]P originally adsorbed to the DPM. Although transformation is
                                          3-35

-------
slow, the long retention of particles, including DPM, in humans may cause the fraction eluted
and metabolized to be considerably higher than this figure.
       In analyzing phagolysosomal dissolution of various ions from particles in the lungs of
Syrian golden hamsters, however, Godleski et al. (1988) demonstrated that solubilization did not
necessarily result in clearance of the ions (and therefore general bioavailability) in that binding of
the solubilized components to cellular and extracellular structures occurred. It is reasonable to
assume that phagocytized DPM particles may be subject to similar processes and that these
processes would be important in determining the rate of bioavailability of the particle-bound
constituents of DPM.
       Alveolar macrophages or macrophage cell lines that were exposed to high concentrations
of DPM in vitro were observed to undergo apoptosis, which was attributed to the generation of
reactive oxygen radicals (ROR) (Hiura et al. 1999). Further experimentation showed that DPM
with the organic constituents extracted was no longer able to induce apoptosis or generate ROR.
The organic extracts alone, however, were able to induce apoptosis as well as the formation of
stress-activated protein kinases that play definitive roles in cellular apoptotic pathways.  The
injurious effects of nonextracted DPM or of DPM extracts were observed to be reversible by the
antioxidant radical scavenger N-acetyl cysteine. These data suggest strongly that, at least at high
concentrations of DPM, the organic constituents contained on DPM play a central role in cellular
toxicity and that this toxicity may  be attributable to the generation of ROR.

3.5.3.  Modeling Studies
       Gerde et al. (1991a,b) described a model simulating the effect of particle aggregation and
PAH content on the rate of PAH release in the lung.  According to this model, particle
aggregation will occur with high exposure concentrations, resulting in a slow release of PAHs
and prolonged exposure to surrounding tissues.  However, large aggregates of particles are
unlikely to form at doses typical of human exposures. Inhaled particles, at low concentrations,
are more likely  to deposit and react with surrounding lung medium without interference from
other particles.  The model predicts that under low-dose exposure conditions, more typical in
humans, particle-associated organics will be released more rapidly from the particles because
they are not aggregated.  Output from this model suggests strongly that sustained exposure of
target tissues to PAHs will result from repeated exposures, not from increased retention due to
association of PAHs with carrier particles. This distinction is important because at low doses
PAH exposure and lung tumor formation would be predicted to occur at sites of deposition rather
than retention, as occurs with high doses.
       The site of release of PAHs influences effective dose to the lungs because,  as noted
previously, at least some free organic compounds deposited in the lungs are rapidly absorbed into
                                           3-36

-------
the bloodstream. Gerde et al. (1991b) predicted PAHs would be retained in the alveoli less than
1 min, whereas they may be retained in the conducting airways for hours. These predictions were
based on an average diffusion distance to capillaries of only about 0.5 pm in the alveoli, as
compared to possibly greater than 50 pm in the conducting airways such as the bronchi. An
experimental study by Gerde et al. (1999) provided support for this prediction. Beagle dogs were
exposed to  3H-B[a]P adsorbed on the carbonaceous core of DPM at a concentration of 15  |_ig
B[a]P/gm particles.  A rapidly eluting fraction from DPM deposited in the alveoli was adsorbed
into the bloodstream and metabolized in the liver, whereas the rapidly eluting fraction from DPM
deposited in the conducting airways was to a large extent retained and metabolized in situ in the
airway epithelium. Thus, organics eluting from DPM depositing in the conducting airways (i.e.,
the TB region) would have a basis for a longer residence time in the tissues (and for consequent
biological activity) than would organics eluting from DPM depositing in the pulmonary
parenchyma.  And, given the same overall deposited dose of DPM to the total pulmonary system,
a deposited dose with a higher proportion in the TB region would incur a higher probability of
tissue interactions with any eluted organics. This may be the case when comparing regional
doses of DPM to humans as compared to rats for two reasons. First, one deposition model
(Freijer et al., 1999) projects that for air concentrations of DPM at either 0.1  or 1.0 mg/m3, a
higher proportion of the total DPM dose to the pulmonary system would be deposited in the TB
area for humans at 31% (TB/Total; 0.098 / 0.318) than for rats at only 16% (0.04 / 0.205).
Second, comparative morphometry data of DPM from chronically exposed rats and primates
showed higher levels of DPM adjacent to conducting airways in primates (i.e., the interstitium of
the respiratory bronchioles) than were present in parallel regions in the rat (interstitium of the
alveolar ducts) (Nikula et al., 1997a,b). The focal nature of this deposition could give rise to
localized high concentrations of any organics eluted.

3.5.4. Summary and Unavailability
       At present, the available data are insufficient to accurately model the effective dose of
organics in the respiratory tract of humans or  animals exposed to DPM. As mentioned above,
though, the following Section (3.6.2.7) does present estimates of the actual amount of organics,
including carcinogenic PAH such as B[a]P, that are deposited in the lung and could become
bioavailable.
       Overall, the results of studies presented in Section 3.6 provide evidence that at least some
of the organic matter adsorbed to DPM deposited in the respiratory tract is eluted. The
percentage  taken up and metabolized to an active form by target cells is, however, uncertain.
Organics eluted from particles deposited in alveoli are likely to rapidly enter the bloodstream via
translocation across endothelial cells, where they may undergo metabolism by enzymes such as
                                          3-37

-------
cytochromes P-450 that are capable of producing reactive species. Organics eluted from particles
deposited in the conducting airways (the bronchioles, bronchi, and trachea) may also undergo
metabolism in other cell types such as the Clara cells with constituent or inducible cytochrome P-
450 species. Risk of harmful effects for particles deposited in the conducting airways is
predicted to be greater because solubilized organic compounds will be retained in the thicker
tissue longer, allowing  for metabolism by epithelial cells lining the airways. Furthermore, since
some deposition in conducting airways occurs primarily at bifurcations, localized higher
concentrations may occur.

3.6.  MODELING THE DEPOSITION AND CLEARANCE OF PARTICLES IN THE
     RESPIRATORY TRACT
3.6.1. Introduction
       The biological effects of inhaled particles are a function of their disposition, i.e., their
deposition  and clearance.  This, in turn, depends on their patterns of deposition (i.e., the sites
within which particles initially come into contact with airway epithelial surfaces and the amount
removed from the inhaled air at these sites) and clearance (i.e., the rates and routes by which
deposited materials are removed from the respiratory tract). Removal of deposited materials
involves the competing processes of macrophage-mediated clearance and dissolution-absorption.
Over the years, mathematical models  for predicting deposition, clearance and, ultimately,
retention of particles in the respiratory tract have been developed.  Such models help interpret
experimental  data and can be used to  make predictions of deposition for cases where data are not
available. A review of various mathematical particle deposition models was given by Morrow
and Yu (1993) and in U.S. EPA (1996).
       Currently available data for long-term inhalation exposures to poorly soluble particles
(e.g., TiO2, carbon black, and DPM) show that pulmonary retention and clearance of these
particles are not adequately described by simple first-order kinetics and a single compartment
representing the alveolar macrophage particle burden. Several investigators have developed
models for deposition, transport, and  clearance of poorly soluble particulate matter in the lungs.
All of these models identify various compartments and associated transport rates, but empirically
derived data are not available to substantiate many of the assumptions made in these models.

3.6.2. Dosimetry Models for DPM
3.6.2.1. Introduction
       The extrapolation of toxicological results from laboratory animals to humans, the goal of
this chapter, requires the use of dosimetry models for both species that include, first, the
deposition  of DPM in various regions of the respiratory tract,  and second, the transport and
                                           3-38

-------
clearance of the particles, including adsorbed constituents, from their deposited sites.  Therefore
the ideal model structure would incorporate both deposition and clearance in animals and
humans.
       Deposition of particles in the respiratory tract, as described above, can be by impaction,
sedimentation, interception, and diffusion, with the contribution from each mechanism a
function of particle size, lung structure, and size and breathing parameters. Because of the size
of diesel particles, under normal breathing conditions most of this deposition takes place by
diffusion, and the fraction of the inhaled mass that is deposited in the thoracic region (i.e., TB
plus A regions) is substantially similar for rats and humans.
       Among deposition models that include aspects of lung structure and breathing dynamics,
the most widely used have been typical-path or single-path models (Yu, 1978; Yu and Diu,
1983). The single-path models are based on an idealized symmetric geometry of the lung,
assuming regular dichotomous branching of the airways and alveolar ducts (Weibel, 1963).  They
lead to modeling the deposition in an average regional sense for a given lung depth. Although
the lower airways of the lung may be reasonably characterized by such a symmetric
representation, there are major asymmetries in the upper airways of the tracheobronchial tree that
in turn lead to different apportionment of airflow and particulate burden to the different lung
lobes.  The rat lung structure is highly asymmetric because of its monopodial nature, leading to
significant errors in a single-path description. This is rectified in the multiple-path model of the
lung, which incorporates asymmetry and heterogeneity in lung branching structure and calculates
deposition at the individual airway level. This model has been developed for the rat lung
(Anjilvel and Asgharian, 1995; Freijer et al., 1999) and, in a limited fashion because of
insufficient morphometric data, for the human lung (Subramaniam et al., 1998; Yeh and Schum,
1980). Such models are particularly relevant for fine and ultrafine particles such as occur in
DPM.  However, models for clearance have not yet been implemented in conjunction with the
use of the multiple-path model.
       Clearance of particles in the respiratory tract takes  place (1) by mechanical processes:
mucociliary transport in the ciliated conducting airways and macrophage phagocytosis and
migration in the nonciliated airways, and (2) by dissolution. The removal of material such as the
carbonaceous core of DPM is largely by mechanical clearance, whereas the clearance of the
organics adsorbed onto the carbon core is principally by dissolution.
       Several models currently exist that integrate both deposition  and clearance, some specific
for humans and others specific for laboratory animals. They differ significantly in the level of
physiological detail that is captured in the model and in the uncertainties associated with the
values of the parameters used.  All of these models identify various compartments and associated
transport rates, but empirically derived data are not available to validate many of the assumptions
                                          3-39

-------
made in the models. A review of the principal human and animal deposition/clearance models,
including candidate models for use in animal-to-human extrapolation in this assessment, are
considered below.

3.6.2.2. Human Models
       The International Commission on Radiological Protection (ICRP) recommends specific
mathematical dosimetry models as a means to calculate the mass deposition and retention by
different parts of the human respiratory tract and, if needed, tissues beyond  the respiratory tract.
The latest ICRP-recommended model, ICRP66 (1994),  considers the human respiratory tract as
four general anatomical regions: the ET region, which is divided into two subregions; the TB
region, which is also subdivided into two regions; and the gas-exchange tissues, which are
further defined as the alveolar-interstitial (AI) region but are exactly comparable to the
pulmonary or A region. The fourth region is the lymph nodes.  The deposition component of the
model for the ET, TB, and A regions is semi-empirical based on equations derived from fitting
experimental deposition data. The dimensional model used for the TB and  A regions was
adopted from several sources (Weibel, 1963; Yeh and Schum, 1980;  and Phalen et al., 1985); the
physical aspects of the individual airway generations for these regions were all averaged after
each source was  adjusted to a standard functional residual capacity.  The equations for estimating
deposition in these areas was empirical, obtained from fitting data obtained from partial human
lung casts or from theoretical calculation for these regions. Deposition in the four regions is
given as a function of particle size with two different types of particle size parameters: activity
median thermodynamic diameter (AMTD) for deposition of particles ranging in size from 0.0005
to 1.0 |_im and the activity median aerodynamic diameter (AMAD) for deposition of particles
from 0.1 to 100|_im. Reference values of regional deposition are provided and guidance is given
for extrapolating to specific individuals and populations under different levels of activity. This
model also includes consideration of particle inhalability, a measure of the degree to which
particles can enter the respiratory tract and be available for deposition.  After deposition occurs in
a given region, two different intrinsic clearance processes act competitively on the deposited
particles: particle transport, including mucociliary clearance from the respiratory tract and
physical clearance of particles to the regional lymph nodes; and absorption, including movement
of material to blood and both dissolution-absorption and transport of ultrafine particles. Rates of
particle clearance derived from studies with human subjects are assumed to be the same for all
types of particles. The  ICRP model provides average concentration or average number values on
a regional basis, i.e., mass or number deposited or retained in the ET, TB, or A regions.
Additionally, while the ICRP66 model was developed primarily for use with airborne radioactive
                                          3-40

-------
particles and gases in humans, its use for describing the dosimetry of inhaled mass of
nonradioactive substances in humans is also appropriate.
       The National Council on Radiation Protection (NCRP) has issued a human respiratory
tract dosimetry model that was developed concurrently with the ICRP model (NCRP, 1997;
Phalen et al.,1991). It addresses (1) inhalability of particles, (2) new subregions of the
respiratory tract, (3) dissolution-absorption as an important aspect of the model, and (4) body size
(and age). The proposed NCRP model defines  the respiratory tract in terms of a naso-oro-
pharyngo-laryngeal (NOPL) region, a TB region, a pulmonary (P) region, and the lung-associated
lymph nodes (LN). Like the ICRP model, the deposition component of the model for the ET
region is semi-empirical, based on equations derived from fitting experimental deposition data.
The dimensional model used for the TB and A regions was that of Yeh and Schum (1980).  The
data from this model were used to estimate physical processes along a typical lung path (vice
multiple-path; see MPPDep model description below) on a generation-by-generation basis. The
rates of dissolution-absorption of particles and their constituents are derived from clearance data
from humans and laboratory animals.  The effect of body growth  on particle deposition is also
considered in the model, although particle clearance rates are assumed to be independent of age.
The NCRP model currently available considers deposition only within these regions of the
respiratory tract. As with the ICRP model, the NCRP model can  be used for evaluating
inhalation exposures to all types of particles. Comparison of regional deposition patterns
estimated by the ICRP66 and the current NCRP models have been reported (Yeh et al., 1996).
One principal difference between the models is the enhanced deposition of ultrafines in the
tracheobronchial region predicted by the NCRP model compared with the ICRP model. This
effect of enhanced deposition is claimed to be due to the entrance configuration of an airway
bifurcation.
       The model of Freijer et al. (1999) is a multiple-path particle deposition model (MPPDep)
for the human respiratory tract that differs fundamentally from the above two models as
described in the Introduction.  Calculations from the model may be based on either single-path or
multiple-path methods for tracking air flow and calculating aerosol deposition in the lung.  The
single-path method calculates deposition for a typical path, whereas the multiple-path method is
capable of incorporating the asymmetry in lung structure and providing lobar-specific and
airway-specific information. Two options are provided for idealizing the geometry of the human
lung; one uses a symmetric geometry for the whole lung and the second option captures the
asymmetry in the lobar structure, but treats the geometry within each lung lobe in a symmetric
fashion. Both models are constructed using morphometric data compiled by Yeh and Schum
(1980). Within each airway, deposition is calculated using theoretically derived efficiencies for
deposition by diffusion (most relevant to DPM), sedimentation, and impaction within the airway

                                         3-41

-------
or airway bifurcation.  Filtration of particulate aerosols by the head is determined using empirical
efficiency functions.  The model calculates deposition of monodisperse and polydisperse aerosols
in the respiratory tract of both humans (and rats) for particles ranging from ultrafine (0.01
microns) to coarse (20 microns) sizes. Various breathing patterns may be simulated:
endotratracheal, nasal, oral, and combined nasal and oral (oronasal).  The exposure scenario may
be constant or variable. For the variable scenario, the user may specify different breathing
patterns either on an hourly basis during the day or activity patterns for variable time durations.
Adjustment for inhalability of the particle is also included as an option. The software in this
model provides results for the deposition fraction and mass deposited in the various regions of
the respiratory tract in graphical and text formats.
       The combined model of Yu et al. (1991) has a human component that will be discussed
below.

3.6.2.3. Animal Models
       Strom et al. (1988) developed a multicompartmental model for particle retention that
partitioned the alveolar region into two compartments on the basis of the physiology of clearance.
The alveolar region has a separate compartment for sequestered macrophages, corresponding to
phagocytic macrophages that are heavily laden with particles and clustered, and consequently
have significantly lowered mobility. The model has the  following compartments:
(1) tracheobronchial tree, (2) free particulate on the alveolar surface, (3) mobile phagocytic
alveolar macrophages, (4) sequestered particle-laden alveolar macrophages, (5) regional lymph
nodes, and (6) gastrointestinal tract.  The model is based on mass-dependent clearance (the rate
coefficients reflect this relationship), which dictates sequestration of particles and their eventual
transfer to the lymph nodes. The transport rates between various compartments were obtained by
fitting the calculated results to lung and lymph node burden experimental data for both exposure
and postexposure periods. Because the number of fitted parameters was large, the model is not
likely to provide unique solutions that would simulate experimental data from various sources
and for different exposure scenarios.  For the same reason, it is not readily possible to use this
model for extrapolating to humans.
       Stober and co-workers have worked extensively in developing models for estimating
retention and clearance of relatively insoluble respirable particles (as DPM) in the lung. Their
most recent work (1994), a revised version of the POCK model, is a rigorous attempt to
incorporate most of the physiologically known aspects of alveolar clearance and retention of
inhaled relatively insoluble particles.  Their multicompartmental kinetics model has five
subcompartments. The transfer of particles between any of the compartments within the alveolar
region is macrophage mediated. There are two compartments that receive particles cleared from
                                          3-42

-------
the alveolar regions: the TB tract and the lymphatic system. The macrophage pool includes both
mobile and particle-laden immobilized macrophages. The model assumes a constant maximum
volume capacity of the macrophages for particle uptake and a material-dependent critical
macrophage load that results in total loss of macrophage mobility. Sequestration of those
macrophages heavily loaded with a particle burden close to a volume load capacity is treated in a
sophisticated manner by approximating the particle load distribution in the macrophages.  The
macrophage pool is compartmentalized in terms of numbers of macrophages that are subject to
discrete particle load intervals. Upon macrophage death, the phagocytized particle is released
back to the alveolar surface; thus phagocytic particle collection competes to some extent with
this release back to the alveolar surface.  This recycled particle load is also divided into particle
clusters of size intervals defining a cluster size distribution on the alveolar surface. The model
yields a time-dependent frequency distribution of loaded macrophages that is sensitive to both
exposure and recovery periods in inhalation studies.
       The POCK model also emphasizes the importance of interstitial burden in the particle
overload phenomenon and indicates that  particle overload (Section 3.4) is a function of a massive
increase in particle burden of the interstitial space rather than total burden of the macrophage
pool.  The relevance of the increased particle burden in the interstitial space lies with the fact that
this compartmental burden is not available for macrophage-mediated clearance and, therefore,
persists even after cessation of exposure.
       Although the POCK model is the most sophisticated in the physiological complexity it
introduces, it suffers from a major disadvantage.  Experimental retention studies provide data
only on total alveolar and lymph node mass burdens of the particles as a function of time.  The
relative fraction of the deposition between the alveolar subcompartments in the Stober model
therefore cannot be obtained experimentally; the model thus uses a large number of parameters
that are simultaneously fit to experimental data.  Although the model predictions are tenable,
experimental data are not currently available to substantiate the proposed compartmental burdens
or the transfer rates associated with these compartments. Thus, overparameterization in the
model leads to the possibility that the model may not provide a unique solution that may be used
for a variety of exposure scenarios, and for the same reason, cannot be used for extrapolation to
humans.  Stober et al. have not developed an equivalent model for humans; therefore the use of
their model in our risk assessment for diesel is not attempted.

3.6.2.4. Combined Models (for interspecies extrapolation)
       Currently available data for long-term inhalation exposures to poorly soluble particles
(e.g., TiO2, carbon black, and DPM) show that pulmonary retention and clearance of these
particles are not adequately described by simple first-order kinetics and a single compartment
                                          3-43

-------
representing the alveolar macrophage particle burden.  A two-compartment lung model that
could be applied to both humans and animals was developed by Smith (1985) and includes
alveolar and interstitial compartments.  For uptake and clearance of particles by alveolar surface
macrophages and interstitial encapsulation of particles (i.e., quartz dust), available experimental
data show that the rate-controlling functions  followed Michaelis-Menton type kinetics, whereas
other processes affecting particle transfer are assumed to be linear. The model was used in an
attempt to estimate interstitial dust and fibrosis levels among a group of 171  silicon carbide
workers; the levels were then compared with evidence of fibrosis from chest radiographs.  A
significant correlation was found between estimated fibrosis and profusion of opacities on the
radiographs. This model provides as many as seven different rate constants derived by various
estimations and under various conditions from both animal and human sources.  The model was
intended for estimation of generalized dust described only as respirable without any other regard
to sizing for establishing the various particle-related rate constants. As most of the described
functions could not be validated with experimental data, the applicability of this model,
especially for particulates in the size range of DPM, was unclear.
       Yu et al. (1991; also reported as Yu and Yoon,  1990) have developed a three-
compartment lung model that consists of tracheobronchial (T), alveolar (A),  and lymph node (L)
compartments  (Appendix A, Figure A-l) and, in addition, considered filtration  by a
nasopharyngeal or head (H) compartment. The tracheobronchial compartment is important for
short-term considerations, whereas long-term clearance takes place via the alveolar compartment.
In contrast to the Stober and Strom approaches, the macrophage compartment in the Yu model
contains all of the phagocytized particles; that is,  there is no  separate (and hypothetical)
sequestered macrophage subcompartment. Absorption by the blood (B) and  gastrointestinal (G)
compartments  was also considered. Although the treatment  of alveolar clearance is
physiologically less sophisticated than that of the Stober et al. model, the Yu model provides a
more comprehensive treatment of clearance by including systemic compartments and the head,
and including the clearance of the organic components of DPM in addition to the relatively
insoluble carbon core.
       In order to progress beyond the classical human ICRP66 retention model, Yu has
addressed the impairment of long-term clearance (the overload effect) by using a set of variable
transport rates  for clearance from the alveolar region as a function of the mass of DPM in the
alveolar compartment.  A functional relationship  for this was derived mathematically (Yu et al.,
1989) based upon Morrow's hypothesis for the macrophage overload effect discussed earlier in
the section on pulmonary overload.  The extent of the impairment depends on the initial particle
burden, with greater particulate concentration leading to slower clearance.
                                          3-44

-------
       Within this model, DPM is treated as being composed of three material components: a
relatively insoluble carbonaceous core, slowly cleared organics (10% particle mass), and fast-
cleared organics (10% particle mass).  Such a partitioning of organics was based on observations
that the retention of particle-associated organics in lungs shows a biphasic decay curve (Sun et
al., 1984; Bond et al., 1986).  For any compartment, each of these components has a different
transport rate.  The total alveolar clearance rate of each material component is the sum of
clearance rates of that material from the alveolar to the tracheobronchial, lymph, and blood
compartments.  In the Strom and Stober models discussed above, the clearance kinetics of DPM
were assumed to be entirely dictated by those of the relatively insoluble carbonaceous core. For
those organic compounds that become dissociated from  the carbon core, clearance rates are likely
to be very different, and some of these compounds may  be metabolized in the pulmonary tissue
or be absorbed by blood.
       The transport rates for the three components were derived from experimental data for rats
using several approximations. The transport rates for the carbonaceous core and the two organic
components were derived by fitting to data from separate experiments.  Lung and lymph node
burdens from the experiment of Strom et  al. (1988) were used to determine the transport rate of
the carbonaceous core. The Yu model incorporates the impairment of clearance by including a
mass dependency in the transport rate.  This mass dependency is easily extracted because the
animals in the experiment were sacrificed over varying periods following the end of exposure.
       It was assumed that the transport rates from the alveolar and lymph compartments to the
blood were equal and independent of the paniculate mass in the alveolar region.  The clearance
rates of particle-associated organics for rats were derived from the retention data of Sun et al.
(1984) for B[a]P and the data of Bond et al. (1986) for nitropyrene adsorbed on diesel particles.
       In their model Yu et al. (1991) make two important assumptions to carry out the
extrapolation in consideration of inadequate human data. First, the transport rates of organics in
the DPM do not change across species. This is based upon lung clearance data of inhaled
lipophilic compounds (Schanker et al.,  1986), where the clearance was seen to be dependent on
the lipid/water partition coefficient. In contrast, the transport rate of the carbonaceous core is
considered to be significantly species dependent (Bailey et al., 1982). DPM clearance rate is
determined by two terms in the model (see Equation  A-82 in Appendix A). The first,
corresponding to macrophage-mediated clearance, is a function of the lung burden and is
assumed to vary significantly across species. The second term, a constant, corresponding to
clearance by dissolution, is assumed to be species independent.  The mass-dependent term for
humans is assumed to vary in the same proportion as in  rats under the same unit surface
particulate dose. The extrapolation is then achieved by using the data of Bailey et al. (1982) for
the low lung burden limit of the clearance rate.  This value of 0.0017/day was lower (i.e., slower)
                                          3-45

-------
than the rat value by a factor of 7.6. This is elaborated further in Appendix A. Other transport
rates that have lung burden dependence are extrapolated in the same manner.
       It should be noted that the Bailey et al. (1982) experiment in humans used fused
monodisperse aluminosilicate particles of 1.9 and 6.1 pm aerodynamic diameters. Yu and co-
workers have used the longer of the half-times observed in this experiment to obtain an alveolar
human clearance rate (A), of 0.00169/day.  In using such data for DPM 0.2 |_im in diameter, they
have assumed the clearance of relatively insoluble particles to be independent of size over the
range in diameter from 6.1 down to 0.2 |_im. This assumption is consistent with observations and
views currently in the literature indicating that clearance mechanisms are not particularly
particle-size dependent (Morrow et al., 1967a,b; Snipes et al., 1983). That the linear dimensions
of an alveolar macrophage, considered to be the principal means of clearance in the A region, are
significantly larger, roughly 10 |_im (Yu et al., 1996), and could therefore accommodate
engulfment of a range of particle sizes also makes this assumption reasonable. Snipes (1979),
however, has reported in rats a A (converted here from half-time values) of 0.0022/day for 1 and
2 |_im particles but a higher value of 0.0039/day for 0.4 |_im particles indicating that clearance
rates may indeed depend on size.  In the absence of reliable data for 0.2 pm particles, the slower
clearance rate pertaining to this larger particle size, i.e., 0.00169/day, is being used.  Such a
choice may underestimate the actual DPM clearance rate in humans. The resulting model output
(i.e., lung DPM burden)  from this slower rate would predict more DPM in the alveolar space
than may actually be present at any given time. Therefore, use of this slower A may be
considered to be more protective of human health. Long-term clearance rates for particle  sizes
more comparable to DPM are available, e.g., iron oxide and polystyrene spheres (Waite and
Ramsden, 1971; Jammet et al., 1978), but these data show a large range in the values obtained
for half-lives or are based upon a very small number of trials, and therefore compare unfavorably
with the quality of data from the Bailey experiment.
       The deposition fractions of particulate matter in the pulmonary  and tracheobronchial
regions of the human lung remain relatively unchanged over the particle size range between
0.2 and 1.0 [am, on the basis of analysis done with the ICRP model (ICRP66, 1994).  As the
clearance of relatively insoluble particles is also likely to remain the same over this range, the
dosimetry results in this  report for the carbonaceous core component of DPM could also be
extended to other particles in this size range within the PM2 5 For respirable particles with
diameters larger than this range, e.g., between 1.0 and 3.5 pm, the extent of the fraction deposited
in the pulmonary region  is unclear.  Results from the ICRP66 (1994) model predict little change
in human deposition for  this diameter range, whereas the earlier model of Yu  and Diu (1983)
predicts a significant  increase  as reported in ICRP66 (1994). It is therefore unclear if either
model would be applicable for particles in this larger-sized range without changing the value for
                                          3-46

-------
the deposition fractions.  As will be presented and discussed below, regional deposition fractions
of DPM-sized particles from the MPPDep, the ICRP66 (1994) and draft NCRP models compare
favorably with the human alveolar deposition in humans specific for DPM, which has been
estimated with the Yu model to be 7% to 13% (Yu and Xu, 1986).
       Although there was good agreement between experimental and modeled results, this
agreement follows a circular logic (as adequately pointed out by Yu and Yoon [1990]) because
the same experimental data that figured into the derivation of transport rates were used in the
model. Nevertheless, even though this agreement is not a validation, it provides an important
consistency check on the model. Further experimental data and policy definitions on what
constitutes validation would be necessary for a more formal validation.
       The model showed that at low lung burdens, alveolar clearance is dominated by
mucociliary transport to the tracheobronchial region, and at high lung burdens, clearance is
dominated by transport to the lymphatic system. The head and tracheobronchial compartments
showed quick clearance of DPM by mucociliary transport and dissolution.  Lung burdens of both
the carbonaceous core and organics were found to be greater in humans than in rats for similar
periods of exposure.
       The Yu and Yoon (1990) version of the model provides a parametric study of the
dosimetry model, examining variation over a range of exposure concentrations, breathing
scenarios, and ventilation parameters; particle mass median aerodynamic diameters; and
geometric standard deviations of the aerosol size distribution. It examines how lung burden
varies with age for exposure over a lifespan, provides dosimetry extrapolations to children, and
examines changes in lung burden with lung volume. The results showed that children would
exhibit more diminished alveolar clearance of DPM at high lung burden than adults when
exposed to equal concentrations of DPM.  These features make the model easy to use in risk
assessment studies.  The reader is referred to Appendix A for further details on the model and for
analyses of the sensitivity of the model to change in parameter values.
       The Yu model presents some uncertainties in addition to those discussed earlier in the
context of particle size dependence of clearance rate. The reports of Yu and Yoon (1990) as well
as Yu et al. (1991) underwent extensive peer review; we list below the most important among the
model uncertainties discussed by the review panel. The experimental data used by the Yu model
for adsorbed organics used passively adsorbed radiolabeled compounds as surrogates for
combustion-derived organics.  These compounds may adhere differently to the carbon core than
do those formed during combustion. Yu has estimated that slowly cleared organics represent
10% of the total particle mass; the actual figure could be substantially less; the reviewers
estimate that the amount of tightly bound organics is probably only 0.1% to 0.25% of the particle
mass.
                                         3-47

-------
       The model was based upon the experimental data of Strom et al. (1988), where
Fischer-344 rats were exposed to DPM at a concentration of 6.0 mg/m3 for 20 h/day and 7
days/week for periods ranging from 3 to 84 days. Such exposures lead to particle overload
effects in rats, whereas human exposure patterns are usually to much lower levels at which
overload will not occur. Parameters obtained by fitting to data under the conditions of the
experimental scenario for rats may not be optimal for the human exposure and concentration of
interest.
       The extrapolation of retained dose from rats to humans assumed that the macrophage-
mediated mechanical clearance of the DPM varies with the specific particulate dose to the
alveolar surface in the same proportion in humans and in rats, whereas clearance rates by
dissolution were assumed to be invariant across  species.  These assumptions have not been
validated.
       It should also be noted that the Yu et al. (1991) model does not possess a formal
interstitial compartment although the lymph nodes, which would be the repository of particles
from the interstitium, are represented.  The work of Nikula et al. (1997a,b) and of Kuempel
(2000) provide compelling information on the significance of an extensive interstitilization
process in primates and in humans. Kuempel (2000) developed a lung dosimetry model to
describe the kinetics of particle clearance and retention in coal miners' lungs. Models with
overloading of lung clearance, as observed in rodent studies, were found to be inadequate to
describe the end-of-life lung dust burdens in those miners. The model that provided the best fit
to the human data included a sequestration process representing the transfer of particles to the
interstitium. These findings are consistent with  a study showing reduced lung clearance of
particles in retired coal miners (Freedman and Robinson, 1988) and with studies showing
increased retention of particles in the lung interstitium of humans and nonhuman primates
compared to rodents exposed to coal dust and/or DE (Nikula et al., 1997a,b). These findings are
also consistent with the established observation that humans and primates clear particles slowly
from the alveolar interstitium compared with rates in rodent species  such as rats and mice (Hsieh
and Yu, 1998). Because several aspects of the Yu model have not been validated on human data
and because it does not include a formal interstitial compartment, it is acknowledged that this
model may therefore have some uncertainty concerning the lung burdens in humans exposed to
occupational levels of dust. However, it is also not known whether the model based on coal
miner data (Kuempel, 2000) would also describe the clearance and retention processes in the
lungs of humans with exposures to particles at lower environmental  concentrations, or to
submicrometer particles such as DE parti culate.  Further investigation of these issues is needed.
                                          3-48

-------
3.6.2.5. Use of the Yu et al (1991) Model for Interspecies Extrapolation
       In addressing the objectives of this chapter, i.e., consideration of what is known and
applicable to DPM concerning particle disposition and the bioavailability of adsorbed organics
on DPM, it is apparent that the database is considerable for both the processes involved in
particle dosimetry and for DPM.  This information makes the goal of predicting a human internal
dose from animal data through a model utilizing this database both feasible and appropriate.
       In their charge to EPA through "Science and Judgment in Risk Assessment" (NRC,
1995), the National Research Council opines that EPA should have principles for judging when
and how to depart from default options. The extensive data presented in this chapter their
scientific validity, and the limitations of the current default procedures provide a basis for
departing from the default options currently identified by the Agency for extrapolating from
animals to humans.  The default option of assuming external concentrations of DPM in animal
studies as being representative of a human concentration (and an equivalent internal dose) is
clearly not adequate given the differences in the basic processes of deposition and clearance
between animals and humans documented by these data. Use of an alternate default option, the
Agency's dosimetric adjustment procedures for inhaled particles in animal-to-human scenarios
(described in U.S. EPA, 1994), is also inadequate as only deposition is predicted and then only
down to an MMAD of 0.5 |_im, whereas the MMAD of DPM is typically 0.2 |_im or smaller.
Models have been described in this section that consider both deposition and retention
specifically for DPM in both laboratory animals and in humans. These points provide
justification for moving away from default options and utilizing the best scientific information
available (i.e., that integrated into deposition/clearance models) in performing the animal-to-
human extrapolation.
       Evaluation of the various models discussed in this chapter should be considered from the
aspect of both the rat and the human.  For rats it is fairly clear that the rat portion of the model of
Yu et al. (1991) is the most appropriate because it is based on data, especially extensive
information on lung burdens, from actual DPM exposures. The model provides for both
deposition and integrated clearance for DPM as well as for two classes of adsorbed organics.
The transport rates in the Yu model are derived directly from experiments with DPM exposed
rats.
       For humans, however, several models are available and discussed above, none of which is
based on DPM-specific data. Deposition, but not clearance, modules are available for all  models,
and Table 3-3  is an attempt to compare deposition projections of the various models to the extent
possible for particles in the range of characteristics of size, distribution, and density of DPM.
Intake parameters such as breathing rates and minute volumes were also matched among the
various models.  As alluded to above and shown in Table 3-3, DPM deposition is predicted to
                                          3-49

-------
Table 3-3. Model comparison for deposition of DPM under equivalent conditions
Compartment Yua
A (model
designation)
ET (model
designation)
TB (model
designation)
Total
13% (A)
8% (H)
8% (TB)
29%
ICRP66"
14.1%(AI)
6% (ETj + ET2)
4% (BB + bb)
24.1%
MPPDepl.ir
16.6% (P)
8.7% (H)
7.2% (TB)
32.5%
NCRP"
17.3% (P)
6.6% (NOPL)
6.2% (TB)
30.1%
aYu and Xu, 1987 (estimated from Figures 1 and 3).
bJarvisetal., 1996.
0 Freijer et al,  1999 (The Yeh-Schum 5-lobe and URT volume of 50 mL options were used.)
dNCRP, 1997.
Note: Particle characteristics were set at 0.2 MMAD, 2.4 sigma g, 1.5 shape factor (equivalent to 0.3 packing
factor), density 1.5 and a concentration of 5 ng/m3. Lung parameters were set at 15 breaths per minute, a tidal
volume of 0.926 L/hr, and a functional residual capacity (FRC) of 3300 mL.

occur in all regions of the respiratory tract but, because diffusion would be the most likely
mechanism of deposition, is most prominent in the alveolar region.  When run under equivalent
conditions, all models show that higher deposition in the alveolar region is higher, generally by a
factor of about 2, than the other regions of the respiratory tract.  The percentages projected by the
different models to be deposited in the alveolar regions were all similar to one another with a
range of only 13% for the Yu model to  17.3 % for the NCRP model. The total deposition of
DPM-like particles predicted by the models was also very similar at around 30%.  Only the ICRP
model differed appreciably from the others in total deposition by a factor of about 1.3 less at
22.9%.  Due to its verity and completeness in representation of the lung, the MPPDep model
could be considered the most theoretically advanced of these deposition models and, presumably,
the most accurate. It can be seen that, at least at the concentration tested, the Yu results and
those of the  MPPDep model could be judged very similar if not the  same in the ET and TB
regions, albeit with the MPPDep predicting slightly more deposition in the A region.  Based on
this limited analysis, total and regional DPM deposition in the human respiratory tract predicted
by the Yu model appear similar to other available human models.
       Further model comparison may be undertaken for those human  models that have
clearance as well as deposition modules available; from Table 3-3, these include the Yu et al.
(1991) and ICRP66 models. Therefore, the human lung burden outputs of these two models were
compared under equivalent physiological parameters, particle characteristics, and duration (70
years) and concentrations of exposure (Figure 3-9).
                                           3-50

-------
             8000
             7000  -
        o)   6000  -
        =i   5000  -
        •|   4000  -
        Q   3000  H
        D)
        E   2000  -
             1000  -
                 0
                    0
1
                                     234
                             mg DPM / m3
Figure 3-9. Modeled estimates of lung burden in humans after a
simulated lifetime exposure to DPM using the Yu et al. (1991; [o]) and
ICRP66 (D) models.  Simulations include both deposition and
clearance. Simulations were run for 70 years using a respiratory
frequency of 15 min ~l and a tidal volume of 0.926 L/breath for a total
daily air intake of 20 m3/day for the various concentrations shown.
Particle characteristics in the ICRP66 model, including MMAD, og,
density, and packing/shape factor were all matched to those used in
the Yu model.
       At DPM concentrations up to about 0.2 mg/m3, the outputs (lung burden) from these two
models are essentially identical (see insert) indicating little if any difference between them in this
concentration range.  This observation is consonant with the minor differences noted in
deposition (Table 3-3).
      Above 0.2 mg/m3 DPM, both models continue to demonstrate a monotonic increase in
lung burden with increasing concentration. However, the output of the Yu et al. (1991) model
begins to diverge markedly from the burdens predicted by  the ICRP model such that the Yu
model predicts a greater burden for a given concentration of DPM  than does the ICRP66 model.
This situation would be predicted based on the assumption in the human portion of the Yu model
of a concentration-dependent macrophage inhibition and particle overload occurring in humans;
such an inhibition would result in impaired clearance processes, thereby allowing for a greater
accumulation of material in the lung with increasing concentrations of DPM. This assumption is
not made in the ICRP model, and materials are therefore not predicted to accumulate in the lung
to the extent predicted by the Yu model.
                                        3-51

-------
       Based on this limited analysis of models and the predictions from them for both
deposition and clearance of DPM in humans, the model of Yu et al. (1991) can be seen to
perform similarly to other available state-of-the-art models.  The Yu model(s) are chosen for
further analysis for the purposes of this document primarily because the animal portion of the
model is based on DPM-specific data and the human components of the model have both
deposition and clearance capacities that do not appear different from other available human
respiratory tract models.

3.6.2.6. Model Variability
       As demonstrated in Table 3-3 and Figure 3-9, there appears to be little variability among
state-of-the-art models available for predicting disposition (both for deposition and for clearance
integrated with deposition) of low levels  of DPM (i.e., up to about 200 i-ig/m3) in the respiratory
tracts of humans.
       Intersubject variability and its relationship to model output, however, is acknowledged in
the ICRP model for deposition efficiencies (ICRP, 1994). This variability, recognized as
substantial by ICRP, is addressed through use of scaling constants derived from estimates of the
upper and lower confidence bounds for regional deposition efficiencies, with the scaling
constants representing the variability in the population. It should be noted that the same
philosophy is inherent in dose-response methodologies such as the RfC, where variability in the
population is accommodated by a 10-fold uncertainty factor rather than by scaling constants.
Inspection of data in ICRP66 (e.g., Figures D-4 through D-7 in the ICRP reference) on nasal and
extrathoracic deposition in adult males shows that these upper and lower boundaries on output
due to intersubject variability are considerably less than 10-fold different from one another.
Thus, dividing model outputs by a factor of 10 such as is done in RfC derivation may well be
inclusive of not only intersubject variability but also of any model-to-model variability as they
exist currently.

3.6.2.7. Model Comparison — Estimations of Deposition of Adsorbed Organics
       The data presented in Table 3-3 may be viewed as single-breath estimates of DPM
deposition patterns in the various regions of the human lung under the breathing patterns and
conditions described in the table for the different models considered in this report. From these
data it is possible to estimate the total mass of DPM deposited in the pulmonary region under a
given set of conditions. Furthermore, if the fraction of organics present on DPM and their ability
to be desorbed or eluted from the DPM are assumed also to be the same, then these deposition
data could be used to estimate the dose of organics to pulmonary tissues. Such a comparison
would not only yield an estimate of the amount of organics but also lend a further comparison
                                          3-52

-------
between the different human models. This exercise was performed for humans breathing 5 |_ig
DPM/m3 continuously, and the results are presented in Table 3-4 below.


Table 3-4.  Comparative model estimates of DPM deposition in human lungs from exposure
to 5 |J,g/m3 continuously for one year

Human
deposition
model
Yuetal. (1991)
ICRP66
MPPDep
NCRP
A
Alveolar
Depa
13%
14.1%
16.6%
17.3%
B
pg DPM
deposited/year13
4745
5147
6059
6315
C
l-ig organics
deposited/year0
598
649
763
796
D
l-ig carcinogenic PAH
deposited/year11
1.82
1.98
2.33
2.43
aAlveolar deposition fractions predicted for DPM (Yu et al, 1991) and for particles with DPM characteristics (from
Table 3-3). No clearance is included in this calculation.
bA total air intake of 20 mVday is assumed. These numbers were obtained by factoring 20 m3 x 5
Alveolar deposition % (column A)  x 365 days/year.
                                                                                 DPM/m3
cln three samples of DPM extract, DPM-associated organics were noted as being 11.1%, 14.7%, and 12. 1% wt.
organics/wt. DPM (Tong and Karasek, 1984) with the average being 12.6%; column B is factored by this average to
generate column C.

dThose seven PAHs identified as being carcinogenic either to humans or to animals (U.S. EPA, 1993) were summed
from the data of Tong and Karasek (1984), where they are reported as a concentration in extract from DPM-
associated organics. In three different samples, the content of these 7 PAHs was noted as 4739, 2054, and 2360
ng/mg of organic extract, with the average being 3051 ng/mg (3.051 ng/mg) organic extract.  This average value
was factored with Column C (in mg) to generate column D.

Note: Estimates from different human deposition models of the total amount of DPM-associated organics deposited
in the pulmonary regions in humans breathing DPM at 5 ng/m3 continuously for 1 year.
       As may be expected, the relatively minor differences (17.3 % / 13% = 1.3) in the
deposition of DPM among the different human models leads to similarly minor differences in
projections of dose of carcinogenic PAHs to the lung at a relatively low concentration of 5 |_ig/m3
DPM. Somewhat unexpected is the small absolute quantity of carcinogenic PAH that may be
delivered to the lung tissues under the conditions of exposure to DPM in this exercise. It should
be noted that exercises similar to this have been carried out by others, e.g., Valberg and Watson
(1999). However, the possibility that high concentrations of DPM may result in localized areas
of deposition (such as the conducting airways), the fact that human exposures may be
                                             3-53

-------
considerably greater than those presupposed in the exercise (e.g., 5 i-ig/m3), the nature of the
assays (i.e., in vitro in Chapter 4 vs. actual inhalation exposures), and the findings that DNA
adducts may result from other known noncarcinogens such as carbon black (Bond et al., 1990)
make the interpretation of such exercises problematic and their meaning unclear.

3.7.  SUMMARY AND DISCUSSION
       The most consistent historical measure of exposure for DE is DPM in units of |_ig or mg
particles/m3, with the underlying assumption that all components of diesel emissions (e.g.,
organics in the form of volatilized liquids or gases) are present in proportion to the DPM mass.
DPM is used as the basic dosimeter for effects from various scenarios such as chronic and acute
exposures as well as for different endpoints such as irritation, fibrosis, or even cancer. There is,
however, little evidence currently available to prove or refute DPM as being the most appropriate
dosimeter.
       DPM dose to the tissue is related to the extent of the deposition and clearance of DPM.
DPM may deposit throughout the respiratory tract via sedimentation or diffusion, with the latter
being prevalent in the alveolar region. Particles that deposit upon airway surfaces may be cleared
from the respiratory tract completely or may be translocated to other sites by regionally distinct
processes that can be categorized as either absorptive (i.e., dissolution) or nonabsorptive (i.e.,
transport of  intact particles via mucociliary transport).  Other mechanisms that can affect
retention of  DPM include endocytosis by alveolar lining cells and interstitialization, which lead
to the accumulation of DPM in the interstitial  compartment of the lung and subsequent
translocation of DPM to lymph nodes; interstitialization of poorly soluble particles may be
prominent in primates and humans compared with rodents, although different rates for this path
could also explain observed results. For poorly  soluble particles such as DPM, species-
dependent rate constants exist for the various clearance pathways that can be modified by factors
such as respiratory tract disease.
       In rats, prolonged exposure to high concentrations of particles will result in particle
overload, a condition that is defined as the overwhelming of macrophage-mediated clearance by
the deposition of particles at a rate exceeding the capacity of that clearance pathway.  This
condition seems to begin to occur in rats when the pulmonary dust burden exceeds about 1 mg
particles/g lung tissue.  On the other hand, there is no clear evidence for particle overload in
humans. Macrophage-mediated clearance is slower in humans than in rats, and kinetics relating
to interstitialization of poorly soluble particulate matter may have a greater consequence in
humans than in rats.
       The degree of bioavailability of the organic fraction of DPM is still somewhat uncertain.
However, reports of DNA alterations in occupationally exposed workers, as well as results of
                                           3-54

-------
animal studies using radiolabeled organics deposited on DPM, indicate that at least a fraction of
the organics present are eluted prior to particle clearance.  Carcinogenic organics eluted in
regions where diffusion may be a relatively long process, such as in the conducting airways vs
the alveolar region, may remain in the lung long enough to be metabolized to an active form or to
interact directly with vital cellular components. The current information suggests that DPM-
associated organics could be involved in a carcinogenic process, although the quantitative data
are far from adequate to make any firm conclusions.
       Use of laboratory animal data in an assessment meant to be applied to humans obligates
some form of interspecies extrapolation. Review and evaluation of the considerable, specific
database in humans and animals on disposition of DPM, its adsorbed organics, and other poorly
soluble particles led to the judgment that default options available for interspecies dosimetry
adjustment could be set aside for more scientifically valid, DPM-specific processes. Refinement
of this process led to the evaluation of several applicable dosimetry models that in turn led to the
identification and choice of the Yu et al. (1991) model to conduct interspecies extrapolation.
This model has a three-compartment lung consisting of tracheobronchial, alveolar, and lymph
node compartments. It treats DPM as being composed of the insoluble carbonaceous core,
slowly cleared organics, and fast-cleared organics, and considers in an integrative manner the
simultaneous processes of both deposition and clearance through empirical data derived from
both laboratory animals and humans.  Also, the model has some limited consideration of model
variability in its outputs describing dose to the lung. Major assumptions made in this model
include that transport rates of organics in DPM do not change across  species and  that the
transport rate of the carbonaceous core is species dependent, with the clearance rate varying with
the dose to the alveolar surface in the same proportion in humans as in rats. Limitations of the
model  include the lack of definitive information on variability and, quite possibly, the lack of a
formal interstitial compartment that may be of consequence in humans. The basis of this model
is to derive an internal dose from an external DPM concentration by utilizing species-specific
physiological and pharmacokinetic parameters and, as such,  is considered to have addressed the
pharmacokinetic aspects of interspecies dosimetry. This aspect of the model addresses some of
the critical data needs for the quantitative analysis of noncancer effects from DPM, the subject
of Chapter 6.
       As parallels have been drawn between DPM and PM2 5 in other chapters, it is perhaps
appropriate to compare them also from the aspect of dosimetry. Obvious comparisons include
the nature of the particle distribution, defined artificially for PM2 5 as compared with the thorough
characterization of DPM for both MMAD (which, at around 0.2 |_im,  is typically more than an
order of magnitude less than the PM25 cutoff and which, more properly, should be termed a mass
median thermodynamic diameter, an MMTD) and geometric standard deviation.  It is  clear that a
                                          3-55

-------
larger portion of PM25 particles than DPM would be above the aerodynamic equivalent diameter

(dae) of 0.5 jam, which is often considered as a boundary between diffusion and aerodynamic

mechanisms of deposition. This would imply that a somewhat larger portion of DPM may pass

on to the lower respiratory tract than would PM2 5. Alveolar deposition in humans specific for

DPM has been estimated with the Yu model to be 7%-13%  (Yu and Xu, 1986), a figure that is

consistent with deposition predictions of other human models (see Table 3-3). This fractional

deposition may be compared to one calculated for PM2 5 and reported in U.S. EPA (1996a);

assuming a MMAD of 2.25 |_im and a geometric standard deviation of 2.4, a fractional alveolar

deposition of 10.2% was reported.  This value is within the range and quite comparable to that

obtained by Yu and Xu (1986), indicating that little difference may exist in alveolar deposition

between DPM and PM2 5,  at least for this assumed geometric standard deviation.


                             REFERENCES FOR CHAPTER 3

Adamson, IYR; Bowden, DH. (1978) Adaptive responses of the pulmonary macrophagic system to carbon: II.
Morphologic studies. Lab Invest 38:430-438.

Adamson, IYR; Bowden, DH. (1981) Dose response of the pulmonary macrophagic system to various particulates
and its relationship to transepithelial passage of free particles. Exp Lung Res 2:165-175.

Anjilvel, S; Asgharian, B. (1995) A multiple-path model of particle deposition in the rat lung. Fundam Appl Toxicol
28:41-50.

Aronson, M. (1963) Bridge formation and cytoplasmic flow between phagocytic cells. J Exp Med 118:1083-1088.

Bailey, MR; Fry, FA; James, AC. (1982) The long-term clearance kinetics of insoluble particles from the human
lung. Ann Occup Hyg 26:273-290.

Bailey, MR; Fry, FA; James, AC. (1985) Long-term retention of particles in the human respiratory tract. J Aerosol
Sci 16:295-305.

Ball, LM; King, LC. (1985) Metabolism, mutagenicity, and activation of 1-nitropyrene in vivo and in vitro. Environ
Int 11:355-361.

Battigelli, MC; Hengstenberg, F; Mannela, RJ; et al. (1966) Mucociliary activity. Arch Environ Health 12:460-466.

Bellmann, B; Muhle, H; Creutzenberg, O; et al. (1990) Recovery behaviour after dust overloading of lungs in rats.  J
Aerosol Sci 21:377-380.

Bevan, DR; Ruggio, DM. (1991) Bioavailability in vivo of benzo[a]pyrene adsorbed to diesel paniculate. Toxicol
Ind Health 7:125-139.

Bohning, DE; Cohn, SH; Lee, HD; et al. (1980) Two-phase deep-lung clearance in man. In: Pulmonary toxicology  of
respirable particles: proceedings of the nineteenth annual Hanford life sciences symposium; October 1979; Richland,
WA. Sanders, CL; Cross, FT; Dagle, GE; et al., eds. Washington, DC: U.S. Department of Energy; pp. 149-161.
Available from: NTIS, Springfield, VA; CONF-791002.
                                             3-56

-------
Bohning, DE; Atkins, HL; Cohn, SH. (1982) Long-term particle clearance in man: normal and impaired. In: Inhaled
particles V: proceedings of an international symposium; September 1980; Cardiff, Wales. Walton, WH, ed. Ann
OccupHyg 26:259-271.

Bond, JA; Mitchell, CE; Li, AP. (1983) Metabolism and macromolecular covalent binding of benzo[a]pyrene in
cultured Fischer-344 rat lung type II epithelial cells. Biochem Pharmacol 32:3771-3776.

Bond, JA; Sun, JD; Medinsky, MA; et al. (1986) Deposition, metabolism, and excretion of l-[14C]nitropyrene and
l-[14C]nitropyrene coated on diesel exhaust particles as influenced by exposure concentration.  Toxicol Appl
Pharmacol 85:102-117.

Bond, JA; Johnson, NF; Snipes, MB; et al. (1990) DNA adduct formation in rat alveolar type II cells: cells
potentially at risk for inhaled diesel exhaust. Environ Mol Mutagen 16:64-69.

Brain, JD; Mensah, GA. (1983)  Comparative toxicology of the respiratory tract. Am Rev RespirDis 128:887-890.

Brooks, AL; Wolff, RK; Royer, RE; et al. (1981) Deposition and biological availability of diesel particles and their
associated mutagenic chemicals. Environ Int 5:263-267.

Chan, TL; Lee, PS; Hering, WE. (1981) Deposition and clearance of inhaled diesel exhaust particles in the
respiratory tract of Fischer rats. J Appl Toxicol 1:77-82.

Chan, TL; Lee, PS; Hering, WE. (1984) Pulmonary retention of inhaled diesel particles after prolonged exposures to
diesel exhaust. Fundam Appl Toxicol 4:624-631.

Cohen, D; Arai, SF; Brain, JD. (1979) Smoking impairs long-term dust clearance from the lung. Science
(Washington, DC) 204:514-517.

Cohen, BS; Xiong, JQ; Fang, CP; et al. (1998) Deposition of charged particles on lung airways. Health Phys
74:554-560.

Crapo, JD; Young, SL; Fram, EK; et al. (1983) Morphometric characteristics of cells in the alveolar region of
mammalian lungs. Am Rev Respir Dis 128:842-846.

Cuddihy, RG; Yeh, HC. (1986) Model analysis of respiratory tract clearance of particles inhaled by people. In:
Annual report of the Inhalation Toxicology Research Institute. Muggenburg, BA; Sun, JD, eds. Albuquerque, NM:
Lovelace Biomedical and Environmental Research Institute; report no. LMF-115; pp. 140-147.

Evans, MJ;  Shami, SG; Martinez, LA. (1986) Enhanced proliferation of pulmonary alveolar macrophages after
carbon instillation in mice depleted of blood monocytes by Strontium-89. Lab Invest 54:154-159.

Felicetti,  SA; Wolff, RK; Muggenburg, BA. (1981) Comparison of trachea! mucous transport in rats, guinea pigs,
rabbits, and dogs. J Appl Physiol: Respir Environ Exercise Physiol 51:1612-1617.

Ferin, J. (1982) Pulmonary alveolar pores and alveolar macrophage-mediated particle clearance. Anat Rec
203:265-272.

Ferin, J; Feldstein, ML. (1978) Pulmonary clearance and hilar lymph node content in rats after particle exposure.
Environ Res 16:342-352.

Freedman, AP; Robinson, SE. (1988) Noninvasive magnetopneumographic studies of lung dust retention and
clearance in coal miners. In: Respirable dust in the mineral industries: health effects, characterization, and control:
proceedings of the international symposium on respirable dust in the mineral industries; October 1986; University
Park, PA. Frantz, RL; Ramani, RV, eds. University Park, PA: Pennsylvania State University Press; pp. 181-186.
                                                 3-57

-------
Freijer, JI; Cassee, FR; Subramaniam, R; et al. (1999) Multiple Path Particle Deposition model, a model for human
and rat airway particle deposition, MPPDep VI. 11. Research for Man and Environment publication number
65001019, Dutch National Institute of Public Health and the Environment, P.O. Box 1, 3720 BA Biltoven,
Netherlands.

Frey, JW; Corn, M. (1967) Physical and chemical characteristics of particulates in a diesel exhaust. Am Ind Hyg
Assoc J 28:468-478.

George, G; Hook, GER. (1984) The pulmonary extracellular lining. Environ Health Perspect 55:227-237.

Gerde, P; Medinsky, MA; Bond, JA. (1991a) Particle-associated polycyclic aromatic hydrocarbons - a reappraisal of
their possible role in pulmonary carcinogenesis. Toxicol Appl Pharmacol 108:1-13.

Gerde, P; Medinsky, MA; Bond, JA. (1991b) The retention of polycyclic aromatic hydrocarbons in the bronchial
airways and in the alveolar region - a theoretical comparison. Toxicol Appl Pharmacol 107:239-252.

Gerde, P; Muggenberg, BA; Dahl, AR. (1999) Bioavailability, absorption and metabolism of diesel soot-absorbed
benzo(a)pyrene after single-breath exposures in dogs. In: Relationships between acute and chronic effects of air
pollution: 7th international inhalation symposium; February; Hannover, Federal Republic of Germany; p. 76.

Gibb, FR; Morrow, PE. (1962) Alveolar clearance in dogs after inhalation of an iron 59 oxide aerosol. J Appl
Physiol 17:429-432.

Godleski, JJ; Stearns, RC; Katler, MR; et al. (1988) Particle dissolution in alveolar macrophages assessed by
electron energy loss analysis using the Zeiss CEM902 electron microscope. J Aerosol Med 1:198-199.

Griffis, LC; Wolff, RK; Henderson, RF; et al. (1983) Clearance of diesel soot particles from rat lung after a
subchronic diesel exhaust exposure. Fundam Appl Toxicol 3:99-103.

Gurley, LR; Spall, WD; Valdez, JG; et al. (1988) An HPLC procedure for the analysis of proteins in lung lavage
fluid. AnalBiochem 172:465-478.

Harmsen, AG; Muggenburg, BA; Snipes, MB; et al. (1985) The role of macrophages in particle translocation from
lungs to lymph nodes. Science (Washington, DC) 230:1277-1280.

Haschek, WM; Wischi, MR. (1991) Respiratory System, Chapter 22 in Handbook of Toxicologic Pathology
(Haschek, WM; Rousseaux, CG; eds.) pp 761-828.  Academic Press, NY.

Health Effects Institute. (1995) Diesel exhaust: a critical analysis of emissions, exposure, and health effects: a special
report of the Institute's Diesel Working Group. Cambridge, MA: Health Effects Institute.

Heinrich, U; Muhle, H; Takenaka, S; et al.  (1986) Chronic effects on the respiratory tract of hamsters, mice, and rats
after long-term inhalation of high concentrations of filtered and unfiltered diesel engine emissions. J Appl Toxicol
6:383-395.

Hemminki, K; Soderling, J; Ericson, P; et al.  (1994) DNA adducts among personnel servicing and loading diesel
vehicles. Carcinogenesis  15:767-769.

Heppleston, AG. (1961) Observations on the  disposal of inhaled dust by means of the double exposure technique. In:
Inhaled particles and vapours: proceedings of an international symposium; March-April 1960; Oxford, United
Kingdom. Davies, CN, ed. New York: Pergamon Press; pp.  320-326.

Heppleston, AG; Young, AE. (1973) Uptake of inert paniculate matter by alveolar cells: an ultrastructural study. J
Pathol 111:159-164.
                                                  3-58

-------
Heyder, J; Gebhart, J; Rudolf, G; et al. (1986) Deposition of particles in the human respiratory tract in the size range
0.005-15 urn. J Aerosol Sci 17:811-825.

Hiura, TS; Kaxzubowski, MP; Li, N; et al.  (1999) Chemicals in diesel exhaust particles generate oxygen radicals
and induce apoptosis in macrophages.  J Immunol 163:5582-5591.

Holt, PF. (1981) Transport of inhaled dust to extrapulmonary sites. J Pathol 133:123-129.

Hsieh, TH; Yu, CP. (1998) Two-phase pulmonary clearance of insoluble particles in mammalian species. Inhal
Toxicol 10:121-130.

International Commission on Radiological Protection (ICRP). (1994) Human respiratory tract model for radiological
protection: a report of a task group of the International Commission on Radiological Protection. Oxford, United
Kingdom: Elsevier Science Ltd. (ICRP publication 66; Annals of the ICRP: v. 24, nos. 1-3).

International Life Sciences Institute (ILSI). (2000) ILSI Risk Science Institute Workshop: the relevance of the rat
lung response to particle overload for human risk assessment. Gardner, DE, ed. Inhal Toxicol 12(1-2): 1-148.

Jammet, H; Drutel, P; Parrot, R; et al. (1978) Etude de 1'epuration pulmonaire chez l'homme apres administration
d'aerosols de particules radioactives [Study of pulmonary function in man after administration of radioactive
paniculate aerosols]. Radioprotection 13:143-166.

Jarvis, NS; Birchall, A; James, AC; et al. (1996) International Commission on Radiological Protection, LUDEP
(LUng DEPosition) 2.07 Personal Computer Program for Calculating Internal Doses Using the ICRP Publication 66
Respiratory Tract Model. NRPB-SR287, National Radiological Protection Board, Chilton, Didcot, Oxon OX11
ORQ, UK.

Keane, MJ; Xing, S-G; Harrison JC; et al. (1991) Genotoxicity of diesel-exhaust particles dispersed in simulated
pulmonary surfactant.  Mutat Res 260:233-238.

King, LC; Kohan, MJ; Austin, AC; etal. (1981) Evaluation of the release of mutagens from diesel particles in the
presence of physiological fluids. Environ Mutagen 3:109-121.

Kittelson, DB; Dolan, DF; Verrant, JA. (1978) Investigation of a diesel exhaust aerosol. Warrendale, PA: Society of
Automotive Engineers, Inc.; technical paper series no. 78109.

Kreyling, WG. (1992) Intracellular particle dissolution in alveolar macrophages. Environ Health Perspect
97:121-126.

Kuempel, ED. (2000). Comparison of human and rodent lung dosimetry models for particle clearance and retention.
Drug Chem Toxicol 23(l):203-222.

Lee, PS;  Chan, TL; Hering, WE. (1983) Long-term clearance of inhaled diesel exhaust particles in rodents.  J Toxicol
Environ Health 12:801-813.

Lee, KP; Trochimowicz, HJ; Reinhardt, CF. (1985) Transmigration of titanium dioxide (TiO2) particles in rats after
inhalation exposure. Exp Mol Pathol 42:331-343.

Lee, KP; Henry, NW, III; Trochimowicz, HJ.; et al. (1986) Pulmonary response to impaired lung clearance in rats
following excessive TiO2 dust deposition. EnvironRes 41:144-167.

Lee, KP; Ulrich, CE; Geil, RG; et al. (1988) Effects of inhaled chromium dioxide dust on rats exposed for two years.
Fundam Appl Toxicol 10:125-145.
                                                  3-59

-------
Lehnert, BE. (1988) Distributions of particles in alveolar macrophages during lung clearance. J Aerosol Med
1:206-207.

Lehnert, BE. (1989) Rates of disappearance of alveolar macrophages during lung clearance as a function of
phagocytized paniculate burden [abstract]. Am Rev Respir Dis 139(suppl.):A161.

Lehnert, BE. (1990) Alveolar macrophages in a particle "overload" condition. J Aerosol Med 3(suppl. 1):S9-S30.

Lehnert, BE; Morrow, PE. (1985) Association of 59iron oxide with alveolar macrophages during alveolar clearance.
Exp Lung Res 9:1-16.

Lehnert, BE; Valdez, YE; Bomalaski, SH. (1988) Analyses of particles in the lung free cell, tracheobronchial lymph
nodal, and pleural space compartments following their deposition in the lung as related to lung clearance
mechanisms. In: Inhaled particles VI: proceedings of an international symposium and workshop on lung dosimetry;
September 1985; Cambridge, United Kingdom. Dodgson, J: McCallum, RI; Bailey, MR; et al., eds. Ann Occup Hyg
32(suppl. 1): 125-140.

Lehnert, BE; Cline, A; London, JE. (1989) Kinetics of appearance of polymorphonuclear leukocytes and their
particle burdens during the alveolar clearance of a high lung burden of particles. Toxicologist 9:77.

Leung, HW; Henderson, RF; Bond, JA; et al. (1988) Studies on the ability of rat lung and liver microsomes to
facilitate transfer and metabolism of benzo[a]pyrene from diesel particles. Toxicology 51:1-9.

Lewis, TR; Green, FHY; Moorman, WJ; et al. (1989) A chronic inhalation toxicity study of diesel engine emissions
and coal dust, alone and combined. J Am Coll Toxicol 8:345-375.

Lippmann, M; Schlesinger, RB. (1984) Interspecies comparisons of particle deposition and mucociliary clearance in
tracheobronchial airways. J Toxicol Environ Health  13:441-469.

Lundborg, M; Lind, B; Camner, P. (1984) Ability of rabbit alveolar macrophages to dissolve metals. Exp Lung Res
7:11-22.

Marafante, E; Lundborg, M;  Vahter, M; et al.  (1987) Dissolution of two arsenic compounds by rabbit alveolar
macrophages in vitro. Fundam Appl Toxicol 8:382-388.

Mauderly, JL. (1996) Lung overload: the dilemma and opportunities for resolution. In: Particle overload in the rat
lung and lung cancer: implications for human risk assessment. Proceedings of a conference; March 1995;
Cambridge, MA. Mauderly, JL; McCunney, RJ; eds. New York: Taylor & Francis.

McClellan, RO. (2000) Particle interactions with the respiratory tract.  In:  Particle-lung interactions.  Geber, P;
Heyder, J, eds. New York: Marcel Dekker, pp. 4-63.

McClellan, RO; Brooks, AL; Cuddihy, RG; et al. (1982) Inhalation toxicology of diesel exhaust particles. In:
Toxicological effects of emissions from diesel engines: proceedings of the EPA diesel emissions symposium;
October 1981; Raleigh, NC. Lewtas, J, ed. New York: Elsevier Biomedical; pp. 99-120. (Developments in
toxicology and environmental science: v.  10)

Morrow, PE. (1966) International Commission on Radiological Protection (ICRP) task group on lung dynamics,
deposition and retention models for internal dosimetry of the human respiratory tract. Health Phys 12:173.

Morrow, PE. (1988) Possible mechanisms to explain dust overloading of the lungs. Fundam Appl Toxicol
10:369-384.

Morrow, PE. (1994) Mechanisms and significance of "particle overload." In: Toxic and carcinogenic effects of solid
particles in the respiratory tract: [proceedings  of the 4th international inhalation symposium]; March 1993;
                                                 3-60

-------
Hannover, Germany. Mohr, U; Dungworth, DL; Mauderly, JL; et al., eds. Washington, DC: International Life
Sciences Institute Press; pp. 17-25.

Morrow, PE; Gibb, FR; Gazioglu, KM. (1967a) The clearance of dust from the lower respiratory tract of man. An
experimental study. In: Inhaled particles and vapors II. Davies, CN, ed. London: Pergamon, pp. 351-359.

Morrow, PE; Gibb, FR; Gazioglu, KM. (1967b) A study of paniculate clearance from the human lungs. Am Rev
RespirDis 96:1209-1221.

Morrow, PE; Yu, CP. (1993) Models of aerosol behavior in airways and alveoli. In: Aerosols in medicine:
principles, diagnosis and therapy. 2nd rev. ed. Moren, F; Dolovich, MB; Newhouse, MT; et al., eds. Amsterdam:
Elsevier; pp. 157-193.

Muhle, H; Bellman, B; Creutzenberg, O; et al. (1990a) Dust overloading of lungs after exposure of rats to particles
of low solubility: comparative studies. J Aerosol Sci 21:374-377.

Muhle, H; Creutzenberg, O; Bellmann, B; et al. (1990b) Dust overloading of lungs: investigations of various
materials, species differences, and irreversibility of effects.  J Aerosol Med 3(suppl. 1): S111-S128.

National Council on Radiation Protection and Measurements (NCRP). (1997) Deposition, retention, and dosimetry
of inhaled radioactive substances, NCRP Report No. 125. Bethesda, MD.

National Research Council (NRC). (1995) Science and judgment in risk assessment. Washington, DC: National
Research Council.

Nielsen, PS; Andreassen, A; Farmer, PB; et al. (1996) Biomonitoring of diesel exhaust-exposed workers. DNA and
hemoglobin adducts and urinary 1-hydroxypyrene as markers of exposure. Toxicol Lett 86:27-37.

Nikula, KJ; Avila, KJ; Griffith, WC; et al. (1997a) Sites of particle retention and lung tissue  responses to chronically
inhaled diesel exhaust and coal dust in rats and cynomolgus monkeys. In: Proceedings of the sixth international
meeting on the toxicology of natural and man-made fibrous and non-fibrous particles; September 1996; Lake Placid,
NY. Driscoll, KE; Oberdorster, G, eds. Environ Health Perspect Suppl 105(5): 1231-1234.

Nikula, KJ; Avila, KJ; Griffith, WC; et al. (1997b) Lung tissue responses and sites of particle retention differ
between rats and cynomolgus monkeys exposed chronically to diesel exhaust and coal dust.  Fundam Appl Toxicol
37:37-53.

Nilsen, A; Nyberg, K; Camner, P. (1988) Intraphagosomal pH in alveolar macrophages after phagocytosis in  vivo
and in vitro of fluorescein-labeled yeast particles. Exp Lung Res 14:197-207.

Oberdorster, G. (1988) Lung clearance of inhaled insoluble and soluble particles. J Aerosol Med 1:289-330.

Oberdorster, G. (1994) Extrapolation of results from animal inhalation studies with particles to humans In: toxic and
carcinogenic effects of solid particles in the  respiratory tract: [proceedings of the 4th international inhalation
symposium]; March 1993 Hannover, Germany. Mohr, U; Dungworth, DL; Mauderly, JL; et al., eds.  Washington,
DC: International Life Sciences Institute Press; pp 335-353.

Oberdorster, G; Ferin, J; Morrow, PE. (1992) Volumetric loading of alveolar macrophages (AM): a possible basis
for diminished AM-mediated particle clearance. Exp Lung Res 18:87-104.

Ohkuma, S; Poole, B. (1978) Fluorescence probe measurement of the intralysosomal pH in living cells and the
perturbation of pH by various agents. Proc Nat. Acad Sci U S A 75:3327-3331.
                                                  3-61

-------
Osterholm, AM; Fait, S; Lambert, B; et al. (1995) Classification of mutations at the human hprt-locus in
T-lymphocytes of bus maintenance workers by multiplex-PCR and reverse transriptase-PCR analysis. Carcinogenesis
16:1909-1995.

Pepelko, WE. (1987) Feasibility of dose adjustment based on differences in long-term clearance rates of inhaled
paniculate matter in humans and laboratory animals. Regul Toxicol Pharmacol 7:236-252.

Phalen, RF; Oldham, MJ. (1983) Tracheobronchial airway structure as revealed by casting techniques. Am Rev
RespirDis 128:S1-S4.

Phalen, RF; Oldham, MJ; Beaucage, CB.; et al. (1985) Postnatal enlargement of human tracheobronchial airways
and implications for particle deposition. Anat. Rec. 212:368-380.

Phalen, RF; Cuddihy, RG; Fisher, GL; et al.  (1991) Main features of the proposed NCRP respiratory tract model.
Radiat. Prot. Dosim. 38:179-184.

Powdrill, J; Buckley, C; Valdez, YE; et al. (1989) Airway intra-luminal macrophages: origin and role in lung
clearance. Toxicologist 9:77.

Pritchard, JN. (1989) Dust overloading causes impairment of pulmonary clearance: evidence from rats and humans.
Exp Pathol 37:39-42.

Raabe, OG. (1982) Deposition and clearance of inhaled aerosols. In: Mechanisms in respiratory toxicology. Witschi,
H, ed. Boca Raton, FL: CRC Press; pp. 27-76.

Raabe, OG; Al-Bayati, MA; Teague, SV; et al. (1988) Regional deposition of inhaled monodisperse, coarse, and fine
aerosol particles in small laboratory animals. In: Inhaled particles VI: proceedings of an international symposium and
workshop on lung dosimetry; September 1985; Cambridge, United Kingdom. Dodgson, J; McCallum, RI; Bailey,
MR; et al., eds. Ann Occup Hyg 32(suppl. l):53-63.

Riley, PA; Dean, RT. (1978) Phagocytosis of latex particles in relation to the cell cycle in 3T3 cells. Exp Cell Biol
46:367-373.

Robertson, B. (1980) Basic morphology of the pulmonary defence system. Eur J Respir Dis 61(suppl. 107):21-40.

Sandusky, CB; Cowden, MW;  Schwartz, SL. (1977) Effect of particle size on regurgitative exocytosis by rabbit
alveolar macrophages. In: Pulmonary macrophage and epithelial cells: proceedings of the sixteenth annual Hanford
biology symposium; September 1976; Richland, WA. Sanders, CL; Schneider, RP; Dagle, GE; et al., eds.
Washington, DC: Energy Research and Development Administration; pp. 85-105. (ERDA symposium series: no.
43). Available from: NTIS, Springfield, VA; CONF-760927.

Schanker, LS; Mitchell, EW; Brown, RA, Jr. (1986) Species comparison of drug absorption from the lung after
aerosol inhalation or intratracheal injection. Drug Metab Dispos 14:79-88.

Schlesinger,  RB.  (1985) Comparative deposition of inhaled aerosols in experimental animals and humans: a review.
J Toxicol Environ Health 15:197-214.

Schlesinger,  R.B. (1995) Deposition and clearance of inhaled particles. In: McClellan, R.O.; Henderson, R.F.; eds.
Concepts in inhalation toxicology. 2nd ed. Washington D.C., Taylor & Francis: pp. 191-224.

Schlesinger,  RB; Ben-Jebria, A; Dahl, AR; et al.  (1997) Disposition of inhaled toxicants. In: Handbook of human
toxicology. Massaro, EJ, ed. Boca Raton, FL: CRC Press; pp. 493-550.

Schum, M; Yeh, HC. (1980) Theoretical evaluation of aerosol deposition in anatomical models of mammalian lung
airways. Bull Math Biol 42:1-15.
                                                 3-62

-------
Siak, JS; Chan, TL; Lee, PS. (1980) Diesel paniculate extracts in bacterial test systems. In: Health effects of diesel
engine emissions: proceedings of an international symposium, v. 1; December 1979; Cincinnati, OH.  Pepelko, WE;
Danner, RM; Clarke, NA, eds. Cincinnati, OH: U.S. EPA, Health Effects Research Laboratory; pp. 245-262; EPA
report no. EPA/600/9-80/057b. Available from: NTIS, Springfield, VA; PB81-173809.

Smith, TJ. (1985) Development and application of a model for estimating alveolar and interstitial dust levels. Ann
OccupHyg 29:495-516.

Snipes, MB. (1979) Long-term retention of monodisperse and polydisperse particles inhaled by beagle dogs, rats and
mice. Albuquerque, NM: Lovelace Biomedical and Environmental Research Institute; Inhalation Toxicology
Research Institute annual report LF-69; pp. 420-423.

Snipes, MB; Clem, MF. (1981) Retention of microspheres in the rat lung after intratracheal instillation. Environ Res
24:33-41.

Snipes, MB; Boecker, BB; McClellan, RO. (1983)  Retention of monodisperse or polydisperse aluminosilicate
particles inhaled by dogs, rats, and mice. Toxicol Appl Pharmacol 69:345-362.

Snipes, MB; Olson, TR; Yeh, HC. (1988) Deposition and retention patterns for 3-, 9-, and 15-|im latex microspheres
inhaled by rats and guinea pigs. Exp Lung Res 14:37-50.

Snipes, MB; McClellan, RO; Mauderly, JL; et al. (1989) Retention patterns for inhaled particles in the lung:
comparisons between laboratory animals and humans for chronic exposures. Health Phys 57(suppl. l):69-78.

Sorokin, SP; Brain, JD. (1975) Pathways of clearance in mouse lungs exposed to iron oxide aerosols.  Anat Rec
181:581-625.

Stahlhofen, W; Gebhart, J; Heyder, J. (1980) Experimental determination of the regional deposition of aerosol
particles in the human respiratory tract. Am Ind Hyg Assoc J41:385-398a.

Stahlhofen, W; Koebrich, R; Rudolf, G; et al.  (1990) Short-term and long-term clearance of particles from the upper
human respiratory tract as function of particle size.  J Aerosol Sci 21(suppl.  1):S407-S410.

Stober, W; Einbrodt, HJ; Klosterkotter, W. (1967) Quantitative studies of dust retention in animal and human lungs
after chronic inhalation. In: Inhaled particles and vapours II: proceedings of an international
symposium; September-October 1965; Cambridge,  United Kingdom. Davies, CN, ed. Oxford, United Kingdom:
Pergamon Press; pp. 409-418.

Stober, W; Morrow, PE; Hoover, MD. (1989) Compartmental modeling of the long-term retention of insoluble
particles deposited in the alveolar region of the lung. Fundam Appl Toxicol 13:823-842.

Stober, W; McClellan, RO; Morrow, PE. (1993) Approaches to modeling disposition of inhaled particles and fibers
in the lung. In: Toxicology of the lung. Gardner, DE; Crapo, JD; McClellan, RO, eds. New York: Raven Press; pp.
527-601.

Stober, W; Morrow, PE; Koch, W; et al. (1994) Alveolar clearance and retention of inhaled insoluble particles in
rats simulated by a model inferring macrophage particle load distributions. J Aerosol Sci 25:975-1002.

Strom, KA. (1984) Response of pulmonary cellular defenses to the inhalation of high concentrations of diesel
exhaust. J Toxicol Environ Health 13:919-944.

Strom, KA; Chan, TL;  Johnson, JT. (1988) Pulmonary retention of inhaled  submicron particles in rats: diesel exhaust
exposures and lung retention model. In: Inhaled particles VI: proceedings of an international symposium and
workshop on lung dosimetry; September 1985; Cambridge, United Kingdom. Dodgson, J; McCallum, RI; Bailey,
MR; et al., eds. Ann Occup Hyg 32(suppl. l):645-657.
                                                 3-63

-------
Subramaniam, RP; Miller, FJ; Asgarian, B. (1998) Numerical predictions of regional deposition of fine particles in
the human lung using a multiple-path model. Proceedings of the American Association of Aerosol Research, annual
meeting; Denver, CO.

Sun, JD; Wolff, RK; Kanapilly, GM. (1982) Deposition, retention, and biological fate of inhaled benzo(a)pyrene
adsorbed onto ultrafine particles and as a pure aerosol. Toxicol Appl Pharmacol 65:231-244.

Sun, JD; Wolff, RK; Kanapilly, GM; et al. (1984) Lung retention and metabolic fate of inhaled benzo(a)pyrene
associated with diesel exhaust particles. Toxicol Appl Pharmacol 73:48-59.

Tong, HY; Karasek, FW.  (1984) Quantitation of polycyclic aromatic hydrocarbons in diesel exhaust paniculate
matter by high-performance liquid chromatography fractionation and high-resolution gas chromatography. Anal
Chem 56:2129-2134.

U.S. Environmental Protection Agency (U.S. EPA). (1982) Air quality criteria for paniculate matter and sulfur
oxides. Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; EPA report no. EPA/600/8-82-029aF-cF.  3v. Available from: NTIS, Springfield, VA;
PB84-156777.

U.S. EPA. (1993) Provisional guidance for quantitative risk assessment of polycyclic aromatic hydrocarbons.  Office
of Health and Environmental Assessment; EPA report no. EPA/600/R-93/089.

U.S. EPA. (1994) Methods for derivation of inhalation reference concentrations and application of inhalation
dosimetry [draft final]. Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental
Criteria and Assessment Office; report no. EPA/600/8-88/066F.

U.S. EPA. (1996) Air quality criteria for paniculate matter. Research Triangle Park, NC: National Center for
Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v. Available from: NTIS, Springfield,
VA; PB96-168224.

Valberg, PA; Watson, AY. (1999) Comparative mutagenic dose of ambient diesel engine exhaust. Inhal Toxicol
ll(3):215-228.

Vastag, E; Matthys, H; Zsamboki, G; et al. (1986) Mucociliary clearance in smokers. Eur J Respir Dis 68:107-113.

Vostal, JJ; Schreck, RM; Lee, PS; et al. (1982) Deposition and clearance of diesel particles from the lung. In:
Toxicological effects of emissions from diesel engines: proceedings of the 1981 EPA diesel emissions symposium;
October 1981; Raleigh, NC. Lewtas, J, ed. New York: Elsevier Biomedical; pp. 143-159. (Developments in
toxicology and environmental science: v. 10).

Waite, DA; Ramsden, D. (1971) The inhalation of insoluble iron oxide particles in the sub-micron range. Parti.
Chromium-51 labelled aerosols. Winfrith, Dorchester, Dorset, United Kingdom: Atomic Energy Establishment;
report no. AEEW-R740.

Warheit, DB; Overby, LH; George, G; et al. (1988) Pulmonary macrophages are attracted to inhaled particles
through complement activation. Exp Lung Res 14:51-66.

Weibel, ER. (1963) Morphometry of the human lung. New York: Academic Press, Inc.

White, HJ; Garg, BD. (1981) Early pulmonary response of the rat lung to inhalation of high concentration of diesel
particles. J Appl Toxicol 1:104-110.

Wolff, RK; Gray, RL. (1980) Trachea! clearance of particles.  In: Inhalation Toxicology Research Institute annual
report: 1979-1980. Diel, JH; Bice, DE; Martinez, BS, eds. Albuquerque, NM: Lovelace Biomedical and
Environmental Research Institute; p. 252; report no. LMF-84.


                                                 3-64

-------
Wolff, RK; Henderson, RF; Snipes, MB; et al. (1986) Lung retention of diesel soot and associated organic
compounds. In: Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the international
satellite symposium on lexicological effects of emissions from diesel engines; July; Tsukuba Science City, Japan.
Ishinishi, N; Koizumi, A; McClellan, R; et al. eds. Amsterdam: Elsevier Science Publishers BV.; pp. 199-211.
(Developments in toxicology and environmental science: v. 13).

Wolff, RK; Henderson, RF; Snipes, MB; et al. (1987) Alterations in particle accumulation and clearance in lungs of
rats chronically exposed to diesel exhaust. Fundam Appl Toxicol 9:154-166.

World Health Organization (WHO). (1996) Diesel fuel and exhaust emissions. Geneva, Switzerland: World Health
Organization, International Programme on Chemical Safety. (Environmental health criteria 171).

Wright, JR; Clements, JA. (1987) Metabolism and turnover of lung surfactant. Am Rev Respir Dis 136:426-444.

Xu, GB; Yu, CP. (1987) Deposition of diesel exhaust particles in mammalian lungs: a comparison between rodents
and man. Aerosol Sci Technol 7:117-123.

Yeh, HC; Schum, GM. (1980) Models of human lung airways and their application to inhaled particle deposition.
Bull Math Biol 42:461-480.

Yeh, H-S; Cuddihy, RG; Phalen, RF; et al. (1996) Comparison of calculated respiratory tract deposition of particles
based on the proposed NCRP model and the new ICRP66 model.  Aerosol Sci Technol 25:134-140.

Yu, CP. (1978) Exact analysis of aerosol deposition during steady breathing. Powder Technol 21:55-62.

Yu, CP; Diu, CK. (1983) Total and regional deposition of inhaled aerosols in humans. J Aerosol Sci 5:599-609.

Yu, CP; Xu, GB. (1986) Predictive models for deposition of DE particulates in human and rat lungs. Aerosol Sci
Technol 5:337-347.

Yu, CP; Xu, GB. (1987) Predictive models for deposition of diesel exhaust particulates in human and rat lungs.
Aerosol Sci Technol 5:337-347.

Yu, CP; Yoon, KJ.  (1990) Retention modeling of diesel exhaust particles in rats and humans. Amherst, NY: State
University of New York at Buffalo (Health Effects Institute research report no. 40).

Yu, CP; Chen, YK; Morrow, PE. (1989)  An analysis of alveolar macrophage mobility kinetics at dust overloading of
the lungs. Fundam Appl Toxicol 13:452-459.

Yu, CP; Yoon, KJ;  Chen, YK. (1991) Retention modeling of diesel exhaust particles in rats and humans. J Aerosol
Med4(2):79-115.

Yu, CP; Ding, YJ; Zhang, L; et al. (1996) A clearance model of refractory ceramic fibers in the rat lung including
fiber dissolution and breakage. J Aerosol Sci 27:151-160.
                                                 3-65

-------
                                  4.  MUTAGENICITY

       The application of mutagenicity data to the question of the potential carcinogenicity of
diesel emissions is based on the premise that genetic alterations are found in all cancers and that
several of the chemicals found in diesel emissions possess mutagenic activity in a variety of
genetic assays.  These genetic alterations can be produced by gene mutations, deletions,
translocations, aneuploidy, or amplification of genes; hence no single genotoxicity assay should
be expected to predict rodent carcinogenicity.  Additionally, because of the inherent biological
differences of measured endpoints, both within genotoxicity assays and between genotoxicity
assays and cancer bioassays, a direct extrapolation should not be expected (see Brusick [1987]
for a more detailed discussion). Indeed, most  genotoxicity data are generated with in vitro
assays that frequently employ concentrations of test agent that may be orders of magnitude
greater than encountered in environmental  situations. With diesel emissions or other mixtures,
additional complications arise because of the complexity of the material being tested.
       Since 1978, more than  100 publications have appeared in which genotoxicity assays were
used with diesel emissions, the volatile  and particulate fractions (including extracts), or
individual chemicals found in diesel emissions. The Huisingh et al. (1978) report not only
identified mutagenic activity in salmonella in several fractions of diesel particular mater (DPM)
extracts, but also indicated that the mutagenic activity, especially quantitatively, was affected by
the extraction solvent as well as method and length of storage. Much of the ensuing research
employed bioassays (most commonly salmonella TA98 without S9) to evaluate (1)  extraction
procedures, (2) fuel modifications,  (3) bioavailability of chemicals from  DPM, and (4) exhaust
filters or other modifications and other variables associated with diesel emissions. The interest
in the contribution of mutagens to carcinogenicity was high in the early 1980s and the lack of
long- term rodent carcinogenicity information on diesel emissions led to the use of
semiquantitative mutagenicity  (and in vitro cell transformation) data from diesel emissions and
epidemiology based cancer potency estimations to derive a comparative potency estimate for
diesel emissions first published by Albert et al. (1983) and more fully discussed in Appendix C
of this report.
       As indicated in Chapter 2, the number  of chemicals in diesel emissions is very large.
Many of these have been determined to exhibit mutagenic activity in a variety of assay systems
(see Table II. in Claxton, 1983).  Although a detailed discussion of those data is beyond the
scope of this document,  some of the mutagenically active compounds found in the gas phase are
ethylene, benzene, 1,3-butadiene, acrolein and several PAHs (see Table 2-21).  Of the particle-
associated chemicals, several PAHs and nitro-PAHs  have been the focus of mutagenic
investigations both in bacteria  and in mammalian cell systems (see Table 2-22).  Several review

                                           4-1

-------
articles, some containing more detailed descriptions of the available studies, are available
(Claxton, 1983; Pepelko andPeirano, 1983; International Agency for Research on Cancer, 1989;
Shirname-More, 1995). Discussions of genotoxicity in the proceedings of several symposia on
the health effects of diesel emissions (U.S. EPA, 1980; Lewtas, 1982; Ishinishi et al., 1986) are
also available.

4.1. GENE MUTATIONS
       Huisingh et al. (1978) demonstrated that dichloromethane extracts from DPM were
mutagenic in strains TA1537,  TA1538, TA98, and TA100 of S. typhimurium, both with and
without rat liver S9 activation.  This report contained data from several fractions as well as DPM
from different vehicles and fuels. Similar results with diesel extracts from various engines and
fuels have been reported by a number of investigators using the salmonella frameshift-sensitive
strains TA1537, TA1538, and TA98 (Siak et al., 1981; Claxton,  1981; Dukovich et al., 1981;
Brooks et al., 1984).  Similarly, mutagenic activity was observed in salmonella forward mutation
assays measuring 8-azaguanine resistance (Claxton and Kohan, 1981) and inE1. coli mutation
assays (Lewtas, 1983).
       One approach to identifying significant mutagens in chemically complex environmental
samples such as diesel exhaust or ambient particulate extracts is  the combination of short-term
bioassays with chemical fractionation (Scheutzle and Lewtas, 1986).  The analysis is most
frequently carried out by sequential extraction with increasingly  polar or binary solvents.
Fractionation by silica-column chromatography separates compounds by polarity or into acidic,
basic, and neutral fractions. The resulting fractions are too complex to characterize by chemical
methods, but the bioassay analysis can be used to determine fractions for further analysis. In
most applications of this concept, salmonella strain TA98 without the addition of S9 has been
used as the indicator for mutagenic activity. Generally, a variety of nitrated polynuclear
aromatic compounds have been found that account for a substantial portion of the mutagenicity
(Liberti et al., 1984; Schuetzle and Frazer, 1986; Schuetzle and Perez, 1983). However, not all
bacterial mutagenicity has been identified in this way, and the identity of the remaining
mutagenic compounds remains unknown. The nitrated aromatics thus far identified in diesel
engine exhaust (DE) were the  subject of review in the IARC monograph on DE (International
Agency for Research on Cancer, 1989).  In addition to the simple qualitative identification of
mutagenic chemicals,  several investigators have used numerical  data to express mutagenic
activity as activity per distance driven or mass of fuel consumed. These types of calculations
have been the basis for estimates that the nitroarenes (both mono- and dinitropyrenes) contribute
a significant amount of the total  mutagenic activity of the whole  extract (Nishioka et al., 1982;
Salmeen et al., 1982; Nakagawa et al., 1983).  In a 1983 review,  Claxton discussed a number of

                                          4-2

-------
factors that affected the mutagenic response in salmonella assays. Citing the data from the
Huisingh et al. (1978) study, the author noted that the mutagenic response could vary by a factor
of 100 using different fuels in a single diesel engine. More recently, Crebelli et al. (1995) used
salmonella to examine the effects of different fuel components. They reported that although
mutagenicity was highly dependent on aromatic content, especially di- or triaromatics, there was
no clear effect of sulfur content of the fuel.  Later, Sjogren et al. (1996) using multivariate
statistical methods with ten diesel fuels concluded that the most influential chemical factors in
salmonella mutagenicity were sulfur contents, certain PAHs (1-nitropyrene) and naphthenes.
       Matsushita et al. (1986) tested particle-free DE gas and a number of benzene nitro-
derivatives and poly cyclic aromatic hydrocarbons (PAHs) (many of which have been identified
as components of DE gas). The particle-free exhaust gas was positive in both TA100 and TA98,
but only without S9 activation.  Of the 94 nitrobenzene derivatives tested, 61 were mutagenic,
and the majority showed greatest activity in TA100 without S9.  Twenty-eight of 50 PAHs tested
were mutagenic, all required the addition of S9 for detection, and most appeared to show a
stronger response in TA100.  When 1,6-dinitropyrene was mixed with various PAHs or an
extract of heavy-duty (HD) DE, the mutagenic activity in TA98 was greatly reduced when S9
was absent but was increased significantly when S9 was present. These latter results suggested
that caution should be used in estimating mutagenicity (or other toxic effects) of complex
mixtures from the specific activity of individual components.
       Mitchell et al. (1981) reported mutagenic activity of DPM extracts of diesel emissions in
the mouse lymphoma L5178Y mutation assay. Positive results were seen both with and without
S9 activation in extracts from several different vehicles, with mutagenic activity only slightly
lower in the presence of S9. These findings have been confirmed in a number of other
mammalian cell systems using several different genetic markers. Casto et al. (1981), Chescheir
et al. (1981), Li and Royer (1982), and Brooks et al.  (1984) all reported positive responses at the
HPRT locus in Chinese hamster ovary (CHO) cells.  Morimoto et al. (1986) used the APRT and
Ouar loci in CHO cells; Curren et al. (1981) used Ouar in BALB/c 3T3 cells.  In all of these
studies, mutagenic activity was observed without S9 activation. Liber et al. (1981) used the
thymidine kinase (TK) locus in the TK6 human lymphoblast cell line and observed induced
mutagenesis only in the presence of rat liver S9 when testing a methylene chloride extract of DE.
Barfknecht et al. (1982) also used the TK6 assay to identify some of the chemicals responsible
for this activation-dependent mutagenicity.  They suggested that fluoranthene, 1-
methylphenanthrene, and 9-methylphenanthrene could account for over 40% of the observed
activity.
       Morimoto et al. (1986) injected DPM extracts (250 to 4,000 mg/kg) into pregnant Syrian
hamsters and measured mutations at the APRT locus in embryo cells cultivated 11 days after i.p.

                                           4-3

-------
injection.  Although neutral fractions from both light-duty (LD) and HD particle extracts resulted
in increased mutation frequency at 2,000 and 4,000 mg/kg, the response at 1,000 mg/kg was not
different from controls.  Also, because the authors did not present data on toxicity or cloning
efficiency, the value of the apparent positive findings at extremely high concentrations is
uncertain at best.  Belisario et al. (1984) applied the Ames test to urine from Sprague-Dawley
rats exposed to single applications of DPM administered by gastric intubation, i.p. injection, or
s.c. gelatin capsules.  In  all cases, dose-related increases were seen in TA98 (without and with
S9) from urine concentrates taken 24 h after particle  administration. Urine from Swiss mice
exposed by inhalation to filtered exhaust (particle concentration 6 to 7 mg/m3) for 7 weeks
(Pereira et al., 1981a) or Fischer 344 rats exposed to  DPM at a concentration of 1.9 mg/m3 for 3
months to 2 years (Ong et al., 1985) was negative in  salmonella strains.
       Schuler and Niemeier (1981) exposed drosophila males in a  stainless steel  chamber
connected to the 3 m3 chamber used for the chronic animal studies at EPA (see Hinners et al.,
1980 for details).  Flies were exposed for 8 h and mated to untreated females 2 days later.
Although the frequency  of sex-linked recessive lethals from treated  males was not different from
that of controls, the limited sample size precluded detecting less than a threefold increase over
controls. The authors noted that, because there were no signs of toxicity, the flies might tolerate
exposures to higher concentrations for longer time periods.
       Driscoll et al. (1996) exposed Fischer 344 male rats to aerosols of carbon black (1.1, 7.1,
and 52.8 mg/m3) or air for 13 weeks (6 hr/day, 5 days/week) and measured hprt mutations in
alveolar type II cells in animals immediately after exposure and at 12 and 32 weeks after the end
of exposure.  Both of the two higher concentrations resulted in significant increases in mutant
frequency. Whereas the mutant frequency from the 7.1 mg/m3 group returned to control levels
by 12 weeks, the mutant frequency of the high-exposure group was  still higher than controls
even after 32 weeks.  Carbon black particles have very little adsorbed PAHs, hence a direct
chemically induced mechanism is highly unlikely. Induction of hprt mutations were also
observed in  rat alveolar epithelial cells after intratracheal instillation with carbon black, a-
quartz, and titanium dioxide (Driscoll et al., 1997). All three types of particles elicited an
inflammatory response as shown by significant increases of neutrophils in bronchoalveolar
lavage (BAL) fluid. Culturing the BAL from exposed rats with a rat lung epithelial cell line
also resulted in elevation of hprt mutational response. This response was effectively eliminated
when catalase was included in the incubation mixture, providing evidence for cell-derived
oxidative damage. Recently, Sato et al. (2000) exposed male Big Blue transgenic F344 rats to
diluted DE (1 and 6 mg/m3 suspended particle concentration) for 4 weeks.  Mutant frequency in
lung DNA was significantly elevated (4.8x control) at 6 mg/m3 but not at 1 mg/m3. Lung DNA
adduct levels measured by 32P-postlabeling and 8-hydroxydeoxyguanosine measured by HPLC

                                           4-4

-------
were elevated at both particle concentrations, but to a lesser extent than mutant frequencies.
Sequence analysis of mutants indicated that some, but not all, of the mutations could be
explained by an oxidative damage mechanism.
       Specific-locus mutations were not induced in (C3H x 101^ male mice exposed to DE 8
h/day, 7 days/week for either 5 or 10 weeks (Russell et al., 1980).  The exhaust was a 1:18
dilution and the average particle concentration was 6 mg/m3. After exposure, males were mated
to T-stock females and matings continued for the reproductive life of the males. The results
were unequivocally negative; no mutants were detected in 10,635 progeny derived from
postspermatogonial cells or in 27,917 progeny derived from spermatogonial cells.
       Hou et al. (1995) measured DNA adducts and hprt mutations in peripheral lymphocytes
of 47 bus maintenance workers and 22 control individuals. All were nonsmoking men from
garages in the Stockholm area and the exposed group consisted of 16 garage workers, 25
mechanics, and 6 other garage workers. There were no exposure data, but the three groups were
considered to be of higher to lower exposure to diesel engine exhaust. Levels of DNA adducts
determined by 32P-postlabeling were significantly higher in workers than controls (3.2 versus 2.3
x 10"8), but hprt mutant frequencies were not different 8.6 versus 8.4 x 10"6). Although group
mean mutant frequencies were not different, both adduct level and mutagenicity were highest
among the 16 most exposed and mutant frequency was significantly correlated with adduct level.
All individuals were genotyped for glutathione transferase GSTM1 and aromatic amino
transferase NAT2 polymorphism.  Neither GSTM1 nulls nor NAT2 slow acetylators exhibited
effects on either DNA adducts or hprt mutant frequencies.

4.2. CHROMOSOME EFFECTS
       Mitchell et al. (1981) and Brooks et al. (1984) reported increases in sister chromatid
exchanges (SCE) in CHO cells exposed to DPM extracts of emissions from both LD and HD
diesel engines. Morimoto et al. (1986) observed increased SCE from both LD and HD DPM
extracts in PAH-stimulated human lymphocyte cultures. Tucker et al. (1986) exposed human
peripheral lymphocyte cultures from four donors to  direct DE for up to 3 h. Exhaust was cooled
by pumping through a plastic tube about 20 feet long; airflow was  1.5 L/min. Samples were
taken at 16, 48, and 160 min of exposure.  Cell cycle delay was observed in all  cultures;
significantly increased SCE levels were reported for two of the four cultures. Structural
chromosome aberrations were induced in CHO cells by DPM extracts from a Nissan diesel
engine (Lewtas, 1983) but not by similar extracts from an Oldsmobile diesel engine (Brooks et
al., 1984).
       DPM dispersed in an aqueous mixture containing dipalmitoyl lecithin (DPL), a
component of pulmonary surfactant or extracted with dichloromethane (DCM)  induced similar

                                          4-5

-------
responses in SCE assays in Chinese hamster V79 cells (Keane et al., 1991), micronucleus tests in
V79 and CHO cells (Gu et al., 1992), and unscheduled DNA synthesis (UDS) in V79 cells (Gu
et al., 1994).  After separating the samples into supernatant and sediment fractions, mutagenic
activity was confined to the sediment fraction of the DPL sample and the supernatant of the
DCM sample. These findings suggest that the mutagenic activity of DPM inhaled into the lungs
could be made bioavailable through solubilization and dispersion of pulmonary surfactants.  In a
later study in the same laboratory, Liu et al. (1996) found increased micronuclei in V79 cells
treated with crystalline quartz and a noncrystalline silica, but response was reduced after
pretreatment of the particles with the simulated pulmonary surfactant.
       Pereira et al. (1981a) exposed female Swiss mice to DE 8 h/day,  5 days/week for 1, 3,
and 7 weeks.  The incidence of micronuclei and structural aberrations was similar in bone
marrow cells  of both control and exposed mice.  Increased incidences of micronuclei, but not
SCE, were observed in bone marrow cells of male Chinese hamsters after 6 months of exposure
to DE (Pereira et al., 1981b).
       Guerrero et al. (1981) observed a linear concentration-related increase in SCE in lung
cells cultured after intratracheal instillation of DPM at doses up to 20 mg/hamster. However,
they did not observe any increase in SCE after 3 months of inhalation exposure to DE particles
(6 mg/m3).
       Pereira et al. (1982) measured SCE in embryonic liver cells of Syrian hamsters. Pregnant
females were exposed to DE diluted with air 1:9 to contain about 12 mg/m3 particles from days 5
to 13 of gestation or injected intraperitoneally with diesel particles or particle extracts on
gestational day 13 (18 h before sacrifice). Neither the incidence of SCE nor mitotic index was
affected by exposure to DE. The injection of DPM extracts but not DPM resulted in a dose-
related increase in SCE; however, the toxicity of the DPM was about twofold greater than the
DPM extract.
       In the only studies with mammalian germ cells, Russell et al. (1980) reported no increase
in either dominant lethals or heritable translocations in males of T-stock mice exposed by
inhalation to diesel emissions. In the dominant lethal test, T-stock males were exposed for 7.5
weeks and immediately mated to females of different genetic backgrounds (T-stock; [C3H x
101]; [C3H x C57BL/6]; [SEC x C57BL/6]). There were no differences from controls in any of
the parameters measured in this assay.  For heritable translocation analysis,  T-stock males were
exposed for 4.5 weeks and mated to (SEC x C57BL/6) females, and the Fx males were tested for
the presence of heritable translocations. Although no translocations were detected among 358
progeny tested, the historical control incidence is less than 1/1,000.
                                          4-6

-------
4.3.  OTHER GENOTOXIC EFFECTS
       Pereira et al. (1981b) exposed male strain A mice to DE emissions for 31 or 39 weeks
using the same exposure regimen noted in the previous section. Analyses of caudal sperm for
sperm-head abnormalities were conducted independently in three separate laboratories.
Although the incidence of sperm abnormalities was not significantly above controls in any of the
three laboratories, there were extremely large differences in scoring among the three (control
values were 9.2%, 14.9%, and 27.8% in the three laboratories). Conversely, male Chinese
hamsters exposed for 6 mo (Pereira et al., 1981c) exhibited almost a threefold increase in sperm-
head abnormalities.  It is noted that the control incidence in the Chinese hamsters was less than
0.5%. Hence, it is not clear whether the differing responses reflect true species  differences or
experimental artifacts.
       A number of studies measuring DNA adducts in animals exposed to DPM, carbon black
or other particles have been reported and are reviewed by Shirname-More (1995). Although
modest increases in DNA adducts have been observed in lung tissue of rats after inhalation of
DPM (Wong et al., 1986; Bond et al., 1990), the magnitude of the increases is small in
comparison with those induced by  chemical carcinogens present in DE (Smith et al., 1993).
While Gallagher et al. (1994) found no increases in total DNA adducts in lung tissue of rats
exposed to DE, carbon black, or titanium dioxide they did  observe an increase in an adduct with
migration properties similar to nitrochrysene and nitro-benzo(a)pyrene adducts from diesel but
not carbon black or titanium dioxide exposures. The majority of the studies used the 32P-
postlabeling assay to detect adducts.  Although this method is sensitive, chemical identity of
adducts can only be inferred if an adduct spot migrates to the same location as a known prepared
adduct.
       DNA adducts have also been measured in humans occupationally exposed to DE.
Distinct adduct patterns were found among garage workers occupationally exposed to DE when
compared to nonexposed controls (Nielsen and Autrup, 1994). Furthermore, the findings were
concordant with the adduct patterns observed in groups exposed to low concentrations of PAHs
from combustion processes. Hemminki et al. (1994) also reported significantly elevated levels
of DNA adducts in lymphocytes from garage workers with known DE exposure compared with
unexposed mechanics. Hou et al. (1995) found elevated adduct levels in bus maintenance
workers exposed to DE. Although no difference in mutant frequency was observed between the
groups, the adduct levels were significantly different (3.2 vs. 2.3 x 10"8). Nielsen et al. (1996)
reported  significantly increased levels of three biomarkers (lymphocyte DNA adducts,
hydroxyethylvaline adducts in hemoglobin, and 1-hydroxypyrene in urine) in DE-exposed bus
garage workers.
                                          4-7

-------
       The role of oxidative damage in causing mutations has received increasing focus
recently. More than 50 different chemicals have been studied in rodents usually measuring the
formation of 8-hydroxydeoxyguanosine (8-OH-dG), a highly mutagenic adduct (Loft et al.,
1998). Increases in the mutagenic DNA adduct 8-hydroxydeoxyguanosine were found in mouse
lung DNA after intratracheal instillation of diesel particles (Nagashima et al., 1995). The
response was dose dependent. Mice fed on a high-fat diet showed an increased response
whereas the responses were partially reduced when the antioxidant,  p-carotene, was included in
the diet (Ichinose et al., 1997). Oxidative damage also has been measured in rat lung tissue after
intratracheal instillation of quartz (Nehls  et al., 1997) and in rat alveolar macrophages after in
vitro treatment with silica dust (Zhang et al., 2000).  Arimoto et al. (1999) demonstrated that
redissolved methanol extracts of DPM also induced the formation of 8-OH-dG adducts in L120
mouse cells.  The response was dependent on both DPM concentration and P450 reductase. A
detailed discussion of the potential role of oxidative damage in DE carcinogenesis is presented in
Chapter 7, Section 7.4.

4.4.  SUMMARY AND DISCUSSION
       Extensive studies with salmonella have unequivocally demonstrated mutagenic activity
in both particulate and gaseous fractions of DE. In most of the studies using salmonella,  DPM
extracts and individual nitropyrenes exhibited the strongest responses  in strain TA98 when no
exogenous activation was provided. Gaseous fractions reportedly showed greater response in
TA100, whereas benzo[a]pyrene and other unsubstituted PAHs are mutagenic only in the
presence of S9 fractions. The induction of gene mutations has been reported in several in vitro
mammalian cell lines after exposure to extracts of DPM. Note that only the TK6 human  cell line
did not give a positive response to DPM extracts in the absence of S9  activation. Mutagenic
activity was recovered in urine from animals treated with DPM by gastric intubation and i.p. and
s.c. implants, but not by inhalation of DPM or diluted diesel exhaust.  Dilutions of whole diesel
exhaust did not induce sex-linked recessive lethals in drosophila or specific-locus mutations in
male mouse germ cells.
       Structural chromosome aberrations and SCE in mammalian cells have been induced by
particles and extracts. Whole exhaust induced micronuclei but not SCE or structural aberrations
in bone marrow of male Chinese hamsters exposed to whole diesel emissions for 6 mo. In a
shorter exposure (7 weeks), neither micronuclei nor  structural aberrations were increased in bone
marrow of female Swiss mice. Likewise, whole DE did not induce dominant lethals or heritable
translocations in male mice exposed for 7.5 and 4.5 weeks, respectively.
       The application of mutagenicity data to the question of the potential carcinogenicity of
diesel emissions is based on the premise that genetic alterations are found in all cancers and that

                                           4-8

-------
several of the chemicals found in diesel emissions possess mutagenic activity in a variety of
genetic assays.  These genetic alterations can be produce by gene mutations, deletions,
translocations, aneuploidy, or amplification of genes, hence no single genotoxicity assay should
be expected to either qualitatively or quantitatively predict rodent carcinogenicity. With diesel
emissions or other mixtures, additional complications arise because of the complexity of the
material being tested. Exercises that combined the salmonella mutagenic potency with the total
concentration of mutagenic chemicals deposited in the lungs could not account for the observed
tumor incidence in exposed rats (Rosenkranz, 1993; Goldstein et al., 1998).  However, such
calculations ignored the contribution of gaseous phase chemicals which have been estimated to
contribute from less than 50% (Rannug et al., 1983) to over 90% (Matsushita et al., 1986) of the
total mutagenicity.  This wide range is partly reflective of the differences in material tested,
semivolatile extracts in the former and whole gaseous emission in the latter.  Of greater
importance is that these calculations are based on a reverse mutation assay in bacteria with
metabolic processes strikingly different from mammals.  This is at least partly reflected in the
observations that  different nitro-PAHs give different responses in bacteria and in CHO cells (Li
and Butcher, 1983) or in human hepatoma-derived cells (Eddy et al.,  1986).

                             REFERENCES FOR CHAPTER 4
Albert, RE; Lewtas, J; Nesnow, S; et al. (1983) Comparative potency method for cancer risk assessment:
application to diesel paniculate emissions. Risk Anal 3:101-117.
Arimoto, T; Yoshikawa, T; Takano, H; et al. (1999) Generation of reactive oxygen species and 8-hydroxy-2'-
deoxyguanosine formation from diesel exhaust particle components in L1210 cells. Jpn J Pharmacol 80:49-54.
Barfknecht, TR; Hites, RA; Cavaliers, EL; et al. (1982) Human cell mutagenicity of polycyclic aromatic
hydrocarbon components of diesel emissions. In: Toxicological effects of emissions from diesel engines:
proceedings of the Environmental Protection Agency 1981 diesel emissions symposium; October 1981; Raleigh, NC.
(Developments on toxicology and environmental science: v. 10.) Lewtas, J, ed. New York: Elsevier Biomedical; pp.
277-294.
Belisario, MA; Buonocore, V; De Marinis, E; et al. (1984) Biological availability of mutagenic compounds adsorbed
onto diesel exhaust paniculate. Mutat Res 135:1-9.
Bond, JA; Mauderly, JL; Wolff, RK. (1990) Concentration- and time-dependent formation of DNA adducts in lungs
of rats exposed to diesel exhaust. Toxicology 60:127-135.
Brooks, A; Li, AP; Butcher, JS; et al. (1984) A comparison of genotoxicity of automobile exhaust particles from
laboratory and environmental sources. Environ Mutagen 6:651-668.
Brusick, D. (1987) Principles of genetic toxicology. New York; Plenum Press.
Casto, BC; Hatch, GG; Huang, SL; et al. (1981) Mutagenic and carcinogenic potency of extracts of diesel and
related environmental emissions: in vitro mutagenesis and oncogenic transformation. Environ Int 5:403-409.
                                              4-9

-------
Chescheir, GM, III; Garrett, NE; Shelburne, JD; et al. (1981) Mutagenic effects of environmental particulates in the
CHO/HGPRT system. In: Application of short-term bioassays in the fractionation and analysis of complex
environmental mixtures. Waters, MD; Sandhu, SS; Huisingh, JL; et al., eds. New York: Plenum Press; pp. 337-350.

Claxton, LD.  (1981) Mutagenic and carcinogenic potency of diesel and related environmental emissions: Salmonella
bioassay. Environ Int 5:389-391.

Claxton, LD.  (1983) Characterization of automotive emissions by bacterial mutagenesis bioassay: a review. Environ
Mutagen 5:609-631.

Claxton, L; Kohan, M. (1981) Bacterial mutagenesis and the evaluation of mobile-source emissions. In: Short-term
bioassays in the analysis of complex environmental mixtures II: proceedings of the second symposium on the
application of short-term bioassays in the fractionation and analysis of complex environmental mixtures; March
1980; Williamsburg, VA. (Hollaender, A; Welch, BL; Probstein, RF, eds. Environmental science research series: v.
22.) Waters, MD; Sandhu, SS; Huisingh, JL; et al., eds. New York: Plenum Press;  pp. 299-317.

Crebelli, R; Conti, L; Crochi, B; et al. (1995) The effect of fuel composition on the mutagenicity of diesel engine
exhaust. MutatRes 346:167-172.

Curren, RD; Kouri, RE; Kim, DM; et al. (1981) Mutagenic and carcinogenic potency of extracts from diesel related
environmental emissions: simultaneous morphological transformation and mutagenesis in BALB/c 3T3 cells.
Environ Int 5:411-415.

Driscoll, KE;  Carter, JM; Howard, BW; et al. (1996) Pulmonary inflammatory, chemokine, and mutagenic responses
in rats after subchronic inhalation of carbon black.  Toxicol Appl Pharmacol 136:372-380.

Driscoll, KE;  Deyo, KC;  Carter, JM; et al. (1997) Effects of particle exposure and  particle-elicited inflammatory
cells on mutation in rat alveolar epithelial cells. Carcinogenesis 18:423-430.

Dukovich, M; Yasbin, RE; Lestz, SS; et al. (1981)  The mutagenic and SOS-inducing potential of the soluble organic
fraction collected from diesel paniculate emissions. Environ Mutagen 3:253-264.

Eddy, EP; McCoy, EC; Rosenkranz, HS; et al. (1986) Dichotomy in the mutagenicity and genotoxicity of
nitropyrenes:  apparent effect of the number of electrons involved in nitroreduction. Mutat Res 161:109-111.

Gallagher, J; Heinrich, U; George, M; et al.  (1994) Formation of DNA adducts in  rat lung following chronic
inhalation of diesel emissions, carbon black and titanium dioxide particles. Carcinogenesis 15:1291-1299.

Goldstein, LS; Weyand, EH; Safe, S; et al. (1998) Tumors and DNA adducts in mice exposed to benzo[a]pyrene and
coal tars: implications for risk assessment. Environ Health Perspect 106 Suppl 6:1325-1330.

Gu, ZW; Zhong, BZ; Nath, B; et al. (1992) Micronucleus induction and phagocytosis in mammalian cells treated
with diesel emission particles. Mutat Res 279:55-60.

Gu, ZW; Zhong, BZ; Keane, MJ; et al. (1994) Induction of unscheduled DNA synthesis in V70 cells by diesel
emission particles dispersed in simulated pulmonary surfactant. Ann Occup Hyg 38:345-349.

Guerrero, RR; Rounds, DE; Orthoefer, J. (1981) Sister chromatid exchange analysis  of Syrian hamster lung cells
treated in vivo with diesel exhaust particulates. Environ Int 5:445-454.

Hemminki, K; Soderling, J; Ericson, P; et al. (1994) DNA adducts among personnel  servicing and loading diesel
vehicles. Carcinogenesis  15:767-769.

Hinners, RG;  Burkart, JK; Malanchuk, M. (1980) Facilities for diesel exhaust studies. In: Health effects of diesel
engine emissions: proceedings of an international symposium; December 1979. Pepelko, WE; Danner, RM; Clarke,
                                                  4-10

-------
NA, eds. Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp. 681-697;
EPA report no. EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.

Hou, S; Lambert, B; Hemminki, K. (1995) Relationship between hprt mutant frequency, aromatic DNA adducts and
genotypes for GSTM1 andNAT2 in bus maintenance workers.  Carcinogensis 16:1913-1917.

Huisingh, J; Bradow, R; lungers, R; et al. (1978) Application of bioassay to the characterization of diesel particle
emissions. In: Application of short-term bioassays in the fractionation and analysis of complex environmental
mixtures: [proceedings of a symposium; February; Williamsburg, VA].  (Hollaender, A; Probstein, F; Welch, BL,
eds. Environmental science research: v. 15.) Waters, MD; Nesnow, S; Huisingh, JL; et al., eds. New York: Plenum
Press; pp. 383-418.

IARC (International Agency for Research on Cancer). (1989) Diesel and gasoline engine exhausts and some
nitroarenes. (IARC monographs on the evaluation of carcinogenic risks to humans: v. 46).  Lyon, France: World
Health Organization; pp. 41-185.

Ichinose, T; Yajima, Y; Nagashima, M; et al. (1997) Lung carcinogenesis and formation of 8-hydroxy-
deoxyguanosine in mice by diesel exhaust particles.  Carcinogenesis 18:185-192.

Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. (1986) Carcinogenic and mutagenic effects of diesel engine
exhaust: proceedings of the international satellite symposium on lexicological effects of emissions from diesel
engines; July; Tsukuba Science City, Japan. (Developments in toxicology and environmental science: v. 13.)
Amsterdam: Elsevier Science Publishers BV.

Keane, MJ; Xing, SG; Harrison, JC; et al. (1991) Genotoxicity of diesel-exhaust particles dispersed in simulated
pulmonary surfactant. MutatRes 260:233-238.

Lewtas, J. (1982) Mutagenic activity of diesel emissions. In: lexicological effects of emissions from diesel engines:
proceedings of the Environmental Protection Agency 1981  diesel emissions symposium; October 1981; Raleigh, NC.
(Developments in toxicology and environmental science: v. 10.) Lewtas, J, ed.  New York: Elsevier Biomedical; pp.
243-264.

Lewtas, J. (1983) Evaluation of the mutagenicity and carcinogenicity of motor vehicle emissions in short-term
bioassays. Environ Health Perspect 47:141-152.

Li, AP; Dutcher, JS. (1983) Mutagenicity of mono-,  di-, and tri-nitropyrenes in Chinese hamster ovary cells. Mutat
Res 119:387-392.

Li, AP; Royer, RE. (1982) Diesel-exhaust-particle extract enhancement of chemical-induced mutagenesis in cultured
Chinese hamster ovary cells: possible interaction of  diesel exhaust with environmental chemicals. Mutat Res
103:349-355.

Liber, HL; Andon, BM; Kites, RA; et al. (1981) Diesel soot: mutation measurements in bacterial and human cells.
Environ Int 5:281-284.

Liberti, A; Ciccioli, P; Cecinato,  A; et al. (1984) Determination of nitrated-polyaromatic hydrocarbons (nitro-PAHs)
in environmental samples by high resolution chromatographic techniques. J High Resolut Chromatogr Commun
7:389-397.

Liu, X; Keane, MJ; Zong, BZ; et al. (1996) Micronucleus formation in V79 cells treated with respirable silica
dispersed in medium and simulated pulmonary surfactant. Mutat Res 361:89-94.

Loft, S; Deng, XS; Tuo, J; et al. (1998) Experimental study of oxidative DNA damage. Free Radical Res 29:525-
539.
                                                 4-11

-------
Matsushita, H; Goto, S; Endo, O; et al. (1986) Mutagenicity of diesel exhaust and related chemicals. In:
Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the international satellite symposium on
toxicological effects of emissions from diesel engines; July; Tsukuba Science City, Japan. (Developments on
toxicology and environmental science: v. 13.) Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. Amsterdam:
Elsevier Science Publishers BV; pp. 103-118.

Mitchell, AD; Evans, EL; Jotz, MM; et al. (1981) Mutagenic and carcinogenic potency of extracts of diesel and
related environmental emissions: in vitro mutagenesis and DNA damage. Environ Int 5:393-401.

Morimoto, K; Kitamura, M; Kondo, H; et al. (1986) Genotoxicity of diesel exhaust emissions in a battery of in-vitro
short-term bioassays. In: Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the
international satellite symposium on toxicological effects of emissions from diesel engines; July; Tsukuba Science
City, Japan. (Developments in toxicology and environmental science: v. 13.) Ishinishi, N; Koizumi, A; McClellan,
RO; et al., eds. Amsterdam: Elsevier Science Publishers BV; pp. 85-102.

Nakagawa, R; Kitamori, S; Horikawa, K; et al.  (1983) Identification of dinitropyrenes in diesel-exhaust particles:
their probable presence as the major mutagens.  Mutat Res 124:201-211.

Nagashima, M; Kasai, H; Yokota, J; et al.  (1995) Formation of an oxidative DNA damage, 8-
hydroxydeoxyguanosine, in mouse lung DNA after intratracheal instillation of diesel exhaust particles and effects of
high dietary fat and beta-carotene on this process. Carcinogenesis 16:1441-1445.

Nehls, P; Seiler, F; Rehn, B; et al. (1997) Formation and persistence of 8-oxoguanine in rat lung cells as an
important determinant in tumor formation following particle exposure. Environ Health Perspect 105(5):1291-1296.

Nielsen, PS; Autrup, H. (1994) Diesel exhaust-related DNA adducts in garage workers. Clin Chem 40:1456-1458.

Nielsen, PS; Andreassen, A; Farmer, PB; et al.  (1996) Biomonitoring of diesel-exhaust exposed workers. DNA and
hemoglobin adducts and urinary  1-hydroxyproline as markers of exposure. Toxicol Lett 86:27-37.

Nishioka, MG; Petersen, BA; Lewtas, J. (1982) Comparison of nitro-aromatic content and direct-acting mutagenicity
of diesel emissions. In: Polynuclear aromatic hydrocarbons: physical and biological chemistry. Cooke, M; Dennis,
AJ; Fisher, GL, eds. Columbus, OH: Battelle Press; pp. 603-613.

Ong, T; Whong,  WZ; Xu, J; et al. (1985) Genotoxicity studies of rodents exposed to coal dust and diesel emission
particulates. Environ Res 37:399-409.

Pepelko, WE; Peirano, WB. (1983) Health effects of exposure to diesel engine emissions: a summary of animal
studies conducted by the U.S. Environmental Protection Agency's Health Effects Research Laboratories at
Cincinnati, Ohio. J Am Coll Toxicol 2:253-306.

Pereira, MA; Connor, TH; Meyne, J; et al. (1981a) Metaphase analysis, micronucleus assay and urinary
mutagenicity assay of mice exposed to diesel emissions. Environ Int 5:435-438.

Pereira, MA; Sabharwal, PS; Gordon,  L; et al. (1981b) The effect of diesel exhaust on sperm-shape abnormalities in
mice.  Environ Int 5:459-460.

Pereira, MA; Sabharwal, PS; Kaur, P; etal. (1981c) In vivo detection of mutagenic effects of diesel exhaust by
short-term mammalian bioassays. Environ Int 5:439-443.

Pereira, MA; McMillan, L; Kaur, P; et al. (1982) Effect of diesel exhaust emissions, particulates, and extract on
sister  chromatid exchange in transplacentally exposed fetal  hamster liver. Environ Mutagen 4:215-220.

Rannug, U; Sundvall, A; Westerholm, R; et al.  (1983) Some aspects of mutagenicity testing of the paniculate phase
and the gas phase of diluted and  undiluted automobile exhaust. Environ Sci Res 27:3-16.
                                                  4-12

-------
Rosenkranz, HS. (1993) Revisiting the role of mutagenesis in the induction of lung tumors in rats by diesel
emissions. Mutat Res 303:91-95.

Russell, LB; Generoso, WM; Oakberg, EF; et al. (1980) Tests for heritable effects induced by diesel exhaust in the
mouse. Martin Marietta Energy Systems, Inc., Oak Ridge National Laboratory; report no. ORNL-5685.

Sato, H; Sone, H; Sagai, M; et al. (2000) Increase in mutation frequency in lung of Big Blue(R) rat by exposure to
diesel exhaust. Carcinogenesis 21:653-661.

Salmeen, I; Durisin, AM; Prater, TJ; et al. (1982) Contribution of 1-nitropyrene to direct-acting Ames assay
mutagenicities of diesel paniculate extracts. Mutat Res 104:17-23.

Schuetzle, D; Frazier, JA. (1986) Factors influencing the emission of vapor and paniculate phase components from
diesel engines. In: Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the international
satellite symposium on lexicological effects of emissions from diesel engines; July; Tsukuba Science City, Japan.
(Developments in toxicology and environmental science: v. 13.) Ishinishi, N; Koizumi, A; McClellan, RO; et al.,
eds. Amsterdam: Elsevier Science Publishers BV; pp. 41-63.

Schuetzle, D; Lewtas, J. (1986) Bioassay-directed chemical analysis in environmental research. Anal Chem
58:1060A-1076A.

Schuetzle, D; Perez, JM. (1983) Factors influencing the emissions of nitrated-polynuclear aromatic hydrocarbons
(nitro-PAH)  from diesel engines. J Air Pollut Control Assoc 33:751-755.

Schuler, RL; Niemeier, RW. (1981) A study of diesel emissions on Drosophila. Environ Int 5:431-434.

Shirname-More, L.  (1995) Genotoxicity of diesel emissions. Part I: Mutagenicity and other genetic effects. Diesel
exhaust: a critical analysis of emissions, exposure, and health effects. A special report of the Institute's Diesel
Working Group.  Cambridge, MA: Health Effects Institute, pp. 222-242.

Siak, JS; Chan, TL; Lees, PS. (1981) Diesel paniculate extracts in bacterial test systems. Environ Int 5:243-248.

Sjogren, M;  Li, H; Banner, C; et al.  (1996) Influence of physical and chemical characteristics of diesel fuels and
exhaust emissions on biological effects of particle extracts: a multivariate statistical analysis often diesel fuels.
Chem Res Toxicol 9:197-207.

Smith, BA; Fullerton, NF; Aidoo, A; et al. (1993) DNA adduct formation in relation to lymphocyte mutations and
lung tumor induction in F344 rats treated with the environmental pollutant 1,6- dinitropyrene. Environ Health
Perspect 99:277-280.

Tucker, JD; Xu, J; Stewart, J; et al. (1986) Detection of sister chromatid exchanges induced by volatile
genotoxicants. Teratogen Carcinogen Mutagen 6:15-21.

U.S. Environmental Protection Agency (EPA). (1980) Health effects of diesel engine emissions: proceedings of an
international symposium. Cincinnati, OH: Office of Research and Development; EPA 600/9-80/057b.

Wong, D; Mitchell, CE; Wolff, RK; etal. (1986) Identification of DNA damage as a result of exposure of rats to
diesel engine exhaust. Carcinogenesis  7:1595-1597.

Zhang, Z; Shen, HM; Zhang, QF; et al. (2000) Involvement of oxidative stress in crystalline silica-induced
cytotoxicity  and genotoxicity in rat alveolar macrophages. Environ Res 82:245-252.
                                                  4-13

-------
             5. NONCANCER HEALTH EFFECTS OF DIESEL EXHAUST

       The objective of this chapter is to review and evaluate potential health effects other than
cancer associated with inhalation exposure to diesel engine exhaust (DE). Data have been
obtained from diverse human, laboratory animal, and in vitro test systems.  The human studies
comprise both occupational and human experimental exposures, the former consisting of
exposure to DE in the occupational environment, and the latter consisting of exposure to diluted
DE or diesel particulate matter (DPM) under controlled conditions. The laboratory animal
studies consist of both acute and chronic exposures of laboratory animals to DE or DPM.
Diverse in vitro test systems composed of human and laboratory animal cells treated with DPM
or components of DPM have also been used to investigate the effects of DPM at the cellular and
molecular levels. DPM mass (mg/m3) has been used almost exclusively as a measure of DE
exposure in human and experimental studies.  The noncancer health effects of DPM have been
reviewed previously by the Health Effects Institute (HEI, 1995) and in the Air Quality for
Particulate Matter Criteria Document, the PM CD  (U.S. EPA, 1996).  The noncancer health
effects attributable to ambient particulate matter (PM), which is composed in part of DPM, as
well as the potential mechanisms underlying these effects have also been previously reviewed in
the PM CD (U.S. EPA, 1996) and have been summarized in this document in Chapter 6, Section
6.4.
       This chapter begins with  descriptions of studies that have shown various health effects
occurring as a result of exposure to DE/DPM (Section 5.1). The human studies portion of this
section (5.1.1) discusses results from both short-term and long-term studies as well as
specialized studies such as those of populations contiguous to major highways (5.1.2). Studies
using laboratory animals are ordered into various subsections under Section 5.1.3.
Investigations devoted to elucidating the possible modes of action of DE/DPM are covered in
Section 5.2; the mode-of-action issue of particle overload in animals is discussed elsewhere in
the document (Chapter 3,  Section 3.4). Section 5.3 describes evidence for the various
interactions of DPM with  other conditions such as  disease. Other sections address issues such as
species-comparative responses to DE/DPM (Section  5.4) and influence of dose rate (Section
5.5).  The summary/conclusion of this chapter, relating the totality of this information to
possible human effects of DE/DPM, is in Section 5.6.
                                          5-1

-------
5.1. HEALTH EFFECTS OF WHOLE DIESEL EXHAUST
5.1.1. Human Studies
5.1.1.1. Short- Term Exposures
       In a controlled human study, Rudell et al. (1990, 1994) exposed eight healthy subjects in
an exposure chamber to diluted exhaust from a diesel engine for 1 h, with intermittent exercise.
Dilution of the DE was controlled to provide a median NO2 level of approximately 1.6 ppm.
Median particle number was 4.3 x lOVcm3, and median levels of NO and CO were 3.7 and 27
ppm, respectively (particle size and mass concentration were not provided). There were no
effects on spirometry or on closing volume using nitrogen washout. Five of eight subjects
experienced unpleasant smell, eye irritation, and nasal irritation during exposure.
Bronchoalveolar  lavage (BAL) was performed 18 hours after exposure and was compared with a
control BAL performed 3 weeks prior to exposure; there was no control air exposure.  Small but
statistically significant reductions were seen in numbers of BAL mast cells, extent of AM
phagocytosis of opsonized yeast particles, and lymphocyte CD4/CD8 ratios. A small increase in
recovery of polymorphonuclear cells (PMNs) was also observed. These findings suggest that
DE may induce mild airway inflammation in the absence of spirometric changes.  This study
provides an intriguing glimpse of the effect of DE exposure in humans, but only one exposure
level was used, the number of subjects was low, and a limited range of endpoints was reported,
so the data are inadequate to generalize about the human response.
       Rudell et  al. (1996) exposed  volunteers to DE for 1 h in an exposure chamber. Light
work on a bicycle ergometer was performed during exposure. Exposures included either DE or
exhaust with particle numbers reduced 46% by a particle trap. The engine used was a new
Volvo model 1990, a six-cylinder  direct-injection turbocharged diesel with an intercooler, which
was run at a steady speed of 900 rpm during the exposures.  Comparison of this study with
others was difficult because neither exhaust dilution ratios nor particle concentrations were
reported. Carbon monoxide concentrations of 27-30 ppm and NO of 2.6-2.7 ppm, however,
suggested DPM concentrations may  have equaled several mg/m3. The most prominent
symptoms during exposure were irritation of the eyes and nose and an unpleasant smell.  Both
airway resistance and specific airway resistance increased significantly during the exposures.
Despite the 46%  reduction in particle numbers by the trap,  effects on symptoms and lung
function were not significantly attenuated.
       Nordenhall et al. (2000) had  15 healthy human subjects (13 males, 2 females) breathe in
an exposure chamber diluted DE from an idling diesel engine to give a PM10 concentration of
300 |ig/m3, which was also associated with a median steady-state NO2 concentration of 1.6 ppm.
                                          5-2

-------
Exposures were for 1 h, with each individual serving as their own control by being exposed to
filtered air, also for 1 h but at a different time. Sputum production was then induced and sputum
examined at 6 and 24 hr postexposure (for both air and DPM) with differential cell counts and
soluble protein counts performed. In comparing the same individual's results after exposure to
air and after exposure to DE, increases were found in the percentage of sputum neutrophiles
(37.7% vs. 26.2%) after 6 hr, along with increases in concentrations of the soluble proteins
interleukin-6 (12.0 vs. 6.3 pg/mL) and methylhistamine (0.11 vs. 0.12 ug/L).  These differences
between air and DPM were not present at 24 hr.  Thus, breath exposure to DE produces early
induction of an inflammatory response in  healthy humans that can be detected using sputum
analysis.
       Wade and Newman (1993) describe the situation of three railroad workers who
developed persistent asthma associated with overexposure to DE from locomotives. The
overexposure was a consequence of multiple hours of high levels of diesel exposure from riding
in locomotive units trailing immediately behind the lead locomotive. Lines of evidence
supporting railroad locomotive DE inducing asthma in these individuals include, (1) all three
exhibited clear signs of asthma leading (in two of the three cases) to immediate first-time
hospitalization and treatment for asthma, (2) all three developed symptoms within a few hours of
the overexposure, and (3) all three experienced exacerbation of symptoms upon reexposure to
locomotive DE. Although this report and that of Kahn et al. (1988) described below both
provide supporting evidence for DE being able to cause asthma in humans under extreme but
uncharacterized conditions, both suffer from the same limitations, including no reliable data on
the concentration of diesel emissions and associated gaseous components, the duration of the
exposures, or information on others that were exposed under these conditions but who did not
develop asthma symptoms.
       Kahn et al. (1988) reported the occurrence of 13 cases of acute overexposure to DE
among Utah and Colorado coal miners. Twelve miners had symptoms of mucous membrane
irritation, headache, and lightheadedness.  Eight individuals reported nausea; four reported a
sensation of unreality; four reported heartburn; three reported weakness, numbness, and tingling
in their extremities; three reported vomiting; two reported chest tightness; and two others
reported wheezing. Each miner lost time from work because of these symptoms, which resolved
within 24 to 48 h. No air monitoring data were presented; poor work practices were described as
the predisposing conditions for overexposure.  No follow-up was available for these exposed
individuals.
       El Batawi and Noweir (1966) reported that among 161 workers from two garages where
diesel-powered buses were serviced and repaired, 42% complained of eye irritation, 37% of
headaches, 30% of dizziness, 19% of throat irritation, and 11% of cough and phlegm.  Ranges of
                                          5-3

-------
mean concentrations of DE components in the two diesel bus garages were as follows:  0.4 to 1.4
ppm NO2, 0.13 to 0.81 ppm SO2, 0.6 to 44.1 ppm aldehydes, and 1.34 to 4.51 mg/m3 of DPM;
the highest concentrations were obtained close to the exhaust systems of the buses.
       Eye irritation was reported by Battigelli (1965) in six subjects after 40 s of chamber
exposure to diluted DE containing 4.2 ppm NO2, 1 ppm SO2,  55 ppm CO, 3.2 ppm total
hydrocarbons, and 1 to 2 ppm total aldehydes; after 3 min and 20 s of exposure to diluted DE
containing 2.8 ppm NO2, 0.5 ppm SO2, 30 ppm CO, 2.5 ppm total hydrocarbons, and <1 to 2
ppm total aldehydes; and after 6 min of exposure to diluted DE containing 1.3 ppm NO2, 0.2
ppm SO2, <20 ppm CO, <2.0 ppm total hydrocarbons, and <1.0 ppm total aldehydes. The
concentration of DPM was not reported.
       Katz et al. (1960) described the experience of 14 chemists and their assistants monitoring
the environment of a train tunnel used by diesel-powered locomotives.  Although workers
complained on three occasions of minor eye and throat irritation, no correlation was established
with concentrations of any particular component of DE.
       The role of radicals generated from particulate matter, including DPM, in producing
toxicity has been discussed in the literature (Valavanidis et al., 2000), as has the role of
antioxidant defenses in protecting against species such as radicals that may arise from acute DE
exposure. Blomberg et al. (1998) investigated changes in the antioxidant defense network
within the respiratory tract lining fluids of human subjects following DE exposure. Fifteen
healthy, nonsmoking, asymptomatic subjects were exposed to filtered air or DE (DPM  300
|ig/m3) for 1 h on two separate occasions at least 3 weeks  apart. Nasal lavage fluid and blood
samples were collected prior to, immediately after, and 5 1A h post exposure. Bronchoscopy was
performed 6 h after the end of DE exposure. Nasal lavage ascorbic acid concentration increased
tenfold during DE exposure, but returned to basal levels 5.5 h postexposure. DE had no
significant effects on nasal lavage uric acid or GSH concentrations, and did not affect plasma,
bronchial wash, or bronchoalveolar lavage antioxidant concentrations, nor malondialdehyde or
protein carbonyl concentrations. The authors concluded that the physiological response to acute
DE exposure is an acute increase in the level of the antioxidant ascorbic acid in the nasal cavity.

5.1.1.1.1. Diesel exhaust odor. The odor of DE is considered by most people to be
objectionable; at high intensities, it may produce sufficient physiological and psychological
effects to warrant concern for public health.  The intensity of the odor of DE is an exponential
function of its concentration such that a tenfold change in the concentration will alter the
intensity of the odor by one unit.  Two human panel rating scales have been used to measure DE
odor intensity. In the first (Turk,  1967), combinations of odorous materials were selected to
simulate DE odor; a set of 12 mixtures, each having twice the concentration of that of the
                                          5-4

-------
previous mixture, is the basis of the diesel odor intensity scale (D-scale).  The second method is
the TIA (total intensity of aroma) scale based on seven steps, ranging from 0 to 3, with 0 being
undetectable, 1A very slight, and 1 slight and increasing in one-half units up to 3, strong (Odor
Panel of the CRC-APRAC Program Group on Composition of Diesel Exhaust, 1979; Levins,
1981).
       Surveys, utilizing volunteer panelists, have been taken to evaluate the general public's
response to the odor of DE. Hare and Springer (1971) and Hare et al. (1974) found that at a D
rating of about 2 (TIA = 0.9, slight odor intensity), about 90% of the participants perceived the
odor, and almost 60% found it objectionable. At a D rating of 3.2 (TIA = 1.2, slight to moderate
odor intensity), about 95% perceived the odor, and 75%  objected to it, and, at a D rating of 5
(TIA =1.8, almost moderate), about 95% objected to it.
       Linnell and Scott (1962) reported odor threshold  measurement in six subjects and found
that the dilution factor needed to reach the threshold ranged from 140 to 475 for this small
sample of people. At these dilutions, the concentrations  of formaldehyde ranged from 0.012 to
0.088 ppm.

5.1.1.1.2. Pulmonary and respiratory effects. Battigelli (1965) exposed 13 volunteers to three
dilutions of DE obtained from a one-cylinder, four-cycle, 7-hp diesel engine (fuel type
unspecified) and found that 15-min to 1-h exposures had no significant effects on pulmonary
resistance.  Pulmonary resistance was measured by plethysmography utilizing the simultaneous
recording of esophageal pressure and airflow determined by electrical differentiation of the
volume signal from a spirometer.  The concentrations of the constituents in the three diluted
exhausts were 1.3, 2.8, and 6.2  ppm NO2; 0.2, 0.5, and 1 ppm SO2; <20, 30, and 55 ppm CO;
and <1.0, <1 to 2, and 1 to 2 ppm total aldehydes, respectively. DPM concentrations were not
reported.
       A number of studies have evaluated changes in pulmonary function occurring over a
workshift in workers occupationally exposed to DE (specific  time period not always reported but
assumed to be 8 h). In a study of coal miners, Reger (1979) found that both forced expiratory
volume in 1 s (FEVj) and forced vital capacity (FVC) decreased by 0.05 L in 60 diesel-exposed
miners, an amount not substantially different from reductions seen in non-diesel-exposed miners
(0.02 and 0.04 L, respectively). Decrements in peak expiratory flow rates were similar between
diesel and non-DE-exposed miners. Although the monitoring data were not reported, the authors
stated that there was no relationship between the low concentrations of measured respirable dust
or NO2 (personal samplers) when compared with shift changes for any lung function parameter
measured for the diesel-exposed miners.  In summary, this study (available as an abstract only)
states that no evidence was found for additional lung function effect over a shift for miners
                                          5-5

-------
exposed to diesel emissions as compared with controls, i.e., nonexposed office workers and coal
miners not exposed to diesel emissions.
       Ames et al. (1982) compared the pulmonary function of 60 coal miners exposed to DE
with that of a control group of 90 coal miners not exposed to DE for evidence of acute
respiratory effects associated with exposure to DE. Changes over the workshift in FVC, FEVl3
and forced expiratory flow rate at 50% FVC (FEF50) were the indices for acute respiratory
effects.  The environmental concentrations of the primary pollutants were 2.0 mg/m3 respirable
dust (<10 |im MMAD), 0.2 ppm NO2, 12 ppm CO, and 0.3 ppm formaldehyde. The
investigators reported a statistically significant decline in FVC and FEVj over the workshift in
both the diesel-exposed and comparison groups.  Current smokers had greater decrements in
FVC, FEVl5 and FEF50 than did ex-smokers and nonsmokers. There was a marked disparity
between the ages and the time spent underground for the two study groups. Diesel-exposed
miners were about 15 years younger and had worked underground for 15 fewer years (4.8 versus
20.7 years) than miners not exposed to DE. The  significance to the results of these differences
between the populations is difficult to ascertain.
       Except for the expected differences related to age, 120 underground iron ore miners
exposed to DE had no workshift changes in FVC and FEVj when compared with 120 matched
surface miners (Jorgensen and Svensson, 1970).  Both groups had equal numbers (30) of
smokers and nonsmokers. The frequency of bronchitis was higher among underground workers,
much higher among smokers than nonsmokers, and also higher among older than younger
workers.  The authors reported that the underground miners had exposures of 0.5 to 1.5 ppm
NO2 and between 3 and 9 mg/m3 particulate matter, with 20% to 30% of the particles <5 jim
MMAD.  The majority of the particles were iron ore; quartz was 6% to 7% of the fraction
<5 |om MMAD.
       Gamble et al. (1979) measured preshift FEVj and FVC in 187 salt miners and obtained
peak flow forced expiratory flow rates at 25%, 50%, and 75% of FVC (FEF25, FEF50, or FEF75).
Postshift pulmonary function values were determined from total lung capacity and flows at
preshift percentages of FVC.  The miners were exposed to mean NO2 levels of 1.5 ppm and
mean respirable particulate levels of 0.7 mg/m3.  No statistically significant changes were found
between changes in pulmonary function and in NO2 and respirable particles combined.  Slopes
of the regression of NO2 and changes in FEVl3 FEF25, FEF50, and FEF75 were significantly
different from zero. The authors concluded that these small reductions in pulmonary function
were attributable to variations in NO2 within each of the five salt mines that contributed to the
cohort.
       Gamble et al. (1987a) investigated the acute effects of DE in 232 workers in four diesel
bus garages using an acute respiratory questionnaire and before and after workshift spirometry.
                                          5-6

-------
The prevalence of burning eyes, headaches, difficult or labored breathing, nausea, and wheeze
experienced at work was higher in the diesel bus garage workers than in a comparison
population of lead/acid battery workers who had not previously shown a statistically significant
association of acute symptoms with acid exposure.  Comparisons between the two groups were
made without adjustment for age and smoking. There was no detectable association of exposure
to NO2 (0.23 ppm ± 0.24 S.D.) or inhalable (less than  10 |im MMAD) particles (0.24 mg/m3 ±
0.26 S.D.) and acute reductions in FVC,  FEVj, peak flows, FEF50, and FEF75. Workers who had
respiratory symptoms had slightly greater but statistically insignificant reductions in FEVj and
FEF50.
       Ulfvarson et al. (1987) evaluated workshift changes in the pulmonary function of 17 bus
garage workers, 25 crew members of two types of car ferries, and 37 workers on roll-on/roll-off
ships.  The latter group was exposed primarily to DE;  the first two groups were exposed to both
gasoline and DE.  The diesel-only exposures that averaged 8 h consisted of 0.13 to 1.0 mg/m3
paniculate matter, 0.02 to 0.8 mg/m3 (0.016 to 0.65 ppm) NO, 0.06 to 2.3 mg/m3 (0.03 to  1.2
ppm) NO2,  1.1 to 5.1 mg/m3 (0.96 to 4.45 ppm) CO, and up to 0.5 mg/m3 (0.4 ppm)
formaldehyde. The largest decrement in pulmonary function was observed during a workshift
following no exposure to DE for 10 days. Forced vital capacity and FEVj were significantly
reduced over the workshift (0.44 L and 0.30 L,/><0.01 and/><0.001, respectively).  There  was
no difference between smokers and nonsmokers. Maximal midexpiratory flow,  closing volume
expressed as the percentage of expiratory vital capacity, and alveolar plateau gradient (phase 3)
were not affected. Similar but less pronounced effects on FVC (-0.16 L) were found in a
second, subsequent study of stevedores (n = 24) only following 5 days of no  exposure to diesel
truck exhaust.  Pulmonary function returned to normal after 3 days without occupational
exposure to DE.  No exposure-related correlation was  found between the observed pulmonary
effects and concentrations of NO, NO2, CO, or formaldehyde; however, it was suggested that
NO2 adsorbed onto the DE particles may have contributed to the overall  dose of NO2 to the
lungs.  In a related study, six workers (job category not defined) were placed in an exposure
chamber and exposed to  diluted DE containing 0.6 mg/m3 DPM and 3.9  mg/m3 (2.1 ppm)  NO2.
The exhaust was generated by a 6-cylinder, 2.38-L diesel engine, operated for 3 h and 40 min at
constant speed, equivalent to 60 km/h, and at about one-half full engine load. No effect on
pulmonary function was  observed.
       In a hypothesis-generating study, Kilburn (2000) examined neurobehavioral and
pulmonary function of a  small group of workers exposed to DE either as railroad workers  (n=10)
over a range of 15 to 50 years or as electricians (n=6)  over a range of 0.6 to 1.5 years.
Neurobehavioral and visual functions batteries showed nearly all of these individuals to be
neurobehaviorally impaired in relation to a referent population in one or more areas, including
                                          5-7

-------
reaction time, balance, blink reflex latency, verbal recall, and color vision confusion indices.
Pulmonary function tests also showed that 10 of the 16 had airway obstruction and another
group of 10 of the 16 had chronic bronchitis, chest pain, tightness, and hyperreactive airways.
This work implies that with sufficiently sensitive methods, noncancer effects from DPM/DE
exposure may be detectable in sufficiently exposed human populations.

5.1.1.1.3. Imnwnological effects. Salvi  et al. (1999) exposed healthy human subjects to diluted
DE (DPM 300 |ig/m3) for 1 h with intermittent exercise.  Although there were no changes in
pulmonary function, there were significant increases in neutrophils and B lymphocytes as well as
histamine and fibronectin in airway lavage fluid. Bronchial biopsies obtained 6 h after DE
exposure showed a significant increase in neutrophils, mast cells, and CD4+ and CD8+ T
lymphocytes, along with upregulation of the endothelial adhesion molecules ICAM-1 and
VCAM-1 and increases in the number of LFA-1+ in the bronchial tissue.  Significant increases
in neutrophils and platelets were observed in peripheral blood following exposure to DE.
       In a follow-up investigation of potential  mechanisms underlying the DE-induced airway
leukocyte infiltration, Salvi et al. (2000) exposed healthy human volunteers to diluted DE, on
two separate occasions for 1 h each, in an exposure chamber. Fiber-optic bronchoscopy was
performed 6 h after each exposure to obtain endobronchial biopsies and bronchial wash (BW)
cells. These workers observed that DE exposure enhanced gene transcription of  IL-8 in the
bronchial tissue and BW cells and increased growth-regulated oncogene-a protein expression and
IL-8 in the bronchial  epithelium; there was also  a trend toward an increase in IL-5 mRNA gene
transcripts in the bronchial tissue.
       In an attempt to evaluate the potential  allergenic effects of DPM in humans, Diaz-
Sanchez and associates carried out a series of clinical investigations.  In the first of these (Diaz-
Sanchez et al.,  1994), healthy human volunteers were challenged by spraying either saline or
0.30 mg (300 jig) DPM into their nostrils. The authors considered this  dose to be equivalent to
breathing the outdoor air in Los Angeles for a 24-h period on an average day. Enhanced IgE
levels were noted in nasal lavage cells in  as little as 24 h, with peak production observed 4 days
after DPM challenge. The effects seemed to be  somewhat isotype-specific, because in contrast
to IgE results, DPM challenge had no effect on the levels of IgG, IgA, IgM, or albumin. The
selective enhancement of local IgE production was demonstrated by a dramatic increase in IgE-
secreting cells.
       Although direct effects of DPM on B-cells have been demonstrated by in vitro studies,  it
was considered likely that other cells regulating the IgE response may also be affected.
Cytokine production was therefore measured in  nasal lavage cells from healthy human
volunteers challenged with DPM (0 or 0.15 mg in 200 jiL saline) sprayed into each nostril (Diaz-
                                          5-8

-------
Sanchez et al., 1996). Before challenge with DPM, most subjects' nasal lavage cells had
detectable levels of only interferon-y, IL-2, and IL-13 wRNA. After challenge with DPM, the
cells produced readily detectable levels of mRNA for IL-2, IL-4, IL-5, IL-6, IL-10, IL-13, and
interferon-y.  Although the cells in the nasal lavage before and after challenge do not necessarily
represent the same ones either in number or type, the broad increase in cytokine production was
considered by the  authors not to be simply the result of an increase in T cells recovered in the
lavage fluid. On the basis of these findings, the authors concluded that the increase in nasal
cytokine expression after exposure to DPM can be predicted to contribute to enhanced local IgE
production and thus play a role  in pollutant-induced airway disease.
       The ability of DPM to act as an adjuvant to the ragweed allergen Amb a I was also
examined by nasal provocation  in ragweed-allergic subjects using 0.3 mg (300 jig) DPM, Amb a
I, or both (Diaz-Sanchez et al.,  1997). Although allergen and DPM each enhanced ragweed-
specific IgE, DPM plus allergen promoted a 16-times greater antigen-specific IgE production.
Nasal challenge with  DPM also influenced cytokine production. Ragweed challenge resulted in
a weak response, DPM challenge caused a strong but nonspecific response, and allergen plus
DPM caused a significant increase in the expression of mRNA for THO and TH2-type cytokines
(IL-4, IL-5, IL-6, IL-10, IL-13), with a pronounced inhibitory effect on IFN-y gene expression.
The author concluded that DPM can enhance B-cell differentiation and, by initiating and
elevating IgE production, may be a factor in the increased incidence of allergic airway disease.
       In a further extension of these studies, Diaz-Sanchez et al. (1999) examined the potential
for DPM to lead to primary sensitization of humans by driving a de novo mucosal IgE response
to  a neoantigen, keyhole limpet hemocyanin (KLH). Ten atopic subjects were given an initial
nasal immunization of KLH followed by two biweekly nasal challenges with KLH. Fifteen
different atopic subjects were treated identically, except that DPM was administered 24 h before
each KLH exposure.  Intranasal administration of KLH alone led to the generation of an anti-
KLH IgG and IgA humoral response, which was detected in nasal fluid samples.  No anti-KLH
IgE was observed  in any of these  subjects. In contrast, when challenged with KLH preceded by
DPM, 9 of the 15  subjects produced anti-KLH-specific IgE. KLH-specific IgG and IgA at levels
similar to those seen with KLH alone were also detected. Subjects who received DPM and KLH
had significantly increased IL-4, but not IFN-gamma, levels in nasal lavage fluid, whereas these
levels were unchanged in subjects receiving KLH alone. These investigators concluded that
DPM can function as a mucosal adjuvant to a de novo IgE response and may increase allergic
sensitization among atopic individuals.

5.1.1.1.4.  Human cell culture  studies. The potential mechanisms by which DPM may act to
cause allergenic effects has been examined in human cell culture studies. Takenaka et al. (1995)
                                          5-9

-------
reported that DPM extracts enhanced IgE production from purified human B cells.  IgE
production in these cells (stimulated by exogenous addition of interleukin-4 plus monoclonal
antibody) was enhanced (i.e., further stimulated) 20% to 360% by the addition of DPM extracts
(1-50 ng/mL) over a period of 10-14 days.  DPM extracts in the absence of exogenously added
IL-4 and/or monoclonal antibodies did not themselves induce IgE production or synergize with
interleukin-4 alone to induce IgE from purified B cells, suggesting that the extracts were
enhancing ongoing IgE production rather than inducing germline transcription or isotype
switching. The authors concluded that enhancement of IgE production in the human airway
resulting from the organic fraction of DPM may be an important factor in the increasing
incidence of allergic airway disease.
       Terada et al. (1997) examined the effects of DPM and DPM extract on eosinophil
adhesion, survival rate, and degranulation. Eosinophils, human mucosal microvascular
endothelial cells (HMMECs), and human nasal epithelial cells (HNECs) were preincubated in
the presence of DPM and DPM extract. 35S-labeled eosinophils were allowed to adhere to
monolayers of HMMECs and HNECs. Although neither DPM nor DPM extract affected the
adhesiveness of HMMECs and HNECs to eosinophils, DPM and DPM extract each significantly
increased eosinophil adhesiveness to HNECs; neither affected eosinophil adhesiveness to
HMMECs. DPM extract also induced eosinophil degranulation without changing the eosinophil
survival rate. These results indicate that DPM may play an important role in promoting the nasal
hypersensitivity induced by  enhanced eosinophil infiltration of epithelium and eosinophil
degranulation. It should also be noted that eosinophils are major components of allergic
inflammatory disorders, including asthma and nasal allergy.
       Terada et al. (1999) examined the effects of DPM extract on the expression of histamine
HI receptor (H1R) mRNA in HNECs and HMMECs, and on the production of IL-8 and GM-
CSF induced by histamine.  HNECs and HMMECs, isolated from human nasal mucosa
specimens, were cultured with DPM extract. DPM extract increased the expression of H1R
mRNA in both HNECs and HMMECs.  The amount of IL-8 and GM-CSF induced by histamine
was also significantly higher in HNECs and HMMECs treated with DPM extract. These results
strongly suggest that DPM accelerates the inflammatory change by not only directly
upregulating H1R expression but also by increasing histamine-induced IL-8 and GM-CSF
production. Histamine is the most important chemical mediator in the pathogenesis of nasal
allergy.
       Steerenberg et al. (1998) studied the effects of exposure to DPM on airway epithelial
cells, the first line of defense against inhaled pollutants. Cells from a human bronchial cell line
(BEAS-2B) were cultured in vitro and exposed to DPM (0.04-0.33 mg/mL) and the effects on
IL-6 and IL-8 production were observed. Increases in IL-6 and IL-8 production compared to the
                                         5-10

-------
nonexposed cells (11- and 4-fold, respectively) were found after 24 or 48 h exposure to DPM.
This increase was lower (17- and 3.3-fold) compared to silica and higher compared to titanium
dioxide, which showed no increase for either IL-6 or IL-8. The study was  extended to observe
the effects of DPM on inflammation-primed cells. BEAS-2B cells were exposed to TNF-a
followed by DPM.  Additive effects on IL-6 and IL-8 production by BEAS-2B cells were found
after TNF-a priming and subsequent exposure to DPM only at a low dose  of DPM and TNF-a
(0.05-0.2 ng/mL). The investigators concluded that BEAS-2B phagocytized DPM and produced
an increased amount of IL-6 and IL-8, and that in TNF-a-primed BEAS-2B cells DPM
increased interleukin production only at low concentrations of DPM and TNF-a.
       Ohtoshi et al. (1998) studied the effect of suspended particulate matter (SPM), obtained
from high-volume air samplers, and DPM obtained from exhaust of a stationary diesel engine on
the production of IL-8 and granulocyte-colony stimulating factor (GM-CSF) by human airway
epithelial cells in vitro. Nontoxic doses of DPMs stimulated  production of IL-8 and GM-CSF by
three kinds of human epithelial cells (nasal polyp-derived upper airway, normal bronchial, and
transformed bronchial epithelial cells) in a dose- and time-dependent fashion at a DPM
concentration as low as 10 |ig/mL.  SPM applied at 250 and 2,500 |ig/mL had  a stimulatory
effect on GM-CSF, but not on IL-8 production. The effects could be blocked  with a protein
synthesis inhibitor, suggesting that the process required de novo protein synthesis, and appeared
to be due to an extractable component because neither charcoal nor graphite showed such
stimulatory effects. The authors concluded that SPM and DPM,  a component  of SPM, may be
important air pollutants in the activation of airway cells for the release of cytokines relevant to
allergic airway inflammation.
       The mechanisms underlying DPM-induced injury to airway cells were  investigated in
human bronchial epithelial cells (HBECs) in culture (Bayram et al., 1998a). HBECs from
bronchial explants obtained at surgery were cultured and exposed to DPM  (10-100 |ig/mL)
suspended in a serum-free supplemented medium (SF-medium) or to a SF-medium filtrate of
DPM.  The filtrate was obtained by incubating DPM (50 jig/mL) in SF-medium for 24 h. The
effects of DPM and DPM filtrate on permeability, ciliary beat frequency (CBF), and release of
inflammatory mediators were observed. DPM and filtered solution of DPM significantly
increased the electrical resistance of the cultures but did not affect movement of bovine serum
albumin across cell cultures. DPM and filtered DPM solution significantly attenuated the CBF
of these cultures and significantly increased the release of IL-8. DPM also increased the release
by these cultures of GM-CSF and soluble intercellular adhesion molecule-1 (sICAM-1). These
authors also observed that activated charcoal was not able to induce changes in electrical
resistance, attenuate CBF, and increase the release of inflammatory mediators  from HBEC, and
proposed that these effects were due most likely to the compounds adsorbed onto the DPM
                                         5-11

-------
rather than the size of DPM. The authors concluded that exposure of airway cells to DPM may
lead to functional changes and release of proinflammatory mediators and that these effects may
influence the development of airway disease.
       Bayram et al. (1998b) investigated the sensitivity of cultured airway cells from asthmatic
patients to DPM. Incubation with DPM (10-100 jig/mL) significantly attenuated the CBF in
both the asthmatic and nonasthmatic bronchial epithelial cell cultures. Cultured airway cells
from asthmatic patients constitutively released significantly greater amounts of IL-8, GM-CSF,
and sICAM-1 than  cell cultures from nonasthmatic subjects. Only cultures from asthmatic
patients additionally released RANTES. The authors concluded that cultured airway cells from
asthmatic subjects differ with regard to the amounts and types of proinflammatory mediators
they can release and that the increased sensitivity of bronchial epithelial cells of asthmatic
subjects to DPM may result in  exacerbation of their disease symptoms.
       Devalia et al. (1999) investigated the potential sensitivity of HBECs biopsied from atopic
mild asthmatic patients and non-atopic nonasthmatic subjects to DPM. HBECs from asthmatic
patients constitutively released significantly greater amounts of IL-8, GM-CSF, and sICAM-1
than HBECs from nonasthmatic subjects.  RANTES was only released by HBECs of asthmatic
patients. Incubation of the asthmatic cultures with 10 |lg/mL DPM significantly increased the
release of IL-8, GM-CSF, and  sICAM-1 after 24 h.  In contrast, only higher concentrations (50-
100 |lg/mL DPM) significantly increased the release of IL-8 and GM-CSF from HBECs of
nonasthmatics. The authors conclude that the increased sensitivity of the airways of asthmatics
to DPM may be, at least in part, a consequence of greater constitutive and DPM-induced release
of specific pro-inflammatory mediators from bronchial epithelial cells.
       Abe and co-workers have demonstrated formation of increased cytokine levels in
cultured human bronchial epithelial cells exposed to freshly generated DE, but not to filtered
DE, i.e., particle-free DE (Abe et al., 2000). Cytokine IL-8 protein as well as transforming
growth factor (TGF)-PI mRNAs were induced in a time-dependent manner (from 0.5 to 14 h of
exposure)  in BET-1A human bronchial epithelial cells in response to exposure to freshly
generated, cooled, humidified DE that was diluted to 2.9 mg DPM/m3. The gas obtained by
filtration of DE alone did not show any sustained increase in these indicators, suggesting that DE
particles play a more important role in eliciting these responses than do the accompanying gases
(10.6 ppm CO, 7.3  ppm NO2, and 3.3 ppm SO2).
       To elucidate the intracellular signal transduction pathway regulating IL-8 and RANTES
production, Hashimoto et al. (2000) examined the role of p38 mitogen-activated protein (MAP)
kinase in DPM-induced (DPM = 10, 50, or 100 |lg/mL) IL-8 and RANTES production by
HBECs. They also examined the effect of a thiol-reducing agent, N-acetylcysteine (NAC), on
DPM-induced p38 MAP kinase activation and cytokine production. The authors conclude that
                                         5-12

-------
p38 MAP kinase plays an important role in the DPM-activated signaling pathway that regulates
IL-8 and RANTES production by HBECs and that the cellular redox state is critical for DPM-
induced p38 MAP kinase activation leading to IL-8 and RANTES production.
       Boland et al. (1999) compared the biological effects of carbon black and DPM (2.5
|ig/cm2 culture surface) collected from catalyst- and noncatalyst-equipped diesel vehicles in
cultures of both human bronchial epithelial cells and human nasal epithelial cells. Transmission
electron microscopy indicated that DPM was phagocytosed by epithelial cells and translocated
through the epithelial cell sheet. The time and dose dependency of phagocytosis and its
nonspecificity for different particles (DPM, carbon black, and latex particles) were established
by flow cytometry. DPM also induced a time-dependent increase in interleukin-8,  GM-CSF,
and interleukin-1 P release.  The inflammatory response occurred later than phagocytosis and,
because carbon black had no effect on cytokine release, its extent appeared to depend on the
content of adsorbed organic compounds. Furthermore, treatment of the exhaust gas to decrease
the adsorbed organic fraction reduced the DPM-induced increase in GM-CSF factor release.
These results indicate that DPM can be phagocytosed by and induce a specific inflammatory
response in airway epithelial cells.

5.1.1.1.5. Summary. In the available exposure studies, considerable variability is reported in
DE detection threshold.  The odor scales described in some of these studies have no general use
at present because they are not objectively defined; however, the studies do clearly indicate
substantial interindividual variability in the ability to detect odor and the level at which it
becomes objectionable. Much of what is known about the acute effects of DE comes from case
reports that lack clear measurements of exposure concentrations. The studies of pulmonary
function changes in exposed humans have looked for changes occurring over a workshift or after
a short-term exposure. The overall conclusion of these studies is that reversible changes in
pulmonary function in humans can occur in relation to DE exposure, although it is not possible
to relate these changes to specific exposure levels. Numerous studies described in this section,
conducted in humans and in isolated cell systems derived from humans exposed to DPM,
revealed various biochemical and pathophysiological alterations, such as IgE changes, altered
levels of cytokines/chemokines, and goblet-cell hyperplasia, with nearly all these responses
being key changes and markers of allergic inflammatory disorders of the airways such as  asthma
and nasal allergies (Nel et al., 1998). Thus, a major point of significance about these findings is
that they indicate that DPM could be viewed as having the potential to elicit inflammatory and
immunological responses and responses typical of asthma, and that DPM may be a likely factor
in the increasing incidence of allergic hypersensitivity. These studies have also shown that
effects are due primarily to the organic fraction and that DPM enhances the allergic response to
                                          5-13

-------
known allergens. Results from these studies, including those wih laboratory animals, indicate
that DPM could be viewed as having the potential to influence the development of airway
inflammation and disease through its adjuvant properties and by causing the release of
proinflammatory mediators.

5.1.1.2. Long-Term Exposures
       Several epidemiologic studies have evaluated the effects of chronic exposure to DE on
occupationally exposed workers.
       Battigelli et al. (1964) measured several indices of pulmonary function, including vital
capacity, FEVl3 peak flow, nitrogen washout, and diffusion capacity in 210 locomotive
repairmen exposed to DE in 3 engine houses. The average exposure of these locomotive
repairmen to DE was 9.6 years.  When compared with a control group matched for age, body
size, "past extrapulmonary medical history"  (no explanation given), and job status (154 railroad
yard workers), no significant clinical differences were found in pulmonary function or in the
prevalence of dyspnea, cough, or sputum between the DE-exposed and nonexposed groups.
Exposure to DE showed marked seasonal variations because the doors of the engine house were
open in the summer and closed in the winter. For the exposed group, the maximum daily
workplace concentrations of air pollutants measured were 1.8 ppm NO2, 1.7 ppm total
aldehydes, 0.15 ppm acrolein, 4.0 ppm SO2,  and 5.0 ppm total hydrocarbons. The concentration
of airborne particles was not reported.
       Gamble et al. (1987b) examined 283  diesel bus garage workers from four garages in two
cities to determine if there was excess chronic respiratory morbidity associated with exposure to
DE.  Tenure of employment was used as a surrogate of exposure; mean tenure of the study
population was 9 years ± 10 years S.D.  Exposure-effect relationships within the study
population showed no detectable associations of symptoms with tenure. Reductions in FVC,
FEVj, peak flow, and FEF50 (but not FEF75)  were associated with increasing tenure.  Compared
with a control population (716 nonexposed blue-collar workers) and after indirect adjustment for
age,  race, and smoking, the exposed workers had a higher incidence of cough, phlegm, and
wheezing; however, there was no correlation between symptoms and length of employment.
Dyspnea showed an exposure-response trend but no apparent increase in prevalence. Mean
FEVj, FVC, FEF50, and peak flow were not reduced in the total cohort compared with the
reference population, but were reduced in workers with 10 years or more tenure.
       Purdham et al. (1987) evaluated respiratory symptoms and pulmonary function in
17 stevedores employed in car ferry operations who were exposed to both diesel and gasoline
exhausts and in a control group of 11 on-site office workers. Twenty-four percent of the
exposed group and 36% of the controls were smokers. If a particular symptom was considered
                                         5-14

-------
to be influenced by smoking, smoking status was used as a covariate in the logistic regression
analysis; pack-years smoked was a covariate for lung function indices. The frequency of
respiratory symptoms was not significantly different between the two groups; however, baseline
pulmonary function measurements were significantly different.  The latter comparisons were
measured by multiple regression analysis using the actual (not percentage predicted) results and
correcting for age, height, and pack-years smoked. The stevedores had significantly lower
FEVj, FEVj/FVC, FEF50, and FEF75 (p<0.021,/K0.023,/K0.001, and/K0.008, respectively),
but not FVC.  The results from the stevedores were also compared with those obtained from a
study of the respiratory health status of Sydney, Nova Scotia, residents.  These comparisons
showed that the dock workers had higher FVC, similar FEVl3 but lower FEVj/FVC and flow
rates than the residents of Sydney.  Based on these consistent findings, the authors concluded
that the lower baseline function measurements in the stevedores provided evidence of an
obstructive ventilatory defect, but caution in interpretation was warranted because of the small
sample size.  There were no significant changes in lung function over the workshift,  nor was
there a difference between the two groups. The stevedores were exposed to significantly
(p<0.04) higher concentrations of particulate matter (0.06 to 1.72 mg/m3, mean 0.50 mg/m3)
than the controls (0.13 to 0.58 mg/m3, mean not reported). Exposures of stevedores  to SO2,
NO2, aldehydes, and PAHs were very low; occasional CO concentrations in the 20 to 100 ppm
range could be detected for periods up to 1 h in areas where blockers were chaining gasoline-
powered vehicles.
       Additional epidemiologic studies on the health hazards posed by exposure to DE have
been conducted for mining operations.  Reger et al. (1982) evaluated the respiratory  health status
of 823 male coal miners from six diesel-equipped mines compared with 823 matched coal
miners not exposed to DE.  The average tenure of underground work for the underground miners
and their controls was only about 5 years; on average, the underground workers in diesel mines
spent only 3 of those 5 years underground in diesel-use mines.  Underground miners exposed to
DE reported a higher incidence of symptoms of cough and phlegm but proportionally fewer
symptoms of moderate to severe dyspnea than their matched counterparts. These differences in
prevalence of symptoms were not statistically significant. The diesel-exposed underground
miners, on the average, had lower FVC, FEVl3 FEF50, FEF75, and FEF90 but higher peak flow
and FEF25 than their matched controls.  These differences, however,  were not statistically
significant. Health indicators for surface workers and their matched controls were directionally
the same as for matched underground workers.  There were no consistent relationships between
the findings of increased respiratory symptoms, decreased pulmonary function,  smoking history,
years of exposure, or monitored atmosphere pollutants (NOX, CO, particles, and aldehydes).
Mean concentrations of NOX at the six mines ranged from 0 to 0.6 ppm for short-term area
                                         5-15

-------
samples, 0.13 to 0.28 ppm for full-shift personal samples, and 0.03 to 0.80 for full-shift area
samples. Inhalable particles (less than 10 |im MMAD) averaged 0.93 to 2.73 mg/m3 for
personal samples and 0 to 16.1 mg/m3 for full-shift area samples. Ames et al. (1984), using a
portion of the miners studied by Reger, examined 280 diesel-exposed underground miners in
1977 and again in 1982.  Each miner in this group had at least 1 year of underground mining
work history in 1977.  The control group was 838 miners with no exposure to DE. The miners
were evaluated for prevalence of respiratory symptoms, chronic cough, phlegm, dyspnea, and
changes in FVC, FEVl3 and FEF50.  No air monitoring data were reported; exposure to DE gases
and mine dust particles were described as very low.  These authors found no decrements in
pulmonary function or increased prevalence of respiratory symptoms attributable to exposure to
DE. In fact, the 5-year incidences of cough, phlegm, and dyspnea were greater in miners
without exposure to DE.
       Attfield (1978) studied 2,659 miners from 21 mines (8 metal, 6  potash, 5 salt, and
2 trona). Diesels were employed in  only 18 of the mines, but the 3 mines not using diesels were
not identified. The years of diesel usage,  ranging from  8 in trona mines to 16 in potash mines,
were used as a surrogate for exposure to DE.  Based on a questionnaire, an increased prevalence
of persistent cough was associated with exposure to aldehydes; this finding, however, was not
supported by the pulmonary function data. No adverse  respiratory symptoms or pulmonary
function impairments were related to CO2, CO, NO2, inhalable dust, or  inhalable quartz. The
author failed to comment on whether the prevalence of  cough was related to the high incidence
(70%) of smokers in the cohort.
       Questionnaire,  chest radiograph, and spirometric data were collected by Attfield et al.
(1982) on 630 potash miners from six potash mines.  These miners were exposed for an average
of 10 years (range 5 to 14 years) to 0.1 to 3.3 ppm NO2, 0.1 to 4.0 ppm aldehyde, 5 to 9 ppm
CO, and total dust concentrations of 9 to 23 mg/m3.  No attempt was made to measure diesel-
derived particles separately from other dusts. The ratio of total to inhalable (<10 |im MMAD)
dust ranged from 2 to 11.  An increased prevalence of respiratory symptoms was related solely
to smoking. No association was found between symptoms and tenure of employment, dust
exposure, NO2, CO, or aldehydes. A higher prevalence of symptoms of cough and phlegm was
found, but no differences in pulmonary function (FVC and FEVj) were found in these
diesel-exposed potash miners when compared with the predicted values derived from a logistics
model based on blue-collar workers  working in nondusty jobs.
       Gamble et al. (1983) investigated respiratory morbidity in 259 miners from 5 salt mines
in terms of increased respiratory symptoms, radiographic findings, and  reduced pulmonary
function associated with exposure to NO2, inhalable particles (<10 |lm MMAD), or years
worked underground.  Two of the mines used diesel extensively; no diesels were used in one salt
                                         5-16

-------
mine. Diesels were introduced into each mine in 1956, 1957, 1963, or 1963 through 1967.
Several working populations were compared with the salt miner cohort. After adjustment for
age and smoking, the salt miners showed no increased prevalence of cough, phlegm, dyspnea, or
airway obstruction (FEVj/FVC) compared with aboveground coal miners, potash miners, or
blue-collar workers. The underground coal miners consistently had an elevated level of
symptoms. Forced expiratory volume at 1 s, FVC, FEF50, and FEF75 were uniformly lower for
salt miners in relation to all the comparison populations. There was, however, no association
between changes in pulmonary function and years worked, estimated cumulative inhalable
particles, or estimated NO2 exposure. The highest average exposure to paniculate matter was
1.4 mg/m3 (particle size not reported, measurement includes NaCl). Mean NO2  exposure was
1.3 ppm,  with a range of 0.17 ppm to 2.5 ppm.  In a continuation of these studies, Gamble and
Jones (1983) grouped the salt miners into low-, intermediate-, and high-exposure categories
based on  tenure in jobs with DE exposure. Average concentrations of inhalable particles and
NO2 were 0.40, 0.60, and 0.82 mg/m3 and 0.64, 1.77, and 2.21  ppm for the three diesel exposure
categories, respectively.  A statistically significant concentration-response association was found
between the prevalence of phlegm in the salt miners and exposure to DE (p<0.0001) and a
similar, but nonsignificant, trend for cough and dyspnea.  Changes in pulmonary function
showed no association with diesel tenure. In a comparison with the control group of
nonexposed, blue-collar workers, adjusted for age and smoking, the overall prevalence of cough
and phlegm (but not dyspnea) was elevated in the diesel-exposed workers.  Forced expiratory
volumes at 1 s and FVC were within 4% of expected, which was considered to be within the
normal range of variation for a nonexposed population.
       In a preliminary study of three subcohorts from bus company personnel (clerks [lowest
exposure], bus drivers [intermediate exposure], and bus garage workers [highest exposure])
representing different levels of exposure to DE, Edling and Axelson (1984) found a fourfold
higher risk ratio for cardiovascular mortality in bus garage workers, even after adjusting for
smoking history and allowing for at least 10 years of exposure and 15 years or more of induction
latency. Carbon monoxide was hypothesized as the etiologic agent for the increased
cardiovascular disease but was not measured. However, in a more comprehensive epidemiologic
study, Edling et al. (1987) evaluated mortality data covering a 32-year period for a cohort of 694
bus garage employees and found no significant differences between the observed and expected
number of deaths from cardiovascular disease.  Information on exposure components and their
concentrations was not reported.
       The absence of reported noncancerous human health effects, other than infrequently
occurring effects related to respiratory symptoms and pulmonary function changes, is notable.
Unlike studies in laboratory animals, to be described later in this chapter, studies of the impact
                                         5-17

-------
of DE on the defense mechanisms of the human lung have not been performed.  No direct
evidence is available in humans regarding doses of DE, gas phase, particulate phase, or total
exhaust that lead to impaired particle clearance or enhanced susceptibility to infection. A
summary of epidemiologic studies is presented in Table 5-1.
                Table  5-1.  Human studies of exposure to diesel exhaust
      Study
           Description
                   Findings
                                          Acute exposures
 Kahn et al.
 (1988)

 El Batawi and
 Noweir(1966)
 Battigelli
 (1965)

 Katz et al.
 (1960)

 Hare and
 Springer (1971)
 Hare et al.
 (1974)
 Linnell and
 Scott (1962)

 Rudell et al.
 (1990, 1994)
 Rudell et al.
 (1996)
 Battigelli
 (1965)
13 cases of acute exposure, Utah and
Colorado coal miners.
161 workers, two diesel bus garages.
Six subjects, eye exposure chamber,
three dilutions.
14 persons monitoring DE in a train
tunnel.
Volunteer panelists who evaluated
general public's response to odor of
DE.

Odor panel under highly controlled
conditions determined odor threshold
for DE.
Eight healthy nonsmoking subjects
exposed for 60 min in chamber to DE
(3.7 ppm NO, 1.5 ppm NO2, 27 ppm
CO, 0.5 mg/m3 formaldehyde,
particles (4.3 x 106/cm3). Exercise,
10 of each 20 min (75 W).
Volunteers exposed to DE for 1 h
while doing light work. Exposure
concentrations uncertain.
13 volunteers exposed to three
dilutions of DE for 15 min to 1 h.
Acute reversible sensory irritation, headache,
nervous system effects, bronchoconstriction were
reported at unknown exposures.
Eye irritation (42%), headache (37%), dizziness
(30%), throat irritation (19%), and cough and
phlegm (11%) were reported in this order of
incidence by workers exposed in the service and
repair of diesel-powered buses.
Time to onset was inversely related and severity of
eye irritation was associated with the level of
exposure to DE.
Three occasions of minor eye and throat irritation;
no correlation established with concentrations of
DE components.
Slight odor intensity, 90% perceived, 60%
objected; slight to moderate odor intensity, 95%
perceived, 75% objected; moderate odor intensity,
100% perceived, almost 95% objected.
In six panelists, the volume of air required to dilute
raw DE to an odor threshold ranged from a factor of
140 to 475.
Odor, eye and nasal irritation in 5/8 subjects.  BAL
findings: small decrease in mast cells, lymphocyte
subsets and macrophage phagocytosis; small
increase in PMNs.
Unpleasant smell along with irritation of eyes and
nose reported. Airway resistance increased.
Reduction of particle concentration by trapping did
not affect results.
No significant effects on pulmonary resistance were
observed as measured by plethysmography.
 Wade and
 Newman
 (1993)
 Diaz-Sanchez
 et al. (1994)
Three railroad workers acutely
exposed to DE.

Volunteers challenged by a nasal
sprayofO.SOmgDPM.
The workers developed symptoms of asthma.
Enhancement of IgE production reported due to a
dramatic increase in IgE-secreting cells.
                                               5-18

-------
            Table 5-1. Human studies of exposure to diesel exhaust (continued)
    Study
           Description
                   Findings
Takenaka et al.
(1995)
Diaz-Sanchez
et al. (1996)

Diaz-Sanchez
et al. (1997)
Salvi et al.
(1999)
Salvi et al.
(2000)
Nightingale et
al. (2000)
Volunteers challenged by a nasal
spray of 0.30 mg DPM.
Volunteers challenged by a nasal
spray of 0.30 mg DPM.

Ragweed-sensitive volunteers
challenged by a nasal spray of 0.30
mg DPM alone or in combination
with ragweed allergen.


Volunteers exposed to diluted DE
(DPM 300 |-ig/m3) for 1 h with
intermittent exercise.
Volunteers exposed to diluted DE
(DPM 300 |-ig/m3) for 1 h.
Volunteers exposed to resuspended
DPM (200 ug/m3) for 2 h at rest
DPM extracts enhanced interleukin-4 plus
monoclonal antibody-stimulated IgE production as
much as 360%, suggesting an enhancement of
ongoing IgE production rather than inducing
germline transcription or isotype switching.

A broad increase in cytokine expression predicted
to contribute to enhanced local IgE production.

Ragweed allergen plus DPM-stimulated ragweed-
specific IgE to a much greater degree than ragweed
alone, suggesting DPM may be a key feature in
stimulating allergen-induced respiratory allergic
disease.

        No changes in pulmonary function, but
        significant increases in neutrophils, B
        lymphocytes, histamine, and fibronectin in
        airway lavage fluid.
        Bronchial biopsies 6 h after exposure
        showed significant increase in neutrophils,
        mast  cells, CD4+ and CD8+ T
        lymphocytes; upregulation of ICAM-1 and
        VCAM-1; increases in the number of LFA-
        1+ in bronchial tissue.
        Significant increases in neutrophils and
        platelets observed in peripheral blood.

•       DPM enhanced gene transcription of IL-8
        in bronchial tissue and bronchial wash
        cells
        Increased expression of growth-regulated
        oncogene-oc and IL-8 in bronchial
        epithelium; trend towards increased IL-5
        mRNA gene transcripts.

        DPM increased exhaled levels of CO
        DPM increased sputum neutrophils and
        myeloperoxidase
                                    Studies of cross-shift changes
Reger (1979)      Five or more VC maneuvers by each
                 of 60 coal miners exposed to DE at
                 the beginning and end of a workshift.
                                    FEVb FVC, and PEFR were similar between diesel
                                    and non-diesel-exposed miners.  Smokers had an
                                    increased number of decrements over shift than
                                    nonsmokers.
                                               5-19

-------
            Table 5-1. Human studies of exposure to diesel exhaust (continued)
    Study
           Description
                   Findings
Ames et al.
(1982)
Jorgensen and
Svensson
(1970)
Gamble et al.
(1979)
Gamble et al.
(1987a)
UlfVarson et al.
(1987)
Battigelli et al.
(1964)
Pulmonary function of 60 diesel-
exposed compared with 90 non-
diesel-exposed coal miners over
workshift.

240 iron ore miners matched for
diesel exposure, smoking, and age
were given bronchitis questionnaires
and spirometry pre- and
postworkshift.

200 salt miners performed before-
and after-workshift spirometry.
Personal environmental NO2 and
inhalable particle samples were
collected.

232 workers in 4 diesel bus garages
administered acute respiratory
questionnaire and before and after
workshift spirometry. Compared to
lead/acid battery workers previously
found to be unaffected by their
exposures.

Workshift changes in pulmonary
function were evaluated in crews of
roll-on/ roll-off ships and car ferries
and bus garage staff.  Pulmonary
function was evaluated in six
volunteers exposed to diluted DE, 2.1
ppm NO2, and 0.6 mg/m3 paniculate
matter.
Significant workshift decrements occurred in
miners in both groups who smoked; no significant
differences in ventilatory function changes between
miners exposed to DE and those not exposed.

Among underground (surrogate for diesel exposure)
miners, smokers, and older age groups, frequency
of bronchitis was higher. Pulmonary function was
similar between groups and subgroups except for
differences accountable to age.

Smokers had greater but not significant reductions
in spirometry than ex- or nonsmokers. NO2 but not
paniculate levels significantly decreased FEV1,
FEF25, FEF50, and FEF75 over the workshift.


Prevalence of burning eyes, headache, difficult or
labored breathing, nausea, and wheeze were higher
in diesel bus workers than in comparison
population.
Pulmonary function was affected during a workshift
exposure to DE, but it normalized after a few days
with no exposure. Decrements were greater with
increasing intervals between exposures. No effect
on pulmonary function was observed in the
experimental exposure study.
                               Cross-sectional and longitudinal studies
210 locomotive repairmen exposed to
DE for an average of 9.6 years in
railroad engine houses were
compared with 154 railroad yard
workers of comparable job  status but
no exposure to DE.
No significant differences in VC, FEVb peak flow,
nitrogen washout, or diffusion capacity or in the
prevalence of dyspnea, cough, or sputum were
found between the DE-exposed and nonexposed
groups.
                                               5-20

-------
            Table 5-1. Human studies of exposure to diesel exhaust (continued)
    Study
Description
Findings
Gamble et al.     283 male diesel bus garage workers
(1987b)          from four garages in two cities were
                 examined for impaired pulmonary
                 function (FVC, FEVb and flow
                 rates). Study population with a mean
                 tenure of 9 ± 10 years S.D. was
                 compared to a nonexposed blue-
                 collar population.

Purdham et al.    Respiratory symptoms and
(1987)           pulmonary function were evaluated
                 in 17 stevedores exposed to both
                 diesel and gasoline exhausts in car
                 ferry operations; control group was
                 11 on-site office workers.
Reger et al.       Differences in respiratory symptoms
(1982)           and pulmonary function were
                 assessed in 823 coal miners from 6
                 diesel-equipped mines compared to
                 823 matched coal miners not exposed
                 toDE.

Ames et al.       Changes in respiratory symptoms and
(1984)           function were measured during a 5-
                 year period in 280 diesel-exposed
                 and 838 nonexposed U.S.
                 underground coal miners.

Attfield (1978)    Respiratory symptoms and function
                 were assessed in 2,659 miners from
                 21 underground metal mines (1,709
                 miners) and nonmetal mines (950
                 miners).  Years of diesel usage in the
                 mines were surrogate for exposure to
                 DE.
                        Analyses within the study population showed no
                        association of respiratory symptoms with tenure.
                        Reduced FEVj and FEF50 (but not FEF75) were
                        associated with increasing tenure. The study
                        population had a higher incidence of cough,
                        phlegm, and wheezing unrelated to tenure.
                        Pulmonary function was not affected in the total
                        cohort of diesel-exposed but was reduced with 10
                        or more years of tenure.
                        No differences between the two groups for respira-
                        tory symptoms.  Stevedores had lower baseline lung
                        function consistent with an obstructive ventilatory
                        defect compared with controls and those of Sydney,
                        Nova Scotia, residents.  Caution in interpretation is
                        warranted because of small sample size. No
                        significant changes in lung function over workshift
                        or difference between two groups.
                        Underground miners in diesel-use mines reported
                        more symptoms of cough and phlegm and had
                        lower pulmonary function. Similar trends were
                        noted for surface workers at diesel-use mines.
                        Pattern was consistent with small airway disease
                        but factors other than exposure to DE thought to be
                        responsible.
                        No decrements in pulmonary function or increased
                        prevalence of respiratory symptoms were found
                        attributable to DE. In fact, 5-year incidences of
                        cough, phlegm, and dyspnea were greater  in miners
                        without exposure to DE than in miners exposed to
                        DE.
                        Questionnaire found an association between an
                        increased prevalence of cough and aldehyde
                        exposure; this finding was not substantiated by
                        spirometry data. No adverse symptoms or
                        pulmonary function decrements were related to
                        exposure to NO2, CO, CO2, dust, or quartz.
                                               5-21

-------
            Table 5-1. Human studies of exposure to diesel exhaust (continued)
    Study
           Description
                   Findings
Attfield et al.
(1982)
Gamble et al.
(1983)
Gamble and
Jones (1983)
Edling and
Axelson (1984)
Edling et al.
(1987)
Respiratory symptoms and function
were assessed in 630 potash miners
from 6 potash mines through a
questionnaire, chest radiographs, and
spirometry. A thorough assessment
of the environment of each mine was
made concurrently.

Respiratory morbidity was assessed
in 259 miners in 5 salt mines by
respiratory symptoms, radiographic
findings, and spirometry. Two mines
used diesels extensively, two had
limited use, and one used no diesels
in 1956, 1957, 1963, or 1963 through
1967. Several working populations
were compared with the salt-mine
cohort.
Same as above.  Salt miners were
grouped into low-, intermediate-, and
high-exposure categories based on
tenure in jobs with diesel exposure.
Pilot study of 129 bus company
employees classified into 3 diesel-
exhaust exposure categories: clerks
(0), bus drivers (1), and bus garage
workers.

Cohort of 694 male bus garage
employees followed from 1951
through 1983 was evaluated for
mortality from cardiovascular
disease. Subcohorts categorized by
levels of exposure were clerks (0),
bus drivers (1), and bus garage
employees (2).	
No obvious association indicative of diesel
exposure was found between health indices, dust
exposure, and pollutants. Higher prevalences of
cough and phlegm but no differences in FVC and
FEVj were found in these diesel-exposed potash
workers when compared with predicted values from
a logistic model based on blue-collar staff working
innondusty jobs.
After adjustment for age and smoking, salt miners
showed no symptoms or increased prevalence of
cough, phlegm, dyspnea, or air obstruction
(FEVj/FVC) compared with aboveground coal
miners, potash workers, or blue-collar workers.
FEVb FVC, FEF50, and FEF75 were uniformly
lower for salt miners in comparison with all the
comparison populations. No changes in pulmonary
function were associated with years of exposure or
cumulative exposure to inhalable particles or NO2.
A statistically significant dose-related association
of phlegm and diesel exposure was noted.  Changes
in pulmonary function showed no association with
diesel tenure. Age- and smoking-adjusted rates  of
cough, phlegm, and dyspnea were 145%, 169%,
and 93% of an external comparison population.
Predicted pulmonary function indices showed small
but significant reductions; there was no dose-
response relationship.

The most heavily exposed group (bus garage
workers) had a fourfold increase in risk of dying
from cardiovascular disease, even after correction
for smoking and allowing for  10 years of exposure
and 14 years or more of induction latency  time.

No increased mortality from cardiovascular disease
was found among the members of these five bus
companies when compared with the general
population or grouped as subcohorts with different
levels of exposure.
                                               5-22

-------
       To date, no large-scale epidemiologic study has looked for effects of chronic exposure to
DE on pulmonary function. In the long-term longitudinal and cross-sectional studies, a
relationship was generally observed between work in a job with diesel exposure and respiratory
symptoms (such as cough and phlegm), but there was no consistent effect on pulmonary
function. The interpretation of these results is hampered by lack of measured DE exposure
levels and the short duration of exposure in these cohorts.  The studies are further limited in that
only active workers were included, and it is possible that workers who have developed
symptoms or severe respiratory disease are likely to have moved away from these jobs. The
relationship between work in a job with diesel exposure and respiratory  symptoms may be due to
short-term exposure.

5.1.2.  Traffic Studies
       The relationship between traffic density and respiratory health in children has been
examined in a series of studies in Holland in children attending schools located near major
freeways. Cough, wheeze, runny nose, and doctor-diagnosed asthma were reported more often
for children living within 100 m of freeways carrying between 80,000 and  150,000 vehicles per
day (van Vliet et al., 1997). Separate counts for truck traffic indicated a range from 8,000 to
17,500 trucks per day. Truck traffic intensity and concentration of "black smoke," considered
by the authors to be a proxy for DPM, measured in schools were found to be significantly
associated with chronic respiratory symptoms, with the relationships being more pronounced in
girls than in boys.
       Brunekreef et al.  (1997) measured lung function in children in six areas located near
major motorways and assessed their exposure to traffic-related air pollution using separate traffic
counts for automobiles and trucks. They also measured air pollution in the children's schools.
Although lung function was associated with truck traffic density, there was a lesser association
with automobile traffic density.  The association was stronger in those children living closest
(300 m) to the roadways. Lung function  was also associated with  concentration of "black
smoke" (source and constitution unclear from the  study) measured inside the schools. The
associations were stronger in girls than in boys. The authors conclude that exposure to vehicular
pollution, in particular DPM, may lead to reduced lung function in children living near major
motorways.
       In a follow-up study of traffic-related air pollution and its effect  on the respiratory health
of children living near roadways, Brunekreef et al. (2000) showed that the intensity of truck
traffic was significantly associated with the prevalence of wheeze, phlegm, bronchitis, eye
symptoms, and allergy to dust and pets. Associations with yearly averaged PM2 5 and "soot"
concentrations measured inside and outside the schools showed similar patterns.  Truck traffic
                                          5-23

-------
intensity was also significantly associated with a positive skin prick test or elevated IgE for
outdoor allergens. There were no associations between traffic intensity or PM2 5 and "soot"
concentrations and lung function, bronchial responsiveness, and allergic reactions to indoor
allergens.  Further analysis of the data showed that the associations between traffic-related air
pollution and symptoms were almost entirely related to children with bronchial hyperreactivity
or sensitization to common allergens.

5.1.3. Laboratory Animal Studies
       Because humans and laboratory animals show similar nonneoplastic responses to inhaled
particles (ILSI, 2000), animal studies have been conducted to assess the pathophysiologic effects
of DPM. Because of the large number of statistical comparisons made in the laboratory animal
studies, and to permit uniform, objective evaluations within and among studies, data will be
reported as significantly different (i.e.,/><0.05) unless otherwise specified. The exposure
regimens used and the resultant exposure conditions employed in the laboratory animal
inhalation studies are summarized in Tables 5-2 through 5-16.  Other than the pulmonary
function studies performed by Wiester et al. (1980) on guinea pigs during their exposure in
inhalation chambers,  the pulmonary function studies performed by other investigators, although
sometimes unreported, were interpreted as being conducted on the following day or thereafter
and not immediately following exposure.

5.1.3.1. Acute Exposures
       The acute toxicity of undiluted DE  to rabbits, guinea pigs, and mice was assessed by
Pattle et al. (1957). Four engine operating conditions were used, and 4 rabbits, 10 guinea pigs,
and 40 mice were tested under each exposure  condition for 5 h (no controls were used).
Mortality was  assessed up to 7 days after exposure. With the engine operating under light load,
the exhaust was highly irritating but not lethal to the test species, and only mild tracheal and
lung damage was observed in the exposed  animals.  The exhaust contained 74 mg/m3 DPM
(particle size not reported), 560 ppm CO, 23 ppm NO2, and 16 ppm aldehydes.  Exhaust
containing 5 mg/m3 DPM, 380 ppm CO, 43 ppm NO2, and 6.4 ppm aldehydes resulted in low
mortality rates (mostly below 10%) and moderate lung damage. Exhaust containing 122 mg/m3
DPM, 418 ppm CO, 51 ppm NO2, and 6.0  ppm aldehydes produced high mortality rates (mostly
above 50%) and severe lung damage.  Exhaust containing 1,070 mg/m3 DPM,  1,700 ppm CO,
12 ppm NO2, and 154 ppm aldehydes resulted in 100% mortality in all three species.  High CO
levels, which resulted in a carboxyhemoglobin value of 60% in mice and 50% in rabbits and
guinea pigs, were considered to be the main cause of death in the latter case. High NO2 levels
                                          5-24

-------
were considered to be the main cause of lung damage and mortality seen in the other three tests.
Aldehydes and NO2 were considered to be the main irritants in the light load test.
       Kobayashi and Ito (1995) administered 1, 10, or 20 mg/kg DPM in phosphate-buffered
saline to the nasal mucosa of guinea pigs.  The administration increased nasal airway resistance,
augmented increased airway resistance and nasal secretion induced by a histamine aerosol,
increased vascular permeability in dorsal skin,  and augmented vascular permeability induced by
histamine. The increases in nasal airway resistance and secretion are considered typical
responses  of nasal mucosa against allergic stimulation.  Similar results were reported for guinea
pigs exposed via inhalation for 3 h to DE diluted to DPM concentrations of either 1 or 3.2
mg/m3 (Kobayashi et al., 1997).  These studies show that short-term exposure to DPM augments
nasal mucosal hyperresponsiveness induced by histamine in guinea pigs.

5.1.3.2. Short-Term and Subchronic Exposures
       A number of inhalation studies have employed a regimen of 20 h/day, 7 days/week for
varying exposure periods up to 20 weeks to differing concentrations of airborne particulate
matter, vapor, and gas concentrations of diluted DE. Exposure regimens and characterization of
gas-phase  components for these studies are summarized in Table 5-2.
Pepelko et al. (1980a) evaluated the pulmonary function of cats exposed under these conditions
for 28 days to 6.4 mg/m3 DPM. The only significant functional change observed was a decrease
in maximum expiratory flow rate at 10% vital capacity.   The excised lungs of the exposed cats
appeared charcoal gray, with focal  black spots visible on the pleural surface.  Pathologic changes
included a predominantly peribronchial localization of black-pigmented macrophages within the
alveoli characteristic  of focal pneumonitis  or alveolitis.
       The effects of a short-term DE exposure on arterial blood gases, pH, blood buffering,
body weight changes, lung volumes, and deflation pressure-volume (PV) curves of young adult
rats were evaluated by Pepelko (1982a). Exposures were 20 h/day, 7 days/week for 8 days to a
concentration of 6.4 mg/m3 DPM in the nonirradiated exhaust (RE) and 6.75 mg/m3 in the
irradiated  exhaust (IE). In spite of the irradiation, levels of gaseous compounds were not
substantially different between the two groups  (Table 5-2). Body weight gains were
significantly reduced in the RE-exposed rats and to an even greater degree in rats exposed to IE.
Arterial blood gases and standard bicarbonate were unaffected, but arterial blood pH was
significantly reduced in rats exposed to IE.  Residual volume and wet lung weight were not
affected by either exposure, but vital capacity and total lung capacity were increased
significantly following exposure to RE.  The shape of the deflation PV curves were nearly
identical for the control, RE, and IE groups.
                                          5-25

-------
Table 5-2. Short-term effects of diesel exhaust on laboratory animals
Species/sex
Rat, F344, M;
Mouse, A/3, M; Hamster,
Syrian, M
Rat, F344, M, F; Mouse,
CD-I, M, F
Cat, Inbred, M
Rat, Sprague-
Dawley, M
|\j Guinea Pig,
0-1 Hartley, M, F
Rat, F344,
M
Guinea Pig, Hartley, M
Guinea Pig, Hartley, M
Exposure period
20 h/day
7 days/week
10-13 weeks
7 h/day
5 days/week
19 weeks
20 h/day
7 days/week
4 weeks
20 h/day
7 days/week
4 weeks
20 h/day
7 days/week
4 weeks
20 h/day
5.5 days/week
4 weeks
30 min
3h
Particles C x T
(mg/m3) (mg-h/m3)
1.5 2,100 to 2,730
0.19|JmMMD
0.21 140
1.0 665
4.4 2,926
6.4 3,584
6.4 3,584
6.8° 3,808
6.8° 3,808
6.0 2,640
6.8 |_lm MMD
1-2 mg DPM —
Intranasally
1 0.5
3.2 1.6
CO NO2 SO2
(ppm) (PPm) (PPm) Effects
6.9 0.49 — Increase in lung wt; increase in
thickness of alveolar walls;
minimal species difference
— — — No effects on lung function in rats
— — — (not done in mice); increase in
— — — PMNs and proteases and AM
aggregation in both species
14.6 2.1 2.1 Few effects on lung function; focal
pneumonitis or alveolitis
16.9 2.49 2.10 Decreased body wt; arterial blood
16.1° 2.76° 1.86° pH reduced; vital capacity, total
lung capacities increased
(<0.01 ppm O3)°
16.7 2.9 1.9 Exposure started when animals
were 4 days old; increase in
(<0.01 ppm O3)° pulmonary flow; bardycardia
— — — Macrophage aggregation; increase
in PMNs; Type II cell
proliferation; thickened alveolar
walls
— — — Augmented increases in nasal
airway resistance and vascular
permeability induced by a
histamine aerosol
5.9 1.4 0.13 Similar results to those reported in
12.9 4.4 0.34 the previous study using intranasal
challenge
Study
Kaplan etal. (1982)
Mauderly et al. (1981)
Pepelko et al. (1980a)
Pepelko (1982a)
Wiesteretal. (1980)
White and Garg (1981)
Kobayashi and Ito
(1995)
Kobayashi et al. (1997)

-------
             Table 5-2.  Short-term effects of diesel exhaust on laboratory animals (continued)
Species/sex
Guinea Pig, Hartley, M, F



Mouse ICR, M




Rat, Sprague-Dawley,
M

Exposure period
20 h/day
7 days/week
8 weeks

6 weeks




24 h


Particles C x T
(mg/m3) (mg-h/m3)
6.3 7,056



100 ng DPM —
intranasally



5-100 Hg/106 —
AM/mL of
DPM
CO NO2 SO2
(ppin) (ppni) (PPm) Effects
17.4 2.3 2.1 Increase in relative lung wt. AM
aggregation; hypertrophy of goblet
cells; focal hyperplasia of alveolar
epithelium
— — — DPM aggravated ovalbumin-
induced airway inflammation and
provided evidence that DPM can
enhance manifestations of allergic
asthma
— — — Unchanged, but not organic-free
DPM enhanced production of
proinflammatory cytokines
Study
Wiesteretal. (1980)



Takanoetal. (1997)




Yangetal. (1997)


     "Irradiated exhaust.
     PMN = Polymorphonuclear leukocyte.
L/i   AM = Alveolar macrophage.
K>

-------
       In related studies, Wiester et al. (1980) evaluated pulmonary function in 4-day-old
guinea pigs exposed for 20 h/day, 7 days/week for 28 days to IE having a concentration of 6.3
mg/m3 DPM. When housed in the exposure chamber, pulmonary flow resistance increased 35%,
and a small but significant sinus bradycardia occurred as compared with controls housed and
measured in control air chambers (p<0.002). Respiratory rate, tidal volume, minute volume, and
dynamic compliance were unaffected, as were lead-1 electrocardiograms.
       A separate group of adult guinea pigs was necropsied after 56 days of exposure to IE, to
diluted RE, or to clean air (Wiester et al., 1980). Exposure resulted in a significant increase in
the ratio of lung weight to body weight (0.68% for controls, 0.78% for IE, and 0.82% for RE).
Heart/body weight ratios were not affected by exposure. Microscopically, there was a marked
accumulation of black pigment-laden AMs throughout the lung, with a slight to moderate
accumulation in bronchial and carinal lymph nodes. Hypertrophy of goblet cells in the
tracheobronchial tree was frequently observed, and focal hyperplasia of alveolar lining cells was
occasionally observed.  No evidence of squamous metaplasia of the tracheobronchial  tree,
emphysema, peribronchitis, or peribronchiolitis was noted.
       White and Garg (1981) studied pathologic alterations in the lungs of rats (16 exposed and
8 controls) after exposure to DE containing 6 mg/m3 DPM.  Two rats from the exposed group
and one rat from the control group (filtered room air) were sacrificed after each exposure
interval of 6 h and 1,  3, 7, 14, 28, 42, and 63 days; daily exposures were for 20 h and were 5.5
days/week. Evidence of AM recruitment and phagocytosis of diesel particles was found at the 6-
h sacrifice; after 24 h of exposure there was a focal, scattered increase in the number  of Type II
cells. After 4 weeks of exposure, there were morphologic changes in size, content, and shape of
AM, septal thickening adjacent to clusters of AMs, and an appearance of inflammatory cells,
primarily within the septa. At 9 weeks of exposure, focal aggregations of particle-laden
macrophages developed near the terminal bronchi, along with an influx of PMNs, Type II cell
proliferation, and thickening of alveolar walls.  The affected alveoli occurred in clusters that, for
the most part, were located near the terminal bronchioles, but occasionally were focally located
in the lung parenchyma. Hypertrophy of goblet cells in the tracheobronchial tree was frequently
observed, and focal hyperplasia  of alveolar lining cells was occasionally observed. No evidence
of squamous metaplasia of the tracheobronchial tree, emphysema, peribronchitis, or
peribronchiolitis was noted.
       Mauderly et al. (1981) exposed rats and mice by inhalation to diluted DE for 545 h over
a 19-week period on a regimen of 7 h/day, 5 days/week at concentrations of 0, 0.21, 1.02, or
4.38 mg/m3 DPM.  Indices of health effects were minimal following  19 weeks of exposure.
There were no significant exposure-related differences in mortality or body weights of the rats or
mice. There also were no significant differences in respiratory function (breathing patterns,
                                          5-28

-------
dynamic lung mechanics, lung volumes, quasi-static PV relationships, forced expirograms, and
CO-diffusing capacity) in rats; pulmonary function was not measured in mice.  No effect on
tracheal mucociliary or deep lung clearances were observed in the exposed groups. Rats, but not
mice, had elevated immune responses in lung-associated lymph nodes at the two higher exposure
levels. Inflammation in the lungs of rats exposed to 4.38 mg/m3 DPM was indicated by
increases in PMNs and lung tissue proteases.  Histopathologic findings included AMs that
contained DPM, an increase in Type II cells, and the presence of particles in the interstitium and
tracheobronchial lymph nodes.
       Kaplan et al. (1982) evaluated the effects of subchronic exposure to DE on rats,
hamsters, and mice.  The exhaust was diluted to a concentration  of 1.5 mg/m3 DPM; exposures
were 20 h/day, 7 days/week. Hamsters were exposed for 86 days, rats and mice for 90 days.
There were no significant differences in mortality or growth rates between exposed and control
animals.  Lung weight relative to body weight of rats exposed for 90 days was  significantly
higher than the mean for the control group. Histological examination of tissues of all three
species indicated particle accumulation in the lungs and mediastinal lymph nodes. Associated
with the larger accumulations, there was a minimal increase in the thickness of the alveolar
walls, but the vast majority of the particles elicited no response.  After 6 mo of recovery,
considerable clearance of the DPM from the lungs occurred in all three species, as evaluated by
gross pathology and histopathology.  However, no quantitative estimate of clearance was
provided.
       Toxic effects in animals from acute exposure to DE appear to be primarily attributable to
the gaseous components (i.e., mortality from CO intoxication and lung injury caused by cellular
damage resulting from NO2 exposure). The results from short-term exposures  indicate that rats
experience minimal lung function impairment even at DE levels sufficiently high to cause
histological and cytological changes in the lung.  In subchronic studies of durations of 4 weeks
or more, frank adverse health effects are not readily apparent and, when found, are mild and
result from exposure to concentrations of about 6 mg/m3 DPM and durations of exposures of 20
h/day.  There is ample evidence that subchronic exposure to lower levels of DE affects the lung,
as indicated by accumulation of particles, evidence of inflammatory response, AM aggregation
and accumulation near the terminal bronchioles, Type II cell proliferation, and thickening of
alveolar walls adjacent to AM aggregates. Little evidence exists, however,  that subchronic
exposure to DE impairs lung function.

5.1.3.3. Chronic Exposures
5.1.3.3.1.  Effects on growth and longevity.  Changes in growth, body weight, absolute or
relative organ weights, and longevity can be measurable indicators of chronic toxic effects.
                                          5-29

-------
Such effects have been observed in some, but not all, of the long-term studies conducted on
laboratory animals exposed to DE. There was limited evidence for an effect on survival in the
published chronic animal studies; deaths occurred intermittently early in one study in female rats
exposed to 3.7 mg/m3 DPM; however, the death rate began to decrease after 15 mo, and the
survival rate after 30 mo was slightly higher than that of the control group (Ishinishi et al.,
1988).  Studies of the effects of chronic exposure to DE on survival and body weight or growth
are detailed in Table 5-3.
       Increased lung weights and lung-to-body weight ratios have been reported in rats, mice,
and hamsters.  These data are summarized in Table 5-4. In rats exposed for up to 36 weeks to
0.25 or 1.5 mg/m3 DPM, lung wet weights (normalized to body weight) were significantly
higher in the 1.5 mg/m3 exposure group than control values after 12 weeks of exposure
(Misiorowski et al., 1980). Rats and Syrian hamsters were exposed for 2 years (five 16-h
periods per week) to DE diluted to achieve concentrations of 0.7, 2.2, and 6.6 mg/m3 DPM
(Brightwell et al., 1986). At necropsy, a significant increase in lung weight was seen in both rats
and hamsters exposed to DE compared with controls.  This finding was more pronounced in the
rats in which the increase was progressive with both duration of exposure and particulate matter
level.  The increase was greatest at 30 mo (after the end of a 6-mo observation period in the
high-concentration male group where the lung weight was 2.7 times the control and at 24 mo in
the high-concentration female group [3.9 times control]). Heinrich et al. (1986a,b; see also
Stober, 1986) found a significant increase in wet and dry weights of the lungs of rats and mice
exposed at 4.24 mg/m3 DPM for 1 year in comparison with controls. After 2 years, the
difference was a factor of 2 (mice) or 3 (rats). After the same exposure periods, the hamsters
showed increases of 50% to 75%, respectively. Exposure to equivalent filtered DE (i.e., without
DPM) caused no significant effects in any of the species. Vinegar et al. (1980, 1981a,b) exposed
hamsters to two levels of DE with resultant concentrations of about 6 and  12 mg/m3 DPM for 8
h/day, 7 days/week for 6 mo. Both exposures significantly increased lung weight and lung-
weight to body-weight ratios. The difference between lung weights of exposed and control
hamsters exposed to 12 mg/m3 DPM was approximately twice that of those exposed to 6 mg/m3.
      Heinrich et al. (1995) reported that rats exposed to 2.5 and 7 mg/m3 DPM for 18 h/day,
5 days/week for 24 mo showed significantly lower body weights than controls starting at day
200 in the high-concentration group and at day 440 in the low-concentration group.  Body
weight in the low-concentration group was unaffected, as was mortality in any group. Lung
weight was increased in the 7 mg/m3 group starting at 3 mo and persisting throughout the study,
while the 2.5 mg/m3 group showed increased lung weight only at 22 and 24 mo of exposure.
Mice (NMRI strain) exposed to 7 mg/m3 in this study for 13.5 mo  had no increase in mortality
and insignificant decreases in body weight.  Lung weights were dramatically affected, with
                                         5-30

-------
Table 5-3. Effects of chronic exposures to diesel exhaust on survival and growth of laboratory animals
Species/sex
Rat, F344, M, F;
Monkey, Cynomolgus, M

Rat, F344, M;
Guinea Pig, Hartley, M


Hamster, Chinese, M


Rat, Wistar, M


Rat, F344, M, F;
Mouse, CD-I, M, F


Rat, Wistar, F;
Mouse, MMRI, F

Rat, F344
M, F

Raf
F344/Jcl.




Rat, Wistar, F;
Mouse, NMRI, F
(7 mg/m3 only)
Exposure
period
7 h/day
5 days/week
104 weeks
20 h/day
5 days/week
106 weeks

8 h/day
7 days/week
26 weeks
6 h/day
5 days/week
87 weeks
7 h/day
5 days/week
130 weeks

19 h/day
5 days/week
104 weeks
16 h/day
5 days/week
104 weeks
16 h/day
6 days/week
130 weeks



18 h/day
5 days/week
24 mo
Particles
(mg/m3)
2.0
0.23-0.36|J m MMD

0.25
0.75
1.5
0.19|JmMMD
6.0
12.0

8.3
0.71 |Jm MMD

0.35
3.5
7.1
0.25 |Jm MMD
4.24
0.35 |Jm MMD

0.7
2.2
6.6
0.1 ld
0.41d
1.08d
2.31d
3.72'
0.2-0.3 |Jm MMD
0.84
2.5
6.98
CxT
(mg-h/m3)
7,280


2,650
7,950
15,900

8,736
17,472

21,663


1,592
15,925
31,850

41,891


5,824
18,304
54,912
1,373
5,117
13,478
28,829
46,426

7,400
21,800
61,700
CO
(ppm)
11.5


2.7"
4.4"
7.1"

—
—

50.0


2.9
16.5
29.7

12.5


	
—
32.0
1.23
2.12
3.96
7.10
12.9

2.6
8.3
21.2
NO2
(ppm)
1.5


O.lb
0.27b
0.5b

—
—

4.0-6.0


0.05
0.34
0.68

1.5


	
—
—
0.08
0.26
0.70
1.41
3.00

0.3
1.2
3.8
S02
(ppm) Effects
0.8 No effects on growth or survival


— Reduced body weight in rats at
— 1.5 mg/m3
—

— No effect on growth
—

— No effect on growth or mortality
rates

— No effect on growth or mortality
— rates
—

1 . 1 Reduced body wts; increased
mortality in mice

— Growth reduced at 2.2 and
— 6.6 mg/m3
—
0.38 Concentration-dependent
1.06 decrease in body weight; earlier
2.42 deaths in females exposed to 3.72
4.70 mg/m3, stabilized by 1 5 mo
4.57

0.3 Reduced body weight in rats at
1.1 2.5 and 6.98 mg/m3 and no effect
3.4 in mice
Study
Lewis et al.
(1989)

Schreck et al.
(1981)


Vinegar et al.
(1981a,b)

Karagianes
etal. (1981)

Mauderly et al.
(1984, 1987a)


Heinrich et al.
(1986a)

Brightwell et al.
(1986)

Research
Committee for
HERP Studies
(1988)


Heinrich et al.
(1995)


-------
     Table 5-3.  Effects of chronic exposures to diesel exhaust on survival and growth of laboratory animals (continued)

Species/sex
Mice, NMRI, F;
C57BL/6N, F




Rats, F344, M


Mouse, CD-I,
M,F


Exposure
period
18 h/day
5 days/week
13. 5 mo
(NMRI)
24 mo
(C57BL/N)
16 h/day
5 days/week
23 mo
7 h/day
5 days/week
104 weeks

Particles
(mg/m3)
6.98





2.44
6.33
C x T CO NO2
(mg-h/m3) (ppm) (PPm)
35,500 - NMRI 14.2 2.3
38,300 - C57




19,520 — —
50,640 — —
S02

(ppm) Effects
2.8 Reduced body
weight in NMRI
mice but not in C57BL/6N mice




— Reduced survi
— after 300 days




val in 6.33 mg/m3
. Body weight

Study
Heinrich et al.
(1995)




Nikula et al.
(1995)
significantly lower at 6.33 mg/m3
0.35
3.5
7.1
0.25 urn MOD
1,274 3 0.1
12,740 17 0.3
25,844 30 0.7

— No effect on growth or mortality
— rates
—




Mauderly et al.
(1996)


"Estimated from graphically depicted mass concentration data.
L/i ""Estimated from graphically
presented mass concentration data for NO2
(assuming 90% NO and 10% NO2).



LtJ °Data for tests with light-duty engine; similar results with heavy-duty engine.
""Light-duty engine.
'Heavy-duty engine.

-------
Table 5-4. Effects of chronic exposures to diesel exhaust on organ weights and organ-to-body-weight ratios
Species/sex
Rat, F344, M;
Mouse, A/3, M;
Hamster, Syrian, M
Rat, F344, M, F


Rat, F344, M


Rat, F344, F


Rat, F344; M
Guinea Pig,
Hartley, M

Hamster, Chinese,
M

Rat, Wistar, F;
Hamster, Syrian,
M, F
Mouse, NMRI, F
Rat, F344, M, F;
Hamster, Syrian,
M, F
Cat, inbred, M


Mouse, NMRI, F
(7 mg/m3 only)

Exposure
period
20 h/day
7 days/week
12-13 weeks
7 h/day
5 days/week
52 weeks
20 h/day
5.5 days/ week
36 weeks
7 h/day
5 days/week
104 weeks
20 h/day
5.5 days/ week
78 weeks

8 h/day
7 days/week
26 weeks
19 h/day
5 days/week
120- 140 weeks

16 h/day
5 days/week
104 weeks
8 h/day
7 days/week
124 weeks
18 h/day
5 days/week
24 mo
Particles
(mg/m3)
1.5
0.19|lmMMD

2.0
0.23-0.36 |lm
MMD
0.25
1.5
0.19|lmMMD
2.0
0.23-0.36 |lm
MMD
0.25
0.75
1.5
0.19|lmMMD
6.0
12.0

4.24
0.35 |lm MMD


0.7"
2.2"
6.6
6.0"
12.0b

0.84
2.5
6.98
CxT CO
(mg-h/m3) (ppm)
2,520-2,730 —


3,640 12.7


990 —
5,940 —

7,280 11.5


2,145 —
6,435 —
12,870 —

8,736 —
17,472 —

48,336-56,392 12.5



5,824 —
18,304 —
54,912 32.0
41,664 20.2
83,328 33.2

7,400 2.6
21,800 8.3
61,700 21.2
NO2 SO2
(ppm) (ppm) Effects
— — No effect on liver, kidney, spleen, or
heart weights

1.6 0.83 No effects on weights of lungs, liver,
heart, spleen, kidneys, and testes

— — Increase in relative lung weight at
— — 1.5 mg/m3 only initially seen at
12 weeks
1.5 0.81 No effects on heart weights


— — No effects on heart mass
— —
— —

— — Increase in lung weight and lung/body
— — weight ratio

1.5 1.1 Increase in rat, mouse, and hamster
lung weight and dry weights


— — Increase in lung weight concentration
— — related in rats; heart weight/body
— — weight ratio greater at 6.6 mg/m3
2.7 2.7 Decrease in lung and kidney weights
4.4 5.0

0.3 0.3 Increased rat and mouse lung weight at
1.2 1.1 7 mg/m3 from 6 mo and at 2. 5 mg/m3
3.8 3.4 at 22 and 24 mo
Study
Kaplan etal. (1982)


Green etal. (1983)


Misiorowski et al.
(1980)

Vallyathan et al. (1986)


Penney etal. (1981)



Vinegar etal. (1981a,b)


Heinrich et al.
(1986a,b)
Stober(1986)

Brightwell et al. (1986)


Pepelkoetal. (1980b,
1981)
Moorman etal. (1985)
Heinrich et al. (1995)



-------
  Table 5-4. Effects of chronic exposures to diesel exhaust on organ weights and organ-to-body-weight ratios (continued)
Species/sex
Mouse, NMRI, F;
C57BL/6N, F

Rats, F344, M
Rat
Mouse

Exposure
period
18h/day
5 days/week
13.5 mo (NMRI)
24 mo
(C57BL/N)
16 h/day
5 days/week
23 mo



Particles C x T
(mg/m3) (mg-h/m3)
6.98 35,500 - NMRI
38,300 - C57

2.44 19,520
6.33 50,640
0.8
2.5
6.98
6.98
4.5
CO NO2 SO2
(ppin) (ppm) (PPm) Effects Study
14.2 2.3 2.8 Increased lung weight Heinrich et al. (1995)

— — — Increase in lung weight was significant Nikulaet al. (1995)
— — — at 2 and 6 mg/m3
Increased lung weight in rats and mice Henderson et al.
at 3. 5 and 7.1 mg/m3 (1988a)


"1 to 61 weeks of exposure.
b62 to 124 weeks of exposure.

-------
 increases progressing throughout the study from 1.5-fold at 3 mo to 3-fold at 12 mo.  Mice
(NMRI and C57BL/6N strains) were also exposed to 4.5 mg/m3 for 23 mo. In NMRI mice, the
body weights were reported to be significantly lower than controls, but the magnitude of the
change is not reported, so biological significance cannot be assessed.  Mortality was slightly
increased, but statistical significance is not reported. The C57BL/6N mice showed minimal
effects on body weight and mortality, which were not statistically significant. Lung weights
were dramatically affected in both strains.
       Nikula et al. (1995) exposed male and female F344  rats to DPM concentrations of 2.4
and 6.3 mg/m3 for 16 h/day, 5 days/week for 23 mo in a study designed to compare the effects of
DPM with those of carbon black. Significantly reduced survival was observed in males exposed
to 6.3 mg/m3 but not in females or at the lower concentration. Body weights were decreased by
exposure to 6.3 mg/m3 DPM in both male and female rats throughout the exposure period.
Significant increases in lung weight were first seen at 6  mo in the high-exposure group and at
12 to 18 mo in the low-exposure group.
       No evidence was found in the published literature that chronic exposure to DE affected
the weight of body organs other than the lung and heart (e.g., liver, kidney, spleen, or testes)
(Table 5-4).  Morphometric analysis  of hearts from rats  and guinea pigs exposed to 0.25, 0.75, or
1.5 mg/m3 DPM 20 h/day, 5.5 days/week for 78 weeks revealed no significant alteration in mass
at any exposure level or duration of exposure (Penney et al., 1981). The analysis included
relative wet weights of the right ventricle, left ventricle, combined atria, and ratio of right to left
ventricle. Vallyathan et al. (1986) found no significant  differences in heart weights and the ratio
of heart weight to body weight between rats exposed to  2 mg/m3 DPM for 7 h/day, 5 days/week
for 24 mo and their respective clean-air chamber controls. No significant differences were found
in the lungs, heart, liver, spleen, kidney, and testes of rats exposed for 52 weeks, 7 h/day, 5
days/week to diluted DE containing 2 mg/m3 DPM compared with their respective controls
(Green etal., 1983).

5.1.3.3.2. Effects on pulmonary function.  The effect of long-term exposure to DE on
pulmonary function has been evaluated in laboratory studies of rats, hamsters, cats, and
monkeys.  These studies are summarized in Table 5-5, along with more details on the exposure
characteristics, in general order of increasing dose (C x  T) of DPM. The text will be presented
using the same approach.
       Lewis et al. (1989) evaluated functional residual capacity and airway resistance and
conductance in 10 control and 10 diesel-exposed rats (2 mg/m3 DPM, 7 h/day, 5 days/week for
52 or 104 weeks). At the 104-week evaluation, the rats  were also examined for maximum flow
volume impairments. No evidence of impaired pulmonary  function as a result of the exposure to
                                          5-35

-------
       Table 5-5. Effects of diesel exhaust on pulmonary function of laboratory animals
Exposure
Species/sex period
Rat, F344, M, F 7 h/day
5 days/week
104 weeks
Monkey, 7 h/day
Cynomolgus, M 5 days/week
104 weeks
Rat, F344, M 20 h/day
5.5 days/week
87 weeks
Rat, Wistar, F 7-8 h/day
5 days/week
104 weeks
Hamster, Chinese, M 8 h/day
7 days/week
Y1 26 weeks
OJ
ON
Rat, F344, 7 h/day
M, F 5 days/week
130 weeks

Rat, F344, M, F; 16 h/day
Hamster, Syrian, M, F 5 days/week
104 weeks
Hamster, Syrian, M, F 19 h/day
5 days/week
120 weeks
Rat, Wistar, F 19 h/day
5 days/week
140 weeks
Cat, inbred, M 8 h/day
7 days/week
124 weeks
Particles
(mg/m3)
2.0
0.23-0.36 |lm
MMD
2.0
0.23-0.36 |lm
MMD
1.5
0.19|lmMMD
3.9
0.1 |J,mMMD
6.0
12.0
0.35
3.5
7.1
0.23-0.26 |lm
MMD
0.7
2.2
6.6
4.24
0.35 |lm MMD
4.24
0.35 |lm MMD
6.0"
12.0b
CXT
(mg-h/m3)
7,280
7,280
14,355
14,196-16,224
8,736
17,472
1,593
15,925
31,850

5,824
18,304
54,912
48,336
56,392
41,664
83,328
CO N02 S02
(ppin) (ppm) (PPm) Effects
11.5 1.5 0.8 No effect on pulmonary function
11.5 1.5 0.8 Decreased expiratory flow; no effect
on vital or diffusing capacities
7.0 0.5 — Increased functional residual capacity,
expiratory volume, and flow
18.5 1.2 3.1 No effect on minute volume,
compliance, or resistance
— — — Decrease in vital capacity, residual
— — — volume, and diffusing capacity;
increase in static deflation lung
volume
2.9 0.05 — Diffusing capacity, lung compliance
16.5 0.34 — reduced at 3. 5 and 7.1 mg/m3
29.7 0.68 —

— — — Large number of pulmonary function
— — — changes consistent with obstructive
— — — and restrictive airway diseases at
6.6 mg/m3 (no specific data provided)
12.5 1.5 1.1 Significant increase in airway
resistance
12.5 1.5 1.1 Decrease in dynamic lung compliance;
increase in airway resistance
20.2 2.7 2.1 Decrease in vital capacity, total lung
33.3 4.4 5.0 capacity, and diffusing capacity after
2 years; no effect on expiratory flow
Study
Lewis etal. (1989)
Lewis etal. (1989)
Gross (1981)
Heinrich et al. (1982)
Vinegar etal. (1980,
1981a,b)
Mauderly et al. (1988)
McClellan et al. (1986)

Brightwell et al. (1986)
Heinrich et al. (1986a)
Heinrich et al. (1986a)
Pepelko et al. (1980b,
1981)
Moorman etal. (1985)
"1 to 61 weeks exposure.
b62 to 124 weeks of exposure.

-------
DE was found in rats. Lewis et al. (1989) exposed male cynomolgus monkeys to DE for 7
h/day, 5 days/week for 24 mo. Groups of 15 monkeys were exposed to air, DE (2 mg/m3), coal
dust, or combined coal dust and DE.  Pulmonary function was evaluated prior to exposure and at
6-mo intervals during the 2-year exposure, including compliance and resistance, static and
dynamic lung volumes, distribution of ventilation, diffusing capacity, and maximum ventilatory
performance. There were no effects on lung volumes, diffusing capacity, or ventilation
distribution, so there was no evidence of restrictive disease. There was, however,  evidence of
obstructive airway disease as measured by low maximal flow rates in diesel-exposed monkeys.
At 18 mo of exposure, forced expiratory flow at 25% of vital capacity and forced expiratory
flow normalized to FVC were decreased.  The measurement of forced expiratory flow at 40% of
total lung capacity was significantly decreased at 12, 18, and 24 mo of exposure.  The finding of
an obstructive effect in monkeys contrasts with the finding of restrictive type effects in other
laboratory animal species (Vinegar et al., 1980, 1981a; Mauderly et al., 1988; Pepelko et al.,
1980b, 1981) and suggests a possible difference in effect between primate and small animal
respiratory tracts.  In these monkeys there were no specific histopathological effects reported
(see next section), although particle aggregates were reported in the distal airways, suggesting
more small airway deposition.
       Gross (1981) exposed rats for 20 h/day, 5.5 days/week for 87 weeks to DE containing 1.5
mg/m3 DPM. When the data were normalized (e.g., indices expressed in units of airflow or
volume for each animal by its own forced expiratory volume), there were no apparent
functionally significant changes occurring in the lungs at 38 weeks of exposure that might be
attributable to the inhalation of DE.  After 87 weeks of exposure, functional residual capacity
(FRC) and its component volumes (expiratory reserve [ER] and residual volume [RV]),
maximum expiratory flow (MEF) at 40% FVC, MEF at 20% FVC, and FEVai were
significantly greater in the diesel-exposed rats. An observed increase in airflow at the end of the
forced expiratory maneuver when a decreased airflow would be expected from the increased
FRC, ER,  and RV data (the typical scenario of human pulmonary disease) showed these data  to
be inconsistent with known clinically significant health effects.  Furthermore, although the lung
volume changes in the diesel-exposed rats could have been indicative of emphysema or chronic
obstructive lung disease, this interpretation was contradicted by the airflow data, which suggest
simultaneous lowering of the resistance of the distal airways.
       Heinrich et al. (1982) evaluated the pulmonary function of rats exposed to  a
concentration of 3.9 mg/m3 DPM for 7 to 8 h/day, 5 days/week for 2 years. When compared
with a control group, no significant changes in respiratory rate, minute volume, compliance, or
resistance  occurred in the exposed group (number of rats per group was not stated).
       Chinese hamsters (eight or nine per group) were exposed 8 h/day, 7 days/week, for 6 mo
to concentrations of either about 6 mg/m3 or about 12 mg/m3 DPM (Vinegar et al., 1980,
                                         5-37

-------
1981a,b).  Vital capacity, vital capacity/lung weight ratio, residual lung volume by water
displacement, and CO2 diffusing capacity decreased significantly in hamsters exposed to
6 mg/m3 DPM.  Static deflation volume-pressure curves showed depressed deflation volumes for
diesel-exposed hamsters when volumes were corrected for body weight and even greater
depressed volumes when volumes were corrected for lung weight.  However, when volumes
were expressed as percentage of vital capacity, the diesel-exposed hamsters had higher lung
volumes at 0 and 5 cm H2O. In the absence of confirmatory histopathology, the authors
tentatively concluded that these elevated lung volumes and the significantly reduced diffusing
capacity in the same hamsters were indicative of possible emphysematous changes in the lung.
Similar lung function changes were reported in hamsters exposed at 12 mg/m3 DPM, but detailed
information was not reported.  It was stated, however, that the decrease in vital capacity was
176% greater in the second experiment than in the first.
       Mauderly et al. (1988; see also McClellan et al., 1986) examined the impairment of
respiratory function in rats exposed for 7 h/day, 5 days/week for 24 mo to diluted DE with 0.35,
3.5, or 7.1 mg/m3 DPM. After 12 mo of exposure to the highest concentration of DE, the
exposed rats (n = 22) had lower total lung capacity (TLC), dynamic lung  compliance (Cdyn),
FVC, and CO diffusing capacity than controls (n = 23). After 24 mo of exposure to 7.1 mg/m3
DPM, mean TLC, Cdyn, quasi-static chord compliance, and CO diffusing capacity were
significantly lower than control values. Nitrogen  washout and percentage of FVC expired in 0.1
s were significantly greater than control values. There was no evidence of airflow obstruction.
The functional alterations were attributed to focal fibrotic and emphysematous lesions and
thickened alveolar membranes observed by histological examination.  Similar functional
alterations and histopathologic lesions were observed in the rats exposed to 3.5 mg/m3 DPM, but
such changes usually occurred later in the exposure period and were generally less pronounced.
There were no significant decrements in pulmonary function for the 0.35  mg/m3 group at any
time during the study nor were there reported histopathologic changes in this group.
       Mauderly et al. (1989) examined the effects of DE on normal rats and on rats with
experimentally induced pulmonary emphysema to see if emphysematous rats have increased
susceptibility to DPM.  The results from parallel lifetime exposures of these 2 groups  of rats at
3.5 mg/m3 DPM showed that only possibly 1 of 65 measured parameters gave results suggesting
that rats with emphysematous lungs might be more susceptible than rats with normal lungs to the
effects of DE exposure.
       Additional studies were conducted by Heinrich et al. (1986a,b; see also Stober, 1986) on
the effects of long-term exposure to DE on the pulmonary function of hamsters and rats. The
exhaust was diluted to achieve a concentration of 4.24 mg/m3 DPM; exposures were for 19
h/day, 5 days/week for a maximum of 120 weeks  (hamsters) or 140 weeks (rats). After 1 year of
exposure to the DE, the hamsters exhibited a significant increase in airway resistance and a
                                         5-38

-------
nonsignificant reduction in lung compliance.  For the same time period, rats showed increased
lung weights, a significant decrease in Cdyn, and a significant increase in airway resistance.
These indices did not change during the second year of exposure.
       Syrian hamsters and rats were exposed to 0.7, 2.2, or 6.6 mg/m3 DPM for five 16-h
periods per week for 2 years (Brightwell et al., 1986). There were no treatment-related changes
in pulmonary function in the hamster. Rats exposed to the highest concentration of DE
exhibited changes in pulmonary function (data not presented) that were reported to be consistent
with a concentration-related obstructive and restrictive disease.
       Pepelko et al. (1980b;  1981; see also Pepelko,  1982b) and Moorman et al. (1985)
measured the lung function of adult cats chronically exposed to DE.  The cats were exposed for
8 h/day and 7 days/week for 124 weeks. Exposures were at 6 mg/m3 for the first 61 weeks and
12 mg/m3 from weeks 62 to 124. No definitive pattern of pulmonary function changes was
observed following 61 weeks  of exposure; however, a classic pattern of restrictive lung disease
was found at 124 weeks. The significantly reduced lung volumes (TLC, FVC, FRC, and
inspiratory  capacity [1C]) and the significantly lower single-breath diffusing capacity, coupled
with normal values for dynamic ventilatory function (mechanics of breathing), indicate the
presence of a lesion that restricts inspiration but does not cause airway  obstruction or loss of
elasticity.  This pulmonary physiological syndrome is consistent with an interstitial fibrotic
response that was later verified by histopathology (Plopper et al., 1983).
       Pulmonary function impairment has been reported in rats, hamsters, cats,  and monkeys
chronically exposed to DE.  In all species but the monkey, the pulmonary function testing results
have been consistent with restrictive lung disease. The monkeys demonstrated evidence of small
airway obstructive responses.  The disparity between the findings in monkeys and those in rats,
hamsters, and cats could be in part the result of increased particle retention in the smaller species
resulting from (1) exposure to DE that has higher airborne concentrations of gases, vapors, and
particles and/or (2) longer duration of exposure.  The nature of the pulmonary impairment is also
dependent on the site of deposition and routes of clearance, which are determined by the
anatomy and physiology of the test laboratory species and the exposure regimen. The data on
pulmonary  function effects raise the possibility that DE produces small airway disease in
primates compared with primarily alveolar effects in small animals and that similar changes
might be expected in humans  and monkeys. The findings of Nikula et  al. (1997a,b) suggest that
a larger fraction of particles are translocated to the interstitium of the respiratory tract in
primates that are heavily exposed than in rats  that are heavily exposed,  including the interstitium
of the respiratory bronchioles, an anatomical site absent in rats.  Nikula and co-workers'
pulmonary  histopathological findings may have a relationship to these  functional findings (see
Chapter 3 for a complete discussion).  Unfortunately, the  available data in primates are too
limited to draw clear conclusions.
                                          5-39

-------
5.1.3.3.3.  Lung morphology, biochemistry, and lung lavage analysis. Several studies have
examined the morphological, histological, and histochemical changes occurring in the lungs of
laboratory animals chronically exposed to DE. The histopathological effects of diesel exposure
in the lungs of laboratory animals are summarized in Table 5-6, ranked in order of C x T. Table
5-6 also contains an expanded description of exposures.
       Kaplan et al. (1982) performed macroscopic and microscopic examinations of the lungs
of rats, mice, and hamsters exposed for 20 h/day, 7 days/week for 3 mo to DE containing 1.5
mg/m3 DPM. Gross examination revealed diffuse and focal deposition of the diesel particles that
produced a grayish overall appearance of the lungs with scattered, denser black areas. There was
clearance of particles via the lymphatics to regional lymph nodes. Microscopic examination
revealed no anatomic changes in the upper respiratory tract; the mucociliary border was normal
in appearance. Most of the particles were in macrophages, but some were free  as small
aggregates on alveolar and bronchiolar surfaces.  The particle-laden macrophages were often in
masses near the entrances of the lymphatic drainage and respiratory ducts. Associated with these
masses was a minimal increase in the thickness of the alveolar walls; however,  the vast majority
of the particles elicited no response.  After 6 mo of recovery, the lungs of all three species
contained considerably less pigment, as assessed by gross pathological and histopathological
examinations.
       Lewis et al. (1989;  see also Green et al.,  1983) performed serial histological
examinations of rat lung tissue exposed to DE containing 2 mg/m3 DPM for 7 h/day, 7
days/week for 2 years. Accumulations of black-pigmented AMs were seen in the alveolar ducts
adjacent to terminal bronchioles as early as 3 mo of exposure, and particles were seen within the
interstitium of the alveolar ducts.  These macular lesions increased in size up to 12 mo of
exposure.  Collagen or reticulum fibers were seen only rarely in association with deposited
particles; the vast majority of lesions showed no evidence of fibrosis. There was no evidence of
focal emphysema with the  macules. Multifocal histiocytosis (24% of exposed rats) was
observed only after 24 mo  of exposure. These lesions were most commonly observed
subpleurally and were composed of collections of degenerating macrophages and amorphous
granular material within alveoli, together with fibrosis and chronic inflammatory cells in the
interstitium.
                                          5-40

-------
Table 5-6. Histopathological effects of diesel exhaust in the lungs of laboratory animals
Species/sex
Rat, F344, M;
Mouse, A/3, M;
Hamster, Syrian, M
Monkey, Cynomolgus,
M
Rat, F344, M, F
Rat, Sprague-Dawley,
M; Mouse, A/HEJ, M
Hamster, Chinese, M
Hamster, Syrian, M, F
Rat, Wistar, M
Rat, F344, F
Rat, F344, M, F;
Mouse, CD-I,
M, F
Exposure
period
20 h/day
7 days/week
12-13 weeks
7 h/day
5 days/week
104 weeks
7 h/day
5 days/week
104 weeks
8 h/day
7 days/week
39 weeks
8 h/day
5 days/week
26 weeks
7-8 h/day
5 days/week
120 weeks
6 h/day
5 days/week
87 weeks
8 h/day
7 days/week
104 weeks
7 h/day
5 days/week
130 weeks
Particles C x T
(mg/m3) (mg-h/m3)
1.5 2,520-2,730
0.19|lmMDD
2.0 7,280
0.23-0.36 |lm
MOD
2.0 3,640
0.23-0.36 |lm
MOD
6.0 13,104
6.0 6,240
12.0 12,480
3.9 16,380-18,720
0.1 |J,mMDD
8.3 21,663
0.71 |lm MOD
4.9 28,538
0.35 1,592
3.5 15,925
7.1 31,850
0.23 |lm MOD
CO NO2 SO2
(ppm) (ppm) (PPm) Effects
— — — Inflammatory changes, increase in lung
weight, increase in thickness of alveolar
walls
11.5 1.5 0.8 AM aggregation; no fibrosis,
inflammation, or emphysema
11.5 1.5 0.8 Multifocal histiocytosis, inflammatory
changes, Type II cell proliferation,
fibrosis
— — — Increase in lung protein content and
collagen synthesis but a decrease in
overall lung protein synthesis in both
species; prolylhydroxylase activity
increased in rats in utero
— — — Inflammatory changes, AM accumu-
— — — lation, thickened alveolar lining, Type II
cell hyperplasia,edema, increase in
collagen
18.5 1.2 3.1 Inflammatory changes, 60%
adenomatous cell proliferation
50.0 4.0-6.0 — Inflammatory changes, AM aggregation,
alveolar cell hypertrophy, interstitial
fibrosis, emphysema (diagnostic method-
ology not described)
7.0 1.8 13.1 Type II cell proliferation, inflammatory
changes, bronchial hyperplasia, fibrosis
2.9 0.05 — Alveolar and bronchiolar epithelial
16.5 0.34 — metaplasia in rats at 3.5 and 7.0 mg/m3,
29.7 0.68 — fibrosis at 7.0 mg/m3 in rats and mice,
inflammatory changes
Study
Kaplan etal. (1982)
Lewis etal. (1989)
Bhatnagar et al.
(1980)
Pepelko (1982a)
Bhatnagar et al.
(1980)
Pepelko (1982a)
Pepelko (1982b)
Heinrich et al. (1982)
Karagianes et al.
(1981)
Iwai et al. (1986)
Mauderly et al.
(1987a)
Henderson et al.
(1988a)

-------
                  Table 5-6. Histopathological effects of diesel exhaust in the lungs of laboratory animals (continued)
Species/sex
Rats, SPF 344

Exposure
period
7h/day
5days/week
104 weeks

Particles
(mg/m3)
2 mg/m3 coal
dust (CD)
2 mg/m3 DPM
1 mg/m3
CD + 1 mg/m3
DPM
C x T CO NO2 SO2
(mg-h/m3) (Ppm) (PPm) (PPm) Effects
— — — — • Assessed pharmacological
responses of rat airway
smooth muscle in vitro
• Maximal contractile
responses to acetylcholine of
tissues from CD-, DPM-, and
Study
Fedenetal. (1985)

to
CD + DPM- exposed animals
significantly increased;
effects of CD and DPM were
additive
Maximal relaxation response
to isoproterenol increased
significantly by CD + DPM
exposure, but not by
individual treatments
The results indicate that
chronic exposure to CD,
DPM, and CD + DPM
produce differential
modifications in the behavior
of rat airway smooth muscle
Rat, Wistar, F;
Mouse, NMRI, F
(7 mg/m3 only)

Mouse, NMRI, F;
C57BL/6N, F


Mouse

Rat, M, F,
F344/Jcl.

Mouse, NMRI, F
18h/day
5 days/week
24 mo

18h/day
5 days/week
13.5 mo (NMRI)
24 mo
(C57BL/N)


16 h/day
6 days/week
130 weeks

19 h/day
5 days/week
120 weeks
0.8
2.5
6.98

6.98


4.5

o.ir
0.41"
1.08"
2.31"
3.72b
4.24
7,400
21,800
61,700

35,500 - NMRI
38,300 - C57




1,373
5,117
13,478
28,829
46,336
48,336
2.6
8.3
21.2

14.2




1.23
2.12
3.96
7.10
12.9
12.5
0.3
1.2
3.8

2.3




0.08
0.26
0.70
1.41
3.00
1.5
0.3
1.1
3.4

2.8




0.38
1.06
2.42
4.70
4.57
1.1
Bronchioalveolar hyperplasia, interstitial
fibrosis in all groups. Severity and
incidence increase with exposure
concentration
No increase in tumors. Noncancer
effects not discussed


No increase in tumors
Noncancer effects not discussed
Inflammatory changes Type II cell
hyperplasia and lung tumors seen at
>0.4 mg/m3; shortening and loss of cilia
in trachea and bronchi

Inflammatory changes, bronchiole-
alveolar hyperplasia, alveolar lipo-
proteinosis, fibrosis
Heinrich et al. (1995)






Research Committee
for HERP Studies
(1988)

Heinrich et al.
(1986a)

-------
       Table 5-6.  Histopathological effects of diesel exhaust in the lungs of laboratory animals (continued)
Exposure
Species/sex period
Rat, Wistar, F 19 h/day
5 days/week
140 weeks
Guinea Pig, Hartley, M 20 h/day
5.5 days/week
104 weeks



Cat, inbred, M 8 h/day
7 days/week
124 weeks
Rat, F344, M 16 h/day
5 days/week
23 mo
Mouse, CD- 1 , M,F 7 h/day
5 days/week
104 weeks


"Light-duty engine.
""Heavy-duty engine.
°1 to 61 weeks exposure.
d62 to 124 weeks of exposure.
Particles
(mg/m3)
4.24


0.25
0.75
1.5
6.0


6.0C
12.0d

2.44
6.33

0.35
3.5
7.1
0.25 |lm MOD





CxT
(mg-h/m3)
56,392


2,860
8,580
17,160
68,640


41,664
83,328

19,520
50,640

1,274
12,740
25,844






CO NO2 SO2
(ppm) (ppm) (PPm) Effects
12.5 1.5 1.1 Thickened alveolar septa; AM
aggregation; inflammatory changes;
hyperplasia; lung tumors
— — — Minimal response at 0.25 and
— — — ultrastructural changes at 0.75 mg/m3;
— — — thickened alveolar membranes; cell
— — — proliferation; fibrosis at 6.0 mg/m3;
increase in PMN at 0.75 mg/m3 and
1.5 mg/m3
20.2 2.7 2.1 Inflammatory changes, AM aggregation,
33.2 4.4 5.0 bronchiolar epithelial metaplasia, Type II
cell hyperplasia, peribronchiolar fibrosis
— — — AM hyperplasia, epithelial hyperplasia,
— — — inflammation, septal fibrosis,
bronchoalveolar metaplasia
3 0.1 — Exposure-related increase in lung soot,
17 0.3 — pigment-laden macrophages, lung
30 0.7 — lesions.
Bronchiolization in alveolar ducts at
7.1 mg/m3





Study
Heinrich et al.
(1986a)

Barnhart et al. (1981,
1982)
Vostaletal. (1981)
Wallace et al. (1987)


Plopperetal. (1983)
Hyde etal. (1985)

Nikulaetal. (1995)


Mauderly et al.
(1996)







AM = Alveolar macrophage.
PMN = Polymorphonuclear leukocyte.

-------
Epithelial lining cells adjacent to collections of pigmented macrophages showed a marked
Type II cell hyperplasia; degenerative changes were not observed in Type I cells.  Histological
examination of lung tissue from monkeys exposed for 24 mo in the same regimen as  used
for rats revealed aggregates of black particles, principally in the distal airways of the lung.
Particles were present within the cytoplasm of macrophages in the alveolar spaces as well
as the interstitium. Fibrosis, focal emphysema, or inflammation was not observed. No specific
histopathological lesions were reported for the monkey.
       Nikula et al. (1997a,b) reevaluated the lung tissue from this study.  They concluded that
there were no significant differences in the amount of retained particulate matter between
monkeys and rats exposed under the same conditions. The rats, however, retained a greater
portion of the parti culate matter in lumens of the alveolar ducts and alveoli than did the
monkeys. Conversely, monkeys retained a greater portion of the parti culate material in the
interstitium than did rats.  Aggregations of particle-laden macrophages in the alveoli were rare,
and there were few signs of particle-associated inflammation in the monkeys.  Minimal
histopathologic lesions were detected in the interstitium.
       Histopathological  effects of DE on the lungs of rats have been investigated by the Health
Effects Research Program on Diesel Exhaust (HERP) in Japan (Ishinishi et al., 1986,  1988).
Both light-duty (LD) and  heavy-duty (HD) diesel engines were used. The exhaust was diluted to
achieve nominal concentrations of 0.1 (LD only), 0.4 (LD and HD), 1 (LD and HD), 2 (LD and
HD), and 4 (HD only) mg/m3 DPM. Rats were exposed for 16 h/day, 6 days/week for 30 mo.
No histopathological changes were observed in the lungs of rats exposed to 0.4 mg/m3 DPM or
less. At concentrations above 0.4 mg/m3 DPM, severe morphological changes were observed.
These changes consisted of shortened and absent cilia in the tracheal and bronchial epithelium,
marked hyperplasia of the bronchiolar epithelium, and swelling of the Type II cellular
epithelium. These lesions appeared to increase in severity with increases in exhaust
concentration and duration of exposure. There was no difference in the degree of changes in
pulmonary pathology at the same concentrations  between the LD and the HD series.
       Heinrich et al. (1982) investigated histological changes occurring in the respiratory tract
of hamsters exposed to DE. Exposures were for 7 to 8  h/day, 5 days/week for 104 weeks to DE
diluted to achieve a concentration of 3.9 mg/m3 DPM. Significantly higher numbers of hamsters
in the group exposed to DE exhibited definite proliferative changes in the lungs compared with
the groups exposed to  particle-free DE or clean air.  Sixty percent of these changes were
described as adenomatous proliferations.
       Heinrich et al. (1995) reported increased incidence and severity of bronchioloalveolar
hyperplasia in rats exposed to 0.8, 2.5, and 7 mg/m3.  The lesion in the lowest concentration
group was described as very slight to moderate.  Slight to moderate interstitial fibrosis also
increased in incidence and severity in all exposed groups, but incidences were not reported.  This

                                          5-44

-------
chronic study also exposed NMRI mice to 7 mg/m3 for 13.5 mo and both NMRI and C56BL/6N
mice to 4.5 mg/m3 for 24 mo.  Noncancer histological endpoints are not discussed in any detail
in the report, which is focused on the carcinogenicity of diesel as compared with titanium
dioxide and carbon black.
       Iwai et al. (1986) performed serial histopathology on the lungs of rats at 1, 3, 6, 12, and
24 mo of exposure to DE. Exposures were for 8 h/day, 7 days/week for 24 mo; the exposure
atmosphere contained 4.9 mg/m3 DPM. At 1 and 3 mo of exposure, there were minimal
histological changes in the lungs of the exposed rats. After 6 mo of exposure, there were
particle-laden macrophages distributed irregularly throughout the lung and a proliferation of
Type II cells with adenomatous metaplasia in areas where the macrophages had accumulated.
After  1 year of exposure, foci of heterotrophic hyperplasia of ciliated or nonciliated bronchiolar
epithelium on the adjacent alveolar walls were more common, the quantity of deposited
particulate matter increased, and the number of degenerative AMs and proliferative lesions of
Type II or bronchiolar epithelial cells increased. After 2 years of exposure, there was a fibrous
thickening of the alveolar walls, mast-cell infiltration with  epithelial hyperplasia in areas where
the macrophages had accumulated, and neoplasms.
       Heinrich et al. (1986a; see also  Stober, 1986) performed histopathologic examinations of
the respiratory tract of hamsters, mice,  and rats exposed to DE that had 4 mg/m3 DPM.
Exposures were for 19 h/day, 5 days/week; the maximum exposure period was 120 weeks for
hamsters and mice and 140 weeks for rats.  Histological examination revealed different levels of
response among the three species. In hamsters, the exhaust produced thickened alveolar septa,
bronchioloalveolar hyperplasia, and what were termed emphysematous lesions (diagnostic
methodology not described).  In mice, bronchoalveolar hyperplasia occurred in 64% of the mice
exposed to the exhaust and in 5% of the controls.  Multifocal alveolar lipoproteinosis occurred in
71% and multifocal interstitial fibrosis  occurred in 43% of the mice exposed to exhaust but in
only 4% of the controls.  In exposed rats, there were severe inflammatory changes in the lungs,
as well as thickened septa, foci of macrophages, and hyperplastic and metaplastic lesions.
       Nikula et al. (1995) reported in  detail the nonneoplastic effects in male  and female
F344 rats exposed to 2.4 or 6.3 mg/m3 of DPM. At 3 mo in the low-concentration group,
enlarged particle-containing macrophages were found with minimal aggregation. With higher
concentration and longer duration of exposure, the number and size of macrophages and
aggregates increased. Alveolar epithelial hyperplasia was found starting at 3 mo and in all rats
at 6 mo.  These lesions progressed to chronic active inflammation, alveolar proteinosis, and
septal fibrosis  at 12 mo.  Other lesions  observed late in the study included bronchiolar-alveolar
metaplasia, squamous metaplasia, and squamous cysts. This study reports in detail the
progression of lesions in DE exposure and finds relatively little difference between the lesions
caused by DE  exposure and exposure to similar levels of carbon black particles.

                                          5-45

-------
       The effects of DE on the lungs of rats exposed to 8.3 ± 2.0 mg/m3 DPM were
investigated by Karagianes et al. (1981). Exposures were for 6 h/day, 5 days/week, for 4, 8, 16,
or 20 mo. Histological examinations of lung tissue noted focal aggregation of particle-laden
AMs, alveolar histiocytosis, interstitial fibrosis, and alveolar emphysema (diagnostic
methodology not described). Lesion severity was related to length of exposure. No significant
differences were noted in lesion severity among the DE, the DE plus coal dust (5.8 ± 3.5
mg/m3), or the high-concentration (14.9 ± 6.2 mg/m3) coal dust exposure groups following 20
mo of exposure.
       Histological changes in the lungs of guinea pigs exposed to diluted DE containing either
0.25, 0.75, 1.5, or 6.0 mg/m3 DPM were reported by Barnhart et al. (1981; 1982). Exposures at
0.75 and 1.5 mg/m3 for 2 weeks to 6 mo resulted in an uptake of exhaust particles by three
alveolar cell types (AMs, Type I cells, and interstitial macrophages) and also by granulocytic
leukocytes (eosinophils). The alveolar-capillary membrane increased in thickness as a result of
an increase in the absolute tissue volume of interstitium and Type II cells.  In a continuation of
these studies, guinea  pigs were exposed to DE (up to 6.0 mg/m3 DPM) for 2 years (Barnhart et
al., 1982). A minimal tissue response occurred at a concentration of 0.25 mg/m3.  After 9 mo of
exposure, there was a significant increase, about 30%, in Type I and II cells, endothelial cells,
and interstitial cells over concurrent age-matched controls; by 24  mo only macrophages and
Type II cells were significantly increased. As in the earlier study, ultrastructural evaluation
showed that Type I cells, AMs, and eosinophils phagocytized the diesel particles.  Exposure to
0.75 mg/m3 for 6 mo resulted in fibrosis in regions of macrophage clusters and in focal Type II
cell proliferation.  No additional information was provided regarding the fibrotic changes with
increasing concentration or duration of exposure.  With increasing concentration/duration of DE
exposure, Type II cell clusters occurred in some alveoli.  Intraalveolar debris was particularly
prominent after exposures at 1.5 and 6.0 mg/m3 and consisted of secretory products from Type II
cells.
       In studies conducted on hamsters, Pepelko (1982b) found  that the lungs of hamsters
exposed for 8 h/day,  7 days/week for 6 mo to 6 or 12 mg/m3 DPM were characterized by large
numbers of black AMs in the alveolar spaces, thickening of the alveolar epithelium, hyperplasia
of Type II cells, and edema.
       Lungs from rats and mice exposed to 0.35, 3.5, or 7.1 mg/m3 (0.23 to 0.26 |lm mass
median diameter [MMD]) for 7 h/day and 5  days/week showed pathologic lesions (Mauderly
et al., 1987a; Henderson et al.,  1988a). After 1 year of exposure at 7.1 mg/m3, the lungs of the
rats exhibited focal areas of fibrosis; fibrosis increased with increasing duration of exposure and
was observable in the 3.5-mg/m3 group of rats at 18 mo. The severity of inflammatory
responses and fibrosis was directly related to the exposure level.  In the 0.35 mg/m3 group of
rats, there was no inflammation or fibrosis. Although the mouse lungs contained higher burdens

                                          5-46

-------
of diesel particles per gram of lung weight at each equivalent exposure concentration, there was
substantially less inflammatory reaction and fibrosis than was the case in rats.  Fibrosis was
observed only in the lungs of mice exposed at 7.1 mg/m3 and consisted of fine fibrillar
thickening of occasional alveolar septa.
       Histological examinations were performed on the lungs of cats initially exposed to
6 mg/m3 DPM for 61 weeks and subsequently increased to 12 mg/m3 for Weeks 62 to 124 of
exposure.  Plopper et al. (1983; see also Hyde et al., 1985) concluded from the results of this
study that exposure to DE produced changes in both epithelial and interstitial tissue
compartments and that the focus of these lesions in the peripheral lung was the centriacinar
region where the alveolar ducts join the terminal conducting airways.  This conclusion was based
on the following evidence. The epithelium of the terminal and respiratory bronchioles  in
exposed cats consisted of three cell types (ciliated, basal, and Clara cells) compared with only
one type (Clara cells) in the controls.  The proximal acinar region showed evidence of
peribronchial fibrosis and bronchiolar epithelial metaplasia.  Type II cell hyperplasia was present
in the proximal interalveolar septa. The more distal alveolar ducts and the majority of the rest of
the parenchyma were unchanged from controls.  Peribronchial fibrosis was greater at the end of
6 mo in clean air following exposure, whereas the bronchiolar epithelial metaplasia was most
severe at the end of exposure. Following an additional 6 mo in clean air, the bronchiolar
epithelium more closely resembled the control epithelial cell population.
       Wallace et al. (1987) used transmission electron microscopy (TEM) to determine the
effect of DE on the intravascular and interstitial cellular populations of the lungs of exposed rats
and guinea pigs. Exposed animals and matched  controls were exposed to 0.25, 0.75, 1.5, or 6.0
mg/m3 DPM for 2, 6, or 10 weeks or 18 mo. The results inferred the following:  (1) exposure to
6.0 mg/m3 for 2 weeks was insufficient to elicit any cellular response, (2) both species
demonstrated an adaptive multicellular response to DE, (3) increased numbers of fibroblasts
were found in the interstitium from week 6 of exposure through month 18, and (4) there was no
significant difference in either cell type or number in alveolar capillaries, but there was a
significant increase at 18 mo in the mononuclear population in the interstitium of both species.
       Additional means for assessing the adverse effects of DE on the lung are to examine
biochemical and cytological changes in bronchoalveolar lavage fluid (BALF) and in lung tissue.
Fedan et al. (1985) performed studies to determine whether chronic exposure of rats affected the
pharmacologic characteristics of rat airway smooth muscle. Concentration-response
relationships for tension changes induced with acetylcholine, 5-hydroxytryptamine, potassium
chloride, and isoproterenol were assessed in vitro on isolated preparations of airway smooth
muscle (trachealis). Chronic exposure to DE significantly increased the maximal contractile
responses to acetylcholine compared with control values; exposure did not alter the  sensitivity
                                          5-47

-------
(EC50 values) of the muscles to the agonists. Exposures were to DE containing 2 mg/m3 DPM
for 7 h/day, 5 days/week for 2 years.
       Biochemical studies of BALF obtained from hamsters and rats revealed that exposures to
DE caused significant increases in lactic dehydrogenase, alkaline phosphatase, glucose-6-
phosphate dehydrogenase (G6P-DH), total protein, collagen, and protease (pH 5.1) after
approximately 1 year and 2 years of exposure (Heinrich et al., 1986a). These responses were
generally much greater in rats than in hamsters. Exposures were to DE containing 4.24 mg/m3
DPM for 19 h/day, 5 days/week for 120 (hamsters) to 140 (rats) weeks.
       Protein, P-glucuronidase activity, and acid phosphatase activity were significantly
elevated in BALF obtained from rats exposed to DE containing 0.75 or 1.5 mg/m3 DPM for
12 mo (Strom, 1984).  Exposure for 6 mo resulted in significant increases in acid phosphatase
activity at 0.75 mg/m3 and in protein, P-glucuronidase, and acid phosphatase activity at the
1.5 mg/m3 concentration.  Exposure at 0.25 mg/m3 DPM did not affect the three indices
measured at either time period. The exposures were for 20 h/day, 5.5 days/week for 52 weeks.
       Additional biochemical studies (Misiorowski et al., 1980) were conducted on laboratory
animals exposed under the same conditions and at the same  site as reported  on by Strom (1984).
In most cases, exposures at 0.25 mg/m3 did not cause any significant changes. The DNA content
in lung tissue and the rate of collagen synthesis were significantly increased at 1.5  mg/m3 DPM
after 6 mo. Collagen deposition was not affected. Total lung collagen content increased in
proportion to the increase in lung weight. The activity of prolyl hydroxylase was significantly
increased at 12 weeks at 0.25 and 1.5 mg/m3; it then decreased with age. Lysal oxidase activity
did not change. After 9 mo of exposure, there were significant increases in lung phospholipids
in rats and guinea pigs exposed to 0.75 mg/m3 and in lung cholesterol in rats and guinea pigs
exposed to 1.5 mg/m3. Pulmonary prostaglandin dehydrogenase activity was stimulated by an
exposure at 0.25 mg/m3 but was not affected by exposure at 1.5 mg/m3 (Chaudhari et al., 1980,
1981). Exposures for 12 or 24 weeks resulted in a concentration-dependent lowering of this
enzyme activity. Exposure of male rats and guinea pigs at 0.75 mg/m3 for 12 weeks did not
cause any changes in glutathione levels of the lung, heart,  or liver. Rats exposed for 2 mo at
6 mg/m3 showed a significant depletion of hepatic glutathione, whereas the lung showed an
increase of glutathione (Chaudhari and Dutta, 1982).  Schneider and Felt (1981) reported that
similar exposures did not substantially change adenylate cyclase and guanylate cyclase activities
in lung or liver tissue of exposed rats and guinea pigs.
       Bhatnagar et al. (1980; see also Pepelko, 1982a) evaluated changes in the biochemistry of
lung connective tissue of diesel-exposed rats and mice.  The mice were exposed for 8 h/day and
7 days/week for up to 9 mo to exhaust containing 6 mg/m3 DPM.  Total lung protein content was
measured, as  was labeled proline and labeled leucine. Leucine incorporation is an index of total
protein synthesis, although collagen is very low in leucine. Proline incorporation reflects

                                          5-48

-------
collagen synthesis. Amino acid incorporation was measured in vivo in the rat and in short-term
organ culture in mice. Both rats and mice showed a large increase in total protein (41% to 47%
in rats), while leucine incorporation declined and proline incorporation was unchanged. These
data are consistent with an overall depression of protein synthesis in diesel-exposed animals and
also with a relative increase in collagen synthesis compared to other proteins. The increase in
collagen synthesis suggested proliferation of connective tissue and possible fibrosis (Pepelko,
1982a).
       A number of reports (McClellan et al., 1986; Mauderly et al., 1987a, 1990a; Henderson
et al., 1988a) have addressed biochemical and cytological changes in lung tissue and BALF of
rodents exposed for 7 h/day, 5 days/week for up to 30 mo at concentrations of 0, 0.35, 3.5, or
7.1  mg/m3 DPM.  At the lowest exposure level (0.35 mg/m3), no biochemical or cytological
changes occurred in the BALF  or in lung tissue in either Fischer 344 rats or CD-I  mice.
Henderson et al. (1988a) provide considerable time-course information on inflammatory events
taking place throughout a chronic exposure.  A chronic inflammatory response was seen at the
two higher exposure levels in both species, as evidenced by increases in inflammatory cells
(macrophages and neutrophils), cytoplasmic and lysosomal enzymes (lactate dehydrogenase,
glutathione reductase, and P-glucuronidase), and protein (hydroxyproline) in BALF.  Analysis
of lung tissue indicated similar  changes in enzyme levels as well as an increase in total lung
collagen content.  After 18 mo  of exposure, lung tissue glutathione was depleted in a
concentration-dependent fashion in rats but was slightly increased in mice.  Lavage fluid levels
of glutathione and glutathione reductase activity increased in a concentration-dependent manner
and were higher in mice than in rats.
       Rats exposed for up to 17 days to diluted DE (3.5 mg/m3 DPM) had a fivefold increase in
the  bronchoconstrictive prostaglandin PGF2,, and a twofold increase in the inflammatory
leukotriene LTB4. In similarly  exposed mice, there was a twofold increase in both parameters.
These investigators (Henderson et al., 1988a,b) concluded that the release of larger amounts of
such mediators of inflammation from the alveolar phagocytic cells of rats accounted for the
greater fibrogenic response seen in that species.
       Biochemical analysis of lung tissue from cats exposed for 124 weeks and held in clean
air for an additional 26 weeks indicated increases of lung collagen; this finding was confirmed
by an observed increase in total lung wet weight and in connective tissue fibers estimated
morphometrically (Hyde et al.,  1985).  Exposures were for 7 h/day, 5 days/week at 6 mg/m3
DPM for 61 weeks and at 12 mg/m3 for weeks 62 to 124.
       Heinrich et al. (1995) reported on bronchoalveolar lavage in animals exposed for 24 mo
and found exposure-related increases in lactate dehydrogenase, P-glucuronidase, protein, and
hydroxyproline in groups exposed to 2.5  or 7 mg/m3, although detailed data are not presented.
Lavage analyses were not carried out in concurrent studies in mice.

                                          5-49

-------
       The pathogenic sequence following the inhalation of DE as determined
histopathologically and biochemically begins with the interaction of diesel particles with airway
epithelial cells and phagocytosis by AMs. The airway epithelial cells and activated macrophages
release chemotactic factors that attract neutrophils and additional AMs.  As the lung burden of
DPM increases, there is an aggregation of particle-laden AMs in alveoli adjacent to terminal
bronchioles, increases in the number of Type II cells lining particle-laden alveoli, and the
presence of particles within alveolar and peribronchial interstitial tissues and associated lymph
nodes.  The neutrophils and macrophages release mediators of inflammation and oxygen radicals
that deplete a biochemical defense mechanism of the lung (i.e., glutathione). As will be
described later in more detail, other defense mechanisms are affected, particularly the decreased
viability of AMs, which leads to decreased  phagocytic activity and death of the macrophage.
The latter series of events may result in the presence of pulmonary inflammatory, fibrotic, or
emphysematous lesions.  The data suggest that there may be a threshold of exposure to DE
below which adverse structural and biochemical effects may not occur in the lung; however,
differences in the anatomy and pathological responses of laboratory animals coupled with their
lifespans compared with humans make a  determination  of human levels of exposure to DE
without resultant pulmonary injury a difficult and challenging endeavor.

5.1.3.3.4. Effects on pulmonary defense mechanisms.   The respiratory system has a number of
defense mechanisms that negate or compensate for the effects produced by the injurious
substances that repeatedly insult the upper respiratory tract,  the tracheobronchial airways, and
the alveoli. The effects of exposure to DE on the pulmonary defense mechanisms of laboratory
animals as well as more details on exposure atmosphere are summarized in Table 5-7 and ranked
by cumulative exposure (C * T).
       Several studies have been conducted investigating the effect of inhaled DE on the
deposition and fate of inert tracer particles or diesel particles themselves. Lung clearance of
deposited particles occurs in two distinct phases: a rapid phase (hours to days) from the
tracheobronchial region via the mucociliary escalator and a much slower phase (weeks to
months) from the nonciliated pulmonary  region via, primarily but not solely, AMs. Battigelli et
al. (1966) reported impaired tracheal mucociliary clearance  in vitro in excised trachea from rats
exposed for single or repeated exposures of 4 to 6 h at two dilutions of DE that resulted in
exposures of approximately 8 and  17 mg/m3 DPM. The exposure to 17  mg/m3 resulted in
                                          5-50

-------
Table 5-7. Effects of exposure to diesel exhaust on the pulmonary defense mechanisms of laboratory animals
Species/sex

Guinea Pig,
Hartley


Rat, F344, M





Rat, F344, M







Rat F344/Crl,
M, F
Mouse, CD, M, F





Rat, Wistar, F


Rat, F344/Crl, M


Exposure
period

20 h/day
5.5
days/week
8 weeks
7 h/day
5 days/week
104 weeks



20 h/day
5.5
days/week
26, 48, or
52 weeks



7 h/day
5 days/week
104 weeks
(rat),
78 weeks
(mouse)


18 h/day
5 days/week
24 mo
7 h/day
5 days/week
24 mo
Particles
(mg/m3)

0.25
1.5
0.19|lmMDD

2.0
0.23-0.36 |lm
MOD



0.25"
0.75"
1.5"
0.19|lmMDD




0.35
3.5
7.0
0.25 |lm MOD




0.8
2.5
7.1
3.49


CxT
(mg-h/m3)

220
1,320


7,280





715-8,580







1,274C
12,740°
25,480°





7,400
21,800
61,700
12,704


CO NO2 SO2
(ppm) (ppm) (ppm)
Alveolar macrophage status
2.9 — —
7.5 — —


11.5 1.5 0.81





2.9 — —
4.8 — —
7.5 — —





2.9 0.05 —
16.5 0.34 —
29.7 0.68 —





2.6 0.3 —
8.3 1.1 —
21.2 3.4 —
9.8 1.2 —


Effects

No significant changes in absolute numbers
of AMs


Little effect on viability, cell number, oxygen
consumption, membrane integrity, lyzomal
enzyme activity, or protein content of AMs;
decreased cell volume and ruffling of cell
membrane and depressed luminescence of
AM
AM cell counts proportional to concentration
of DPM at 0.75 and 1.5 mg/m3; AM
increased in lungs in response to rate of DPM
mass entering lung rather than total DPM
burden in lung; increased PMNs were
proportional to inhaled concentrations and/or
duration of exposure; PMNs affiliated with
clusters of aggregated AM rather than DPM
Significant increases of AM in rats and mice
exposed to 7.0 mg/m3 DPM for 24 and 18
mo, respectively, but not at concentrations of
3.5 or 0.35 mg/m3 DPM for the same expos-
ure durations; PMNs increased in a dose-
dependent fashion in both rats and mice
exposed to 3.5 or 7.0 mg/m3 DPM and were
greater in mice than in rats
Changes in differential cell counts in lung
lavage

Significantly reduced AM in lavage at 24 mo


Study

Chen et. al. (1980)



Castranova et al. (1985)





Strom (1984)
Vostal et al. (1982)






Henderson et al. (1988a)







Heinrich et al. (1995)


Mauderly et al. (1990a)



-------
             Table 5-7. Effects of exposure to diesel exhaust on the pulmonary defense mechanisms of laboratory animals
             (continued)
to
Species/sex

Rat, M, F






Rat, Wistar, F


Rat, F344, M,
developing 0-6
mo
adult 6-12 mo
Rat, F344, M, F



Rat, F344, M





Rat, Sprague-
Dawley, M


Exposure
period

7h/day
5 days/week
12 weeks




18h/day
5 days/week
24 mo
7h/day
5 days/week
6 mo

7h/day
5 days/week
1 8 weeks

7h/day
5 days/week
26-104
weeks


4-6 h/day
7 days/week
0.1 to 14.3
weeks
Particles
(mg/m3)

0.2
1.0
4.5
0.25|lm MOD



0.8
2.5
7.1
3.55



0.15
0.94
4.1
<0.5 |J,mMDD
2.0
0.23-0.36 |lm
MOD



0.9
8.0
17.0

C x T CO NO2 SO2
(mg-h/m3) (Ppm) (PPm) (PPm) Effects
Clearance
84 — — — Evidence of apparent speeding of tracheal
420 — — — clearance at the 4. 5 mg/m3 level after 1 week
1,890 — — — of 99mTc macroaggregated-albumin and
reduced clearance of tracer aerosol in each of
the three exposure levels at 12 weeks; indica-
tion of a lower percentage of ciliated cells at
the 1.0 and 4.5 mg/m3 levels
7,400 2.6 0.3 0.3 Significant increase in clearance half-time of
21,800 8.3 1.2 1.1 inhaled labeled aerosols in all groups at 3- 18
61,700 21.2 3.8 3.4 mo
3,321 7.9 9.5 Clearance of 2 |_lm, aluminosilicate particles.
Half-time significantly increased in adult, not
different in developing rats

94.5 — — — Lung burdens of DPM were concentration-
592 — — — related; clearance half-time of DPM almost
2,583 — — — double in 4. 1 mg/m3 group compared to 0. 1 5
mg/m3 group
1,820-7,280 11.5 1.5 0.8 No difference in clearance of 59Fe3O4
particles 1 day after tracer aerosol
administration; 120 days after exposure
tracer aerosol clearance was enhanced; lung
burden of DPM increased significantly
between 12 and 24 mo of exposure
2.5-10,210 — 5.0 0.2 Impairment of tracheal mucociliary clearance
— 2.7 0.6 in a concentration-response manner
— 8.0 1.0

Study

Wolff and Gray (1980)






Heinrich et al. (1995)


Mauderly et al. (1987b)



Griffisetal. (1983)



Lewis etal. (1989)





Battigelli et al. (1966)




-------
        Table 5-7. Effects of exposure to diesel exhaust on the pulmonary defense mechanisms of laboratory animals
        (continued)
Species/sex
Rat, F344,
M, F






Rat, F344/Crl, M




MiceCD-l,F







Mice, CR/CD-1, F







Exposure
period
7h/day
5 days/week
130 weeks





7h/day
5 days/week
24 mo


7h/day
5 days/week
4, 12, or
26 weeks




8h/day
7 days/week
2 h up to
46 weeks




Particles
(ing/in3)
0.35
3.5
7.1
0.25 |lm MOD




3.49




2.0
0.23-0.36 |lm
MOD





5.3 to 7.9







C x T CO NO2 SO2
(mg-h/m3) (ppni) (PPm) (PPm) Effects Study
1,593 2.9 0.1 — No changes in tracheal mucociliary clearance Wolff et al. (1987)
15,925 16.5 0.3 — after 6, 12, 18, 24, or 30 mo of exposure;
31,850 29.7 0.7 — increases in lung clearance half-times as
early as 6 mo at 7.0 mg/m3 level and 18 mo
at 3.5 mg/m3 level; no changes seen at 0.35
mg/m3 level; after 24 mo of diesel exposure,
long-term clearance half-times were
increased in the 3.5 and 7.0 mg/m3 groups
12,704 9.8 1.2 — Doubling of long-term clearance half-time Mauderly et al. (1990a)
for clearance of 1.0 |_lm aluminosilicate
particles. Less effect on clearance in animals
with experimentally induced emphysema
Microbial-induced mortality
280-1,820 11.5 1.5 0.8 Mortality similar at each exposure duration Hahon et al. (1985)
when challenged with Ao/PR/8/34 influenza
virus; in mice exposed for 3 and 6 mo, but
not 1 mo, there were increases in the
percentages of mice having lung
consolidation, higher virus growth, depressed
interferon levels, and a fourfold reduction in
hemagglutinin antibody levels
11-20,350 19 1.8 0.9 Enhanced susceptibility to lethal effects of Campbell et al. (1980, 1981)
to to to S. pyogenes infections at all exposure
22 3.6 2.8 durations (2 and 6 h; 8, 15, 16, 307, and 321
days); inconclusive results with
S. typhimurium because of high mortality
rates in controls; no enhanced mortality
when challenged with A/PR8-3 influenza
virus
"Chronic exposure lasted 52 weeks.
""Chronic exposure lasted 48 weeks.
"Calculated for 104-week exposure.
DPM = Diesel particulate matter.
AM = Alveolar macrophage.
PMN = Polymorphonuclear leukocyte.

-------
decreased clearance after a single exposure as well as after a cumulative exposure of 34 or 100 h.
Clearance was reduced to a lesser extent and in fewer tracheas from animals exposed to 8 mg/m3
for a cumulative exposure of 40 h. Lewis et al. (1989) found no difference in the clearance of
59Fe3O4 particles (1.5  |im MMAD, ag 1.8)  1 day after dosing control and DE-exposed rats (2
mg/m3, 7 h/day, 5 days/week for 8 weeks).
      Wolff et al. (1987) and Wolff and Gray (1980) studied the effects of both subchronic and
chronic DE exposure on the tracheal  clearance of particles. Tracheal clearance assessments were
made by measuring the retention of radiolabeled technetium macroaggregated-albumin
remaining 1 h after instillation in the distal trachea of rats. In the subchronic studies, rats were
exposed to 0.2, 1.0, or 4.5 mg/m3 DPM on a 7 h/day, 5 days/week schedule for up to 12 weeks.
After 1 week there was an apparent speeding of tracheal clearance at the 4.5 mg/m3 exposure
level (p=0.10), which returned toward baseline after 6 weeks and was slightly below the baseline
rate at 12 weeks. In the 1.0 mg/m3 group, there was a progressive significant reduction in the
clearance rate at 6 and 12 weeks of exposure. There was a trend toward reduced clearance in the
0.2 mg/m3 group.  Scanning electron micrographs indicated minimal changes in ciliary
morphology; however, there was an indication of a lower percentage of ciliated cells at the 1.0
and 4.5 mg/m3 levels.  In the chronic studies, rats were exposed to 0, 0.35, 3.5, or 7.1 mg/m3 for
7 h/day, 5 days/week for 30 mo. There were no significant differences  in tracheal clearance
rates between the control group and any of the exposure groups after 6, 12, 18, 24, or 30 mo of
exposure.  The preexposure measurements for all groups, however, were significantly lower than
those during the exposure period, suggesting a possible age effect.  The preexposure value for
the 3.5-mg/m3 group was also significantly lower than the control group.
      There is a substantial body of evidence for an impairment of particle clearance from the
bronchiole-alveolar region  of rats following exposure to DE.  Griffis et al. (1983) exposed rats 7
h/day, 5 days/week for 18 weeks to DE at 0.15, 0.94,  or 4.1 mg/m3 DPM. Lung burdens of the
0.15, 0.94, and 4.1 mg/m3 levels were 35, 220, and 1,890 |lg/g lung, respectively, 1  day after the
18-week exposure. The clearance half-time of the DPM was significantly greater, almost
double, for the 4.1 mg/m3 exposure group than for those of the lower exposure groups, 165 ± 8
days versus 99 ± 8 days (0.94 mg/m3) and 87 ± 28 days (0.15 mg/m3), respectively.
      Chan et al. (1981) showed a dose-related slowing of 14C-diesel particle clearance in rats
preexposed to DE at 0.25 or 6 mg/m3 particulate matter for 20 h/day, 7  days/week for 7 to 112
days.  Clearance was inhibited in the 6 mg/m3 group when compared by length of exposure or
compared with the 0.25 mg/m3 or control rats at the same time periods.
      Heinrich et al. (1982) evaluated lung clearance in rats exposed for approximately 18 mo
at 3.9 mg/m3 DPM for 7 to 8 h/day, 5 days/week. Following exposure to 59Fe2O3-aerosol, the
rats were returned to the DE exposure and the radioactivity was measured over the thoracic area
                                          5-54

-------
at subsequent times. The biological half-life of the iron oxide deposited in the rats' lungs was
nearly twice that of controls.
       Heinrich also used labeled iron oxide aerosols to study clearance in rats exposed to 0.8,
2.5, or 7 mg/m3 diesel DPM for 24 mo (Heinrich et al., 1995).  Clearance measurements were
carried out at 3, 12, and 18 mo of exposure. Half-times of clearance were increased in a
concentration- and duration-related manner in all exposed groups, with a range of a 50%
increase in the 0.8 mg/m3 group at  3 mo to an 11-fold increase in the 7 mg/m3 group at 19 mo.
The differential cell counts in these animals were stated to have shown clear effects in the 2.5
and 7 mg/m3 groups, but specific information about the changes is not reported.
       Wolff et al. (1987) investigated alterations in DPM clearance from the lungs of rats
chronically exposed to DE at 0, 0.35, 3.5, or 7.1 mg/m3 DPM for 7 h/day,  5 days/week for up to
24 mo. Progressive increases in lung burdens were observed over time in  all groups; levels of
DPM in terms of milligrams per lung were 0.60, 11.5, and 20.5 after 24 mo of exposure at the
0.35, 3.5, or 7.1 mg/m3 exposure levels, respectively. There were significant increases in 16-day
clearance half-times of inhaled radiolabeled particles of 67Ga2O3 (0.1 |im MMD) as early as
6 mo at the 7.1 mg/m3 level and 18 mo at the 3.5 mg/m3 level; no significant changes were seen
at the 0.35 mg/m3 level at any time point examined.  Rats inhaled fused aluminosilicate particles
(2 |im MMAD) labeled with 134Cs  after 24 mo of DE exposure; long-term clearance half-times
were 79, 81, 264, and 240 days for the 0,  0.35, 3.5, and 7.1 mg/m3 groups, respectively.
Differences were significant between the  control and the 3.5 and 7.1 mg/m3 groups (p < 0.01),
but not between the control and the 0.35 mg/m3 group.
       Mauderly et al. (1987b) compared the effects of DE in the developing lung to the adult
lung by exposing groups of male F344 rats to 3.5 mg/m3 for 7 h/day, 5 days/week for 6 mo.   One
group (adult) was exposed between 6 and 12  mo of age, and the other was exposed beginning in
utero and until 6 mo of age. Clearance of an inhaled monodisperse 2 |im aluminosilicate
particle was measured after exposure for 6 mo. The  clearance half-time of the slow phase was
found to be doubled in the diesel-exposed adult rats compared with age-matched controls and
was not significantly affected in developing rat lungs.
       Mauderly et al. (1990a) compared the effects of DE in normal lungs with rats in which
emphysema had been induced experimentally by instillation of elastase 6 weeks before DE
exposures.  The rats were exposed  to 3.5 mg/m3 DPM for 7 h/day, 5 days/week for 24 mo.
Measurements included histopathology, clearance, pulmonary function, lung lavage, and
immune response. In the rats that were not pretreated with elastase, there was a significant
reduction in the number of macrophages recovered by pulmonary lavage in contrast to the
increases in macrophages reported  by Strom (1984) and Henderson et al. (1988). The half-time
of the slow phase of clearance of inhaled, 1 |im, monodisperse particles was doubled in the
animals without elastase pretreatment. The elastase pretreatment did not affect clearance  in

                                          5-55

-------
unexposed animals but significantly reduced the effect of diesel. The clearance half-time was
significantly less in elastase-pretreated, diesel-exposed animals than in diesel-exposed normal
animals.  Many other effects measured in this study were also less affected by diesel exposure in
elastase-treated animals. Measurements of lung burden of DPM showed that elastase-pretreated
animals accumulated less than half as much DPM mass as normal animals exposed at the same
time, suggesting that the difference in effect could be explained by differences in dose to the
lung.  The composite results of this study indicate that, at least in a murine laboratory animal
species, the presence of a pulmonary restrictive disease such as emphysema does not seem to
exacerbate the effects of chronic exposure to diesel.
       Lewis et al. (1989) conducted lung burden and 59Fe3O4 tracer studies in rats exposed for
12 and 24 mo to 2 mg/m3 DPM (7 h/day, 5 days/week). The slope of the Fe3O4 clearance curve
of the DPM-exposed animals was significantly steeper than that of the controls, indicating a
more rapid alveolar clearance of the deposited 59Fe3O4.  After 120 days from the inhalation of the
tracer particle, 19% and 8% of the initially deposited 59Fe3O4 were present in the lungs of control
and DE-exposed rats, respectively. The lung burden of DPM,  however, increased significantly
between 12 and 24 mo of exposure (0.52 to 0.97% lung dry weight), indicating a later dose-
dependent inhibition of clearance.
       Alveolar macrophages, because of their phagocytic and digestive capabilities, are one of
the prime defense mechanisms of the alveolar region of the lung against inhaled particles. Thus,
characterization of the effects of DE  on various properties of AMs provides information on the
integrity or compromise of a key pulmonary defense mechanism.  The physiological viability of
AMs from diesel-exposed rats was assessed after 2 years of exposure by Castranova et al.
(1985). The 7 h/day, 5 days/week exposure at 2 mg/m3 DPM had little effect on the following:
viability, cell number, oxygen consumption, membrane integrity, lysosomal enzyme activity, or
protein content of the AMs. A slight decrease in cell volume, a decrease in chemiluminescence
indicative of a decreased secretion of reactive oxygen species,  and a decrease in ruffling of the
cell membrane were observed. These latter findings could be reflective of an overall reduction
in phagocytic activity.
       Exposure to DE has been reported both to increase the number of recoverable AMs from
the lung (Strom, 1984; Vostal et al.,  1982; Henderson et al., 1988a) or to produce no change in
numbers (Chen et al., 1980; Castranova et al., 1985). Strom (1984) found that in rats exposed  to
0.25 mg/m3 DPM for 20 h/day, 5.5 days/week for 6 mo or 1 year, as well as in the controls, BAL
cells consisted entirely of AMs, with no differences in the cell counts in the lavage fluid. At the
higher concentrations, 0.75  or 1.5 mg DPM/m3, the count of AM increased proportionally with
the exposure concentration; the results were identical for AMs at both 6 and 11 or 12 mo of
exposure.  The increase in AM counts was much larger after exposure to 1.5 mg/m3 DPM for
6 mo than after exposure to 0.75 mg/m3 for 1 year, although the total mass (calculated as C x T)

                                          5-56

-------
of deposited particulate burden was the same.  These data suggested to the authors that the
number of lavaged AMs was proportional to the mass influx of particles rather than to the actual
DPM burden in the lung. These results further implied that there may be a threshold for the rate
of mass influx of DPM into the lungs of rats above which there was an increased recruitment of
AMs. Henderson  et al. (1988a) reported similar findings of significant increases of AMs in rats
and mice exposed to 7.1 mg/m3 DPM for 18 and 24 mo, respectively, for 7 h/day, 5 days/week,
but not at concentrations of 3.5 or 0.35 mg/m3 for the same exposure durations. Chen et al.
(1980), using an exposure regimen of 0.25 and 1.5 mg/m3 DPM for 2 mo and 20 h/day and
5.5 days/week, found no  significant changes in absolute numbers of AMs from guinea pig
BALF, nor did Castranova et al. (1985) in rat BALF following exposure to 2 mg/m3 DPM for
7 h/day, 5 days/week for 2  years.
       A similar inflammatory response was noted by Henderson et al. (1988a) and Strom
(1984), as evidenced by an increased number of PMNs present in BALF from rodents exposed to
DE.  Henderson et al. (1988) found these changes in rats and mice exposed to 7.1 and 3.5  mg/m3
DPM for 7 h/day,  5 days/week. Significant increases in BALF PMNs were observed in mice at
6 mo of exposure and thereafter at the 7.1 and 3.5 mg/m3 exposure levels, but in rats only  the 7.1
mg/m3 exposure level showed  an increase in BALF PMNs at 6 mo of exposure and thereafter.
Significant increases in BALF PMNs occurred in rats at 12, 18,  and 24 mo of exposure to 3.5
mg/m3 DPM. Although increases in PMNs were usually greater in mice in terms of absolute
numbers, the PMN response in terms of increase relative to controls was only about one-third
that of rats.  Strom (1984) reported that the increased numbers of PMNs in BALF were
proportional to the inhaled  concentrations and/or duration of exposure. The PMNs also appeared
to be affiliated with clusters of aggregated AMs rather than to the diesel particles per  se.
Proliferation of Type II cells likewise occurred in response to the formed aggregates of AMs
(White and Garg,  1981).
       The integrity of pulmonary defense mechanisms can also be ascertained by assessing if
exposure to DE affects colonization and clearance of pathogens  and alters the response of the
challenged animals to respiratory tract infections. Campbell et al. (1980, 1981) exposed mice to
DE followed by infectious  challenge with Salmonella typhimurium, Streptococcuspyogenes, or
A/PR8-3  influenza virus  and measured microbial-induced mortality. Exposures to DE were to 6
mg/m3 DPM for 8 h/day, 7 days/week for up to 321 days. Exposure to DE resulted in enhanced
susceptibility to the lethal effects of S. pyogenes infection at all exposure durations (2 h, 6 h; 8,
15, 16, 307,  and 321 days).  Tests with S. typhimurium were inconclusive because of high
mortality rates in the controls.  Mice exposed to DE did not exhibit an enhanced mortality when
challenged with the influenza virus. Hatch et al. (1985) found no changes in the susceptibility of
mice to Group C Streptococcus sp. infection following intratracheal injection of 100 |lg of DPM
suspended in unbuffered  saline.

                                         5-57

-------
       Hahon et al. (1985) assessed virus-induced mortality, virus multiplication with
concomitant IFN levels (lungs and sera), antibody response, and lung histopathology in mice
exposed to DE prior to infectious challenge with Ao/PR/8/34 influenza virus.  Weanling mice
were exposed to DE containing 2 mg/m3 DPM for 7 h/day, 5 days/week. In mice exposed for 1,
3, and 6 mo, mortality was similar between the exposed and control mice.  In mice exposed for 3
and 6 mo, however, there were significant increases in the percentage of mice having lung
consolidation, higher virus growth, depressed IFN levels, and a fourfold reduction in
hemagglutinin antibody levels; these effects were not seen after the  1-mo exposure.
       The effects of DE on the pulmonary defense mechanisms appear to be determined by
three critical factors related to exposure: the concentrations of the pollutants, the exposure
duration, and the exposure pattern. Higher doses of DE as determined by an increase in one or
more of these three variables have been reported to increase the numbers of AMs, PMNs, and
Type II cells in the lung, whereas lower doses fail to produce such changes. In rats, the single
most significant contributor to the impairment of the pulmonary defense mechanisms appears to
be an excessive accumulation of DPM, particularly as particle-laden aggregates of AMs.  Such
an accumulation would result from an increase in deposition and/or a reduction in clearance.
The deposition of particles does not appear to change significantly following exposure to
equivalent DE doses over time. Because of the significant nonlinearity in particle accumulation
between low and high doses of DE exposure, coupled with no evidence of increased particle
deposition, an impairment in one or more of the mechanisms of pulmonary defense appears to be
responsible for the DPM accumulation and subsequent pathological sequelae.  The time of onset
of pulmonary clearance impairment was dependent both on the magnitude and on the duration of
exposures. For example, for rats exposed for 7 h/day, 5 days/week for 104 weeks, the
concentration needed to induce pulmonary clearance impairment appears to lie between 0.35 and
2.0 mg/m3 DPM.

5.1.3.3.5. Effects on the immune system—inhalation studies. The effects of DE on the
immune system of guinea  pigs were investigated by Dziedzic (1981). Exposures were to
1.5 mg/m3 DPM for 20 h/day,  5.5  days/week for up to 8 weeks.  There was no effect of diesel
exposure when compared with matched controls for the number of B and T lymphocytes and
null cells isolated from the tracheobronchial lymph nodes, spleen, and blood.  Cell viability as
measured by trypan blue exclusion was comparable between the exposed and control groups.
The results of this study and others on the effects of exposure to DE on the immune system are
summarized in Table 5-8.
       Mentnech et al. (1984) examined the effect of DE on the immune system of rats.
Exposures were to 2 mg/m3 DPM for 7 h/day, 5 days/week for up to 2 years. Rats exposed for
12 and 24 mo were tested for immunocompetency by determining antibody-producing cells in

                                         5-58

-------
the spleen 4 days after immunization with sheep erythrocytes. The proliferative response of
splenic T-lymphocytes to the mitogens concanavalin A and phytohemagglutinin was assessed in
rats exposed for 24 mo. There were no significant differences between the exposed and control
animals.  Results obtained from these two assays indicate that neither humoral immunity
(assessed by enumerating antibody-producing cells) nor cellular immunity (assessed by the
lymphocyte blast transformation assay) were markedly affected by the exposures.
       Bice et al. (1985) evaluated whether or not exposure to DE would alter antibody immune
responses induced after lung immunization of rats and mice.  Exposures were to 0.35, 3.5, or 7.1
mg/m3 DPM for 7 h/day, 5 days/week for 24 mo. Chamber controls and exposed animals were
immunized by intratracheal instillation of SRBCs after 6, 12, 18, or 24 mo of exposure. No
suppression in the immune response occurred in either species. After 12, 18,  and 24 mo of
exposure, the total number of anti-SRBC IgM antibody forming cells (AFCs) was elevated in
rats, but not in mice, exposed to 3.5 or 7.1 mg/m3 DPM; after 6 mo of exposure, only the 7.1
mg/m3 level was found to have caused this response in rats. The number of AFCs per 106
lymphoid cells in lung-associated lymph nodes and the levels of specific IgM, IgG, or IgA in rat
sera were not significantly altered.  The investigators concluded that the increased cellularity and
the presence of DPM in the lung-associated lymph nodes had only a minimal  effect on the
immune and antigen filtration function of these tissues.
       The effects of inhaled DE and DPM have been studied in a murine model of allergic
asthma (Takano et al., 1998a,b). ICR mice were exposed for  12 h/day, 7 days/week for 40 weeks
to DE (0.3, 1.0, or 3.0 mg/m3).  The mice were sensitized with ovalbumin  (OA) after 16 weeks
exposure and subsequently challenged with  aerosol allergen (1% OA in isotonic saline for 6
min) at 3-week intervals during the last 24 weeks of exposure. Exposure to DE enhanced
allergen-related eosinophil recruitment to the submucosal layers of the airways and to the
bronchoalveolar space, and increased protein levels of GM-CSF and IL-5 in the lung in a
dose-dependent manner. In  the DE-exposed mice, increases in eosinophil  recruitment and local
cytokine expression were accompanied by goblet-cell proliferation in the bronchial epithelium
and airwayhyperresponsiveness to inhaled acetylcholine.  In contrast, mice exposed to clean air or
DE without allergen provocation showed no eosinophil recruitment to the submucosal layers of
the airways or to the bronchoalveolar space, and few goblet-cells in the bronchial epithelium. The
                                         5-59

-------
         Table 5-8.  Effects of inhalation of diesel exhaust on the immune system of laboratory animals
Species/sex
Guinea Pig,
Hartley, M


Rat, F344, M



Rat, F344;
Mouse, CD-I




Mouse,
BALB/C, M
Exposure
period
20 h/day
5.5 days/week
4 or 8 weeks

7 h/day
5 days/week
52 or 104 weeks

7 h/day
5 days/week
104 weeks



12 h/day,
7 days/week,
Particles
(mg/m3)
1.5
0.19 |lm
MOD

2.0
0.23-0.36
|J,mMDD

0.35
3.5
7.1
0.25 |lm
MOD

3.0
6.0
CxT CO
(mg-h/m3) (ppm)
660 or 7,280 7.5



3,640 or 11.5
7,280


1,274 2.9
12,740 16.5
25,480 29.7



756 —
1,512 —
NO2
(ppm)
	



1.5



0.05
0.34
0.68



2.8
4.1
SO2
(ppm) Effects
— No alterations in numbers of B, T, and null
lymphocytes or cell viability among lymphocytes
isolated from tracheobronchial lymph nodes, spleen,
or blood
0.8 Neither humoral immunity (assessed by enumerating
antibody-producing cells) nor cellular immunity
(assessed by the lymphocyte blast transformation
assay) were markedly affected
— Total number of anti-sheep red blood cell IgM AFC in
— the lung-associated lymph nodes was elevated in rats
— exposed to 3.5 or 7.0 mg/m3 DPM (no such effects in
mice); total number of AFC per 106 lymphoid cells in
lung-associated lymph nodes and level of specific
IgM, IgG, or IgA in rat sera were not altered
1.7 Spleen weights in mice exposed to DE (6 mg/m3)
2.7 increased significantly. Serum anti-OA IgE antibody
Study
Dziedzic
(1981)


Mentnech et
al. (1984)


Bice et al.
(1985)




Fujimaki et al.
(1997)
3 weeks
Mice administered OA
intranasally before,
immediately after, and
3 weeks after exposure
                                            0.4|lm
                                                            liters in mice exposed to 6 mg/m3 significanlly higher
                                                            lhan conlrol. Anligen-slimulaled IL-4 and IL-10
                                                            produclion increased while IFN-g produclion
                                                            decreased significanlly in spleen cells from DE-
                                                            exposed (6 mg/m3) mice stimulated wilh OA in vilro.
                                                            DE inhalation may affecl antigen-specific IgE
                                                            antibody produclion Ihrough alteration of Ihe
                                                            cylokine nelwork.
Mouse,
C3H/Hen, M
12 h/day,
for 12 weeks.  Before
exposure mice injected IP
with OA. After 3 weeks
and every 3 weeks
thereafter, mice
challenged with OA
aerosol.
1.0
3.0
1,008
3,024
1.42        0.87      DE + antigen challenge induced airway
4.02        1.83      hyperresponsiveness and inflammation with increased
                     eosinophils, mast cells, and goblet cells.
                     DE alone induced airway hyperresponsiveness, but
                     not eosinophilic infiltration or increased goblet cells.
                     DE inhalation enhanced airway hyperresponsiveness
                     and airway inflammation caused by OA sensitization.
Miyabara et
al. (1998a)

-------
            Table 5-8.  Effects of inhalation of diesel exhaust on the immune system of laboratory animals (continued)
Species/sex
Mouse,
C3H/HeN,
M






Mouse,
ICR
(murine model
Exposure
period
12 h/day,
for 5 weeks. After 7 days
mice injected IP with OA.
At end of exposure mice
challenged with OA
aerosol for 15 minutes.



12 h/day, 7days/week,
40 weeks.
After 16 weeks sensitized
Particles
(mg/m3)
3.0








0.3
1.0
3.0
CxT CO
(mg-h/m3) (ppm)
1,260 —








1,008 —
3,360
10,080
NO2 SO2
(ppm) (ppm) Effects
4.08 1.26 DE alone increased neutrophils and macrophages in
BAL fluid; after DE + OA challenge eosinophils
increased.
OA alone increased eosinophils but the increase was
enhanced by DE.
DE + OA, but not DE alone, increased goblet cells,
respiratory resistance, production of OA-specific IgE
and Igl in the serum, and overexpression of IL-5 in
lung tissue.
— — DE exposure enhanced allergen-related recruitment to
the submucosal layers of the airways and the
bronchoalveolar space, and increased GM-CSF and
Study
Miyabara et
al. (1998b)







Takano et al.
(1998a)

of allergic       to OA and challenged
asthma)         with OA aerosol for
               6 min, at 3-week intervals
               during the last 24 weeks
               of exposure.
IL-5 in the lung in a dose-dependent manner.
Increases in eosinophil recruitment and local cytosine
expression accompanied by goblet cell proliferation
in the bronchial epithelium and airway
hyperresponsiveness to inhaled acetylcholine. Mice
exposed to clean air or DE without allergen
provocation showed no eosinophil recruitment to the
submucosal layers of the airways nor to the
bronchoalveolar space, and few goblet cells in the
bronchial epithelium. Daily inhalation of DE may
enhance allergen-related respiratory diseases such as
allergic asthma, and effect may be mediated by the
enhanced local expression of IL-5 and GM-CSF.	
DPM = Diesel particulate matter.
AFC = Antibody-forming cells.

-------
authors concluded that daily inhalation of DE can enhance allergen-related respiratory diseases
such as allergic asthma, and that this effect may be mediated by the enhanced local expression of
IL-5 and GM-CSF. The effect of DPM on a second characteristic of allergic asthma, airway
hyperresponsiveness, was examined by Takano et al. (1998b).  Laboratory mice were
administered OA, DPM, or OA and DPM combined by intratracheal instillation for 6 wk.
Respiratory resistance (Rrs) after acetylcholine challenge was measured 24 h after the final
instillation. Rrs was significantly greater in the mice treated with OA and DPM than in the other
treatments. The authors concluded that DPM can enhance airway responsiveness associated with
allergen exposure.
      In a series of inhalation studies following earlier instillation studies, Miyabara and
co-workers investigated whether inhalation of DE could enhance allergic reactions in laboratory
mice. C3H/HeN mice were exposed to DE (3 mg DPM/m3) by inhalation for 5 weeks (Miyabara
et al., 1998b) and, after 7 days of exposure, were sensitized to OA injected intraperitoneally. At
the end of the DE exposure, the mice were challenged with an  OA aerosol for  15 min. DE
caused an increase in the numbers of neutrophils and macrophages in bronchoalveolar lavage
fluid independent of OA sensitization, whereas a significant increase in eosinophil numbers
occurred only after DE exposure was combined with antigen challenge. Even  though OA alone
caused an increase in eosinophil numbers in lung tissue, this response was enhanced further by
DE.  DE exposure combined with OA sensitization enhanced the number of goblet-cells in lung
tissue, respiratory resistance, production of OA-specific IgE and IgGj in the serum, and
overexpression of IL-5 in lung tissue.  In a second study, C3H/HeN mice were sensitized with
OA injected intraperitoneally and then exposed to DE by inhalation for 12 h/day for 3 mo at
either 1  or 3 mg/m3 (Miyabara et al., 1998a). After 3 weeks of DE exposure, and every 3 weeks
thereafter, the mice were challenged with an OA aerosol. Exposure to DE with antigen
challenge induced airway hyperresponsiveness and airway inflammation, which was
characterized by increased numbers of eosinophils and mast cells in lung tissue.  The increase in
inflammatory cells was accompanied by an increase in goblet cells in the bronchial epithelium.
Airway hyperresponsiveness, but not eosinophilic infiltration or increased goblet cells, was
increased by DE exposure alone. These workers concluded that inhalation of DE can enhance
airway hyperresponsiveness and airway inflammation caused by OA sensitization in mice.
      The effects of DE on IgE antibody production were investigated in BALB/c mice
sensitized with OA and exposed by inhalation to DE (3.0 and 6.0 mg/m3) for 3 weeks (Fujimaki
et al., 1997). The mice were sensitized by intranasal administration of OA alone before,
immediately after, and 3 weeks  after DE inhalation. While body and thymus weights were
unchanged in the DE-exposed and control mice, spleen weights in mice exposed to 6 mg/m3 DE
increased significantly.  Anti-OA IgE  antibody liters in the sera of mice exposed to 6 mg/m3 DE
were significantly higher than control.  Total IgE and anti-OA  IgG in sera from DE-exposed and

                                          5-62

-------
control mice remained unchanged. Cytokine production was measured in vitro stimulated with
OA in spleen cells from mice exposed to DE (6 mg/m3). Antigen-stimulated interleukin-4 (IL-4)
and -10 (IL-10) production increased significantly in vitro in spleen cells from DE-exposed mice
compared with controls, while ZFN-y production decreased markedly.  The authors concluded
that DE inhalation in mice may affect antigen-specific IgE antibody production through
alteration of the cytokine network.

5.1.3.3.6. Effects on the immune system—noninhalation studies.  The immune response of
laboratory animals to DPM has been studied in various noninhalation models, and the results of
these studies are presented in Table 5-9.  Takafuji et al. (1987) evaluated the IgE antibody
response of mice inoculated intranasally at intervals of 3 weeks with either 0.5 or 25 |lg of DPM
in ovalbumin per mouse. Antiovalbumin IgE antibody liters, assayed by passive cutaneous
anaphylaxis, were enhanced by doses as low as 1 |lg of particles compared with immunization
with ovalbumin alone.
       Muranaka et al. (1986) studied the effects  of DPM on IgE antibody production in
immunized mice. A greater IgE antibody response was noted in mice immunized by ip injection
of ovalbumin (OA) mixed with DPM, either 0.02, 0.2, or 2mg per mouse, than in animals
immunized with OA alone.  This effect of DPM on IgE antibody production in mice was also
demonstrated in mice immunized with repeated injections of dinitrophenylated-OA. Moreover,
a persistent IgE-antibody response to Japanese cedar pollen (JCPA), a common pollen allergen
causing allergic rhinitis in Japan, was observed in mice immunized with JCPA mixed with DPM
but not in animals immunized with JCPA alone. The results suggest an association between the
adjuvant activity of DPM and allergic rhinitis caused by JCPA.
       Takano et al. (1997) designed a study to evaluate the effects of DPM on the
manifestations of allergic asthma in mice, with emphasis on antigen-induced airway
inflammation; the local expression of IL-5, GM-CSF, IL-2, and IFN-y; and the production of
antigen-specific IgE and IgG.  Male ICR mice were intratracheally instilled with ovalbumin
(OVA), DPM,  and DPM+OVA. DPM was obtained from a 4JBl-type, light-duty 2.74 L, four-
cylinder Isuzu diesel engine operated at a steady speed of 1,500 rpm under a load of 10 torque
(kg/m).  The OVA-group mice were instilled with 1 |lg OVA at 3 and 6 weeks.  The mice
receiving DPM alone were instilled with 100 |lg DPM weekly for 6 weeks. The OVA + DPM
group received the combined treatment in the same protocol as the OVA and the DPM groups,
respectively. Additional groups were exposed for 9 weeks.  DPM aggravated OVA-induced
airway inflammation, characterized by infiltration of eosinophils and lymphocytes and an
                                         5-63

-------
          Table 5-9.  Effects of diesel particulate matter on the immune response of laboratory animals
Model
                                Treatment
                                                                                                        Effects
                                                                                                                                                                Reference
Mouse,
BDFI, F

Mouse,
ICR, w/w", M
Mouse,
A/3, M
Mouse,
BDF1: M
Mouse,
BALB/C,
nu/nu, F


Mouse,
BALB/cA, F
Mouse,
ICR, M
Intratracheal instillation of DPM, once/week for
16 weeks
Mice immunized intranasally with Der f II +
pyrene, or Der f II + DPM 7 times at 2-week
intervals
Mice were administered 25 mg of each of 5 fine
particles (Kanto loam dust, fly ash, CB, DPM,
and aluminum hydroxide [alum]) intranasally
and exposed to aerosolized Japanese cedar
pollen allergens (JCPA) for intervals up to 18
wk
Inoculated OA with DPM or CB into hind
footpad measured response using popliteal
lymph node assay


Intranasal administration of DPM. Mice
immunized with OA or OA combined with
DPM or CB
Intratracheal instillation of OA, DPM, or OVA
and DPM combined, once/week for 6 wk
Intranasally delivered doses of DPM as low as 1 mg exerted an adjuvant activity for IgE antibody          Takafuji et al.
production.                                                                                     (1987)

Infiltration of inflammatory cells, proliferation of goblet cells, increased mucus secretion, respiratory       Sagai et al. (1996)
resistance, and airway constriction. Increased eosinophils in the submucosa of the proximal bronchi and
medium bronchioles.  Eosinophil infiltration suppressed by pretreatment with PEG-SOD. Bound sialic
acid, an index of mucus secretion, in bronchial alveolar lavage fluids increased, but was suppressed by
PEG-SOD.  Increased respiratory resistance suppressed by PEG-SOD.  Oxygen radicals produced by
instilled DPM may cause features characteristic of bronchial asthma in mice.

IgE antibody responses to Der f II enhanced in mice immunized with Der f 11+ pyrene or Der f II +         Suzuki  et al. (1996)
DPM compared with Der f II alone. Response was dose related. DPM and pyrene contained in DPM
have adjuvant activity on IgE and IgGl antibody production in mice immunized with house dust mite
allergen.

Measurements  were made of JCPA-specific IgE and IgG antibody liters, the protein-adsorbing capacity    Maejima et al.
of each type of particle, and nasal rubbing movements (a parameter of allergic rhinitis in mice). The        (1997)
increases in anti-JPCA IgE and IgG antibody liters were significantly greater in mice Irealed wilh
particles and aerosolized JCPA lhan in mice Irealed wilh aerosolized JCPA alone. In a subsequenl
experimenl, Ihe mice received Ihe particles as before, bul aboul 160,000 grains of Japanese cedar pollen
(JCP) were dropped onlo Ihe lip of Ihe nose of each mouse Iwice a week for 16 wk. After 18 wk Ihere
were no significanl differences in Ihe anli-JCPA IgE and IgG production, nasal rubbing, or
hislopalhological changes. The workers concluded lhal Ihe nalure of Ihe particle, Ihe ability of Ihe
particle lo absorb antigens, and/or particle size is nol related lo Ihe enhancemenl of IgE antibody
production or symptoms of allergic rhinilis. However, IgE anlibody production did appear lo occur
earlier in mice Irealed wilh particles lhan in mice immunized wilh allergens alone.

Increased response (increased weighl, cell numbers, cell proliferation) and longer response observed       L0vik el al. (1997)
wilh DPM and OA, compared lo DPM or OA alone. Response was specific and nol an unspecific
inflammatory response. CB was slighlly less polenllhan DPM. Nonexlraclable carbon core conlribules
subslanlially lo adjuvanl activity of DPM.

Increased response lo antigen in animals receiving DPM or CB. Increased number of responding          Nilsen el al. (1997)
animals and increased serum anli OA IgE anlibody. Bolh DPM and CB have adjuvanl activity for IgE
production. DPM response more pronounced lhan CB, indicaling bolh organic matter adsorbed lo DPM
and Ihe nonexlraclable carbon core responsible for adjuvanl activity.

Respiratory resistance (Rrs) measured 24 h after Ihe final instillation. Rrs after acelylcholine challenge      Takano el al.
was significanlly greater in Ihe mice Irealed wilh OVA and DPM lhan olher Irealmenls. DPM can          (1998b)
enhance airway responsiveness associaled wilh allergen exposure.
OA - Ovalbumin.
DPM - Diesel particulale matter.
CB - Carbon black.
                                              PEG-SOD - Polyelhyleneglycol-conjugaled superoxide dismulase.
                                              IL-4 - Inlerleukin-4.
                                              IL-5 - Inlerleukin-5.
                                              IL-10-Inlerleukin-10.
                                              IFN - Inlerferon-g.
                                              GM-CSF  -Granulocyle-colony stimulating factor.
                                              IP - Inlraperiloneally.

-------
increase in goblet cells in the bronchial epithelium. DPM in combination with antigen markedly
increased IL-5 protein levels in lung tissue and bronchoalveolar lavage supernatants compared
with either antigen or DPM alone. The combination of DPM and antigen induced significant
increases in local expression of IL-4, GM-CSF, and IL-2, whereas expression of ZFN-y was not
affected.  In addition, DPM exhibited adjuvant activity for the antigen-specific production of
IgG and IgE.
       The potential role of oxygen radicals in injury caused by DPM was investigated by Sagai
et al. (1996).  These workers reported that repeated intratracheal instillation of DPM (either 0.1
or 0.2 mg per mouse, once/week for 16 weeks) in mice caused marked infiltration of
inflammatory cells, proliferation of goblet cells, increased mucus  secretion, respiratory
resistance, and airway constriction. Eosinophils in the submucosa of the proximal bronchi and
medium bronchioles increased eightfold following instillation. Eosinophil infiltration was
significantly suppressed by pretreatment with polyethyleneglycol-conjugated superoxide
dismutase (PEG-SOD), an inhibitor of oxygen radicals. Bound sialic acid concentrations in
bronchial alveolar lavage fluids, an index of mucus secretion,  increased with DPM, but were
also suppressed by pretreatment with PEG-SOD. Goblet cell hyperplasia, airway narrowing, and
airway constriction also were observed with DPM.
       Respiratory resistance to acetylcholine in the DPM group was 11 times higher than in
controls, and the increased resistance was significantly suppressed by PEG-SOD pretreatment.
These findings indicate that oxygen radicals caused by intratracheally instilled DPM elicit
responses characteristic of bronchial asthma.
       Potential adjuvant effects of DPM on the response to the model allergen OA were
investigated in BALB/c mice using the popliteal lymph node (PLN) assay (L0vik et al., 1997).
DPM inoculated together with OA into one hind footpad (0.02 mL of a 5 mg/mL DPM
suspension) gave a significantly augmented response (increase in weight, cell numbers, and  cell
proliferation) in the draining popliteal lymph node as compared to DPM or OA alone. The
duration of the local lymph node response was also longer when DPM was given with the
allergen.  The lymph node response appeared to be of a specific immunologic character and not
an unspecific inflammatory reaction. The OA-specific response IgE was increased in mice
receiving OA together with DPM as compared with the response in mice receiving OA alone.
Further studies using carbon black (CB) as a surrogate for the  nonextractable core of DPM found
that while CB resembled DPM in its capacity to increase the local lymph node response and
serum-specific IgE response to OA, CB appeared to be slightly less potent than DPM.  The
results indicate that the nonextractable particle core contributes substantially to the adjuvant
activity of DPM.
       Nilsen et al. (1997) investigated which part of the particle  was responsible, the carbon
core and/or the adsorbed organic substances, for the adjuvant activity of DPM. Female

                                          5-65

-------
BALB/cA mice were immunized with OA alone or in combination with DPM or CB particles by
intranasal administration a total of four times, once weekly, at 25 jig/inoculation. There was an
increased response to the antigen in animals receiving OA together with DPM or CB, compared
with animals receiving OA alone. The response was seen as both an increased number of
responding animals and increased serum anti OA IgE response.  The workers concluded that
both DPM and CB have an adjuvant activity for specific IgE production, but that the activity of
DPM may be more pronounced than that of CB. The results suggest that both the organic matter
adsorbed to DPM and the nonextractable carbon are responsible for the observed adjuvant effect
of DPM.
       The effects of DPM and its components (extracted particles and particle extracts) on the
release of proinflammatory cytokines, interleukin-1 (IL-1), and tumor necrosis factor-a (TNF-
a) by alveolar macrophages (AMs) were investigated by Yang et al. (1997). Rat AMs were
incubated with 0,  5,  10, 20, 50, or 100 |lg/106 AM/mL of DPM, methanol-extracted DPM, or
equivalent concentrations of DPM at 37 °C for 24 h.  At high concentrations, both DPM and
DPM extracts were shown to increase IL-1-like activity secreted by AMs, whereas extracted
particles had no effect. Neither particles, particle extracts, or extracted particles stimulated
secretion of TNF-a.  DPM inhibited lipid polysaccharide (LPS)-stimulated production of IL-1
and TNF-a. In contrast, interferon (IFN)-Y-stimulated production of TNF-a was not affected
by DPM. Results of this study indicate that the organic fraction of exhaust particles is
responsible for the effects noted.  Stimulation of IL-1 but not TNF-a suggests that IL-1, but not
TNF-a, may play an important role in the development of DPM-induced inflammatory and
immune responses.  The cellular mechanism involved in inhibiting increased release of IL-1 and
TNF-a by LPS is unknown, but may be a contributing factor to  the decreased AM phagocytic
activity and increased susceptibility to pulmonary infection after prolonged exposure to DPM.
       Fujimaki et al. (1994) investigated the relationship between DPM and IgE antibody
production, interleukin 4 (IL-4) production in BALB/c mice  treated with DPM mixed with
antigen OA or JCP antigen by intratracheal instillation. BALB/c mice were injected with DPM
(300 |ig) plus OA or OA alone and, after the last instillation, the proliferative response and
lymphokine production by mediastinal lymph node cells (LNC)  were examined in vitro. The
proliferative response to OA in mediastinal LNC from mice injected with DPM plus OA was
enhanced to 4-17 times that of control mice.  IL-4 production by OA stimulation was also
enhanced in mediastinal LNC  from mice injected with DPM plus OA.  A significantly larger
amount of anti-OA IgE antibody was detected in sera from DPM- and OA-injected mice
compared with those from control mice. The levels of IL-4,  estimated by JCP antigen in
mediastinal LNC, from mice injected with DPM plus JCP antigen were twofold higher than
those from mice injected with  JCP antigen alone.  These results  suggest that intratracheal
                                          5-66

-------
instillation of DPM affects antigen-specific IgE antibody responses via local T-cell activation,
especially enhanced IL-4 production.
       Suzuki et al. (1993) investigated the adjuvant activity of pyrene, one of many PAHs
contained in DPM, on IgE antibody production in mice. In the first experiment, mice were
immunized with 1 mg of OA alone, 1  mg of OA plus 1 mg of pyrene, or 1 mg of OA plus 1 mg
of DPM, respectively.  The IgE antibody responses to OA in mice immunized with OA plus
pyrene or OA plus DPM were enhanced as compared to those in mice immunized with OA
alone; the highest responses were observed in mice immunized with OA plus DPM. In the
second experiment,  mice were immunized with 10 mg of JCPA alone or 10 mg of JCPA plus 5
mg of pyrene. The IgE antibody responses to JCPA in mice immunized with JCPA plus pyrene
were higher than those in mice immunized with JCPA alone. The results indicate that pyrene
contained in DPM acts as an adjuvant in IgE antibody production in immunized mice.
       Suzuki et al. (1996) investigated the effect of pyrene on IgE and IgGl antibody
production in mice to clarify the relation between mite allergy and adjuvancy of the chemical
compounds in DPM. The mite allergen was Der f II, one of the major allergens of house dust
mite (Dermatophagoidesfarinae). Allergen mice were grouped and immunized with Der f II
(5 |lg), Der f II (5 |lg)  plus pyrene (200 |lg), and Der f II (5 |lg) plus  DPM (100 |lg)
intranasally seven times at 2-week intervals.  The separate groups of mice were also immunized
with Der f II (10 |ig) plus the same dose of adjuvants in the same way. The IgE antibody
responses to Der f II in mice immunized with Der f II plus pyrene or Der f II plus DPM were
markedly enhanced compared with those immunized with Der f II alone.  The anti-Der f II IgE
antibody production increased with increasing the dose of Der f II from 5 |lg to  10 |lg in mice
immunized with Der f II plus the same dose of adjuvants.  The  IgGl antibody responses to Der f
II in mice immunized with Der f II (10 |lg) plus pyrene  (200 |lg) or Der f II (10 |lg) plus DPM
(100 |lg) were greater than those immunized with 10 |lg of Der f II alone. In addition, when
peritoneal macrophages obtained from normal mice were incubated with pyrene or DPM in
vitro, an enhanced IL-la production by the macrophages was observed. When spleen
lymphocytes obtained from the mice immunized with Der f II (10 |lg) plus DPM (100 |lg) or
Der f II (10 |lg) plus pyrene (200 |lg) were stimulated with 10  |lg of Der f II in vitro,  an
enhanced IL-4 production of the lymphocytes was also observed compared with those
immunized with Der f II alone. This study indicates that DPM and pyrene (one of the many
PAHs adsorbed onto DPM) have an adjuvant activity on IgE and IgGl antibody production in
mice immunized intranasally with a house dust mite allergen.
       Maejima et al. (1997) examined the potential adjuvant activity of several different fine
particles. These workers administered 25 |ig of each of 5 particles (Kanto loam dust, fly ash,
CB, DPM, and aluminum hydroxide [alum]) intranasally in mice and exposed them to
aerosolized JCPA for intervals up to 18 weeks. Measurements  were made of JCPA-specific IgE

                                         5-67

-------
and IgG antibody liters, the protein-adsorbing capacity of each type of particle, and nasal
rubbing movements (a parameter of allergic rhinitis in mice).  The increases in anti-JPCA IgE
and IgG antibody liters were significantly greater in mice treated with particles and plus
aerosolized JCPA than in mice treated with aerosolized JCPA alone. In a subsequent
experiment, the mice received the particles as before, but about 160,000 grains of JCP were
dropped onto the tip of the nose of each mouse twice a week for 16 weeks.  After 18 weeks there
were no significant differences in the anti-JCPA IgE and IgG production, nasal rubbing, or
histopathological changes. The workers concluded that the nature of the particle, the ability of
the particle to absorb antigens, and particle size are not related to the enhancement of IgE
antibody production or symptoms of allergic rhinitis. However, IgE antibody production did
appear to occur earlier in mice treated with particles than in mice immunized with allergens
alone.
       The potential for DPM to modulate cytokine production has been demonstrated in
cultured mouse bone marrow-derived mast cells (BMMC). Saneyoshi et al. (1997) examined the
production of cytokines in BMMC treated with DPM (0.8, 2 and 4 mg/mL).  Production of
interleukin-4  (IL-4) and IL-6 was higher in BMMC stimulated with A23187 and treated with
low concentrations of DPM than in controls, but no increase was seen in BMMC treated with
high DPM. After pretreatment with low DPM for 24 h, IL-4 production in BMMC stimulated
with A23187  was lower than in controls. Antigen-induced IL-4 production increased
significantly in BMMC treated with 0.4 or 0.8 mg/mL DPM, but did not increase with low
DPM.  Although the enhancement of IL-4 production of BMMC stimulated with A23187 plus
DPM was not completely inhibited by 2-mercaptoethanol, treatment with dexamethasone
inhibited further IL-4 production. Thus, DPM may affect the immune response via the
modulation of cytokine production in mast cells.
       Ormstad et al. (1998) investigated the potential for DPM as well as other suspended
particulate matter (SPM) to act as a carrier for allergens into the airways.  These investigators
found both Can f 1  (dog) and Bet v 1 (birch pollen) on the surface of SPM collected in air from
different homes. In an extension of the study, they found that DPM adhered to polycarbonate
filters had the potential of binding both of these allergens as well as Fel d 1 (cat) and Der p 1
(house mite).  The authors conclude that soot particles in indoor air house dust may act as carrier
of several allergens in indoor air.
       Knox  et al. (1997) investigated whether free grass pollen allergen molecules, released
from pollen grains by osmotic shock (Suphioglu et al., 1992) and dispersed in microdroplets of
water in aerosols, can bind to DPM mounted on copper grids in air. Using natural highly
purified Lol p 1, immunogold labeling with specific monoclonal antibodies, and a high-voltage
transmission electron-microscopic imaging technique, these workers demonstrated binding of the
major grass pollen allergen, Lol p 1, to DPM in vitro. These workers conclude that binding of

                                          5-68

-------
DPM with Lol p 1 might be a mechanism by which allergens can become concentrated in air and
trigger attacks of asthma.
       Murphy et al. (1999) examined the comparative toxicities to the lung of four different-
sized CB particles and DPM, in primary cultures of mouse Clara and rat type II epithelial cells.
Particle toxicity was assessed by cell attachment to an extracellular matrix substratum. The CB
particles varied in toxicity to Clara and type II cells.  DPM stored for 2 weeks was equally toxic
to both cell types. DPM became progressively less toxic to type II cells with time of storage.
Both primary epithelial cell types internalized the particles in culture.  These workers concluded
that bioreactivity was related to CB particle size and surface area, with the smaller particles
having the larger surface area being the more toxic. Although freshly prepared DPM was
equally toxic to type II and Clara cells, DPM became progressively less toxic to the type II cells
with time.
       Exposure studies in laboratory  animals and isolated cell  systems derived from animals
also  indicate that DPM can elicit both inflammatory and immunological changes.  Moreover, the
effects appear to be due to both the nonextractable carbon core and the adsorbed organic fraction
of the diesel particle. Changes in IgE, goblet cell hyperplasia, mast cell influx, and cytokines in
various animal models and in vitro model systems are all key markers of asthma. The data
further indicate a role for oxygen radicals in DPM injury because the extent of the injury can be
reduced by treatment with antioxidants. DPM also has the capacity  to bind and transport
airborne allergens.

5.1.3.3.7.  Effects on the liver.  Meiss et al. (1981) examined alterations in the hepatic
parenchyma of hamsters by using thin-section and freeze-fracture histological techniques.
Exposures to DE were for 7 to 8 h/day, 5  days/week, for 5 mo at about 4 or 11 mg/m3 DPM.
The  livers of the hamsters exposed to both concentrations of DE exhibited moderate dilatation of
the sinusoids, with activation of the Kupffer cells and slight changes in the cell nuclei. Fatty
deposits were observed in the sinusoids, and small fat droplets were occasionally observed in the
peripheral hepatocytes. Mitochondria often had a loss of cristae and exhibited a pleomorphic
character.  Giant microbodies were seen in the hepatocytes, which were moderately enlarged,
and gap junctions between hepatocytes exhibited a wide range in structural diversity. The results
of this study and others on the effect of exposure of DE on the liver of laboratory animals are
summarized in Table 5-10.
                                          5-69

-------
       Table 5-10.  Effects of exposure to diesel exhaust on the liver of laboratory animals
Species/sex
Rat, F344, M, F
Hamster, Syrian
Cat, inbred, M

Exposure
period
7 h/day
5 days/week
52 weeks
7-8 h/day
5 days/week
22 weeks
8 h/day
7 days/week
124 weeks
Particles
(mg/m3)
2.0
0.23-0.36 |lm
MOD
4.0
8.0
11.0
6.0"
12.0b

CxT
(mg-h/m3)
3,640
3,080-9,680
41,664
83,328

CO
(ppm)
12.7
12.0
19.0
25.0
20.2
33.3

NO2
(ppm)
1.6
0.5
1.0
1.5
2.7
4.4

SO2
(ppm)
0.83
3.0
6.0
7.0
2.1
5.0

Effects
No changes in absolute liver weight or
liver/body weight ratio
Enlarged sinusoids, with activated Kupffer's
cells and slight changes of nuclei; fatty
deposits; mitochondria, loss of cristae and
pleomorphic character; gap junctions between
hepatocytes had wide range in structural
diversity
No change in the absolute liver weight

Study
Green et al. (1983)
Meissetal. (1981)
Hopper etal. (1983)

"1 to 61 weeks of exposure.
b62 to 124 weeks of exposure.

-------
       Green et al. (1983) and Plopper et al. (1983) reported no changes in liver weights of rats
exposed to 2 mg/m3 DPM for 7 h/day, 5 days/week for 52 weeks or of cats exposed to 6 to
12 mg/m3, 8 h/day, 7 days/week for 124 weeks. The use of light and electron microscopy
revealed that long-term inhalation of varying high concentrations of DE caused numerous
alterations to the hepatic parenchyma of guinea pigs. A less sensitive index of liver toxicity,
increased liver weight, failed to detect an effect of DE on the liver of the rat and cat following
long-term exposure to DE. These results are too limited to understand potential impacts on the
liver.

5.1.3.3.8.  Blood and cardiovascular systems.  Several studies have evaluated the effects of DE
exposure on hematological and cardiovascular parameters of laboratory  animals.  These studies
are summarized in Table 5-11. Standard hematological indices of toxicological effects on red
and white blood cells failed to detect dramatic  and consistent responses. Erythrocyte (RBC)
counts were reported as being unaffected in cats (Pepelko and Peirano, 1983), rats and monkeys
(Lewis et al.,  1989), guinea pigs and rats (Penney et al., 1981), and rats  (Karagianes et al.,
1981); lowered in rats (Heinrich et al., 1982); and elevated in rats (Ishinishi et al., 1988;
Brightwell et  al., 1986).  Mean corpuscular volume was significantly increased in monkeys, 69
versus 64 (Lewis et al., 1989), and hamsters (Heinrich et al., 1982), and lowered in rats
(Ishinishi et al., 1988).  The only other parameters of erythrocyte status  and related events were
lowered mean corpuscular hemoglobin and mean corpuscular hemoglobin concentration inl rats
(Ishinishi et al., 1988), a 3% to 5% increase in carboxyhemoglobin saturation in rats
(Karagianes et al.,  1981), and a suggestion of an increase in prothrombin time (Brightwell et al.,
1986). The biological significance of these findings regarding adverse health effects is deemed
to be inconsequential.
       Three investigators (Pepelko and  Peirano,  1983; Lewis et al., 1989; Brightwell et al.,
1986) reported an increase in the percentage of banded neutrophils in cats and rats. This effect
was not observed in monkeys (Lewis et al., 1989). The health implications of an increase in
abnormal maturation of circulating neutrophils are uncertain but indicate a toxic response of
leukocytes following exposures to DE. Leukocyte counts were reported to be reduced in
hamsters (Heinrich et al., 1982); increased in rats (Brightwell et al., 1986); and unaffected in
cats, rats, and monkeys (Pepelko and Peirano,  1983; Ishinishi  et al., 1988; Lewis et al., 1989).
These inconsistent findings indicate that the  leukocyte counts are more indicative of the clinical
status of the laboratory animals than any  direct effect of exposure to DE.
       No significant changes in heart mass were found in guinea pigs or rats exposed to DE
(Wiester et al.,  1980; Penney et al.,  1981; Lewis et al., 1989).  Rats exposed to DE showed a
greater increase in the medial wall thickness of pulmonary arteries of differing diameters and
                                          5-71

-------
            Table 5-11.  Effects of exposure to diesel exhaust on the hematological and cardiovascular systems of laboratory animals
to
Exposure Particles
Species/sex
Monkey,
Cynomolgus, M

Rat, F344, M, F


Guinea Pig,
Hartley, M, F

Hamster, Syrian,
M, F

Rat, F344;
Guinea Pig,
Hartley

Rat, Wistar, M


Rat, F3444/Jcl,
M, F




Rat, F344




Cat, Inbred, M


"Nonirradiated DE.
'irradiated DE.
"Light-duty engine.
period
7 h/day
5 days/week
104 weeks
7 h/day
5 days/week
104 weeks
20 h/day
7 days/week
8 weeks
7-8 h/day
5 days/week
75 weeks
20 h/day
(mg/m3)
2
0.23-0.36 |lm MOD

2
0.23-0.36 |lm MOD

6.3"
6.8b

3.9
0.1 |J,mMDD

0.25
5.5 days/week 0.75
78 weeks

6 h/day
5 days/week
78 weeks
16 h/day
6 days/week
130 weeks



16 h/day
5 days/week
104 weeks


8 h/day
7 days/week
124 weeks



1.5
0.19|lmMDD
8.3
0.71 |lm MOD

o.ir
0.41C
1.08C
2.31C
3.72d
0.1 |lttiMDD
0.7
2.2
6.6


6.0e
12.0f

dHeavy-duty engine.
el to 61 weeks of exposure.
f62 to 124 weeks of exposure.
CXT
(mg-h/m3)
7,280


7,280


7,056
7,616

10,238-11,700


2,145
6,435
12,870

19,422


1,373
5,117
13,478
28,829
46,426

5,824
18,304
54,912


41,664
83,328




CO
(ppm)
11.5


11.5


17.4
16.7

18.5


3.0
4.8
6.9

50.0


1.23
2.12
3.96
7.10
12.9

	
—
32.0


20.2
33.3




N02
(ppm)
1.5


1.5


2.3
2.9

1.2


0.11
0.27
0.49

4-6


0.08
0.26
0.70
1.41
3.00

	
—
—


2.7
4.4




S02
(ppm)
0.8


0.8


2.1
1.9

3.1


	
—
—

—


0.38
1.06
2.42
4.70
4.57

	
—
—


2.1
5.0





Effects
Increased MCV


Increase in banded neutrophils; no effect on heart
or pulmonary arteries

No effect on heart mass or ECG; small decrease in
heart rate (IE only)

At 29 weeks, lower erythrocyte count; increased
MCV; reduced leukocyte count

No changes in heart mass or hematology at any
exhaust level or duration of exposure in either
species

3% increase in COHb


At higher concentrations, RBC, Hb, Hct slightly
elevated; MCV and mean corpuscular hemoglobin
and concentration were lowered



Increases in RBC, Hb, Hct, and WBC, primarily
banded neutrophils; suggestion of an increase in
prothrombin time; increased heart/body weight
and right ventricular/heart ratios and decreased left
ventricular contractility in 6.6 mg/m3 group
Increases in banded neutrophils; significant at 12
mo, but not 24 mo





Study
Lewis et al. (1989)


Lewis etal. (1989)
Vallyathan et al.
(1986)
Wiester et al.
(1980)

Heinrich et al.
(1982)

Penney et al.
(1981)


Karagianes et al.
(1981)

Ishinishi et al.
(1988)




Brightwell et al.
(1986)



Pepelko and
Peirano (1983)




       Key: MCV = Mean corpuscular volume.

-------
right ventricular wall thickness; these increases, however, did not achieve statistically significant
levels (Vallyathan et al., 1986). Brightwell et al. (1986) reported increased heart/body weight
and right ventricular/heart weight ratios and decreased left ventricular contractility in rats
exposed to 6.6 mg/m3 DPM for 16 h/day, 5 days/week for 104 weeks.
       The effects of DPM on the endothelium-dependent relaxation (EDR) of vascular smooth
muscle cells have been investigated (Ikeda et al., 1995, 1998). Incubation of rat thoracic aortae
with suspensions of DPM (10-100 |lg/mL) markedly attenuated acetylcholine-induced EDR.
The mechanism of this effect was studied further in cultured porcine endothelial cells (CPE).
A 10-min incubation of CPE with DPM (0.1-100 |lg/mL) inhibited endothelium-dependent
relaxing factor (EDRF) or nitric oxide (NO) release. A 10-min incubation of DPM with NO
synthase inhibited formation of NO2", a product of NO metabolism.  The authors concluded that
DPM, at the concentrations tested, neither induced cell damage nor inhibited EDRF release from
CPE, but scavenged and thereby blocked the physiological action of NO.

5.1.3.3.9.  Serum chemistry. A number of investigators have studied the effects of exposure to
DE on serum biochemistry, and no consistent effects have been found. Such studies are
summarized in Table 5-12.
       The biological significance of changes in serum chemistry reported by Lewis et al.
(1989) in female but not male rats  exposed at 2 mg/m3 DPM for 7 h/day, 5 days/week for 104
weeks is difficult to interpret. Not only were the effects noted in one sex (females) only, but the
serum enzymes, lactate dehydrogenase (LDH), serum glutamic-oxaloacetic transaminase
(SGOT), and serum glutamic-pyruvic transaminase (SGPT), were elevated in the control group,
a circumstance contrary to denoting organ damage in the exposed female rats. The elevations of
liver-related serum enzymes in the control versus the exposed female rats appear to be a random
event among these aged subjects. The incidence of age-related disease, such as mononuclear cell
leukemia, can markedly affect such enzyme levels, seriously compromising the usefulness of a
comparison to historical  controls.  The serum sodium values of 144 versus 148 mmol/L in
control and exposed rats, respectively, although statistically different, would have no biological
significance.
       The increased serum enzyme activities, alkaline phosphatase, SGOT,  SGPT,
gamma-glutamyl transpeptidase, and decreased cholinesterase activity suggest an impaired liver;
however, such an impairment was  not established histopathologically (Heinrich et al., 1982;
Ishinishi et al., 1988; Brightwell et al.,  1986).  The increased urea nitrogen, electrolyte levels,
and gamma globulin concentration and reduction in total blood proteins are indicative of
impaired kidney function.  Again, there was no histopathological confirmation of impaired
kidneys in these studies.
                                          5-73

-------
Table 5-12. Effects of chronic exposures to diesel exhaust on serum chemistry of laboratory animals

Species/sex
Rat, F344, M, F


Hamster, Syrian, M, F


Rat, F344/JcL, M, F






Rat, F344; Hamster,
Syrian





Cat inbred, M


"Light-duty engine.
""Heavy-duty engine.
Exposure
period
7 h/day
5 days/week
104 weeks
7-8 h/day
5 days/week
75 weeks
16 h/day
6 days/week
130 weeks




16 h/day
5 days/week
104 weeks




8 h/day
7 days/week
124 weeks


Particles
(mg/m3)
2.0
0.23
0.36 |lm MOD
3.9
0.1 |lniMDD

0.11"
0.41"
1.08"
2.31"
3.72b
0. 19-0.28 |lm
MOD
0.7
2.2
6.6




6.0C
12.0d



CxT
(mg-h/m3)
7,280


10,238-11,700


1,373
5,117
13,478
28,829
46,426


5,824
18,304
54,912




41,664
83,328



CO NO2 SO2
(ppm) (ppm) (ppm) Effects
11.5 1.5 0.8 Decreased phosphate, LDH, SGOT, and SGPT;
increased sodium in females but not males

18.5 1.2 3.1 After 29 weeks, increases in SGOT, LDH, alkaline
phosphatase, gamma-glutamyl transferase, and BUN

1.23 0.08 0.38 Lower cholinesterase activity in males in both the
2.12 0.26 1.06 light-and heavy-duty series and elevated gamma
3.96 3.96 2.42 globulin and electrolyte levels in males and females
7.10 7.10 4.70 in both series
12.9 3.00 4.57


— — — Rats, 6.6 mg/m3, reduction in blood glucose, blood
— — — proteins, triglycerides, and cholesterol; increase in
32.0 — — BUN, alkaline phosphate alamine, and aspartate
aminotransferases (SGPT and SGOT); hamsters, 6.6
mg/m3, decrease in potassium, LDH, aspartate amino-
transferase; increase in albumin and gamma-glutamyl
transferase
20.2 2.7 2. 1 BUN unaltered; SGOT and SGPT unaffected; LHD
33.3 4.4 5.0 increase after 1 year of exposure




Study
Lewis et al.
(1989)

Heinrich et al.
(1982)

Research
Committee for
HERP Studies
(1988)



Brightwell et al.
(1986)





Pepelko and
Peirano (1983)



°1 to 61 weeks of exposure.
d62 to 124 weeks of exposure.
Key: LDH
SGOT
BUN
SGPT
Lactate dehydrogenase.
Serum glutamic-oxaloacetic
transaminase.


Blood urea nitrogen.
Serum glutamic-pyruvic transaminase.

-------
       Clinical chemistry studies suggest impairment of both liver and kidney functions in rats
and hamsters chronically exposed to high concentrations of DE.  The absence of
histopathological confirmation, the appearance of such effects near the end of the lifespan of the
laboratory animal, and the failure to find such biochemical changes in cats exposed to a higher
dose, however, tend to discredit the probability of hepatic  and renal hazards to humans exposed
at atmospheric levels of DE.

5.1.3.3.10. Effects on microsomal enzymes.  Several studies have examined the effects of DE
exposure on microsomal enzymes associated with the metabolism and possible activation of
xenobiotics, especially polynuclear aromatic hydrocarbons (PAH). These studies are
summarized in Table 5-13. Lee et al. (1980) measured the activities of aryl hydrocarbon
hydroxylase (AHH) and epoxide hydrase (EH) in liver, lung, testis, and prostate gland of adult
male rats exposed to 6.32 mg/m3 DPM 20 h/day for 42 days.  Maximal significant AHH
activities (pmol/min/mg microsomal protein) occurred at different times during the exposure
period, and differences between controls and exposed rats, respectively, were as follows:
prostate 0.29 versus 1.31, lung 3.67 versus 5.11, and liver 113.9 versus 164.0. There was no
difference in AHH activity in the testis between exposed and control rats.  Epoxide hydrase
activity was not significantly different from control values for any of the organs tested.
       Pepelko and Peirano (1983) found no statistically significant differences in liver
microsomal cytochrome P448-450 levels and liver microsomal AHH between control  and diesel-
exposed mice at either 6 or 8 mo of exposure. Small differences were noted in the lung
microsomal AHH activities, but these were believed to be  artifactual differences, due to
increases in nonmicrosomal lung protein present in the microsomal preparations.  Exposures to 6
mg/m3 DPM were for 8 h/day, 7 days/week.
       Rabovsky et al. (1984) investigated the effect of chronic exposure to DE on microsomal
cytochrome P450-associated benzo[a]pyrene (B[a]P)  hydroxylase and 7-ethoxycoumarin
deethylase activities in rat lung and liver. Male rats were exposed for 7 h/day, 5 days/week for
104 weeks to 2 mg/m3 DPM. The exposure had no effect  on B[a]P hydroxylase or
7-ethoxycoumarin deethylase activities in lung or liver.  In related studies, Rabovsky et al.
(1986) examined the effects of DE on viral induced enzyme activity and interferon production in
female mice. The mice were exposed for 7 h/day, 5 days/week for 1 mo to DE diluted to
achieve a concentration of 2 mg/m3 DPM.  After the exposure, the mice were inoculated
intranasally with influenza virus.  Changes in serum levels of interferon and liver microsomal
activities of 7-ethoxycoumarin, ethylmorphine demethylase, and nicotinamide adenine dinucleotide
                                          5-75

-------
Table 5-13. Effects of chronic exposures to diesel exhaust on microsomal enzymes of laboratory animals

Species/sex
Rat, F344, M


Mouse, CD-1,F




Rat, Sprague-
Dawley, M



Rat, F344, M







Rat, F344, F



Rat, F344, M





Mouse, A/3, M


AHH = aryl
Exposure
period
—


7h/day
5 days/week 0
4 weeks


20 h/day
7 days/week
1-7 weeks


20 h/day
5.5 days/week
4, 13, 26, or
39 weeks
20 h/day
5.5 days/week
4, 13, 26, or
39 weeks
7 h/day
5 days/week 0.
12, 26, or
104 weeks
20 h/day
5.5 days/week
8-53 weeks



8 days/week
7 days/week
26 or 35 weeks
hydrocarbon hydro xylase.
Particles C x t CO
(ing/in3) (mg-h/m3) (ppm)
— — —


2.0 280 11.5
.2-0.36 |lm mdd



6.3 882-6,174 17.4




0.75 330-6,435 4.8
1.5 7.5
0.19|lmmdd

0.75 330-6,435 4.8
1.5 7.5
0.19|lmmdd

2.0 840-7,280 11.5
23-0.36 |lm mdd


0.25 220-8,745 2.9
1.5 7.5
0.19 |lmmdd



6.0 17.4 17.4



N02 S02
(ppm) (ppm) Effects
— — Intratracheal administration of DPM extract required
doses greater than 6 mg/m3 before the lung AHH was
barely doubled; liver AHH activity was unchanged
1.5 0.8 Mice inoculated intranasally with influenza virus had
smaller increases in ethylmorphine demethylase
activity on days 2 to 4 postvirus infection and abolition
of day 4 postinfection increase in NADPH-dependent
cytochrome c reductase
2.3 2.1 AHH induction occurred in lung, liver, and prostate
gland but not in testes; maximum significant activities
occurred at different times; liver has greatest overall
activity, percent increase highest in prostate; expoxide
hydrase activity was unaffected
— — Inhalation exposure had no significant effect on liver
— — AHH activity; lung AHH activity was slightly reduced
after 6-mo exposure to 1.5 mg/m3 DPM; an ip dose of
dp extract, estimated to be equivalent to inhalation
— — exposure, had no effect on AHH activity in liver and
— — lungs; cyt. P-50 was unchanged in lungs and liver
following inhalation or ip administration

1.5 0.8 No effect on B[a]p hydrolase or 7-exthoxycoumarin
deethylase activities in the liver


— — After 8 weeks, no induction of cyt. P-450, cyt. P-448,
— — or NADPH-dependent cyt. c reductase; after 1 year of
exposure, liver microsomal oxidation of B[a]p was not
increased; 1 year of exposure to either 0.25 or
1.5 mg/m3 DPM impaired lung microsomal metabolism
ofB[a]p
2.3 2.1 No differences in lung and liver AHH activities and
liver P-448, P-450 levels



Study
Chen (1986)


Rabovsky et al. (1986)




Lee et al. (1980)




Chen and Vostal (1981)







Rabovsky et al. (1984)



Navarroetal. (1981)





Pepelko and Peirano
(1983)


B[a]p = benzo[a]pyrene.

-------
phosphate (NADPH)-dependent cytochrome c reductase were measured. In the absence of viral
inoculation, exposure to DE had no significant effects on the activity levels of the two liver
microsomal monooxygenases and NADPH-dependent cytochrome c reductase.  Exposure to DE
produced smaller increases in ethylmorphine demethylase activity on days 2 to 4 postvirus
infection and also abolished the day 4 postinfection increase in NADPH-dependent cytochrome
c reductase when compared with nonexposed mice. These data suggested to the authors that the
relationship that exists between metabolic detoxification and resistance to infection in unexposed
mice was altered during a short-term  exposure to DE.
       Chen and Vostal (1981) measured the activity of AHH and the content of cytochrome
P450 in the lungs and livers of rats exposed by inhalation of DE or intraperitoneal (i.p.) injection
of a dichloromethane extract of DPM. In the inhalation exposures, the exhaust was diluted to
achieve concentrations of 0.75 or 1.5 mg/m3 DPM, and the exposure regimen was 20 h/day,
5.5 days/week for up to 9 mo.  The concentration of total hydrocarbons and particle-phase
hydrocarbons was not reported. Parenteral administration involved repeated injections at several
dose levels for 4 days. Inhalation exposure had no significant effect on liver microsomal AHH
activity; however, lung AHH activity was slightly reduced after 6 mo exposure to 1.5 mg/m3.
An i.p. dose of DPM extract, estimated to be equivalent to the inhalation exposure, had  no effect
on AHH activity in liver  or lungs. No changes were observed in cytochrome P450 contents in
lungs or liver following inhalation exposure or i.p. treatment.  Direct intratracheal administration
of a dichloromethane DPM extract required doses greater than 6 mg/kg body weight before the
activity of induced AHH in the lung was barely doubled; liver AHH activity remained
unchanged (Chen, 1986).
       In related studies, Navarro et al. (1981) evaluated the effect of exposure to DE on rat
hepatic and pulmonary microsomal enzyme activities. The  same exposure regimen was
employed  (20 h/day, 5.5  days/week, for up to 1 year), and the exhaust was diluted to achieve
concentrations of 0.25 and 1.5 mg/m3 DPM (a few studies were also conducted at 0.75 mg/m3).
After 8 weeks of exposure, there was no evidence for the induction of cytochrome P450,
cytochrome P448, or NADPH-dependent cytochrome c reductase in rat liver microsomes. One
year of exposure had little, if any, effect on the hepatic metabolism of B[a]P.  However, 1 year
of exposure to 0.25 and 1.5 mg/m3 significantly impaired the ability of lung microsomes to
metabolize B[a]P (0.15 and 0.02 nmole/30 min/mg protein, respectively, versus
0.32 nmole/30 min/mg protein for the controls).
       There are conflicting results regarding the induction of microsomal AHH activities in the
lungs and liver of rodents exposed to DE. One study reported induction of AHH activity in the
lungs, liver, and prostate of rats exposed to DE containing 6.32 mg/m3 DPM for 20 h/day for 42
days; however, no induction of AHH was observed in the lungs of rats and mice exposed to 6
mg/m3 DPM for 8 h/day, 7 days/week for up to 8 mo or to 0.25 to 2 mg/m3 for periods up to 2

                                         5-77

-------
years. Exposure to DE has not been shown to produce adverse effects on microsomal
cytochrome P450 in the lungs or liver of rats or mice.  The weight of evidence suggests that the
absence of enzyme induction in the rodent lung exposed to DE is caused either by the
unavailability of the adsorbed hydrocarbons or by their presence in quantities insufficient for
enzyme induction.

5.1.3.3.11. Effects on behavior and neurophysiology.  Studies on the effects of exposure to DE
on the behavior and neurophysiology of laboratory animals are summarized in Table 5-14.
Laurie et al. (1978) and Laurie et al. (1980) examined behavioral alterations in adult and
neonatal rats exposed to DE. Exposure for 20 h/day, 7 days/week, for 6 weeks to  exhaust
containing 6 mg/m3 DPM produced a significant reduction in adult spontaneous locomotor
activity  (SLA) and in neonatal pivoting (Laurie et al., 1978). In a follow-up study, Laurie et al.
(1980) found that shorter exposure (8 h/day) to 6 mg/m3 DPM also resulted in a reduction of
SLA in adult rats. Laurie et al. (1980)  conducted additional behavioral tests on adult rats
exposed during their neonatal period. For two of three exposure situations (20 h/day for 17 days
postparturition, or 8 h/day for the first 28 or 42 days postparturition), significantly lower SLA
was observed in the majority of the tests conducted on the adults after week 5 of measurement.
When compared with control rats, adult 15-month-old rats that had been exposed as neonates (20
h/day for 17 days) also exhibited a significantly slower rate of acquisition  of a bar-pressing task
to obtain food.  The investigators noted that the evidence was insufficient to determine whether
the differences were the result of a learning deficit or due to some other cause (e.g., motivational
or arousal differences).
       These data are difficult to interpret in terms of health hazards to humans under ambient
environmental conditions because of the high concentration of DE to which the laboratory rats
were exposed. Additionally, there are no further concentration-response studies to assess at what
exposure levels these observed results persist or abate. A permanent alteration in both learning
ability and activity resulting from exposures early in life is a health hazard whose  significance to
humans should be pursued further.
      Neurophysiological effects from exposure to DE were investigated in rats by Laurie and
Boyes (1980, 1981). Rats were exposed to diluted DE containing 6 mg/m3 DPM for 8 h/day, 7
days/week from birth up until 28 days of age.  Somatosensory evoked potential, as elicited by  a
1 mA electrical pulse to the tibial nerve in the left hind limb, and visual  evoked potential, as
elicited by a flash of light, were the endpoints tested. An  increased pulse latency was reported
for the rats exposed to DE, and this was thought to be caused by a reduction in the degree of
                                          5-78

-------
               Table 5-14.  Effects of chronic exposures to diesel exhaust on behavior and neurophysiology
Species/sex
Rat, Sprague-
Dawley, M
Rat, Sprague
Dawley, F
Rat, Sprague-
Dawley, F


Exposure Particles C x T
period (ing/in3) (mg-h/m3)
8 h/day 6 336-1,344
7 days/week
1-4 weeks
20 h/day 6 5,040
7 days week
6 weeks
8 or 20 h/day 6 1,008-13,440
7 days/week
3, 4, 6, or
16 weeks

CO NO2 SO2
(ppm) (ppm) (PPm) Effects
19 2.5 1.8 Somatosensory and visual evoked
potentials revealed longer pulse
latencies in pups exposed neonatally
19 2.5 1.8 Reduction in adult SLA and in neonatal
pivoting
19 2.5 1.8 Reduction in SLA in adults; neonatal
exposures for 20 or 8 h/day caused
reductions in SLA. Neonatal
exposures for 20 h/day for 17 days
resulted in a slower rate of a
bar-pressing task to obtain food
Study
Laurie and Boyes
(1980, 1981)
Laurie et al. (1978)
Laurie et al. (1980)


        SLA = Spontaneous locomotor activity.
VO

-------
nerve myelinization.  There was no neuropathological examination, however, to confirm this
supposition.
       Based on the data presented, it is not possible to specify the particular neurological
impairment(s) induced by the exposure to DE.  Again, these results occurred following exposure
to a high level of DE and no additional concentration-response studies were performed.

5.1.3.3.12. Effects on reproduction and development. Studies of the effects of exposure to DE
on reproduction and development are summarized in Table 5-15.  Twenty rats were exposed 8
h/day on days 6 through 15 of gestation to diluted DE containing 6 mg/m3 DPM (Werchowski et
al., 1980a,b; Pepelko and Peirano,  1983). There were no signs of maternal toxicity or decreased
fertility. No skeletal  or visceral teratogenic effects were observed in 20-day-old fetuses
(Werchowski et al., 1980a). In a second study, 42 rabbits were exposed to 6 mg/m3 DPM for 8
h/day on gestation days 6 through 18.  No adverse effects on body weight gain or fertility were
seen in the does exposed to DE. No visceral or skeletal developmental abnormalities were
observed in the fetuses (Werchowski et al., 1980b).
       Pepelko and Peirano (1983) evaluated the potential for DE to affect reproductive
performance in mice  exposed from 100 days prior to exposure throughout maturity of the F2
generation. The mice were exposed for 8 h/day, 7 days/week to 12 mg/m3 DPM.  In general,
treatment-related effects were minimal.  Some differences in organ and body weights were
noted, but overall fertility and survival rates were not altered by exposure to DE. The only
consistent change, an increase in lung weights, was accompanied by a gross pathological
diagnosis of anthracosis. These data denoted that exposure to DE at a concentration of 12 mg/m3
did not affect reproduction.  See Section 5.3, which reports a lack of effects of exposure to DE
on rat lung development (Mauderly et al., 1987b).
       Several studies have evaluated the effect of exposure to DE on sperm.  Lewis et al.
(1989) found no adverse sperm effects (sperm motility, velocity, densities, morphology, or
incidence of abnormal sperm) in monkeys exposed for 7 h/day, 5 days/week for 104 weeks to 2
mg/m3 DPM. In another study in which A/Strong mice were exposed to DE containing 6 mg/m3
DPM for 8 h/day for 31 or 38 weeks, no significant differences were observed in sperm
morphology between exposed and control mice (Pereira et al., 1981).   It was noted, however,
that there was a high  rate of spontaneous sperm abnormalities in this strain of mice, and this
may  have masked any small positive effect. Quinto and De Marinis (1984) reported a
statistically significant and dose-related increase in sperm abnormalities in mice injected
intraperitoneally for 5 days with 50, 100, or 200 mg/kg of DPM suspended in corn oil.
A significant decrease in sperm number was seen at the highest dose,  but testicular weight was
unaffected by the treatment.
                                          5-80

-------
           Table 5-15. Effects of chronic exposures to diesel exhaust on reproduction and development in laboratory animals
oo
Species/sex
Mouse,
[C57BL1/
6XC3HPF1!, M
Rat, Sprague-
Dawley, F
Rabbit, New
Zealand Albino,
F
Monkey,
Cynomolgus, M
Mouse,
A/Strong, M
Mouse, CD-I,
M,F
Exposure
period
5 days
8h/day
7 days/week
1.7 weeks
8h/day
7 days/week
1.9 weeks
7h/day
5 days/week
104 weeks
8h/day
7 days/week
31 or
38 weeks
8h/day
7 days/week
6 to 28
weeks
Particles C x T CO
(ing/in3) (mg-h/m3) (ppm)
50, 100, or — —
200 mg/kg
in corn oil;
. .i-P-.
injection
6 571 20
6 638 20
2 7,280 11.5
6 10,416- 20
12,768
12 4,032-18,816 33
N02 S02
(ppm) (ppm) Effects
— — Dose-related increase in
sperm abnormalities;
decrease in sperm number at
highest dose, testicular
weights unaffected
2.7 2.1 No signs of maternal to xicity
or decreased fertility; no
skeletal or visceral
teratogenic effects in 20-day -
old fetuses
2.7 2.1 No adverse effects on
maternal weight gain or
fertility; no skeletal or
visceral teratogenic effects in
the fetuses
1.5 0.8 No effects on sperm motility,
velocity density,
morphology, or incidence of
abnormalities
2.7 2.1 No effect on sperm
morphology; nigh rate of
spontaneous sperm
abnormalities may have
masked small effects
4.4 5.0 Overall fertility and survival
rates were unaffected in the
three-generation
reproductive study; only
consistent change noted, an
increase in lung weights, was
diagnosed as anthracosis
Study
Quinto and De
Marinis (1984)
Werchowski et al.
(1980a)
Pepelko and
Peirano (1983)
Werchowski et al.
(1980a)
Pepelko and
Peirano (1983)
Lewis etal. (1989)
Pereiraetal. (1981)
Pepelko and
Peirano (1983)

-------
       Watanabe and Oonuki (1999) investigated the effects of diesel engine exhaust on
reproductive endocrine function in growing rats. The rats were exposed to whole diesel engine
exhaust (5.63 mg/m3 DPM, 4.10 ppm NO2, and 8.10 ppm NOX); a group was exposed to filtered
exhaust without DPM, and a group was exposed to clean air. Exposures were for 3 mo
beginning at birth (6 hrs/day for 5 days/week).
       Serum levels of testosterone and estradiol were significantly higher and follicle-
stimulating hormone significantly lower in animals exposed to whole DE and filtered exhaust
compared to controls.  Luteinizing hormone was significantly decreased in the whole-exhaust-
exposed group as compared to the control and filtered groups. Sperm production and activity of
testicular hyaluronidase were significantly reduced in both exhaust-exposed groups as compared
to the control group. This study suggests that DE stimulates hormonal secretion of the adrenal
cortex, depresses gonadotropin-releasing hormone, and inhibits spermatogenesis in rats.
Because these effects were not inhibited by filtration, the gaseous phase of the exhaust appears
more responsible than  particulate matter for disrupting the endocrine system.
       The effects of freshly generated DE particles on the reproductive system of male Fischer
344 rats were investigated by Tsukue et al. (2001).  Groups (n=25) of 13-mo. old male rats were
exposed to whole DE diluted to 0.33, 0.99 or 3.24 mg/m3 (MMAD = 0.4 urn) for 8 months
12 hrs/day, 7 days/week. Subsequent to this exposure, evaluation of potential reproductive effect
was performed, including measurement of reproductive organ weights, sperm characteristics and
number, gonadotrophins, testosterone, and inhibin. Results showed either no effect or effects
with an inconsistent dose-response character that typically were not different from controls even
at the highest exposure concentration.
       No teratogenic, embryotoxic, fetotoxic, or female reproductive effects were observed in
mice, rats, or rabbits at exposure levels up to 12 mg/m3 DPM.  Effects on sperm morphology and
number were reported  in hamsters and mice exposed to high doses of DPM; however, no adverse
effects were observed in sperm obtained from monkeys exposed at 2 mg/m3 for 7 hrs/day,
5 days/week for 104 weeks.  Concentrations of 12 mg/m3 DPM did not affect male rat
reproductive fertility in the F0 and Fx generation breeders.  Thus, exposure to DE would not
appear to be a reproductive or developmental hazard.

5.2.  MODE OF ACTION OF DIESEL EXHAUST-INDUCED NONCANCER EFFECTS
5.2.1.  Comparison of Health Effects of Filtered and Unfiltered Diesel Exhaust
       There exist a total of four chronic toxicity studies of DE, in which the experimental
protocol included exposing test animals to exhaust containing no particles. Comparisons were
then made between the effects caused by whole, unfiltered exhaust and those caused by the
gaseous components of the exhaust. Concentrations of components of the exposure atmospheres
in these four studies are given in Table 5-16.

                                         5-82

-------
       Heinrich et al. (1982) compared the toxic effects of whole and filtered DE on hamsters
and rats.  Exposures were at 3.9 mg/m3 for 7 to 8 hrs/day and 5 days/week. Rats exposed for 24
mo to either whole or filtered exhaust exhibited no significant changes in respiratory frequency,
respiratory minute volume, compliance or resistance as measured by a whole-body
plethysmography, or heart rate. In the hamsters, histological changes (adenomatous
proliferations) were seen in the lungs of animals exposed to either whole or filtered exhaust;
however, in all groups exposed to the whole exhaust the number of hamsters exhibiting such
lesions was significantly higher than for the corresponding groups exposed to filtered exhaust or
clean air.  Severity of the lesions was, however, not reported.
       In a second study, Heinrich et al. (1986a,  see also Stober, 1986) compared the toxic
effects of whole and filtered DE on hamsters, rats, and mice. The test animals (96 per test
group) were exposed to 4.24 mg DPM/m3 for 19 hrs/day, 5 days/week for 120 (hamsters and
mice) or  140 (rats) weeks.  Body weights of hamsters were unaffected by either exposure. Body
weights of rats and mice were reduced by the whole exhaust but not by the filtered exhaust.
Exposure-related higher mortality rates occurred in mice after 2 years of exposure to whole
exhaust.  After 1 year of exposure to the whole exhaust, hamsters exhibited increased lung
weights, a significant increase in airway resistance, and a nonsignificant reduction in lung
compliance.  For the same time period, rats exhibited increased lung weights, a significant
decrease  in dynamic lung compliance, and a significant increase in airway resistance. Test
animals exposed to filtered exhaust did not exhibit such effects.  Histopathological examination
indicated that different levels of response occurred in the three species.  In hamsters, filtered
exhaust caused no significant histopathological effects in the lung; whole exhaust caused
thickened alveolar septa, bronchioloalveolar hyperplasia, and emphysematous lesions. In mice,
whole exhaust, but not filtered exhaust, caused multifocal bronchioloalveolar hyperplasia,
multifocal alveolar lipoproteinosis, and multifocal interstitial fibrosis. In rats, there  were no
significant morphological changes in the lungs following exposure to filtered exhaust. In rats
exposed to whole  exhaust, there were severe inflammatory changes in the lungs, thickened
alveolar septa, foci of macrophages,  crystals of cholesterol, and hyperplastic and metaplastic
lesions. Biochemical studies of lung lavage fluids of hamsters and mice indicated that exposure
to filtered exhaust caused fewer changes than did exposure to whole  exhaust. The latter
produced significant increases in lactate dehydrogenase, alkaline phosphatase, glucose-6-
phosphate dehydrogenase (G6PDH), total protein, protease (pH 5.1), and collagen.   The filtered
exhaust had a slight but nonsignificant effect on G6PDH, total protein, and collagen. Similarly,
cytological studies showed that while the filtered exhaust had  no effect on differential cell
                                          5-83

-------
        Table 5-16. Composition of exposure atmospheres in studies comparing unfiltered and filtered diesel exhaust"
Exposure'
Species/sex period
Rat, Wistar, F; 7 h/day
Hamster, Syrian 5 days/week
104 weeks
Rat, F344, F 8 h/day
7 days/week
104 weeks

Rat, F344, M, F; 16 h/day
Hamster, Syrian, 5 days/week
M, F 104 weeks


Rat, Wistar, F; 19 h/day
Hamster, Syrian, F; 5 days/week
Y1 Mouse NMRI, F 120 to
°° 140 weeks



Mouse, NMRI, F, 18 h/day
C57BL/6N, F 5 days/week
23 mo
(NMRI)
24 mo
(C57BL/6N)
Particles
(mg/m3)
Uf 3.9
F —
C —
Uf 4.9
F° —
C —

Uf 0.7
Uf 2.2
Uf 6.6
F" -
C —
Uf 4.24
F" -
C —




Uf 4.5
F 0.01
C 0.01



C x t CO NO2
(mg-h/m3) (ppm) (ppm)
14,196 18.5 1.2
18.0 1.0
— —
28,538 7.0 1.8
— —
— —

5,824 — —
18,304 — —
54,912 32.0 —
32.0 —
1.0 —
48,336 12.5 1.5
56,392 11.1 1.2
0.16 —




40,365 14.2 2.3
14.2 2.9
0.2 0.01



SO2
(ppm) Effects
3.1 No effect on pulmonary function or heart rate in rats; increases in
2.8 pulmonary adenomatous proliferations in hamsters, UF
— significantly higher than F or C
13.1 Body weight decrease after 6 mo in UF, 18 mo in f; lung/body
— rate weight rate higher in both groups at 24 mo; at 2 years,
— fibrosis and epithelial hyperplasia in lungs of uf ; nominal lung
and spleen histologic changes
— Uf: elevated red and white cell counts, hematocrit and hemo-
— globin; increased heart/body weight and right ventricular/heart
— weight ratios; lower left ventricular contractility; changes in blood
— chemistry; obstructive and restrictive lung disease; F: no effects
—
3.1 Uf: decreased body wt in rats and mice but not hamsters; increas-
1.02 ed mortality, mice only; decreased lung compliance and increased
— airway resistance, rats and hamsters; species differences in lung
lavage enzymes and cell counts and lung histopathology and
collagen content, most pronounced in rats; F: no effect on
glucose-6-phosphate dehydrogenase, total protein, and lung
collagen
2.8 Uf: increased lung wet weight starting at 3 mo
2.4
0.1 F: no noncancer effects reported



Study
Heinrich et al.
(1982)

Iwai et al. (1986)



Brightwell et al.
(1986)



Heinrich et al.
(1986a)





Heinrich et al.
(1995)




"Man values.
bUF= unfiltered whole exhaust, F = filtered exhaust, C = control.
"Reported to have the same component concentrations as the unfiltered, except particles were present in undetectable amounts.
""Concentrations reported for high concentration level only.

-------
counts, the whole exhaust resulted in an increase in leukocytes (161 ± 43.3/uL versus 55.7 ±
12.8/uL controls), a decrease in AMs (30.0 ± 12.5 versus 51.3 ± 12.5/uL in the controls),
and an increase in granulocytes (125 ± 39.7 versus 1.23 ± 1.14/uL in the controls).  All
values presented for this study are the mean with its standard deviation. The differences were
significant for each cell type.  There was also a small increase in lymphocytes (5.81 ± 4.72
versus 3.01 ± 1.23 uL in the controls).
       Iwai et al. (1986) exposed rats (24 per group) to whole or filtered DE 8 h/day,
7 days/week for 24 mo.  The whole exhaust was diluted to achieve  a concentration of
4.9 ±1.6 mg/m3 DPM.  Body weights in the whole exhaust group began to decrease after 6 mo
and in both exposed groups began to decrease after 18 mo when compared with controls.
Lung-to-body weight ratios of the rats exposed  to the whole exhaust showed a significant
increase (p<0.01) after 12 mo in comparison with control values. Spleen-to-body weight ratios
of both exposed groups were higher than control values after 24 mo.  After 6 mo of exposure to
whole exhaust, DPM accumulated in AMs, and Type II cell hyperplasia was observed. After
2 years of exposure, the alveolar walls had become fibrotic with mast cell infiltration and
epithelial hyperplasia. In rats exposed to filtered exhaust, after 2 years there were only minimal
histologic changes in the lungs, with slight hyperplasia and stratification of bronchiolar
epithelium and infiltration of atypical lymphocytic cells in the spleen.
       Brightwell  et al. (1986) evaluated the toxic effects of whole and filtered DE on rats and
hamsters.  Three exhaust dilutions were tested,  producing concentrations of 0.7, 2.2, and 6.6
mg/m3 DPM. The test animals (144 rats and 312 hamsters per exposure group) were exposed for
five 16-h periods per week for 2 years.  The four exposure types were gasoline, gasoline catalyst,
diesel, and filtered diesel. The results presented were limited to statistically significant
differences between exhaust-exposed and control animals. The inference from the discussion
section of the paper was that there was a minimum of toxicity in the animals exposed to filtered
DE:  "It is clear from the results presented that statistically significant differences between
exhaust-exposed and control animals are almost exclusively limited to animals exposed to either
gasoline or unfiltered diesel exhaust." Additional results are described in Section 5.1.3.3.
       Heinrich et al. (1995) exposed female NMRI and  C57BL/6N mice to a DE dilution that
resulted in a DPM concentration of 4.5 mg/m3 and to the  same dilution after filtering to  remove
the particles. This study is focused on the carcinogenic effects of DPM exposure, and
inadequate information was presented to compare noncancer effects in filtered versus unfiltered
exhaust.
       A comparison of the toxic responses in laboratory animals exposed to whole exhaust or
filtered exhaust containing no particles demonstrates across studies that when the exhaust is
sufficiently diluted to limit the concentrations of gaseous irritants (NO2 and SO2), irritant vapors
(aldehydes), CO, or other systemic toxicants, the diesel particles are the prime etiologic  agents

                                          5-85

-------
of noncancer health effects, although additivity or synergism with the gases cannot be ruled out.
These toxic responses are both functional and pathological and represent cascading sequelae of
lung pathology based on concentration and species.  The diesel particles plus gas exposures
produced biochemical and cytological changes in the lung that are much more prominent than
those evoked by the gas phase alone. Such marked differences between whole and filtered DE
are also evident from general toxicological indices, such as decreases in body weight and
increases in lung weights, pulmonary function measurements, and pulmonary histopathology
(e.g., proliferative changes in  Type II cells and respiratory bronchiolar epithelium, fibrosis).
Hamsters, under equivalent exposure regimens, have lower levels of retained DPM in their lungs
than rats and mice do and, consequently, less pulmonary function impairment  and pulmonary
pathology. These differences may result from lower DPM inspiration and deposition during
exposure,  greater DPM clearance, or lung tissue less susceptible to the cytotoxicity of deposited
DPM.

5.2.2.  Mode of Action for the Noncarcinogenic Effects of DPM
       As noted in Chapter 2, diesel emissions are a complex mixture that includes both a vapor
phase and a particle phase.  The particle phase consists of poorly soluble carbon particles on the
surfaces of which are adsorbed a large number of organic and inorganic compounds.  Although
the effects to be discussed are considered attributable to the particle phase (termed diesel
particulate matter or DPM), additive or synergistic effects due to the vapor phase cannot be
totally discounted.  This may be especially so in the human studies and the animal toxicology
studies where exposure is to various dilutions of diesel emissions, or in the in vitro studies in
which the test material was captured by filtration.
       The mechanisms by which DPM is inhaled, deposited, and cleared from the respiratory
tract are discussed in Chapter 3.  DPM deposited upon airway surfaces may be cleared from the
respiratory tract completely, or may be translocated to other sites within the respiratory system.
In rats, the pathogenic sequence following the deposition of inhaled DPM begins with the
interaction of DPM with airway epithelial cells and phagocytosis by AMs.  The airway epithelial
cells and activated AMs release chemotactic factors that attract neutrophils and additional AMs.
As the lung burden of DPM increases, there is an aggregation of particle-laden AMs in alveoli
adjacent to terminal bronchioles, increases in the number of Type II cells lining particle-laden
alveoli, and the presence of particles within alveolar and peribronchial interstitial tissues and
associated lymph nodes.
       The macrophages  engulfing the DPM may release cytokines, growth factors, and
proteases, which may  cause inflammation, cell injury, cell proliferation, hyperplasia, and
fibrosis. This is especially true under lung overload conditions occurring in laboratory rats when
the rate of deposition exceeds the rate of alveolar clearance.  This phenomenon is described in

                                           5-86

-------
Chapter 3.  The mechanisms leading to the generation of oxygen radicals and subsequent lung
injury are described in Chapter 7, Section 7.4.3.
       DPM is a poorly soluble particle whose rate of clearance by dissolution is likely
insignificant compared to its rate of clearance as an intact particle. The organic material
adsorbed to the surface is desorbed from the DPM and may enter into metabolic reactions and be
activated and enter into reactions with other macromolecules or be detoxified and excreted
(Figure 7-1).  The diesel particle may be cleared directly by the clearance mechanisms described
in Chapter 3.
       The organic material desorbed from the particle (described in Chapter 7, Section 7.4.7)
appears to be associated with the immunological changes described above.  The potential
adjuvant effects of DPM have also been studied. The results indicate that the nonextractable
particle core and the organic matter adsorbed to the core both contribute to the adjuvant activity
of DPM. Further, it is possible that any of the plethora of compounds present in the organic
fraction of DPM, including various PAH, may elicit this response.
       Thus, the available evidence indicates that DPM has the potential to produce pathological
and immunological changes in the respiratory tract.  Moreover, the magnitude of these responses
is determined by the dose delivered to the respiratory tract and is  attributable to both the carbon
core and the adsorbed organic materials.

5.3.  INTERACTIVE EFFECTS OF DIESEL EXHAUST
       A multitude of factors may influence the susceptibility to  exposure to DE as well as the
resulting response. Some of these have already been discussed in detail (e.g., the composition of
DE and concentration-response data); others will be addressed in  this section (e.g., the
interaction of DE with factors particular to the exposed individual and the interaction of DE
components with other airborne contaminants).
       In a study discussed already in this chapter, Mauderly et al. (1990a)  compared the
susceptibility of normal rats and rats with preexisting laboratory-induced pulmonary emphysema
exposed for 7 h/day, 5 days/week for 24 mo to DE containing 3.5 mg/m3 DPM or to clean air
(controls).  Emphysema was induced in one-half of the  rats by intratracheal instillation of
elastase 6 weeks before exhaust exposure. Measurements included lung burdens of DPM,
respiratory function, bronchoalveolar lavage, clearance of radiolabeled particles, pulmonary
immune responses, lung collagen, excised lung weight and volume, histopathology, and mean
linear intercept of terminal air spaces. None of the data for the 63 parameters measured suggest
that rats with emphysematous lungs were more susceptible than rats with normal lungs to the
effects of DE exposure. In fact, each of the  14 emphysema-exhaust interactions detected by
statistical analysis of variance indicated that emphysema acted to reduce the effects of DE
exposure. DPM accumulated much less rapidly in the lungs of emphysematous rats than in those

                                          5-87

-------
of normal rats. The mean lung burdens of DPM in the emphysematous rats were 39%, 36%, and
37% of the lung burdens of normal rats at 12, 18, and 24 mo, respectively. No significant
interactions were observed among lung morphometric parameters. Emphysema prevented the
exhaust-induced increase for three respiratory indices of expiratory flow rate at low lung
volumes, reduced the exhaust-induced increase in nine lavage fluid indicators of lung damage,
prevented the expression of an exhaust-induced increase in lung collagen, and reduced the
exhaust-induced delay in DPM clearance.
       Mauderly et al. (1987b) evaluated the relative susceptibility of developing and adult rat
lungs to damage by exposure to DE.  Rats (48 per test group) were exposed to DE containing 3.5
mg/m3 DPM and about 0.8 ppm NO2. Exposures were for 7 h/day, 5 days/week through
gestation to the age of 6 mo, or from the age of 6 to 12 mo.  Comparative studies were
conducted on respiratory function, immune response, lung clearance, airway fluid enzymes,
protein and cytology, lung tissue collagen, and proteinases in both age groups. After the 6-mo
exposure, adult rats, compared with controls, exhibited (1) more focal aggregates  of particle-
containing AMs in the alveolar ducts near the terminal bronchioles, (2) a sixfold increase in the
neutrophils (as a percentage of total leukocytes) in the airway fluids, (3) a significantly higher
number of total lymphoid cells in the pulmonary lymph  nodes, (4) delayed clearance of DPM
and radiolabeled particles (t1/2 = 90 days versus 47 days  for controls), and (5) increased lung
weights.  These effects were not seen in the developing rats. On a weight-for-weight
(milligrams of DPM per gram of lung) basis, DPM accumulation in the lungs was similar in
developing and adult rats immediately after the exposure.  During the 6-mo postexposure period,
DPM clearance was much more rapid in the developing  rats, approximately 2.5-fold.  During
postexposure, diesel particle-laden macrophages became aggregated in the developing rats, but
these aggregations were located primarily in a subpleural position. The authors concluded that
exposure to DE, using pulmonary function,  structural (qualitative or quantitative)  biochemistry
as the indices, did not affect the developing rat lung more severely than the adult rat lung.
       As a result of the increasing trend of using diesel-powered equipment in coal mining
operations and the concern for adverse health effects in coal miners exposed to both coal dust or
coal mine dust and DE, Lewis et al.  (1989) and Karagianes et al. (1981) investigated the
interaction of coal dust and DE. Lewis et al. (1989) exposed rats, mice, and cynomolgus
monkeys to (1) filtered ambient air, (2) 2 mg/m3 DPM, (3) 2 mg/m3 respirable coal dust, and (4)
1 mg/m3 of both DPM and respirable coal dust. Gaseous and vapor concentrations were
identical in both DE exposures. Exposures were for 7 h/day, 5 days/week for up to 24 mo.
Synergistic effects between DE and coal dust were not demonstrated;  additive toxic effects were
the predominant effects noted.
       Karagianes et al. (1981) exposed rats (24 per group) to DE containing 8.3  mg/m3 of DPM
alone or in combination with about 6 mg/m3 of coal dust. No synergistic effects were found

-------
between DE and coal dust; additive effects in terms of visual dust burdens in necropsied lungs
were related to dose (i.e., length of exposure and airborne particulate concentrations).
       The health effects of airborne contaminants from sources other than diesel engines may
be altered in the presence of DPM by their adsorption onto the diesel particles.  When adsorbed
onto diesel particles, the gases and vapors can be transported and deposited deeper into the
lungs, and because they are more  concentrated on the particle surface, the resultant cytotoxic
effects or physiological responses may be enhanced. Nitrogen dioxide adsorbed onto carbon
particles caused pulmonary parenchymal lesions in mice, whereas NO2 alone produced edema
and inflammation but no lesions (Boren, 1964).  Exposure to formaldehyde and acrolein
adsorbed onto carbon particles (1  to 4 |im) resulted in the recruitment of PMNs to tracheal and
intrapulmonary epithelial tissues but not when the  aldehydes were tested alone (Kilburn and
McKenzie,  1978).
       Madden et al. (2000) observed that O3 exposure increased the bioactivity of DPM. DPM,
preexposed to O3 for 48 h or nonozone-exposed DPM (1 to 500 |lg), was instilled into the lungs
of laboratory rats. Lung inflammation and injury were examined 24 h after instillation by lung
lavage. DPM pre-exposed to 0.1  PPM O3 was more potent  in increasing neutrophilia, lavage
total protein,  and LDH compared  to unexposed DPM. Treatment of DPM with higher
concentrations of O3 (1.0 PPM) decreased the bioactivity of the particles.
       There is no direct evidence that DE, at concentrations found in the ambient environment,
interacts with other substances in  the exposure environment or the physiological status of the
exposed subject other than impaired resistance to respiratory tract infections. Although there is
experimental evidence that gases  and vapors can be adsorbed onto carbonaceous particles,
enhancing the toxicity of these particles when deposited in the lung, there is no evidence for an
increased health risk from such interactions with DPM under urban atmospheric conditions.
Likewise, there is no experimental evidence in laboratory animals that the youth or preexisting
emphysema of an exposed individual enhances the risk of exposure to DE.

5.4.  COMPARATIVE RESPONSIVENESS AMONG SPECIES TO THE
     HISTOPATHOLOGIC EFFECTS OF DIESEL EXHAUST
       There is some evidence indicating that species may differ in pulmonary responses to DE.
Mauderly (1994) compared the pulmonary histopathology of rats and mice after 18 mo of
exposure to DE.  There was less aggregation of macrophages in mice.  Diffuse septal thickening
was noted in the mice, but there were few inflammatory cells, no focal fibrosis, little epithelial
hyperplasia, and no epithelial metaplasia, as was observed in rats. Heinrich  et al. (1986a)
reported that wet lung weight of hamsters increased only 1.8-fold following chronic exposure to
DE, compared with an increase of 3.4-fold in rats.  Smaller increases in neutrophils, lactic acid
dehydrogenase, collagen, and protein supported the conclusion of a lesser inflammatory response

                                          5-89

-------
in Syrian hamsters. The histopathologic changes in the lungs of Chinese hamsters after 6 mo
exposure to DE, on the other hand, was similar to that of rats (Pepelko and Peirano, 1983).
Guinea pigs respond to chronic DE exposure with a well-defined epithelial proliferation, but it is
based on an eosinophilic response in contrast to the neutrophil-based responses in other species.
Epithelial hyperplasia and metaplasia were quite striking in the terminal and respiratory
bronchioles of cats exposed for 27 mo to DE (Plopper et al., 1983).  This study is of particular
interest because the terminal airways of cats are more similar to those of humans than rodent
species are. It should be noted, however, that exposure concentrations were very high
(12 mg/m3) for most of the period. Lewis et al. (1989) exposed rats and cynomolgus monkeys
8 h per day, 5 days per week for 2 years to DE at a particle concentration of 2 mg/m3.
Unfortunately, this exposure rate was sufficiently low that few effects were noted in either
species other than focal accumulations of particles, primarily in the alveolar macrophages,
interstitium, and lymphoid tissue. It is apparent that species do vary in their pulmonary
responses to DE exposure,  despite the difficulty in making direct comparisons because of
differences in exposure regimes, lifespans, and pulmonary anatomy. Most species do respond,
however, suggesting that humans are likely to be susceptible to induction of pulmonary
pathology during chronic exposure to DE at some level.

5.5. DOSE-RATE AND PARTICULATE CAUSATIVE ISSUES
       The purpose of animal toxicological experimentation is to elucidate mechanisms of
action and identify the hazards and dose-response effects posed by a chemical substance or
complex mixture and to extrapolate these effects to humans for subsequent health assessments.
The cardinal principle in such a process is that the intensity and character of the toxic action are
a function of the dose of the toxic agent(s) that reaches the critical site of action. The
considerable body of evidence reviewed clearly denotes that major noncancerous health hazards
may be presented to the lung following the inhalation of DE. Based on pulmonary function and
histopathological and histochemical effects, a determination can be made concerning which
dose/exposure rates of DE  (expressed in terms of the DPM concentration) result in injury to the
lung and which appear to elicit no effect. The inhalation of poorly soluble particles, such as
those found in DE, increases the pulmonary particulate burden. When the dosing rate exceeds
the ability of the pulmonary defense mechanisms to achieve a steady-state lung burden of
particles, there is a slowing of clearance and the progressive retention of particles in the lung that
can ultimately approach a complete cessation  of lung clearance (Morrow,  1988). This
phenomenon, which is reviewed in Chapter 3, has practical significance both for the
interpretation of experimental inhalation data  and for the prevention of disease in humans
exposed to airborne particles.
                                          5-90

-------
       The data for exposure intensities that cause adverse pulmonary effects demonstrate that
they are less than the exposure intensities reported to be necessary to induce lung tumors. Using
the most widely studied laboratory animal species and the one reported to be the most sensitive
to tumor induction, the laboratory rat, the no-adverse-effect exposure intensity for adverse
pulmonary effects was 56 mg-lrm"3/week (Brightwell et al., 1986). The lowest-observed-effect
level for adverse pulmonary effects (noncancer) in rats was 70 mg-lrm"3/week (Lewis et al.,
1989), and for pulmonary tumors, 122.5 mg-lrm"3/week (Mauderly et al., 1987a).  The results
clearly show that noncancerous pulmonary effects are produced at lower exposure intensities
than are pulmonary tumors. Such data support the position that inflammatory and proliferative
changes in the lung may play a key  role in the etiology of pulmonary tumors in exposed rats
(Mauderly etal., 1990b).
       The effects of DE on the developing lung and on a model of a preexisting disease state
have been studied in rats (Mauderly et al., 1990a,  1987b).  Mauderly et al. (1987b) showed that
diesel did not affect the developing  lung more severely than the adult rat lung, and in fact, that
clearance was faster in the younger  lung. Mauderly et al. (1990a) compared the pulmonary
response to inhalation of DE in rats with elastase-induced emphysema with normal rats. They
found that respiratory tract effects were not more severe in emphysematous rats and that the lung
burden of particles was less in the compromised rat. These studies provide limited evidence that
some factors that are often considered to result in a wider distribution of sensitivity among
members of the population may  not have this effect with diesel exposure. However, these
studies have no counterpart in human studies and extrapolation to humans remains uncertain.
       There is also the issue of whether the noncancerous health effects related to exposure to
DE are caused by the carbonaceous core of the particle or substances adsorbed onto the core, or
both.
       Current understanding, derived primarily from studies in rats, suggests that much of the
toxicity resulting from the inhalation of DE relates to the carbonaceous core of the particles.
Several studies on inhaled aerosols demonstrate that lung reactions characterized by an
appearance of particle-laden AMs and their infiltration into the alveolar ducts, adjoining alveoli,
and tracheobronchial lymph nodes;  hyperplasia of Type II cells; and  the impairment of
pulmonary clearance mechanisms are not limited to exposure to diesel particles. Such responses
have also been observed in rats following the inhalation of coal dust  (Lewis et al., 1989;
Karagianes et al., 1981), titanium dioxide (Heinrich et al., 1995; Lee et al.,  1985), CB (Nikula et
al., 1995; Heinrich et al., 1995),  titanium tetrachloride hydrolysis products (Lee et al., 1986),
quartz (Klosterkotter and Biinemann, 1961), volcanic ash (Wehner et al.,  1986), amosite (Bolton
et al., 1983), and manmade mineral fibers  (Lee et al., 1988) among others.  In more recent
studies, animals have been exposed to CB that is similar to the carbon core  of the DE particle.
Nikula et al. (1995) exposed rats for 24 mo to CB or DE at target exposure  concentrations of 2.5

                                          5-91

-------
and 6 mg/m3 (exposure rates of 200 or 520 mg-lrm"3/week).  Both concentrations induced AM
accumulation, epithelial proliferation, inflammation, and fibrosis. They observed essentially no
difference in potency of nonneoplastic or in tumor responses based on a regression analysis.
       Dungworth et al. (1994) reported moderate to severe inflammation characterized by
multifocal bronchoalveolar hyperplasia, alveolar histiocytosis, and focal segmental fibrosis in
rats exposed to CB for up to 20 mo at exposure rates of 510 to 540 mg-lrm"3/week.  The
observed lung pathology reflects notable dose-response relationships and usually evolves in a
similar manner. With increasing dose, there is an increased accumulation and aggregation of
particle-laden AMs, Type II cell hyperplasia, a foamy (degenerative) macrophage response,
alveolar proteinosis, alveolar bronchiolization, cholesterol granulomas, and often squamous cell
carcinomas and bronchioalveolar adenomas derived from metaplastic squamous cells in the areas
of alveolar bronchiolization.
       Heinrich et al.  (1995) compared effects of diesel exposure in rats and mice with exposure
to titanium dioxide or carbon black.  Exposures to TiO2 and carbon black were adjusted during
the exposure to result in a similar lung burden for the three types of particles. At similar lung
burdens in the rat, DPM, TiO2, and CB had nearly identical effects on  lung weights and on the
incidence of lesions, both noncancer and cancer. Also, a similar effect on clearance of a labeled
test aerosol was measured for the different particles. A comparison of the effect of DPM, TiO2,
and carbon black exposures in mice also showed a similar effect on lung weight, but noncancer
effects were not reported and no significant increase in tumors was observed.
       Murphy et al. (1998) compared the toxicological effects of DPM with three  other
particles chosen for their differing morphology and  surface chemistry.  One mg each of well-
characterized crystalline quartz, amorphous silica, CB, and DPM was administered to laboratory
rats by a single intratracheal instillation.  The laboratory rats were sacrificed at 48 h, and 1, 6,
and 12 weeks after instillation. Crystalline quartz produced significant increases in lung
permeability, persistent surface inflammation, progressive increases in pulmonary surfactant and
activities of epithelial  marker enzymes up to 12 wk  after primary exposure.  Amorphous silica
did not cause progressive effects but did produce initial epithelial damage with permeability
changes that regressed with time after exposure. By contrast, CB had  little if any effect on  lung
permeability, epithelial markers, or inflammation.  Similarly, DPM produced only minimal
changes, although the individual particles were smaller and differed in surface chemistry from
CB.  The authors concluded that DPM is less damaging to the respiratory epithelium than is
silicon dioxide, and that the surface chemistry of the particle is more important than ultrafine
size in explaining biological activity.
       These experiments provide strong support for the idea that DE toxicity results from  a
mechanism that is analogous to that of other relatively inert particles in the lung. This
                                          5-92

-------
qualitative similarity exists along with some apparent quantitative differences in the potency of
various particles for producing effects on the lung or on particle clearance.
       The exact relationship between toxicity and particle size within the ultrafine particle
mode, including DPM (BeruBe et al.,  1999), remains unresolved. Studies reviewed in the PM
CD (U.S. EPA, 1996) suggest a greater inherent potential toxicity of inhaled ultrafine particles.
Exposure to ultrafine particles may increase the release of proinflammatory mediators that could
be involved in lung disease. For example, Driscoll and Maurer (1991) compared the effects  of
fine (0.3 |lm) and ultrafine (0.02 |lm) TiO2 particles instilled into the lungs of laboratory rats.
Although both size modes caused an increase in the numbers of AMs and PMNs in the lungs,
and release of TNF and fibronectin by AMs, the responses were greater and more persistent with
the ultrafine particles. While fine particle exposure resulted in a minimally increased
prominence of particle-laden macrophages associated with alveolar ducts, ultrafine particle
exposure produced a somewhat greater prominence of macrophages, some necrosis of
macrophages, and slight interstitial inflammation of the alveolar duct region.  Moreover,
collagen increased only with exposure to ultrafine particles.
       Oberdorster et al. (1992) compared the effects of fine (0.25  |im) and ultrafine (0.02 |im)
TiO2 particles instilled into the lungs of laboratory rats on various indicators of inflammation.
Instillation of ultrafine particles increased the number of total cells recovered by lavage,
decreased the percentage of AMs, and increased the percentage of PMNs and protein.
Instillation with fine particles did not cause statistically significant effects. Thus,  the ultrafine
particles had greater pulmonary inflammatory potency than did larger sizes of this material.  The
investigators attributed the enhanced toxicity to greater interaction of the ultrafine particles with
their large surface area, with alveolar and interstitial macrophages, which resulted in enhanced
release of inflammatory mediators.  They suggested that ultrafine particles of low in vitro
solubility appear to enter the interstitium more readily than do larger sizes of the same material,
which accounted for the increased contact with macrophages in this compartment of the lung.
Driscoll and Maurer (1991) noted that the pulmonary retention of ultrafine TiO2 particles
instilled into rat lungs was greater than for the same mass of fine-mode TiO2 particles. Thus, the
available evidence tends to suggest a potentially greater toxicity for inhaled ultrafine particles.
       Particle size, volume, surface area, and composition may be the critical elements in the
overload phenomenon following exposure to particles, which could explain those  quantitative
differences. The overloaded AMs secrete a variety of cytokines, oxidants, and proteolytic
enzymes that are responsible for inducing particle aggregation and damaging adjacent epithelial
tissue (Oberdorster, 1994). For a more detailed discussion of mechanism, see Chapter 3.
       On the basis of currently available laboratory animal data, the principal noncancerous
health hazard to humans posed by exposure to DE is a structural or functional injury to the lung.
Such effects are demonstrable at dose rates or cumulative doses  of DPM lower than those

                                           5-93

-------
reported to be necessary to induce lung tumors in rats.  An emerging human health issue
concerning short-term exposure to ambient DE/DPM is the potential for allergenic responses in
several studies.  Heightened allergenic responses including increased cytokine production as well
as increased numbers of inflammatory cells have been detected in nasal lavage from humans
exposed to inhaled or instilled DE/DPM.  In individuals already allergic to ragweed, exposure to
DE/DPM with the allergen was observed to result in an enhanced allergenic response,
particularly IgE production. Current knowledge indicates that the carbonaceous core of diesel
particles is the major causative factor in the injury to the lung and that other factors such as the
cytotoxicity of adsorbed substances on the particles also may play a role.  The lung injury
appears to be mediated through effects on pulmonary AMs.  Because noncancerous pulmonary
effects occur at lower doses than tumor induction does in the rat, and because these effects may
be cofactors in the etiology of DE-induced tumors, noncancerous pulmonary effects must be
considered in the total evaluation of DE, notably the particulate component.

5.6.  SUMMARY AND DISCUSSION
5.6.1. Effects of Diesel Exhaust on Humans
       The most readily identified acute noncancer health effect of DE on humans is its ability
to elicit subjective complaints of eye, throat, and bronchial irritation and neurophysiological
symptoms such as headache, lightheadedness, nausea, vomiting, and numbness and tingling of
the extremities.  Studies of the perception and offensiveness of the  odor of DE and a human
volunteer study in an exposure chamber have demonstrated that the time of onset of the human
subjective symptoms is inversely  related to increasing concentrations of DE and the severity is
directly related to increasing concentrations of DE.  In one study in which a diesel  engine was
operated under varying load conditions, a dilution factor of 140 to 475 was needed to reduce the
exhaust level to an odor-detection threshold level.
       A public health issue is whether short-term exposure to DE might result in an acute
decrement in ventilatory function and whether the frequent repetition of such acute respiratory
effects could result in chronic lung function impairment. One  convenient means of studying
acute decrements in ventilatory function is to monitor differences in pulmonary function in
occupationally exposed workers at the beginning and end of a workshift. In studies of
underground miners, bus garage workers, dockworkers, and locomotive repairmen, increases in
respiratory symptoms (cough, phlegm, and dyspnea) and decreases in lung function (FVC,
FEVl3 PEFR, and FEF25.75) over the course of a workshift were generally found to be minimal
and not statistically significant. In a study of acute respiratory responses in diesel bus garage
workers, there was an increased reporting of cough, labored breathing, chest tightness, and
wheezing, but no reductions in pulmonary function were associated with exposure to DE.
Pulmonary function was affected in stevedores over a workshift exposure to DE but normalized

                                          5-94

-------
after a few days without exposure to DE fumes. In a third study, there was a trend toward
greater ventilatory function changes during a workshift among coal miners, but the decrements
were similar in miners exposed and not exposed to DE.
       Smokers appeared to demonstrate larger workshift respiratory function decrements and
increased incidence of respiratory symptoms.  Acute sensory and respiratory  symptoms were
earlier and more sensitive indicators of potential health risks from diesel exposure than were
decrements in pulmonary function.  Studies on the acute health effects of exposure to DE in
humans, experimental and epidemiologic, have failed to demonstrate a consistent pattern of
adverse effects on respiratory morbidity; the majority of studies offer, at best, equivocal
evidence for an exposure-response relationship. The environmental contaminants have
frequently been below permissible workplace exposure limits; in those few cases where health
effects have been reported, the authors have failed to identify conclusively the individual or
collective  causative agents in the DE.
       Chronic effects of DE exposure have been evaluated in epidemiologic studies of
occupationally exposed workers (metal and nonmetal miners, railroad yard workers, stevedores,
and bus garage mechanics). Most of the epidemiologic data indicate an absence of an excess
risk of chronic respiratory disease associated with exposure to DE. In a few studies, a higher
prevalence of respiratory symptoms, primarily cough, phlegm, or chronic bronchitis, was
observed among the exposed.  These increased symptoms, however, were usually not
accompanied by significant changes in pulmonary function.  Reductions in FEVj and FVC and,
to a lesser extent, FEF50 and FEF75, also have been reported. Two studies detected statistically
significant decrements in baseline pulmonary function consistent with obstructive airway
disease. One study of stevedores had a limited sample size of 17 exposed and 11 controls. The
second study in coal miners showed that both underground and surface workers at diesel-use
mines had somewhat lower pulmonary performance than their matched controls. The proportion
of workers in or at diesel-use mines, however, showed equivalent evidence of obstructive airway
disease, and for this reason the authors of the second paper felt that factors other than diesel
exposure might have been responsible. A doubling of the prevalence of minor restrictive airway
disease was also observed in workers in or at diesel-use mines. These two studies, coupled with
other reported nonsignificant trends in respiratory flow-volume measurements, suggest that
exposure to DE may impair pulmonary function among occupational populations.
Epidemiologic studies of the effects of DE on organ systems other than the pulmonary system
are scant.  Whereas a preliminary study of the association of cardiovascular mortality and
exposure to DE found a fourfold higher risk ratio, a more comprehensive epidemiologic study
by the same investigators found no significant difference between the observed and expected
number of deaths caused by cardiovascular disease.
                                          5-95

-------
       Caution is warranted in the interpretation of results from the epidemiologic studies that
have addressed noncarcinogenic health effects from exposure to DE. These investigations suffer
from myriad methodological problems, including (1) incomplete information on the extent of
exposure to DE, necessitating in some studies estimations of exposures from job titles and
resultant misclassification; (2) the presence of confounding variables such as smoking or
occupational exposures to other toxic substances (e.g., mine dusts); and (3) the short duration
and low intensity of exposures. These limitations restrict drawing definitive conclusions as to
the cause of any noncarcinogenic DE effect, observed or reported.
       It is also apparent that at some level of exposure DE as measured by DPM appears to
have the potential to induce airway inflammation in humans without disease. Also, in one other
study peripheral blood changes were noted.  An emerging area of concern is the immunological
changes that have been documented in response to DE exposure and the potential relationship of
these changes to the explosive growth of asthma in human populations.

5.6.2. Effects of Diesel Exhaust on Laboratory Animals
       Laboratory animal studies of the toxic effects of DE have involved acute, subchronic,
and chronic exposure regimens.  In acute exposure  studies, toxic effects appear to have been
associated primarily with high concentrations of carbon monoxide, nitrogen  dioxide, and
aliphatic aldehydes.  In short- and long-term studies, toxic effects have been associated with
exposure to the complex exhaust mixture.  Effects of DE in various animal species are
summarized in Tables 5-2 to 5-15. In short-term studies, health effects related to function, when
found, are mild and result from extremely  high DPM concentrations of about 6 mg/m3 and
extensive durations of exposure approximating 20 h/day.  There is ample evidence, however,
that other pathophysiological effects such as accumulation of DPM in pulmonary tissues,
evidence of inflammatory response, AM aggregation and accumulation near the terminal
bronchioles, Type II cell proliferation, and the thickening of alveolar walls adjacent to AM
aggregation do occur under short-term exposures at lower levels of DE. Little evidence exists,
however, from short-term studies that exposure to DE impairs lung function. Chronic exposures
cause lung pathology that results in altered pulmonary function and increased DPM retention in
the lung. Exposures to DE have also been associated with increased susceptibility to respiratory
tract infection, neurological or behavioral changes, an increase in banded neutrophils, and
morphological alterations in the liver.

5.6.2.1. Effects on Survival and Growth
       The data presented in Table 5-3 show limited effects on survival in mice and rats and
some evidence of reduced body weight in rats following chronic exposures to concentrations of
1.5 mg/m3 DPM or higher and exposure durations of 16 to 20 h/day, 5 days/week for 104 to

                                          5-96

-------
130 weeks. Increased lung weights and lung to body-weight ratios in rats, mice, and hamsters;
an increased heart to body weight ratio in rats;  and decreased lung and kidney weights in cats
have been reported following chronic exposure to DE. No evidence was found of an effect of
DE on other body organs (Table 5-4). The lowest-observed-effect level in rats approximated 1
to 2 mg/m3 DPM for 7 h/day, 5 days/week for 104 weeks.

5.6.2.2.  Effects on Pulmonary Function
      Pulmonary function impairment has been reported in rats, hamsters, cats, and monkeys
exposed to DE and included lung mechanical properties (compliance and resistance), diffusing
capacity, lung volumes, and ventilatory performance (Table 5-5).  The effects generally
appeared only after prolonged exposures. The  lowest exposure levels (expressed in terms of
DPM concentrations) that resulted in impairment of pulmonary function occurred at 2 mg/m3 in
cynomolgus monkeys (the only level tested), 1.5 and 3.5 mg/m3 in rats, 4.24 and 6 mg/m3 in
hamsters, and 11.7 mg/m3 in cats.  Exposures in monkeys, cats, and rats (3.5 mg/m3) were for
7 to 8 h/day, 5 days/week for 104 to 130 weeks. While this duration is  considered to constitute a
lifetime study in rodents, it is a small part of the lifetime of a monkey or cat.  Exposures in
hamsters and rats (1.5 mg/m3) varied in hours per day (8 to 20) and weeks of exposure (26 to
130).  In all species but the monkey, the testing results were consistent with restrictive lung
disease; alteration in expiratory flow rates indicated that 1.5 mg/m3 DPM was a LOAEL for a
chronic exposure (Gross, 1981). Monkeys demonstrated evidence of obstructive airway disease.
The nature of the pulmonary impairment is dependent on the dose of toxicants delivered to  and
retained in the lung, the site of deposition and effective clearance or repair, and the anatomy and
physiology of the  affected species; these variables appear to be factors in the disparity of the
airway disease in monkey versus the other species tested.

5.6.2.3.  Histopathological andHistochemical Effects
      Histological studies have demonstrated  that chronic exposure to DE can result in effects
on respiratory tract tissue (Table 5-6).  Typical findings include alveolar histiocytosis, AM
aggregation, tissue inflammation, increase in PMNs, hyperplasia of bronchiolar and alveolar
Type II cells, thickened alveolar septa,  edema,  fibrosis, and emphysema. Lesions in the trachea
and bronchi were observed in some studies. Associated with these histopathological findings
were various histochemical changes in  the lung, including increases in lung DNA, total protein,
alkaline and acid phosphatase, glucose-6-phosphate dehydrogenase; increased synthesis of
collagen; and release  of inflammatory mediators such as leukotriene LTB and prostaglandin
PGF2ct. Although  the overall laboratory evidence is that prolonged exposure to DPM results in
histopathological and histochemical changes in the lungs of exposed animals, some studies have
also demonstrated that there may be a threshold of exposure to DPM below which pathologic

                                         5-97

-------
changes do not occur.  These no-observed-adverse-effect levels for histopathological effects
were reported to be 2 mg/m3 for cynomolgus monkeys (the only concentration tested), 0.11 to
0.35 mg/m3 for rats, and 0.25 mg/m3 DPM for guinea pigs exposed for 7 to 20 h/day, 5 to
5.5 days/week for 104 to 130 weeks.

5.6.2.4. Effects on Airway Clearance
       The pathological effects of DPM appear to be strongly dependent on the relative rates of
pulmonary deposition and clearance (Table 5-7). Clearance of particles from the alveolar region
of the lungs is  a multiphasic process involving phagocytosis by AMs.  Chronic exposure to DPM
concentrations of about 1 mg/m3 or above, under varying exposure durations, causes pulmonary
clearance to be reduced, with concomitant focal aggregations of particle-laden AMs, particularly
in the peribronchiolar  and alveolar regions, as well as in the hilar and mediastinal lymph nodes.
The exposure concentration at which focal aggregates of particle-laden AMs occur may vary
from species to species, depending on rate of uptake and pulmonary deposition, pulmonary
clearance rates, the relative size of the AM population per unit of lung tissue, the rate of
recruitment of AMs and leukocytes, and the relative efficiencies for removal of particles by the
mucociliary and lymphatic transport system. The principal means by which PM clearance is
reduced is through a decrease in the function of pulmonary AMs. Impairment of particle
clearance seems to be  nonspecific and applies primarily to dusts that are persistently retained in
the lungs.  Lung dust levels of approximately 0.1 to 1 mg/g lung tissue appear to produce this
effect in the Fischer 344 rat (Health Effects Institute, 1995). Morrow (1988) suggested that the
inability of particle-laden AMs to translocate to the mucociliary escalator is correlated to an
average composite particle volume per AM in the lung. When this particle volume exceeds
approximately 60 |lm3 per AM in the Fischer 344 rat, impairment of clearance appears to be
initiated. When the particulate volume exceeds approximately 600 |im3 per cell, evidence
suggests that AM-mediated particulate clearance virtually ceases, agglomerated particle-laden
macrophages remain in the alveolar region, and increasingly nonphagocytized dust particles
translocate to the pulmonary interstitium. Data for other laboratory animal species and humans
are, unfortunately, limited.

5.6.2.5. Neurological and Behavioral Effects
       Behavioral effects have been observed in rats exposed to DE from birth to 28 days of age
(Table 5-14).  Exposure caused a decreased level of spontaneous locomotor activity and a
detrimental effect on learning in adulthood.  In agreement with the behavioral changes was
physiological evidence for delayed neuronal maturation. Exposures were to 6 mg/m3 DPM for 8
h/day, 7 days/week from birth to about 7, 14, 21, or 28 days of age.
                                          5-98

-------
5.6.2.6. Effects on Immunity and Allergenicity
       Several laboratory animal studies have indicated that exposure to DPM can reduce an
animal's resistance to respiratory infection. This effect, which can occur even after only 2 or 6 h
of exposure to DE containing 5 to 8 mg/m3 DPM, does not appear to be caused by direct
impairment of the lymphoid or splenic immune  systems; however, in one study of influenza
virus infection, interferon levels and hemaglutinin antibody levels were adversely affected in the
exposed mice.
       As with humans, there are animal data suggesting that DPM is a possible factor in the
increasing incidence of allergic hypersensitivity. The effects have been demonstrated primarily
in acute human and laboratory animal studies and appear to be associated with both the
nonextractable carbon core and the organic fraction of DPM. It also appears that synergies with
DPM may increase the potency of known airborne allergens. Both animal  and human cell
culture studies indicate that DPM also has  the potential to act as an adjuvant.

5.6.2.7. Other Noncancer Effects
       Essentially no effects (based on the weight of evidence of a number of studies) were
noted for reproductive and teratogenic effects in mice, rats, rabbits, and monkeys; clinical
chemistry and hematology in the rat, cat, hamster, and monkeys; and enzyme induction in the rat
and mouse (Tables 5-11 through 5-13 and  5-15).

5.6.3. Comparison of Filtered and Unfiltered Diesel Exhaust
       The comparison of the toxic responses in laboratory animals exposed to whole DE or
filtered exhaust containing no particles demonstrates across laboratories that diesel particles are
the principal etiologic agent of noncancerous health effects in laboratory animals exposed to DE
(Table 5-16). Whether the particles act additively or synergistically with the gases cannot be
determined from the designs of the studies. Under equivalent exposure regimens, hamsters have
lower levels of retained DPM in their lungs than rats and mice do and consequently less
pulmonary function impairment and pulmonary pathology. These differences may result from a
lower intake rate of DPM, lower deposition rate and/or more rapid clearance rate, or lung tissue
that is less susceptible to the cytotoxicity of DPM. Observations of a decreased respiration in
hamsters when exposed by inhalation favor lower intake and deposition rates.

5.6.4. Interactive Effects of Diesel Exhaust
       There is no direct evidence that DE interacts with other substances  in an exposure
environment, other than an impaired resistance to respiratory tract infections.  Young animals
were not more susceptible. In several ways, animals with laboratory-induced emphysema were
more resistant. There is experimental evidence that both inorganic and organic compounds can

                                          5-99

-------
be adsorbed onto carbonaceous particles. When such substances become affiliated with
particles, these substances can be carried deeper into the lungs where they might have a more
direct and potent effect on epithelial cells or on AM ingesting the particles.  Few specific studies
to test interactive effects of DE with atmospheric contaminants, other than coal dust, have been
conducted. Coal dust and DPM had an additive effect only.

5.6.5. Conclusions
       Conclusions concerning the principal human hazard from exposure to DE are as follows:

       •      Allergenic inflammatory disorders of the airways to responses typical of asthma
             have been demonstrated under short-term exposure scenarios to either DE or
             DPM.  The evidence indicates that the immunological changes appear to be due
             to the DPM component of DE and that the immunological changes are caused by
             both the nonextractable carbon core and the adsorbed organic fraction of the
             diesel particle.  The toxicological significance of these effects has yet to be
             resolved.
             Some occupational studies of acute exposure to DE during work shifts suggest
             that increased acute sensory and respiratory symptoms (cough, phlegm, chest
             tightness, wheezing) are more sensitive indicators of possible health risks from
             exposure to DE than pulmonary function decrements (which were consistently
             found not to be significantly associated with DE exposure)
       •      Noncancer effects in humans from long-term chronic exposure to DPM are not
             evident.  Noncancer effects from long-term exposure to DPM of several
             laboratory animal species, conducted to assess the pathophysiologic effects of
             DPM in humans showed pulmonary histopathology (principally fibrosis) and
             chronic inflammation.

       Although the mode of action of DE is not clearly evident for any of the effects
documented in this chapter, the respiratory tract effects observed under acute scenarios are
suggestive of an irritant mechanism, while lung effects observed in chronic scenarios indicate an
underlying inflammatory response. Current knowledge indicates that the carbonaceous core of
the diesel particle is the causative agent of the lung effects, with the extent of the injury being
mediated at least in part by a progressive impairment of AMs.  It is noted that lung effects occur
in response to DE exposure in several species and occur in rats at doses lower than those
inducing particle overload and a tumorigenic response (see above); it follows that lung effects
such as inflammation and fibrosis are relevant in the development of risk assessments for DE.
                                         5-100

-------
                                REFERENCES FOR CHAPTER 5

Abe, S; Takizawa, H; Sugawara, I; et al. (2000) Diesel exhaust (DE)-induced cytokine expression in human
bronchial epithelial cells. Am J Respir Cell Mol Biol 22:296-303.

Ames, RG; Attfield, MD; Hankinson, JL; et al. (1982) Acute respiratory effects of exposure to diesel emissions in
coal miners. Am Rev Respir Dis 125:39-42.

Ames, RG; Reger, RB; Hall, DS. (1984) Chronic respiratory effects of exposure to diesel emissions in coal mines.
Arch Environ Health 39:389-394.

Attfield, MD. (1978) The effect of exposure to silica and diesel exhaust in underground metal and nonmetal
miners. In:  Industrial hygiene for mining and tunneling: proceedings of a topical symposium; November; Denver,
CO. Kelley, WD, ed. Cincinnati, OH: The American Conference of Governmental Industrial Hygienists, Inc.; pp.
129-135.

Attfield, MD; Trabant, GD; Wheeler, RW. (1982) Exposure to diesel fumes and dust at six potash mines. Ann
OccupHyg 26:817-831.

Barnhart, MI; Chen, S-T; Salley, SO; et al. (1981) Ultrastructure and morphometry of the alveolar lung of guinea
pigs chronically exposed to diesel engine exhaust: six month's experience. J Appl Toxicol 1:88-103.

Barnhart, MI; Salley, SO; Chen,  S-T; et al. (1982) Morphometric ultrastructural analysis of alveolar lungs of
guinea pigs chronically exposed by inhalation to diesel exhaust (DE). In:  Toxicological effects of emissions from
diesel engines: proceedings of the Environmental Protection Agency diesel emissions symposium; October, 1981;
Raleigh, NC. Lewtas, J., ed. New York: ElsevierBiomedical; pp. 183-200. (Developments in toxicology and
environmental science: v. 10).

Battigelli, MC. (1965) Effects of diesel exhaust. Arch Environ Health 10:165-167.

Battigelli, MC; Mannella, RJ; Hatch, TF. (1964) Environmental and clinical investigation of workmen exposed to
diesel exhaust in railroad engine houses. Ind Med Surg 33:121-124.

Battigelli, MC; Hengstenberg, F; Mannela, RJ; et al. (1966) Mucociliary  activity. Arch Environ Health
12:460-466.

Bayram, H; Devalia, JL;  Sapsford, RJ; et al. (1998a) The effect of diesel  exhaust particles on cell function and
release of inflammatory mediators from human bronchial epithelial cells  in vitro. Am J Respir Cell Mol Biol
18:441-448.

Bayram, H; Devalia, JL; Khair, O; et al. (1998b) Comparison of ciliary activity and inflammatory mediator
release from bronchial epithelial cells of nonatopic nonasthmatic subjects and atopic asthmatic patients and the
effect of diesel exhaust particles in vitro. J Allergy Clin Immunol 102(5): 771-782.

BeruBe, KA; Jones, TP; Williamson, BJ; et al. (1999) Physicochemical characterisation of diesel exhaust
particles: factors for assessing biological activity. Atmos Environ 33:1599-1614.

Bhatnagar,  RS; Hussain, MZ; Sorensen, KR; et al. (1980) Biochemical alterations in lung connective tissue in rats
and mice exposed to diesel emissions. In: Health effects of diesel engine emissions: proceedings of an
international symposium, v. 1; December 1979; Cincinnati, OH. Pepelko, WE; Danner, RM; Clarke, NA, eds.
Cincinnati,
H: U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp. 557-570; EPA report no.
EPA/600/9-80/057a. Available from: NTIS, Springfield, VA; PB81-173809.
                                                5-101

-------
Bice, DE; Mauderly, JL; Jones, RK; et al. (1985) Effects of inhaled diesel exhaust on immune responses after lung
immunization. Fundam Appl Toxicol 5:1075-1086.

Blomberg, A; Sainsbury, C; Rudell, B; et al. (1998) Nasal cavity lining fluid ascorbic acid concentration increases
in healthy human volunteers following short term exposure to diesel exhaust. Free Radical Res 28:59-67.

Boland, S; Baeza-Squiban, A; Fournier, T; et al. (1999) Diesel exhaust particles are taken up by human airway
epithelial cells in vitro and alter cytokine production. Am J Physiol 276:L604-L613.

Bolton, RE; Vincent, JH; Jones, AD; et al. (1983) An overload hypothesis for pulmonary clearance of UICC
amosite fibres inhaled by rats. Br J Ind Med 40:264-272.

Boren, HG. (1964) Carbon as a carrier mechanism for irritant gases. Arch Environ Health 8:119-124.

Brightwell, J; Fouillet, X; Cassano-Zoppi, AL; et al. (1986) Neoplastic and functional changes in rodents after
chronic inhalation of engine exhaust emissions. In: Carcinogenic and mutagenic effects of diesel engine exhaust:
proceedings of the international satellite symposium on toxicological effects of emissions from diesel engines;
July; Tsukuba Science City, Japan. Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. Amsterdam, Holland:
Elsevier Science Publishers B. V.; pp. 471-485. (Developments in toxicology and environmental science: v. 13).

Brunekreef, B; Janssen, NA; de Hartog, J.; et al. (1997) Air pollution from truck traffic and lung function in
children living near motorways. Epidemiology 8(3):298-303.

Brunekreef, B; Janssen, NA; van Vliet, P; et al. (2000) Traffic related air pollution and its effect on respiratory
health of children living near motorways. Presented at: PM 2000: Paniculate matter and health-the scientific
basis for regulatory decision-making.  Sponsored by the Air and Waste Management Association, January 24-28,
2000.

Campbell, KI; George, EL; Washington, IS, Jr. (1980) Enhanced susceptibility  to infection in mice after exposure
to dilute exhaust from light duty diesel engines. In: Health effects of diesel engine emissions: proceedings of an
international symposium, v. 2; December 1979; Cincinnati, OH. Pepelko, WE; Danner, RM; Clarke, NA, eds.
Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp. 772-785;  EPA
report no. EPA/600/9-80/057b. Available from: NTIS,  Springfield, VA; PB81-173817.

Campbell, KI; George, EL; Washington, IS, Jr. (1981) Enhanced susceptibility  to infection in mice after exposure
to dilute exhaust from light duty diesel engines. Environ Int 5:377-382.

Castranova, V; Bowman, L; Reasor, MJ; et al. (1985) The response of rat alveolar macrophages to chronic
inhalation of coal dust and/or diesel exhaust. Environ Res 36:405-419.

Chan, TL; Lee, PS; Hering, WE. (1981) Deposition and clearance of inhaled diesel exhaust particles in the
respiratory tract of Fischer rats. J Appl Toxicol 1:77-82.

Chaudhari, A; Dutta, S. (1982) Alterations in tissue glutathione and angiotensin converting enzyme due to
inhalation of diesel engine exhaust. J Toxicol Environ Health 9:327-337.

Chaudhari, A; Farrer, RG; Dutta, S. (1980) Effect of exposure to diesel exhaust on pulmonary prostaglandin
dehydrogenase (PGDH) activity. In: Health effects of diesel engine emissions: proceedings of an international
symposium, v. 1; December 1979; Cincinnati, OH. Pepelko, WE; Danner, RM; Clarke, NA, eds. Cincinnati, OH:
U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp. 532-537; EPA report no.
EPA/600/9-80-057a. Available from: NTIS, Springfield, VA; PBS 1-173809.

Chaudhari, A; Farrer, RG; Dutta, S. (1981) Effect of exposure to diesel exhaust on pulmonary prostaglandin
dehydrogenase (PGDH) activity. J Appl Toxicol 1:132-134.
                                                 5-102

-------
Chen, KC. (1986) Induction of aryl hydrocarbon hydroxylase in rat tissue following intratracheal instillation of
diesel paniculate extract and benzo[a]pyrene. J Appl Toxicol 6:259-262.

Chen, KC; Vostal, JJ. (1981) Aryl hydrocarbon hydroxylase activity induced by injected diesel paniculate extract
vs inhalation of diluted diesel exhaust. J Appl Toxicol 1:127-131.

Chen, S; Weller, MA; Barnhart, MI. (1980) Effects of diesel engine exhaust on pulmonary alveolar macrophages.
Scanning Electron Microsc 3:327-338.

Devalia, JL; Bayram, H; Abdelaziz, MM; et al. (1999) Differences between cytokine release from bronchial
epithelial cells of asthmatic patients and non-asthmatic subjects: effect of exposure to diesel exhaust particles. Int
Arch Allergy Immunol 118:437-439.

Diaz-Sanchez, D;  Dotson, AR; Takenaka, H; et al. (1994) Diesel exhaust particles induce local IgE production in
vivo and alter the pattern of IgE messenger RNA isoforms. J Clin Invest 94(4): 1417-1425.

Diaz-Sanchez, D;  Tsien, A; Casillas, A; et al. (1996) Enhanced nasal cytokine production in human beings after in
vivo challenge with diesel exhaust particles. J Allergy Clin Immunol 98:114-123.

Diaz-Sanchez, D;  Tsien, A; Fleming, J; et al. (1997) Combined diesel exhaust paniculate and ragweed allergen
challenge markedly enhances human in vivo nasal ragweed-specific IgE and skews cytokine production to a T
helper cell 2-type pattern. J Immunol 158:2406-2413.

Diaz-Sanchez, D;  Garcia, MP; Wang, M; et al. (1999) Nasal challenge with diesel exhaust particles can induce
sensitization to a neoallergen in the human mucosa.  J Allergy Clin Immunol 104(6): 1183-1188.

Driscoll, KE; Maurer, JK. (1991) Cytokine and growth factor release by alveolar macrophages: potential
biomarkers of pulmonary toxicity. Toxicol Pathol 19:398-405.

Dungworth, DL; Mohr, U; Heinrich, U;  et al. (1994) Pathologic effects of inhaled particles in rat lungs:
associations between inflammatory and neoplastic processes. In: Toxic and carcinogenic effects of solid particles
in the respiratory tract: [proceedings of the 4th international inhalation symposium]; March 1993; Hannover,
Germany. Mohr, U; Dungworth, DL; Mauderly, JL, et al., eds. Washington, DC: International Life  Sciences
Institute Press; pp. 75-98.

Dziedzic, D. (1981) Differential counts of B and T lymphocytes in the lymph nodes, circulating blood and spleen
after inhalation of high concentrations of diesel exhaust. J Appl Toxicol 1:111-115.

Edling,  C; Axelson, O. (1984) Risk factors of coronary heart disease among personnel in a bus company. Int Arch
Occup Environ Health 54:181-183.

Edling,  C; Anjou,  CG; Axelson, O; et al. (1987) Mortality among personnel exposed to diesel exhaust. Int Arch
Occup Environ Health 59:559-565.

El Batawi, MA; Noweir, MH.  (1966) Health problems resulting from prolonged exposure to air pollution in diesel
bus garages. Ind Health 4:1-10.

Fedan, JS; Frazier, DG; Moorman, WJ; et al. (1985) Effects of a two-year inhalation exposure of rats to coal dust
and/or diesel exhaust on tension responses  of isolated airway smooth muscle. Am Rev Respir Dis 131:651-655.

Fujimaki, H; Nohara, O; Ichinose, T; et al. (1994) IL-4 production in mediastinal lymph node cells  in mice
intratracheally instilled with diesel exhaust particulates and antigen. Toxicology 92(l-3):261-268.

Fujimaki, H; Saneyoshi,  K; Shiraishi, F; et al. (1997) Inhalation of diesel exhaust enhances antigen-specific IgE
antibody production in mice. Toxicology 116:227-233.


                                                5-103

-------
Gamble, JF; Jones, WG. (1983) Respiratory effects of diesel exhaust in salt miners. Am Rev Respir Dis
128:389-394.

Gamble, J; Jones, W; Hudak, J; et al. (1979) Acute changes in pulmonary function in salt miners. Presented at:
Industrial hygiene for mining and tunneling: proceedings of a topical symposium; November 1978; Denver, CO.
Cincinnati, OH: American Conference of Governmental Industrial Hygienists, Inc.; pp. 119-128.

Gamble, J; Jones, W; Hudak, J. (1983) An epidemiological study of salt miners in diesel and nondiesel mines. Am
JIndMed 4:435-458.

Gamble, J; Jones, W; Minshall, S. (1987a) Epidemiological-environmental study of diesel bus garage workers:
acute effects of NO2 and respirable paniculate on the respiratory system.  Environ Res 42:201-214.

Gamble, J; Jones, W; Minshall, S. (1987b) Epidemiological-environmental study of diesel bus garage workers:
chronic effects of diesel exhaust on the respiratory system. Environ Res 44:6-17.

Green, FHY; Boyd, RL; Danner-Rabovsky, J; et al. (1983) Inhalation studies of diesel exhaust and coal dust in
rats. Scand J Work Environ Health 9:181-188.

Griffis, LC; Wolff, RK; Henderson, RF; et al. (1983) Clearance of diesel soot particles from rat lung after a
subchronic diesel exhaust exposure. Fundam Appl Toxicol 3:99-103.

Gross, KB. (1981) Pulmonary function testing of animals chronically exposed to diluted diesel exhaust. J Appl
Toxicol 1:116-123.

Hahon, N; Booth, JA; Green, F; et al. (1985) Influenza virus infection in mice after exposure to coal dust and
diesel engine emissions. Environ Res 37:44-60.

Hare, CT; Springer, KJ. (1971) Public response to diesel engine exhaust odors [final report]. San Antonio, TX:
Southwest Research Institute; report no.  AR-804. Available from: NTIS,  Springfield, VA; PB-204012.

Hare, CT; Springer, KJ; Somers, JH; et al. (1974) Public opinion of diesel odor. In: Automotive engineering
congress; February-March; Detroit, MI. New York: Society of Automotive Engineers; SAE technical paper no.
740214.

Hashimoto, S; Gon, Y; Takeshita, I; et al. (2000) Diesel exhaust particles activate p38 MAP kinase to produce
interleukin 8 and RANTES by human bronchial epithelial cells and N-acetylcysteine attenuates p38 MAP kinase
activation [In Process Citation]. Am J Respir Crit Care Med  161(l):280-285.

Hatch, GE; Boykin, E; Graham, JA; et al. (1985) Inhalable particles and pulmonary host defense: in vivo and in
vitro effects of ambient air and combustion particles. Environ Res 36:67-80.

Health Effects Institute  (HEI). (1995) Diesel exhaust: a critical analysis of emissions, exposure, and health effects:
a special report of the Institute's Diesel Working Group. Cambridge, MA: Health Effects Institute.

Heinrich, U; Peters, L; Funcke, W; et al. (1982) Investigation of toxic and carcinogenic effects of diesel exhaust
in long-term inhalation exposure of rodents. In:Toxicological effects of emissions from diesel engines:
proceedings of the Environmental Protection Agency diesel emissions symposium; October 1981; Raleigh, NC.
Lewtas, J., ed. New York: Elsevier Biomedical; pp. 225-242. (Developments in toxicology and environmental
science: v. 10).

Heinrich, U; Muhle, H; Takenaka,  S; et al. (1986a) Chronic effects on the respiratory tract of hamsters, mice, and
rats after long-term inhalation of high concentrations of filtered and unfiltered diesel engine emissions. J Appl
Toxicol 6:383-395.
                                                5-104

-------
Heinrich, U; Pott, F; Rittinghausen, S. (1986b) Comparison of chronic inhalation effects in rodents after long-term
exposure to either coal oven flue gas mixed with pyrolized pitch or diesel engine exhaust. In: Carcinogenic and
mutagenic effects of diesel engine exhaust: proceedings of the international satellite syposium on toxicological
effects of emissions from diesel engines; July; Tsukuba Science City, Japan. Ishinishi, N; Koizumi, A; McClellan,
RO; et al., eds. Amsterdam, Holland: Elsevier Science Publishers BV; pp. 441-457. (Developments in toxicology
and environmental science: v. 13).

Heinrich, U; Fuhst, R; Rittinghausen, S; et al. (1995) Chronic inhalation exposure of Wistar rats and two different
strains of mice to diesel engine exhaust, carbon black, and titanium dioxide. Inhal Toxicol 7:533-556.

Henderson, RF; Pickrell, JA; Jones, RK; et al. (1988a) Response of rodents to inhaled diluted diesel exhaust:
biochemical and cytological changes in bronchoalveolar lavage fluid and in lung tissue. Fundam Appl Toxicol
11:546-567.

Henderson, RF; Leung, HW; Harmsen, AG; et al. (1988b) Species differences in release of arachidonate
metabolites in response to inhaled diluted diesel exhaust. Toxicol Lett 42(3):325-32.

Hyde, DM; Plopper, CG; Weir, AJ; et al. (1985) Peribronchiolar fibrosis in lungs of cats chronically exposed to
diesel exhaust. Lab Invest 52:195-206.

Ikeda, M; Suzuki, M; Watarai, K; et al. (1995) Impairment of endothelium-dependent relaxation by diesel exhaust
particles in rat thoracic aorta. Jpn J Pharmacol 68:183-189.

Ikeda, M; Watarai, K; Suzuki, M; et al. (1998) Mechanism of pathophysiological effects of diesel exhaust
particles on endothelial cells. Environ Toxicol Pharmacol 6:117-123.

International Life Sciences Institute (ILSI). (2000) ILSI Risk Science Institute workshop: The relevance of the rat
lung response to particle overload for human risk assessment. Gardner, DE, ed.  Inhal Toxicol: 12(1-2): 1-17.

Ishinishi, N; Kuwabara, N; Nagase, S; et al. (1986) Long-term inhalation studies on effects of exhaust from heavy
and light duty diesel engines onF344 rats. In: Carcinogenic and mutagenic effects of diesel engine exhaust:
proceedings of the international satellite symposium on toxicological effects of emissions from diesel engines;
July; Tsukuba Science City, Japan. Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. Amsterdam, Holland:
Elsevier Science Publishers B. V.; pp. 329-348. (Developments in toxicology and environmental science: v. 13).

Ishinishi, N; Kuwabara, N; Takaki, Y; et al. (1988) Long-term inhalation experiments onDE. In: DE and health
risks: results of the HERP studies. Tsukuba, Ibaraki, Japan: Japan Automobile Research Institute, Inc., Research
Committee for HERP Studies; pp. 11-84.

Iwai, K; Udagawa, T; Yamagishi, M; et al. (1986) Long-term inhalation studies of diesel exhaust on F344 SPF
rats. Incidence of lung cancer and lymphoma. In: Carcinogenic and mutagenic effects  of diesel engine exhaust:
proceedings of the international satellite symposium on toxicological effects of emissions from diesel engines;
July; Tsukuba Science City, Japan. Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. Amsterdam, Holland:
Elsevier Science Publishers B. V.; pp. 349-360. (Developments in toxicology and environmental science: v. 13).

Jorgensen, H; Svensson, A. (1970) Studies on pulmonary function and respiratory tract symptoms of workers in an
iron ore mine where diesel trucks are used underground. J Occup Med 12:348-354.

Kahn, G; Orris, P; Weeks, J. (1988) Acute overexposure to diesel exhaust: report of 13 cases. Am J Ind Med
13:405-406.

Kaplan, HL; MacKenzie, WF; Springer, KJ; et al. (1982) A subchronic study of the effects of exposure of three
species of rodents to diesel exhaust. In: Toxicological effects of emissions from diesel engines: proceedings of the
Environmental Protection Agency diesel emission symposium; October, 1981; Raleigh, NC. Lewtas, J., ed. New
York: Elsevier Biomedical; pp. 161-182. (Developments in toxicology and environmental science: v. 10).


                                                 5-105

-------
Karagianes, MT; Palmer, RF; Busch, RH. (1981) Effects of inhaled diesel emissions and coal dust in rats. Am Ind
HygAssoc 142:382-391.

Katz, M; Rennie, RP; Jegier, Z. (1960) Air pollution and associated health hazards from diesel locomotive traffic
in a railroad tunnel. Occup Health Rev 11:2-15.

Kilburn, KH; McKenzie, WN. (1978) Leukocyte recruitment to airways by aldehyde-carbon combinations that
mimic cigarette smoke. Lab Invest 38:134-142.

Kilburn, KH. (2000) Effects of diesel exhaust on neurobehavioral and pulmonary functions.  Arch Environ Health
55:11-17.

Klosterkotter, W; Btinemann, G. (1961) Animal experiments on the elimination of inhaled dust. In: Inhaled
particles and vapours: proceedings of an international symposium; March-April 1960; Oxford, United Kingdom.
Davies, C. N., ed. New York: Pergamon Press; pp. 327-341.

Knox, RB; Suphioglu, C; Taylor, P; et al.  (1997) Major grass pollen allergen Lol p 1 binds to diesel exhaust
particles: implications for asthma and air pollution. Clin Exp Allergy 27:246-251.

Kobayashi, T; Ito, T. (1995) Diesel exhaust particulates induce nasal mucosal hyperresponsiveness to inhaled
histamine aerosol. Fundam Appl Toxicol 27:195-202.

Kobayashi, T; Ikeue, T; Ito, T; et al. (1997) Short-term exposure to diesel exhaust induces nasal mucosal
hyperresponsiveness to histamine in guinea pigs. Fundam Appl Toxicol 38:166-172.

Laurie, RD; Boyes, WK. (1980) Neurophysiological alterations due to diesel exhaust exposure during the neonatal
life of the rat. In: Health effects of diesel engine emissions: proceedings of an international symposium, v. 2;
December 1979; Cincinnati, OH. Pepelko, WE; Danner, RM; Clarke, NA, eds.  Cincinnati, OH: U.S.
Environmental Protection Agency, Health Effects Research Laboratory; pp. 713-727; EPA report no.
EPA-600/9-80/057b. Available from: NTIS, Springfield, VA; PB81-173817.

Laurie, RD; Boyes, WK. (1981) Neurophysiological alterations due to diesel exhaust exposure during the neonatal
life of the rat. Environ Int 5:363-368.

Laurie, RD; Lewkowski, JP; Cooper, GP; et al. (1978) Effects of diesel exhaust on behavior of the rat. Presented
at: 71st annual meeting of the Air Pollution Control Asscociation; June; Houston, TX. Pittsburgh, PA: Air
Pollution Control Association.

Laurie, RD; Boyes, WK; Wessendarp, T.  (1980) Behavioral alterations due to diesel exhaust exposure. In: Health
effects of diesel engine emissions: proceedings of an international symposium,  v. 2; December 1979; Cincinnati,
OH. Pepelko, WE; Danner, RM; Clarke, NA, eds. Cincinnati, OH: U.S. Environmental Protection Agency, Health
Effects Research Laboratory;  pp. 698-712; EPA report no. EPA-600/9-80-057b. Available from: NTIS,
Springfield, VA; PB81-173817.

Lee, IP; Suzuki, K; Lee, SD; et al. (1980) Aryl hydrocarbon hydroxylase  induction in rat lung, liver, and male
reproductive organs following inhalation exposure to diesel emission. Toxicol Appl Pharmacol 52:181-184.

Lee, KP; Trochimowicz, HJ; Reinhardt, CF. (1985) Pulmonary response of rats exposed to titanium dioxide (TiO2)
by inhalation for two years. Toxicol Appl Pharmacol 79:179-192.

Lee, KP; Kelly, DP; Schneider, PW; et al. (1986) Inhalation toxicity study on rats exposed to titanium
tetrachloride atmospheric hydrolysis products for two years. Toxicol Appl Pharmacol 83:30-45.

Lee, KP; Kelly, DP; O'Neal, FO; et al. (1988) Lung response to ultrafine Kevlar aramid synthetic fibrils following
2-year inhalation exposure in rats. Fundam Appl Toxicol  11:1-20.


                                                5-106

-------
Levins, PL. (1981) Review of diesel odor and toxic vapor emissions. Washington, DC: U.S. Department of
Transportation, National Highway Traffic Safety Administration; report no. DOT-TSC-NHTSA-81-9.

Lewis, TR; Green, FHY; Moorman, WJ; et al. (1989) A chronic inhalation toxicity study of diesel engine
emissions and coal dust, alone and combined. J Am Coll Toxicol 8:345-375.

Linnell, RH; Scott, WE. (1962) Diesel exhaust composition and odor studies. J Air Pollut Control Assoc
12:510-515,545.

Lovik, M; Hegseth, AK; Gaarder, PI; et al.  (1997) Diesel exhaust particles and carbon black have adjuvant
activity on the local lymph node response and systemic IgE production to ovalbumin. Toxicology 121:165-178.

Madden,  MC; Richards, JE; Dailey, LA; et al. (2000) Effect of ozone on diesel exhaust particle toxicity. Toxicol
Appl Pharmacol (submitted)

Maejima, K; Tamura, K; Taniguchi, Y; et al. (1997) Comparison of the effects of various fine particles on IgE
antibody  production in mice inhaling Japanese cedar pollen allergens. J Toxicol Environ Health 52:231-248.

Mauderly, JL. (1994) Noncancer pulmonary effects of chronic inhalation exposure of animals to solid particles.
In: Toxic and carcinogenic effects of solid particles in the respiratory tract: [proceedings of the  4th international
inhalation symposium]; March 1993; Hannover, Germany.  Mohr, U; Dungworth, DL; Mauderly, JL; et al., eds.
Washington, DC: International Life Sciences Institute Press; pp. 43-55.

Mauderly, JL; Benson, JM; Bice, DE; et al. (1981) Observations on rodents exposed for 19 weeks to diluted diesel
exhaust. In: Inhalation Toxicology Research Institute annual report 1980-1981. Albuquerque, NM:  Lovelace
Biomedical and Environmental Research Institute; pp. 305-311; report no. LMF-91.

Mauderly, JL; Benson, JM; Rice, DE; et al. (1984) Life span study of rodents inhaling diesel exhaust: effects on
body weight and survival.  In: Inhalation Toxicology Research Institute annual report. Guilmette, RA; Medinsky,
MA, eds. Albuquerque, NM: Inhalation Toxicology Research Institute; pp. 287-291; report no. LMF-113.
Available from: NTIS, Springfield, VA; DE85-009283.

Mauderly, JL; Bice, DE; Carpenter, RL; et al. (1987a) Effects of inhaled nitrogen dioxide and diesel exhaust  on
developing lung. Cambridge, MA: Health Effects Institute; research report no. 8.

Mauderly, JL; Jones, RK;  Griffith, WC; et al. (1987b)  Diesel exhaust is a pulmonary carcinogen in rats exposed
chronically by inhalation. Fundam Appl Toxicol 9:208-221.

Mauderly, JL; Gillett, NA; Henderson, RF; et al. (1988) Relationships of lung structural and functional changes to
accumulation of diesel exhaust particles. In: Inhaled particles VI: proceedings of an international symposium and
workshop on lung dosimetry; September 1985; Cambridge, United Kingdom. Dodgson, J; McCallum, RI; Bailey,
MR; et al., eds. Ann Occup Hyg 32(suppl. l):659-669.

Mauderly, JL; Bice, DE; Cheng, YS; et al. (1989) Influence of experimental pulmonary emphysema on
toxicological effects from  inhaled nitrogen dioxide and diesel exhaust. Cambridge, MA: Health Effects Institute;
report no. HEI-RR-89/30.  Available from NTIS, Springfield, VA; PB90-247347.

Mauderly, JL; Bice, DE; Cheng, YS; et al. (1990a) Influence of preexisting pulmonary emphysema on
susceptibility of rats to diesel exhaust. Am Rev Respir Dis  141:1333-1341.

Mauderly, JL; Cheng, YS; Snipes, MB. (1990b) Particle overload in toxicology studies: friend or foe? In:
Proceedings of the particle lung interaction symposium; May; Rochester, NY. J Aerosol Med 3(suppl. 1):
S-169-S-187.
                                                5-107

-------
Mauderly, JL; Banas, DA; Griffith, WC; et al. (1996) Diesel exhaust is not a pulmonary carcinogen in CD-I mice
exposed under conditions carcinogenic to F344 rats. Fundam Appl Toxicol 30:233-242.

McClellan, RO; Bice, DE; Cuddihy, RG; et al. (1986) Health effects of diesel exhaust. In: Aerosols: research, risk
assessment and control strategies: proceedings of the second U.S.-Dutch international symposium; May 1985;
Williamsburg, VA. Lee, SD; Schneider, T; Grant, LD; et al., eds. Chelsea, MI: Lewis Publishers, Inc.; pp.
597-615.

Meiss, R; Robenek, H; Schubert, M; et al. (1981) Ultrastructural alterations in the livers of golden hamsters
following experimental chronic inhalation of diluted diesel exhaust emission. Int Arch Occup Environ Health
48:147-157.

Mentnech, MS; Lewis, DM; Olenchock, SA; et al. (1984) Effects of coal dust and diesel exhaust on immune
competence in rats. J Toxicol Environ Health 13:31-41.

Misiorowski, RL; Strom, KA; Vostal, JJ; et al. (1980) Lung biochemistry of rats chronically exposed to diesel
particulates. In: Health effects of diesel engine emissions: proceedings of an international symposium, v. 1;
December 1979. Cincinnati, OH. Pepelko, WE; Danner, RM; Clarke, NA, eds. Cincinnati, OH: U.S.
Environmental Protection Agency, Health Effects Research Laboratory; pp. 465-480; EPA report no.
EPA/600/9-80/057a. Available from: NTIS, Springfield, VA; PB81-173809.

Miyabara, Y; Ichinose, T; Takano, H; et al. (1998a) Diesel exhaust inhalation enhances airway
hyperresponsiveness in mice.  Int Arch Allergy Immunol 116:124-131.

Miyabara, Y; Takano, H; Ichinose, T; et al. (1998b) Diesel exhaust enhances allergic airway inflammation and
hyperresponsiveness in mice.  Am JResp Crit Care Med 157:1138-1144.

Moorman, WJ; Clark, JC; Pepelko, WE; et al. (1985) Pulmonary fuction responses in cats following long-term
exposure to diesel exhaust. J Appl Toxicol 5:301-305.

Morrow, PE. (1988) Possible  mechanisms to explain dust overloading of the lungs. Fundam Appl Toxicol
10:369-384.

Muranaka, M; Suzuki, S; Koizumi, K; et al. (1986) Adjuvant activity of diesel-exhaust particulates for the
production of IgE antibody in mice. J Allergy Clin Immunol 77(4) :616-23.

Murphy, SA; BeruBe, KA; Pooley, FD; et al. (1998) The response of lung epithelium to well characterised fine
particles. Life Sci 62:1789-1799.

Murphy, SA; et al  (1999) Bioreactivity of carbon black and diesel exhaust particles to primary Clara and type II
epithelial cell cultures. Occup Environ Med 56:813-819.

Navarro, C; Charboneau, J; McCauley, R. (1981) The effect of in vivo exposure to diesel exhaust on rat hepatic
and pulmonary microsomal activities. J Appl Toxicol 1:124-126.

Nel, AE; Diaz-Sanchez,D: Ng, D; Hiura, T; and Saxon, A. (1998) Enhancement of allergic  inflammation by the
interaction between diesel exhaust particles and the  immune system. J Allergy Clin Immunology 104(4, pt 1):
539-554.

Nightingale, J A; Maggs, R; Cullinan, P; Donnelly, LE; Rogers, DF; Kinnersley, R; Chung, KF; Barnes, PJ;
Ashmore, M; Newman-Taylor, A. (2000) Airway inflammation after controlled exposure to  diesel exhaust
particulates. Am. J. Respir. Crit. Care Med. 162:161-166.

Nikula, KJ; Snipes, MB; Barr, EB; et al. (1995) Comparative pulmonary toxicities and carcinogenicities of
chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl Toxicol 25:80-94.


                                                5-108

-------
Nikula, KJ; Avila, KJ; Griffith, WC; et al. (1997a) Sites of particle retention and lung tissue responses to
chronically inhaled diesel exhaust and coal dust in rats and cynomolgus monkeys. In: Proceedings of the sixth
international meeting on the toxicology of natural and man-made fibrous and non-fibrous particles; September
1996; Lake Placid, NY. Driscoll, KE; Oberdorster, G, eds. Environ Health Perspect Suppl 105(5): 1231-1234.

Nikula, KJ; Avila, KJ; Griffith, WC; et al. (1997b) Lung tissue responses and sites of particle retention differ
between rats and cynomolgus monkeys exposed chronically to diesel exhaust and coal dust. Fundam Appl Toxicol
o T .o T C"5
37:37-53.

Nilsen, A; Hagemann, R; Eide, I. (1997) The adjunct activity of diesel exhaust particles and carbon black on
systemic IgE production to ovalbumin in mice after intranasal instillation. Toxicology 124:225-232.

Nordenhall, C; Pourazar, J; Blomberg, A; et al. (2000) Airway inflammation following exposure to diesel exhaust:
a study of time kinetics using induced sputum. EurRespir J 15:1046-1051.

Oberdorster, G. (1994) Extrapolation of results from animal inhalation studies with particles to humans? In: Toxic
and carcinogenic effects of solid particles in the respiratory tract: [proceedings of the 4th international inhalation
symposium]; March 1993; Hannover, Germany. Mohr, U; Dungworth, DL; Mauderly, JL; et al., eds. Washington,
DC: International Life Sciences Institute Press; pp. 335-353.

Oberdorster, G; Ferin, J; Gelein, R; et al. (1992) Role of the alveolar macrophage in lung injury: studies with
ultrafine particles. Environ Health Perspect 97:193-199.

Odor Panel of the CRC-APRAC Program Group on Composition of Diesel Exhaust. (1979) Development and
evaluation of a method for measurement of diesel exhaust odor using a diesel odor analysis system (DOAS). New
York: Coordinating Research Council, Inc., Air Pollution Research Advisory Committee; CRC-APRAC project
no.  CAPI-1-64.

Ohtoshi, T; Takizawa, H;  Okazaki, H; et al. (1998) Diesel exhaust particles stimulate human airway epithelial
cells to produce cytokines relevant to airway  inflammation in vitro. J Allergy Clin Immunol 101:778-785.

Ormstad, H; Johansen, BV; Gaarder, PI. (1998) Airborne house dust particles and diesel  exhaust particles as
allergen carriers. Clin Exp Allergy 28:702-708.

Pattle, RE;  Stretch, H; Burgess, F; et al. (1957) The toxicity of fumes from a diesel engine  under four different
running conditions. Br J Ind Med 14:47-55.

Penney, DG; Baylerian, MS; Fanning, KE; et al. (1981)  A study of heart and blood of rodents inhaling diesel
engine exhaust particulates. Environ Res 26:453-462.

Pepelko, WE. (1982a) Effects of 28 days exposure to diesel engine emissions in rats. Environ Res 27:16-23.

Pepelko, WE. (1982b) EPA studies on the toxicological effects of inhaled diesel engine emissions. In:
Toxicological effects of emissions from diesel engines: proceedings  of the Environmental Protection Agency
diesel emissions symposium; October 1981; Raleigh, NC. Lewtas, J, ed. New York: Elsevier Biomedical; pp.
121-142. (Developments in toxicology and environmental science: v. 10).

Pepelko, WE; Peirano, WB. (1983) Health effects of exposure to diesel engine emissions: a summary of animal
studies conducted by the U.S. Environmental Protection Agency's Health Effects Research Laboratories at
Cincinnati, Ohio. J Am Coll Toxicol 2:253-306.

Pepelko, WE; Mattox, JK; Yang, YY; et al. (1980a) Pulmonary function and pathology in cats exposed 28 days to
diesel exhaust. J Environ Pathol Toxicol 4:449-458.
                                                5-109

-------
Pepelko, WE; Mattox, J; Moorman, WJ; et al. (1980b) Pulmonary function evaluation of cats after one year of
exposure to diesel exhaust. In: Health effects of diesel engine emissions: proceedings of an international
symposium, v. 2; December 1979; Cincinnati, OH. Pepelko, WE; Banner, RM; Clarke, NA, eds. Cincinnati, OH:
U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp. 757-765; EPA report no.
EPA/600/9-80-057b. Available from: NTIS,  Springfield, VA; PB81-173817.

Pepelko, WE; Mattox, J; Moorman, WJ; et al. (1981) Pulmonary function evaluation of cats after one year of
exposure to diesel exhaust. Environ Int 5:373-376.

Pereira, MA; Sabharwal, PS; Gordon, L; et al. (1981) The effect of diesel exhaust on sperm-shape abnormalities
in mice. Environ Int 5:459-460.

Plopper, CG; Hyde, DM; Weir, AJ. (1983) Centriacinar alterations in lungs of cats chronically exposed to diesel
exhaust. Lab Invest 49:391-399.

Purdham, JT; Holness, DL; Pilger, CW. (1987) Environmental and medical assessment of stevedores employed in
ferry operations. Appl Ind Hyg 2:133-139.

Quinto, I; De Marinis, E. (1984) Sperm abnormalities in mice exposed to diesel paniculate. Mutat Res 130:242.

Rabovsky, J; Petersen, MR; Lewis, TR; et al. (1984) Chronic inhalation of diesel exhaust and coal dust: effect of
age and exposure on selected enzyme activities associated with microsomal cytochrome P-450 in rat lung and
liver. J Toxicol Environ Health 14:655-666.

Rabovsky, J; Judy, DJ; Rodak, DJ; et al. (1986) Influenza virus-induced alterations of cytochrome P-450 enzyme
activities following exposure of mice to coal and diesel particulates. Environ Res 40:136-144.

Reger, R. (1979) Ventilatory function changes over a work shift for coal miners exposed to diesel emissions. In:
Toxicological and carcinogenic health hazards in the workplace: proceedings of the first annual NIOSH scientific
symposium; 1978; Cincinnati, OH. Bridbord, K; French, J, eds. Park Forest South, IL: Pathotox Publishers, Inc.;
pp.  346-347.

Reger, R; Hancock, J; Hankinson, J; et al.  (1982) Coal miners exposed to diesel exhaust emissions. Ann Occup
Hyg 26:799-815.

Research Committee for HERP Studies. (1988) Diesel exhaust and health risks: results of the HERP studies.
Tsukuba, Ibaraki, Japan: Japan Automobile Research Institute, Inc.

Rudell, B; Sandstrom, T; Stjernberg, N; et al. (1990) Controlled diesel exhaust exposure in an exposure chamber:
pulmonary effects investigated with bronchoalveolar lavage. J Aerosol Sci 21(suppl. I):s411-s414.

Rudell, B; Sandstrom, T; Hammarstrom, U;  et al. (1994) Evaluation of an exposure setup for studying effects of
diesel exhaust in humans.  Int Arch Occup  Environ Health 66:77-83.

Rudell, B; Ledin, MC; Hammarstrom, U; et  al. (1996) Effects on symptoms and lung function in humans
experimentally exposed to diesel exhaust.  Occup Environ Med 53:658-662.

Sagai, M; Furuyama,  A; Ichinose, T. (1996)  Biological effects of diesel exhaust particles (DEP). III. Pathogenesis
of asthma like symptoms in mice. Free Radic Biol Med 21:199-209.

Salvi, S; Blomberg, A; Rudell, B; et al.  (1999) Acute inflammatory responses in the airways and peripheral blood
after short-term exposure to diesel exhaust in healthy human volunteers. Am J Respir Crit Care Med 159:702-709.

Salvi, SS; Nordenhall, C; Blomberg, A; et al. (2000) Acute exposure to diesel exhaust increases IL-8 and GRO-
alpha production in healthy human airways.  Am J Respir Crit Care Med 161(2):550-557.


                                                5-110

-------
Saneyoshi, K; Nohara, O; Imai, T; et al. (1997) IL-4 and IL-6 production of bone marrow-derived mast cells is
enhanced by treatment with environmental pollutants. Int Arch Allergy Immunol 114(3):237-245.

Schneider, DR; Felt, BT. (1981) Effect of diesel paniculate exposure on adenylate and guanylate cyclase of rat
and guinea pig liver and lung. J Appl Toxicol 1:135-139.

Schreck, RM; Soderholm, SC; Chan, TL; et al.  (1981) Experimental conditions in GMR chronic inhalation studies
of diesel exhaust. J Appl Toxicol  1:67-76.

Steerenberg, PA; Zonnenberg, JAJ; Dormans, JAMA; et al. (1998) Diesel exhaust particles induced release of
interleukin 6 and 8 by (primed) human bronchial epithelial cells (BEAS 2B) in vitro. Exp Lung Res 24:85-100.

Stober, W. (1986) Experimental induction of tumors in hamsters, mice and rats after long-term inhalation of
filtered and unfiltered diesel engine exhaust. In: Carcinogenic and mutagenic effects of diesel engine exhaust:
proceedings of the international satellite symposium on toxicological effects of emissions from diesel engines;
July; Tsukuba Science City, Japan. Amsterdam,: Ishinishi, N; Koizumi, A;  McClellan, RO; et al. Elsevier Science
Publishers B. V.; pp. 421-439. (Developments in toxicology and environmental science: v.  13).

Strom, KA. (1984) Response of pulmonary cellular defenses to the inhalation of high concentrations of diesel
exhaust. J Toxicol Environ Health 13:919-944.

Suzuki, T; Kanoh, T; Kanbayashi, M;  et al. (1993) The adjuvant activity of pyrene in diesel exhaust on IgE
antibody production in mice. Arerugi 42(8): 963-8.

Suphioglu, C; Singh, MB; Taylor, P; et al. (1992) Mechanism of grass-pollen-induced asthma. Lancet
339(8793):569-572.

Suzuki, T; Kanoh, T; Ishimori, M; et al. (1996) Adjuvant activity of diesel exhaust particulates (DEP) in
production of anti-IgE and anti-IgGl antibodies to mite allergen in mice.  J Clin Lab Immunol 48:187-199.

Takafuji, S; Suzuki, S; Koizumi, K; et al. (1987) Diesel-exhaust particulates inoculated by the intranasal route
have an adjuvant activity for IgE production in mice. J Allergy Clin Immunol 79:639-645.

Takano, H; Yoshikawa, T; Ichinose, T; et al. (1997) Diesel exhaust particles enhance antigen-induced airway
inflammation and local cytokine expression in mice. Am J Respir Crit Care Med 156:36-42.

Takano, H; Ichinose, T; Miyabara, Y;  et al. (1998a) Inhalation of diesel exhaust enhances allergen-related
eosinophil recruitment and airway hyperresponsiveness in mice. Toxicol Appl Pharmacol 150:328-337.

Takano, H; Ichinose, T; Miyabara, Y;  et al. (1998b) Diesel exhaust particles enhance airway responsiveness
following allergen exposure in mice. Immunopharmacol Immunotoxicol 20:329-336.

Takenaka, H; Zhang, K; Diaz-Sanchez, D; et al. (1995) Enhanced human IgE production results from exposure to
the aromatic hydrocarbons from diesel exhaust: direct effects on B-cell IgE production. J Allergy Clin Immunol
95:103-115.

Terada, N; Maesako, K; Hiruma, K; et al. (1997). Diesel exhaust particulates enhance eosinophil adhesion to nasal
epithelial cells  and cause degranulation. Int Arch Allergy Immunol  114(2): 167-174.

Terada, N; Hamano, N; Maesako, K; et al. (1999) Diesel exhaust particulates upregulate histamine receptor
mRNA and increase histamine-induced IL-8 and GM-CSF production in nasal epithelial cells and endothelial
cells. Clin Exp Allergy 29(l):52-59.

Tsukue, N; Toda, N; Tsubone, H; et al. (2001) Diesel exhaust (DE) affects the regulation of testicular function in
male Fischer 344 rats.  J. Toxicol  Environ Health Part A 63:115-126.


                                                5-111

-------
Turk, A. (1967) Selection and training of judges for sensory evaluation of the intensity and character of diesel
exhaust odors. Cincinnati, OH: U.S. Department of Health, Education, and Welfare, National Center for Air
Pollution Control; Public Health Service publication no. 999-AP-32. Available from: NTIS, Springfield, VA;
PB-174707.

U.S. Environmental Protection Agency (U.S. EPA). (1996) Air quality criteria for paniculate matter. Research
Triangle Park, NC: National Center for Environmental Assessment-RTF Office; report nos.
EPA/600/P-95/001aF-cF. 3v. Available from: NTIS, Springfield, VA; PB96-168224.

UlfVarson, U; Alexandersson, R; Aringer, L; et al. (1987) Effects of exposure to vehicle exhaust on health. Scand
J Work Environ Health 13:505-512.

Valavanidis, A; Salika, A; Theodoropoulou, A. (2000) Generation of hydroxyl radicals by urban suspended
particulate air matter.  The role of iron ions. Atmos Environ 34:2379-2386.

Vallyathan, V; Virmani, R; Rochlani, S; et al. (1986) Effect of diesel emissions and coal dust inhalation on heart
and pulmonary arteries of rats. Br J Ind Med 14:47-55.

van Vliet, P; Knape, M; de Hartog, J; et al. (1997) Motor vehicle exhaust and chronic respiratory symptoms in
children living near freeways. Environ Res 74(2): 122-132.

Vinegar, A; Carson, Al; Pepelko, WE. (1980) Pulmonary function changes in Chinese hamsters exposed six
months to diesel exhaust. In: Health effects of diesel engine emissions: proceedings of an international
symposium, v. 2; December,  1979; Cincinnati, OH. Pepelko, W E; Danner, R M; Clarke, N A, eds. Cincinnati,
OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp. 749-756; EPA report no.
EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.

Vinegar, A; Carson, A; Pepelko, WE; etal. (1981a) Effect of six months of exposure to two levels of diesel
exhaust on pulmonary  function of Chinese hamsters. Fed Proc 40:593.

Vinegar, A; Carson, A; Pepelko, WE. (1981b) Pulmonary function changes in Chinese hamsters exposed six
months to diesel exhaust. Environ Int 5:369-371.

Vostal, JJ; Chan, TL; Garg, BD; et al. (1981) Lymphatic transport of inhaled diesel particles in the lungs of rats
and guinea pigs exposed to diluted diesel exhaust. Environ Int 5:339-347.

Vostal, JJ; White, HJ;  Strom, KA; et al. (1982) Response of the pulmonary defense system to diesel particulate
exposure. In: Toxicological effects of emissions from diesel engines: proceedings of the Environmental Protection
Agency diesel emissions symposium; October 1981; Raleigh, NC. Lewtas, J., ed. New York: Elsevier Biomedical;
pp. 201-221. (Developments in toxicology and environmental science: v.  10).

Wade, JF, III; Newman, LS. (1993) Diesel asthma: reactive airways disease following overexposure to locomotive
exhaust. J Occup Med  35:149-154.

Wallace, MA; Salley, SO; Barnhart, MI. (1987) Analysis of the effects of inhaled diesel exhaust on the alveolar
intravascular and interstitial cellular components of rodent lungs. Scanning Microsc 1:1387-1395.

Watanabe, N; Oonuki, Y. (1999) Inhalation of diesel engine engine exhaust affects  spermatogenesis in growing
male rats. Environ Health Perspect 107:539-544.

Wehner, AP; Dagle, GE; Clark, ML; et al. (1986) Lung changes in rats following inhalation exposure to volcanic
ash for two years. Environ Res 40:499-517.
                                                5-112

-------
Werchowski, KM; Chaffee, VW; Briggs, GB. (1980a) Teratologic effects of long-term exposure to diesel exhaust
emissions (rats). Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory;
EPA report no. EPA-600/1-80-010. Available from: NTIS, Springfield, VA; PB80-159965.

Werchowski, KM; Henne, SP; Briggs, GB. (1980b) Teratologic effects of long-term exposure to diesel exhaust
emissions (rabbits). Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory;
EPA report no. EPA-600/1-80/011. Available from: NTIS, Springfield, VA; PB80-168529.

White, HJ; Garg, BD. (1981) Early pulmonary response of the rat lung to inhalation of high concentration of
diesel particles. J Appl Toxicol 1:104-110.

Wiester, MJ; Iltis, R; Moore, W. (1980) Altered function and histology in guinea pigs after inhalation of diesel
exhaust. Environ Res 22:285-297.

Wolff, RK; Gray, RL. (1980) Tracheal clearance of particles. In: Inhalation Toxicology Research Institute annual
report: 1979-1980. Diel, JH; Bice, DE; Martinez, BS, eds. Albuquerque, NM: Lovelace Biomedical and
Environmental Research Institute; p. 252; report no. LMF-84.

Wolff, RK; Henderson, RF; Snipes, MB; et al. (1987) Alterations in particle accumulation and clearance in lungs
of rats chronically exposed to diesel exhaust. Fundam Appl Toxicol 9:154-166.

Yang, HM; Ma, JY; Castranova, V; et al. (1997) Effects  of diesel exhaust particles on the release of interleukin-1
and tumor necrosis  factor-alpha from rat alveolar macrophages. Exp Lung Res 23:269-284.
                                                5-113

-------
Rudell, B; Sandstrom, T; Hammarstrom, U; et al. (1994) Evaluation of an exposure setup for studying effects of
diesel exhaust in humans. Int Arch Occup Environ Health 66:77-83.

Rudell, B; Ledin, MC; Hammarstrom, U; et al. (1996) Effects on symptoms and lung function in humans
experimentally exposed to diesel exhaust. Occup Environ Med 53:658-662.

Sagai, M; Furuyama, A; Ichinose, T. (1996) Biological effects of diesel exhaust particles (DEP). III. Pathogenesis of
asthma like symptoms in mice. Free Radic Biol Med 21:199-209.

Salvi, S; Blomberg, A; Rudell, B; et al. (1999) Acute inflammatory responses in the airways and peripheral blood
after short-term exposure to diesel exhaust in healthy human volunteers. Am J Respir Crit Care Med 159:702-709.

Salvi, SS; Nordenhall, C; Blomberg, A; et al. (2000) Acute exposure to diesel exhaust increases IL-8 and GRO-alpha
production in healthy human airways. Am J Respir Crit Care Med 161(2):550-557.

Saneyoshi, K; Nohara, O; Imai, T; et al. (1997) IL-4 and IL-6 production of bone marrow-derived mast cells is
enhanced by treatment with environmental pollutants. Int Arch Allergy Immunol  114(3):237-245.

Schneider, DR; Felt, BT. (1981) Effect of diesel paniculate exposure on adenylate and guanylate cyclase of rat and
guinea pig liver and lung. J Appl Toxicol 1:135-139.

Schreck, RM; Soderholm, SC; Chan, TL; et al. (1981) Experimental conditions in GMR chronic inhalation studies of
diesel exhaust. J Appl Toxicol 1:67-76.

Steerenberg, PA; Zonnenberg, JAJ; Dormans, JAMA; et al. (1998) Diesel exhaust particles induced release of
interleukin 6 and 8 by (primed) human bronchial epithelial cells (BEAS 2B) in vitro. Exp Lung Res 24:85-100.

Stober, W. (1986) Experimental induction of tumors in hamsters, mice and rats after long-term inhalation of filtered
and unfiltered diesel engine exhaust. In: Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings
of the international satellite symposium on lexicological  effects of emissions from diesel engines; July; Tsukuba
Science City, Japan. Amsterdam,: Ishinishi, N; Koizumi, A; McClellan, RO; et al. Elsevier Science Publishers B.
V.; pp. 421-439. (Developments in toxicology and environmental science: v. 13).

Strom, KA. (1984) Response of pulmonary cellular defenses to the inhalation of high concentrations of diesel
exhaust. J Toxicol Environ Health 13:919-944.

Suzuki, T; Kanoh, T; Kanbayashi, M; et al. (1993) The adjuvant activity of pyrene in diesel exhaust on IgE antibody
production in mice. Arerugi 42(8):963-8.

Suphioglu, C; Singh, MB; Taylor, P; et al. (1992) Mechanism of grass-pollen-induced asthma. Lancet
339(8793):569-572.

Suzuki, T; Kanoh, T; Ishimori, M; et al. (1996) Adjuvant activity of diesel exhaust particulates (DEP) in production
of anti-IgE and anti-IgGl antibodies to mite allergen in mice. J Clin Lab Immunol 48:187-199.

Takafuji, S; Suzuki, S; Koizumi, K; et al. (1987) Diesel-exhaust particulates inoculated by the intranasal route have
an adjuvant activity for IgE production in mice. J Allergy Clin Immunol 79:639-645.

Takano, H; Yoshikawa, T; Ichinose, T; et al.  (1997) Diesel exhaust particles enhance antigen-induced airway
inflammation and local cytokine expression in mice. Am J Respir Crit Care Med 156:36-42.

Takano, H; Ichinose, T; Miyabara, Y; et al. (1998a) Inhalation of diesel exhaust enhances allergen-related eosinophil
recruitment and airway hyperresponsiveness in mice. Toxicol Appl Pharmacol 150:328-337.

Takano, H; Ichinose, T; Miyabara, Y; et al. (1998b) Diesel exhaust particles enhance airway responsiveness
following allergen exposure in mice. Immunopharmacol Immunotoxicol 20:329-336.


                                                 5-114

-------
Takenaka, H; Zhang, K; Diaz-Sanchez, D; et al. (1995) Enhanced human IgE production results from exposure to
the aromatic hydrocarbons from diesel exhaust: direct effects on B-cell IgE production. J Allergy Clin Immunol
95:103-115.

Terada, N; Maesako, K; Hiruma, K; et al. (1997). Diesel exhaust particulates enhance eosinophil adhesion to nasal
epithelial cells and cause degranulation. Int Arch Allergy Immunol 114(2): 167-174.

Terada, N; Hamano, N; Maesako, K; et al. (1999) Diesel exhaust particulates upregulate histamine receptor mRNA
and increase histamine-induced IL-8 and GM-CSF production in nasal epithelial cells and endothelial cells.  Clin Exp
Allergy 29(l):52-59.

Tsukue, N; Toda, N; Tsubone, H; et al. (2001) Diesel exhaust (DE) affects the regulation of testicular function in
male Fischer 344 rats.  J. Toxicol Environ Health Part A 63:115-126.

Turk, A. (1967) Selection and training of judges for sensory evaluation of the intensity and character of diesel
exhaust odors. Cincinnati, OH: U.S. Department of Health, Education, and Welfare, National Center for Air
Pollution Control; Public Health  Service publication no. 999-AP-32. Available from: NTIS, Springfield, VA;
PB-174707.

U.S. Environmental Protection Agency (U.S. EPA). (1996) Air quality criteria for paniculate matter. Research
Triangle Park, NC: National Center for Environmental Assessment-RTF Office; report nos.
EPA/600/P-95/001aF-cF. 3v. Available from: NTIS, Springfield, VA; PB96-168224.

UlfVarson, U; Alexandersson, R; Aringer, L; et al. (1987) Effects of exposure to vehicle exhaust on health. Scand J
Work Environ Health 13:505-512.

Valavanidis, A; Salika, A; Theodoropoulou, A. (2000) Generation of hydroxyl radicals by urban suspended
paniculate air matter. The role of iron ions.  Atmos Environ 34:2379-2386.

Vallyathan, V; Virmani, R; Rochlani, S; et al. (1986) Effect of diesel emissions and coal dust inhalation on heart and
pulmonary arteries of rats. Br J Ind Med  14:47-55.

van Vliet, P; Knape, M; de Hartog, J; et al. (1997) Motor vehicle exhaust and chronic respiratory symptoms in
children living near freeways. Environ Res 74(2): 122-132.

Vinegar, A; Carson,  Al; Pepelko, WE. (1980) Pulmonary function changes in Chinese hamsters exposed six months
to diesel exhaust. In: Health effects of diesel engine emissions: proceedings of an international symposium, v. 2;
December,  1979; Cincinnati, OH. Pepelko, W E; Danner, R M; Clarke, N A, eds. Cincinnati, OH: U.S.
Environmental Protection Agency, Health Effects Research Laboratory; pp. 749-756; EPA report no.
EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.

Vinegar, A; Carson,  A; Pepelko,  WE; et al. (1981a) Effect of six months of exposure to two levels of diesel exhaust
on pulmonary function of Chinese hamsters. Fed Proc 40:593.

Vinegar, A; Carson,  A; Pepelko,  WE. (1981b) Pulmonary function changes in Chinese hamsters exposed six months
to diesel exhaust. Environ Int 5:369-371.

Vostal, JJ; Chan, TL; Garg, BD;  et al. (1981) Lymphatic transport of inhaled diesel particles in the lungs of rats and
guinea pigs exposed to diluted diesel exhaust. Environ Int 5:339-347.

Vostal, JJ; White, HJ; Strom, KA; et al. (1982) Response of the pulmonary defense system to diesel paniculate
exposure. In: Toxicological effects of emissions from diesel engines: proceedings of the Environmental Protection
Agency diesel emissions symposium; October 1981; Raleigh, NC. Lewtas, J., ed. New York: Elsevier Biomedical;
pp. 201-221. (Developments in toxicology and environmental science: v. 10).
                                                 5-115

-------
Wade, JF, III; Newman, LS. (1993) Diesel asthma: reactive airways disease following overexposure to locomotive
exhaust. J Occup Med 35:149-154.

Wallace, MA; Salley, SO; Barnhart, MI. (1987) Analysis of the effects of inhaled diesel exhaust on the alveolar
intravascular and interstitial cellular components of rodent lungs. Scanning Microsc 1:1387-1395.

Watanabe, N; Oonuki, Y. (1999) Inhalation of diesel engine engine exhaust affects spermatogenesis in growing male
rats. Environ Health Perspect 107:539-544.

Wehner, AP; Dagle, GE; Clark, ML; et al. (1986) Lung changes in rats following inhalation exposure to volcanic ash
for two years. Environ Res 40:499-517.

Werchowski, KM; Chaffee, VW; Briggs, GB. (1980a) Teratologic effects of long-term exposure to diesel exhaust
emissions (rats). Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory; EPA
report no. EPA-600/1-80-010. Available from: NTIS, Springfield, VA; PB80-159965.

Werchowski, KM; Henne, SP; Briggs, GB. (1980b) Teratologic effects of long-term exposure to diesel exhaust
emissions (rabbits). Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory;
EPA report no. EPA-600/1-80/011. Available from: NTIS, Springfield, VA; PB80-168529.

White, HJ; Garg, BD. (1981) Early pulmonary response of the rat lung to inhalation of high concentration of diesel
particles. J Appl Toxicol 1:104-110.

Wiester, MJ; Iltis, R; Moore, W. (1980) Altered function and histology in guinea pigs after inhalation of diesel
exhaust. Environ Res  22:285-297.

Wolff, RK; Gray, RL. (1980) Trachea! clearance of particles. In: Inhalation Toxicology Research Institute annual
report: 1979-1980. Diel, JH; Bice, DE; Martinez, BS, eds. Albuquerque, NM: Lovelace Biomedical and
Environmental Research Institute; p. 252; report no. LMF-84.

Wolff, RK; Henderson, RF; Snipes, MB; et al. (1987) Alterations in particle accumulation and clearance in lungs of
rats chronically exposed to diesel exhaust. Fundam Appl Toxicol 9:154-166.

Yang, HM; Ma, JY; Castranova, V; et al. (1997) Effects of diesel exhaust particles on the release of interleukin-1
and tumor necrosis factor-alpha from rat alveolar macrophages. Exp Lung Res 23:269-284.
                                                 5-116

-------
                              6.  ESTIMATING HUMAN
                NONCANCER HEALTH RISKS OF DIESEL EXHAUST

6.1.  INTRODUCTION
       As discussed earlier in this document (Chapter 2, Section 2.2.7, 2.2.8), diesel engine
exhaust (DE) consists of a complex mixture of gaseous pollutants and particles.  In attempting to
estimate potential health risks associated with human exposure to DE, researchers have focused
attention mostly on the particulate matter (PM) components. They have done so, in part, by
comparing the relative toxicity of unfiltered versus filtered DE (with gaseous components
removed), as discussed in Chapter 5.
       Diesel particulate matter (DPM) consists mainly of: (a) elemental carbon (EC) particles
having relatively large surface areas, (b) soluble organic carbon, including 5-ring or higher
polycyclic aromatic hydrocarbons (PAHs) such as benzo(a)pyrene, and other 3- or 4-ring
compounds distributed between gas and particle phases, and (c) metallic compounds.  DPM also
typically contains small amounts of sulfate/sulfuric acid and nitrates, trace elements, and water,
plus some unidentified components.  DPM is made up almost entirely of fine particles (i.e., all
below 1-3 pm) with a significant subset of ultrafine particles (i.e., those with a mass median
diameter below about 0.1  |_im).
       Health concerns have long focused on DPM.  Toxicological data described in Chapter 5
(Section 5.2) indicate DPM to be the prime etiologic agent of noncancer health effects when DE
is sufficiently diluted to limit the concentrations of gaseous irritants (NO2 and SO2), irritant
vapors (aldehydes), CO, or other systemic toxicants.  The large surface areas of DPM allow for
adsorption of organics from the diesel combustion process and for adsorption of additional
compounds during transport in ambient air. The small size of DPM, combined with their large
surface area, likely enhance the potential for subcellular interactions with important cellular
components of respiratory tissues once the particles are inhaled by humans or other  species
(Johnston  et al., 2000; Oberdorster et al., 2000).
       The content of DPM as described above and in Chapter 2 is of clear toxicological
significance.  The experimental evidence described in Chapter 5 concerning DPM's association
with and etiology of noncancer effects is extensive and compelling. These points, along with the
fact that DPM is easily and most frequently measured and reported in toxicological  studies of
diesel emissions, make DPM a reasonable choice as a measure of diesel emissions.  As a
surrogate,  DPM is as  valid as any other component of DE to show what is currently known—and
probably what is not yet known—about diesel emissions. Therefore, DPM is the quantitative
focus of this chapter.
                                          6-1

-------
       The usual agency approach to evaluating noncancer risks from inhaled exposures to toxic
air pollutants such as ambient DE has been documented by EPA in the methods for derivation of
an inhalation reference concentration (RfC) (U.S. EPA, 1994).  For DPM exposures, this means
combining key elements derived from evaluations of specific DPM noncancer effects in animals
and humans (described in Chapter 5) with the use of quantitative dosimetry models (described in
Chapter 3).  The goal is to estimate DPM concentrations to which humans might be exposed
throughout their lives (i.e., chronically) without experiencing any untoward or adverse effects.
Such an effort can be accomplished through analysis of dose-response relationships where the
adverse response is considered as a function of a corresponding measure of dose.  Chapter 5 is
replete with dose-response information on adverse (but nonlethal) noncancer health effects
observed in long-term (chronic/lifetime) exposure studies to DE in general and to DPM in
particular, albeit mostly in animals.  Chapter 3 analyzes available methods to convert external
exposure concentrations of DPM in animal studies to estimates of a human-equivalent
concentration (HEC).  The following sections of this chapter (Sections 6.2, 6.3, and 6.5) assess
and integrate this information to derive a chronic RfC, using the above-cited methodology in
developing dose-response assessments of the noncancer effects of toxic air pollutants.
       Yet another approach to consider in deriving quantitative estimates of potential human
health risks associated with ambient (nonoccupational) DPM exposures is the extent to which
DPM could contribute to the adverse health effects that have been associated with exposure to
ambient fine PM, PM2 5. Such associations with adverse health effects are based primarily on
epidemiologic studies evaluated in EPA's Air Quality Criteria Document for Particulate Matter
(PM CD) (U.S.  EPA, 1996a).J This PM CD served as the scientific basis for the last periodic
review of the national ambient air quality standards (NAAQS) for PM, which resulted in the
establishment of revised PM standards in 1997, including standards for PM25. DPM is a
component of ambient fine PM (see Chapter 2) and should be considered as a lexicologically
important component of ambient fine PM. Any guidelines established for DPM, then, should be
concordant with information on fine PM in general, as presented in the PM CD. To more fully
consider the implications of the relationship between ambient DPM and fine PM, the
epidemiological evidence on fine PM and the basis for the PM2 5 standards are summarized, and
the relationship between ambient DPM and fine PM is discussed later in this chapter (Section
6.4). This relationship is of interest with respect to the noncancer assessment of DE. As is noted
here , however,  and reflected in Sections 6.2-6.4 below, the definitions, procedures, and statutory
mandates that apply to criteria pollutants such as PM (regulated through the establishment of
       'A new PM CD is now being prepared to reflect the latest scientific studies on ambient PM available since
the last document was completed.

                                           6-2

-------
NAAQS under sections 108 and 109 of the Clean Air Act) are fundamentally different from
those that apply to toxic air pollutants such as DE and to the derivation of RfCs for such
pollutants.  Thus, the ambient PM25 concentrations that are specified as the levels of the PM25
NAAQS should not be compared directly with any RfC that may be derived for DPM. It is
reasonable to observe, however, that the annual PM2 5 standard would be expected to provide a
measure of protection from DPM, reflecting DPM's current approximate proportion to PM25.
       Estimates of DE levels associated with effects occurring under less than lifetime exposure
scenarios (such as acute exposure) are not addressed in this chapter. Studies of acute exposure to
DE are discussed in Chapter 5, but are accompanied by scant dose-response information, with
single-exposure studies for various specialized endpoints (e.g., allergenicity/adjuvancy) and other
multiple-exposure-level studies reporting data on mortality only. Based on currently available
methodologies, these studies do not yet appear to provide a sufficient basis from which to derive
a dose-response assessment for an acute DE exposure scenario.

6.2.  THE INHALATION REFERENCE CONCENTRATION APPROACH
       Historically, approaches such as the Acceptable Daily Intake (ADI) were developed
whereby effect levels, such as no-observed-adverse-effect levels (NOAELs) or lowest-observed-
adverse-effect levels (LOAELs) from human or animal data, were combined with certain "safety
factors" to accommodate areas of uncertainty to make quantitative estimates of a safe dose, i.e., a
level at which no adverse effect would be likely to occur. In response to the National Academy
of Sciences (NAS) report entitled "Risk Assessment in the Federal Government: Managing the
Process" (National Research Council, 1983), EPA developed two approaches similar to the ADI,
i.e., the oral reference dose (RfD) (Barnes and Dourson, 1988) and the parallel inhalation
reference concentration, the RfC, with its formal methodology (U.S. EPA, 1994). Similar to the
ADI in intent, the RfD/C approach is used for dose-response assessment of noncancer effects,
using an explicitly delineated, rigorous methodology that adheres to the principles set forth in the
1983 NRC report.  The RfC methodology includes comprehensive guidance on a number of
complex issues, including consistent application to effect levels of uncertainty factors (UFs)
rather than the ADI safety factors for consideration of uncertainty.  Basically, these approaches
attempt to estimate a likely subthreshold concentration in the human population. Use of the
RfD/C approach is one of the principal current agency methods for deriving dose-response
assessments.
       A chronic RfC is currently defined as:
                                          6-3

-------
       An estimate (with uncertainty spanning perhaps an order of magnitude) of a continuous
       inhalation exposure to the human population (including sensitive subgroups) that is likely
       to be without an appreciable risk of deleterious noncancer effects during a lifetime.

       The RfC approach involves the following general steps:

       •    Identification of a critical effect relevant to humans, i.e., an adverse effect that occurs
           at the lowest exposure/dose in human or animal studies and whose prevention avoids
           the occurrence of all other adverse effects;
           Selection of appropriate dose-response data to derive a point of departure (POD) for
           extrapolation of a key study (or studies) that provides a NOAEL, LOAEL, or
           benchmark concentration (BMCLX)2;
           Estimation of HECs when animal exposure-response data are used (via use of
           PBPK/dosimetry models);
           Application of UFs to the point of departure (e.g., NOAEL, LOAEL, BMCLX) to
           address extrapolation uncertainties (e.g., interindividual variability, interspecies
           differences, adequacy of database); and
           Characterization of the "confidence" in the dose-response assessment and resultant
           RfC.

       The basic quantitative formula for derivation of an RfC, given in Equation 6-1, has as its
basic components an effect level, here a NOAEL, expressed as an HEC, and UFs. The units of
an RfC are typically mg/m3 or |_ig/m3.
       Alternatively, the numerator in Equation 6-1 may be a LOAEL or BMCLX. The
benchmark concentration (BMC) approach and its application in this assessment are documented
in Appendix B and described further below. Also, a modifying factor (MF) may be used in the
denominator of this equation to account for scientific uncertainties, usually relating to the study
chosen as the basis for the RfC. Further specifics of RfC derivation procedures are discussed as
       2BMCLX is defined as the lower 95% confidence limit of the dose that will result in a level of "x" response
(e.g., BMCL10 is the lower 95% confidence limit of a dose for a 10% increase in a particular response). See
Appendix B for further specifics.

                                           6-4

-------
they are used in the following sections.  All such procedures are described in detail in the RfC
Methodology (U.S. EPA, 1994).

6.3.  CHRONIC REFERENCE CONCENTRATION FOR DIESEL EXHAUST
       As concluded in Chapter 5, chronic respiratory effects are the principal noncancer hazard
to humans from long-term environmental exposure to DE. Other effects (e.g., neurological,
liver-related) are observed in animal studies at higher exposures than those producing the
respiratory effects. The human and animal data for the immunological effects of DE are
currently considered inadequate for dose-response evaluation. Thus, the respiratory effects are
considered the "critical effect" for the derivation of a chronic RfC for DE.
       The evidence for chronic respiratory effects is based mainly on animal studies showing
consistent findings of inflammatory, histopathological (including fibrosis), and functional
changes in the pulmonary and tracheobronchial regions of laboratory animals, including the rat,
mouse, hamster, guinea pig, and monkey. Occupational studies of DE provide some
corroborative evidence of possible respiratory effects (e.g., respiratory symptoms and possible
lung function changes), although those studies are generally deficient in exposure information.
       Mode-of-action information about respiratory effects from DE exposure indicates that, at
least in rats, the pathogenic sequence following the inhalation of DPM begins with the
phagocytosis of diesel particles by alveolar macrophages (AMs). These activated AMs release
chemotactic factors that attract neutrophils and additional AMs. As the lung burden of DPM
increases, there are aggregations of particle-laden AMs in alveoli adjacent to terminal
bronchioles, increases in the number of Type n cells lining particle-laden alveoli, and the
presence of particles within alveolar and peribronchial interstitial tissues and associated lymph
nodes. The neutrophils and AMs release mediators of inflammation and oxygen radicals, and
particle-laden macrophages are functionally altered, resulting in decreased viability and impaired
phagocytosis and clearance of particles.  This series of events may result in pulmonary
inflammation, fibrosis, and eventually lesions like those described in the studies reviewed in
Chapter 5.  Although information describing the possible pathogenesis of respiratory effects in
humans is not available, the effects reported in studies of humans exposed to DE are not
inconsistent with the findings in controlled laboratory animal studies.
       Several reasons explain why the dose-response data from rats are considered especially
appropriate for use in characterizing noncancer health effects in humans and deriving a chronic
RfC for DE. First, similar noncancer respiratory effects are seen in other species (mouse,
hamster, guinea pig, and monkey). Second, rats and  humans exhibit similar noncancer responses
(macrophage response and interstitial fibrosis) to other particles such as coal mine dust, silica,
and beryllium (Haschek and Witschi, 1991; Oberdorster, 1994). Third, relative to other species

                                           6-5

-------
there exists a plethora of long-term, specialized, and mechanistic studies in rats. Fourth, an
expert panel convened by the International Life Sciences Institute (ILSI) recommended that
response data on persistent, inflammatory processes may be used to assess nonneoplastic
responses of poorly soluble particles (PSP) such as DPM (ILSI, 2000).

6.3.1.  Principal Studies for Dose-Response Analysis: Chronic, Multiple-Dose Level
       Rat Studies
       The experimental protocols and results from the long-term, repeated-exposure chronic
studies demonstrating and characterizing the critical effects of pulmonary fibrotic changes and
inflammation are discussed in Chapter 5.  Salient points of these studies, including species/sex
of the test species, the exposure regime and concentrations reported in mg DPM/m3, and effect
levels, are abstracted in Table 6-1 for further consideration. The effect levels are designated as
N for no-observed-adverse-effect level, A for adverse-effect level, and BMCL10.
       The purpose of many of the chronic studies listed in this table was not the elucidation of
the concentration-response character of DPM. The studies of Heinrich et al.  (1982, 1986) in
hamsters, mice, and rats; of Iwai et al. (1986) in rats; of Lewis et al. (1989) in monkeys; and of
Pepelko (1982a) in rats are all single-dose-level  analyses that have as their genesis mechanistic or
species-comparative purposes. As discussed in Chapter 5,  many of these studies do provide
valuable supporting information for designation of the critical effect of pulmonary
histopathology. The lack of any clear dose-response data, however, precludes consideration of
these studies as a basis for RfC derivation.
       Likewise, studies of chronic, multiple-level exposure involving species other than rats,
i.e., hamsters (Pepelko, 1982b), cats (Plopper et al., 1983), and guinea pigs (Barnhart et al., 1981,
1982), provide cross-species corroboration of the critical effects of pulmonary histopathology
and inflammatory alteration.
       The remaining studies showing exposure-response relationships in rats for the critical
effects include those of Ishinishi et al. (1986, 1988), Mauderly et al. (1987a), Heinrich et al.
(1995), and Nikula et al. (1995). As described in Chapter 5, all of these studies were conducted
and reported in a thorough, exhaustive manner on the critical effects and little, if any, basis exists
for choosing one over another for purposes of RfC derivation.  One way of taking advantage of
this high degree of methodological  and scientific merit would be to array data from all these
studies and their effect levels (NOAEL, LOAEL, BMCLX) subsequent to normalization of the
exposure conditions, i.e., conversion of the exposure regimes to yield an HEC.  This exercise
would result in an interstudy concentration-response continuum normalized to a continuous
human exposure to DPM that would facilitate the choice of a concentration to use as a point of
departure in deriving an RfC.
                                           6-6

-------
Table 6-1. Histopathological effects of diesel exhaust in the lungs of laboratory animals
Study
Lewis etal. (1989)


Bhatnagar et al.
(1980)
Pepelko (1982a)
Pepelko (1982b)




Heinrich et al.
(1982)

Iwai etal. (1986)


Mauderly et al.
(1987a)
Henderson et al.
(1988)


Heinrich et al.
(1995)











Species/sex
Monkey,
Cynomolgus,
M
Rat, F344,
M,F

Hamster,
Chinese, M



Hamster,
Syrian, M, F

Rat, F344, F


Rat, F344,
M, F; Mouse,
CD-I,
M,F


Rat, Wistar,
F;
Mouse,
NMRI, F
(7 mg/m3
only)
Mouse,
NMRI,F;
C57BL/6N, F




Exposure
period
7 h/day
5 days/wk
104 wks
7 h/day
5 days/wk
104 wks
8 h/day
5 days/wk
26 wks


7-8 h/day
5 days/wk
120 wks
8 h/day
7 days/wk
104 wks
7 h/day
5 days/wk
130 wks



18 h/day
5 days/wk
24 mo



18 h/day
5 days/wk
13. 5 mo
(NMRI)
24 mo
(C57BL/
N)
Particles
(mg/m3)
2.0


2.0


6.0
12.0



3.9


4.9


0.35
3.5
7.1



0.8
2.5
7.0



7.0






Effect level3
N





A




A


A


N
A
A



A
A
A



A






Effects"
AM aggregation; no fibrosis,
inflammation, or emphysema

Multifocal histiocytosis;
inflammatory changes; Type
II cell proliferation; fibrosis
Inflammatory changes; AM
accumulation; thickened
alveolar lining; Type II cell
hyperplasia; edema; increase
in collagen
Inflammatory changes, 60%
adenomatous cell proliferation

Type II cell proliferation;
inflammatory changes;
bronchial hyperplasia; fibrosis
Alveolar and bronchiolar
epithelial metaplasia in rats at
3.5 and 7.0 mg/m3; fibrosis at
7.0 mg/m3 in rats and mice;
inflammatory changes; few
quantitative data given
Bronchioalveolar hyperplasia,
interstitial fibrosis in all
groups; severity and incidence
increase with exposure
concentration; text given only

No increase in tumors;
noncancer effects not
discussed




                                         6-7

-------
Table 6-1.  Histopathological effects of diesel exhaust in the lungs of laboratory animals
(continued)
Study Species/sex
Islmishi et al. (1986, Rat, M, F,
1988) F344, /Jcl.







Heinrich et al. Hamster,
Exposure
period
16 h/day
6 days/wk
130 wks






19 h/day
Particles
(mg/m3)
O.llc
0.41C
1.08C
2.32C

0.46d
0.96d
1.84d
3.72d
4.24
Effect level3
N
N
A
A

N
A
A
A
A
Effects"
Inflammatory changes; Type
II cell hyperplasia and lung
tumors seen at >0.4 mg/m3;
shortening and loss of cilia in
trachea and bronchi; data
given in text only



Inflammatory changes;
 (1986)
 Barnhart et al.
 (1981, 1982);
 Vostaletal. (1981)
Syrian, M, F;
Mouse,
NMRI,F;
Rat, Wistar,
F
Guinea pig,
Hartley, M
5 days/wk
120 wks
 Plopper et al.
 (1983)
 Hyde etal. (1985)
Cat, inbred,
M
 Nikula et al. (1995)   Rat, F344, M
20 h/day
5.5
days/wk
104 wks
8 h/day
7 days/wk
124 wks
               16 h/day
               5 days/wk
               23 mo
0.25
0.75
 1.5
 6.0
6.0C
12.0d
              2.44
              6.33
             thickened alveolar septa;
             bronchioloalveolar
             hyperplasia; alveolar lipo-
             proteinosis; emphysema
             (diagnostic methodology not
             described); hyperplasia; lung
             tumors

    N        Minimal response at 0.25 and
    A        ultrastructural changes at
    A        0.75 mg/m3; thickened
    A        alveolar membranes;  cell
             proliferation; fibrosis at
             6.0 mg/m3; increase in PMN at
             0.75 mg/m3 and 1.5 mg/m3

    A        Inflammatory changes; AM
    A        aggregation; bronchiolar
             epithelial metaplasia; Type II
             cell hyperplasia;
             peribronchiolar fibrosis

   A, A      AM hyperplasia, epithelial
 BMCL10     hyperplasia, inflammation,
             septal fibrosis,
	bronchoalveolar metaplasia
aN= no-observed-adverse-effect level; A = adverse-effect level; BMCL10 = benchmark concentration, lower limit, at
a 10% response level (for incidence); see Appendix A for further specifics.
bAM = Alveolar macrophage; PMN = Polymorphonuclear leukocyte
°Light-duty engine.
dHeavy-duty engine.
°1 to 61 weeks exposure.
d62 to 124 weeks of exposure.
eSee Appendix A.
                                                  6-8

-------
6.3.2.  Derivation of Human Continuous Equivalent Concentrations, HECs
       Pharmacokinetic, or PK, models can be used to estimate across species the external
concentrations of a toxicant that would result in equivalent internal doses. When used for these
purposes, PK models may be termed comparative dosimetric models. Chapter 3 reviewed and
evaluated a number of dosimetric models applicable to DPM. This analysis indicated that
outputs from the human component of the model developed by Yu et al. (1991) specifically for
DPM, such as deposition and estimated lung burden, were not substantially different from other
available models. The analysis also demonstrated that the Yu model accounted for several
diesel-specific phenomena, including particle overload lung clearance rates and interspecies
kinetics of desorption of organics from the carbonaceous core of DPM, both slow- and fast-
cleared. Of importance, the Yu model was parameterized for deposition and clearance in both
animals and humans. Also, the animal component of the model was based on data from rats
actually exposed to DPM, whereas other models analyzed used data based only on generic
particles in  the size range of DPM. It was concluded from this  analysis that the Yu model could
be used to estimate disposition of DPM both in animals and in humans and would therefore be an
acceptable choice in performing animal-to-human extrapolation in deriving a continuous human-
equivalent concentration.  Note, however, that use of this or any other available PK model would
address species differences in dose (i.e., pharmacokinetics, PK), and not necessarily
pharmacodynamics (PD), the other component of uncertainty in animal-to-human or interspecies
extrapolation (U.S. EPA, 1994).
       Guidance on choosing measures of exposure for poorly soluble particles such as DPM
(ILSI, 2000) states that some measures of external dose (e.g., the aerosol exposure parameters of
MMAD, og, particle surface area, and density) should be characterized. Likewise, some
indication of internal dose resulting from the external exposure (e.g., lung burden) should be
measured so that differences in dose metrics may be considered as new mechanistic insights are
developed.  The whole particle, as characterized in this assessment and used in the model of Yu
et al. (1991), meets this recommended guidance, and DPM, in pg/m3, is used as the measure of
external exposure. Internal measures of exposure or dose were also considered in Chapter 3
(Section 3.3.1.1) with the conclusion that the dose metric of lung burden of DPM in terms of
surface area (mg/cm2) at the termination of the exposure period appears to be the most defensible
and appropriate measure of internal dose, especially where clearance is involved. More detailed
specifics are available in Chapter 3 and in Appendix A.
       The logical and operational sequence of deriving a HEC using the Yu model and these
metrics, i.e. external air concentration (in pg/m3) and lung burden (in mg/cm2), is demonstrated
in Figure 6-1. First,  the experimental animal exposures, including external concentration and
                                          6-9

-------
 Animal Exposures
   Cone, and Dur.
    Yu Model
(animal component)
LungBunl,nl^|(huma7^ilgnent)||^
HEC
Figure 6-1. Flow diagram of procedure for calculating HECs.

daily and weekly duration, are entered into the animal component of the Yu model to estimate
the animal lung burden, in mg DPM/cm2, for the specific exposure scenario. The human
component of the Yu model is then used by setting desired exposure conditions (continuous for
70 years) and running the model to find an external exposure DPM concentration that would
result in this same lung burden. The human external DPM concentration matching this lung
burden is the human-equivalent concentration. The step-by-step specifics and results of this
procedure as applied to the various studies in Table 6-1 are shown in Table A-4 and fully
explained in Appendix A.
       The foregoing discussion does not address the variability in outcomes that may be
estimated from the Yu et al. (1991) model from deposition of DPM. The model comparison
exercises in Chapter 3 showed relatively minor differences among the various human models for
one measure,  deposition, and  indicated that human lung burdens estimated by the human
component of the Yu and ICRP66 models were nearly identical at low-exposure concentrations.
Variability in output of their model (lung burden) was also examined by Yu and Yoon (1990),
who studied dependency on tidal volume,  respiration rate, and clearance (in terms of the overall
particle transport rate from the alveolar region,  AA). Analysis indicated that the model output is
sensitive, but not overly so, for these determinative parameters. A ± 20% change in values for
AA, for example, was estimated to result in a 16%-26% change in soot burden at a 0.1 mg/m3
continuous diesel exposure for 10 years. For a ± 10% change in tidal volume, the model
projected changes in soot burden ranging from 14% to 22% for this same exposure scenario. The
fact that the changes in the model outcome were comparable to changes in the input parameters,
such as tidal volume, indicates that the variability of the model when applied to the human
population would reflect the variability of these physiological parameters across that population.
In sum, at low concentrations of DPM (< 0.5 mg/m3), relatively minor differences exist among
the models currently available, and the input parameters in the human population may be a major
source of variability. As discussed below, variability within the human population often is
addressed by applying safety or uncertainty factors, usually in the range of 10 (Renwick and
Lazarus,  1998; U.S. EPA,  1994).
                                         6-10

-------
6.3.3. Dose-Response Analysis—Choice of an Effect Level
       HECs were obtained for the dose levels and exposure scenarios presented in the studies of
Mauderly et al. (1987b), Ishinishi et al. (1986, 1988), Nikula et al. (1995), and Heinrich et al.
(1995), the specifics of which are presented in Appendix A, specifically Table A-4. The HECs,
along with the corresponding specific lung burdens in terms of pg/cm2, were transcribed from
Table A-4 and, along with the accompanying effect level (NOAEL, LOAEL or BMCL10), are
arrayed ordinally in Table 6-2. It is acknowledged that Table 6-2 is by no means a full portrayal
of the dose-response relationship that may exist for DPM and health effects.
       As indicated by the BMCL10 values listed for the Nikula et al. (1995) study in Table 6-2,
the BMC analysis was carried out on the DPM database and is documented in Appendix B. The
chronic rat studies identified in this chapter were analyzed for information suitable for BMC
analysis.  Results yielded only a few datasets of pulmonary toxicity data from a single study, that
of Nikula et al. (1995), that could be used for BMC analysis. These pulmonary data
(histopathology incidence data) were extracted, HEC concentrations were calculated using the
model of Yu, and the BMCs were generated. The results yielded a complex array of BMCL10s
from three different effects in two sexes (both separate and combined) with nine different models
that were evaluated based on the nature of the dataset, on the goodness-of-fit parameters, and on
visual inspection of the graphical outputs.  From among all the benchmark data generated, the
BMCL10 of 0.37 mg/m3 calculated from combined male and female rat pulmonary histopathology
was judged as the most defensible choice.  However, further characterization of this same
benchmark value indicates that it is not a suitable candidate for use as a point of departure for
development of a  dose-response  assessment such as the RfC. Limitations included the excessive
extent of extrapolation from the observed experimental range (see Figure B-l in Appendix B)
and the paucity of data points (there were only two exposure groups) overall. Another serious
limitation is that the high experimental concentrations used (and their C x t product) are well in
the range where the problematic  phenomenon of pulmonary overload in rats occurs (Section
5.1.3.3.4).
       Inspection of Table 6-2 shows that calculating and ordering the HECs created a partial
concentration-response continuum reflected in the estimated internal lung burden also given in
this table. The continuum extends from HECs with no observed adverse effects at concentrations
as low as 0.032 mg/m3 to as high as 0.144 mg/m3 to HECs with an adverse effect level that first
appears definitively in the continuum probably at 0.33 mg/m3 and extends out to 1.95 mg/m3.
       It should be noted that the relationship between HEC and lung burden is not consistently
proportional.  For example, at the lowest HEC listed, 0.032 mg/m3, a lifetime (70 years) of
continuous exposure to this concentration is estimated to result in a specific burden to the lung of
0.0587 |_ig/cm2. At the other end of this spectrum, a lifetime of continuous exposure to 4.4
                                          6-11

-------
Table 6-2. Human equivalent continuous concentrations: 70-year HECs calculated with the model
of Yu et al. (1991) from long-term studies of rats repeatedly exposed to DPMa
Study
Ishinishi et al. (1988) (LDC)
Mauderly et al. (1987a)
Ishinishi et al. (1988) (LDC)
Ishinishi et al. (1988) (HDC)
Heinrich et al. (1995)
Nikulaetal. (1995)
Ishinishi et al. (1988) (HDC)
Ishinishi et al. (1988) (LDC)
Nikulaetal. (1995)
Mauderly etal. (1987a)
Nikulaetal. (1995)
Ishinishi et al. (1988) (HDC)
Heinrich etal. (1995)
Ishinishi et al. (1988) (LDC)
Mauderly etal. (1987a)
Ishinishi et al. (1988) (HDC)
Exposure
concentration
(mg/m3)
0.11
0.35
0.41
0.46
0.84
2.44 & 6.3d
0.96
1.18
2.44 & 6.3d
3.47
2.44
1.84
2.5
2.32
7.08
3.72
Effect levef
NOAEL
NOAEL
NOAEL
NOAEL
LOAEL
BMCL10 -inflam
LOAEL
LOAEL
BMCL10 - fibrosis
LOAEL
LOAEL
AEL
AEL
AEL
AEL
AEL
Lung burden
(modeled)
Oig DPM /cm2)b
0.0587
0.0685
0.245
0.281
0.94
1.34
3.16
4.50
4.70
4.95
7.00
7.63
8.40
9.75
10.9
15.8
HEC
(mg/m3)
0.032
0.038
0.128
0.144
0.33
0.37
0.883
1.25
1.3
1.375
1.95
2.15
2.35
2.75
3.05
4.4
aEffect levels are based on the critical effects of pulmonary histopathology and inflammation as reported in the
 individual studies. NOAEL: no-observed-adverse-effect level; LOAEL: lowest-observed-adverse-effect level;
 AEL: adverse-effect level; BMCL10: lower 95% confidence estimate of the concentration of DPM associated with a
 10% incidence of chronic pulmonary inflammation (inflam) or fibrosis (see Appendices A and B for more
 specifics).
bLung burdens were derived from data generated from the animal portion of the Yu model using the concentration
 and duration scenario of each study. The human portion of the Yu model was then used to estimate the continuous,
 70-year exposures that would result in this same lung burden, i.e., the HEC.  See Table A-4 in Appendix A and
 accompanying text for further specifics on derivation.
°LD/HD = light-duty/heavy-duty diesel engine.
dThese values are the actual exposure levels used in the Nikula study.  These values were converted into HEC and
 entered into BMC equations to obtain the estimate of the BMCL10 listed. The lung burdens for the two BMCL10s
 listed here were derived by interpolation.
                                                6-12

-------
mg/m3 is estimated to result in a specific lung burden of 15.8 |_ig/cm2. This latter lung burden is
disproportionally elevated compared with the burden estimated to result from exposure to the
lowest concentration. Applying the absolute ratio of lung burden/HEC at the lowest HEC
exposure (i.e., 0.0587/ 0.032 = 1.8) to the highest concentration would result in a lower lung
burden, 4.4 x 1.8 = 7.9 |_ig/cm2, which is much lower than the 15.8 |_ig/cm2 indicated.  This
disproportionate increase in lung burden as a function of DPM concentration would be predicted
from the assumption in the Yu model that the overload phenomena occurs in humans, as is
demonstrated in Figure 3-9 in Chapter 3.  Inspection of Table 6-2 shows that this disproportion
between lung burden and HEC begins to be noticeable around 0.33 mg/m3, at the HEC derived
from the Heinrich et al. (1995) study. HECs below this value are not appreciably influenced by
the overload/disproportionate lung burden phenomenon.
       Inspection of the combined interstudy dose-response continuum in Table 6-2 to elucidate
a point of departure for an RfC entails some interpretation.  Exposures at the lower end of this
table show that elevated chronic exposures to DPM consistently result in AELs. Conversely,
entries in the upper portion of this table show that low-level chronic  exposures to DPM have
minimal, if any, effects within the capability  of these studies to detect them. Intermediate
chronic exposures, from 0.128 mg/m3 to 0.9 mg/m3, are, however, less clear and effect levels and
exposures either have no or few observable effects, or effects that are minimally adverse.
In choosing from among levels (e.g., NO AELs, LOAELs, BMCLxs)  as a POD for derivation of
an RfC, the methodology (U.S. EPA, 1994) provides guidance for choice of a highest no-effect
level below an effect level; the interim guidance for the BMC suggests that for use  as a point of
departure, a benchmark (e.g., BMCL10)  should be within the range of the observable response
data so as to avoid excessive extrapolation, and take the shape of the dose-response curve into
consideration (Barnes et al., 1995; U.S. EPA, 1995). The highest no-effect HECs (NOAE!^)
in this table are 0.128 mg/m3 and 0.144  mg/m3 from the Ishinishi et al. (1988) study, nearly
fivefold above other no-effect levels of 0.032 and 0.038 mg/m3.  The lower BMCL10 (0.37
mg/m3) is at nearly the same concentration as the lowest LOAEL of 0.33 mg/m3 and thus may be
too high an estimate for use as a POD based on these data.  As discussed above, the limitations
on this BMCL10, including excessive extrapolation out of the observable range (see Appendix B
for more specifics), make it a less than optimal candidate for consideration as a POD in the
development of dose-response assessments and therefore was not used for this purpose in this
assessment.  However, this BMCL10 (i.e., at a response rate of 0.1 or 10%) was generated directly
from a modeled dose-response curve for chronic inflammation and lends credence to the other
NO AELs in Table 6-2 as being associated with their respective dose-response curve at incidences
of considerably less than 10%. Moreover, the HECs of less than 0.33 mg/m3 are not appreciably
influenced by the overload phenomenon (see above).  Based on this  analysis, the value of 0.144
                                         6-13

-------
mg/m3 is chosen as the POD for development of the RfC, because it is the highest
among those available.

6.3.4. Uncertainty Factors (UF) for the RfC—A Composite Factor of 30
       Areas of uncertainty designated in the RfC that are relevant to the DPM assessment are
interindividual variability and animal-to-human extrapolation. Each shall be addressed in this
section.
       Considerable qualitative but little, if any, quantitative information exists regarding
subgroups that could be sensitive to any respiratory tract effects of DPM. It is acknowledged that
exposure to DPM could be additive to many other daily or lifetime exposures to airborne organic
compounds and nondiesel ambient PM.  It is also likely that individuals who predispose their
lungs to increased particle retention through smoking or other high particulate burdens, who have
existing respiratory tract inflammation or infections, or who have chronic bronchitis, asthma, or
fibrosis could be more susceptible to adverse impacts from DPM exposure (U.S. EPA, 1996a,
Chapter 5 of this document). Also, infants and children could have a greater susceptibility to the
acute/chronic toxicity of DPM because of their greater breathing frequency and consequent
potential for greater particle deposition in the respiratory tract, which has not reached full
development. Increased respiratory symptoms and decreased lung function in children versus
ambient PM levels, of which DPM is  a part, have been observed (U.S. EPA,  1996a).  Thus, even
though the limited evidence currently available (see Chapter 5) produces no clear evidence that
children are especially sensitive to effects from breathing DPM, the possibility that they actually
may be more susceptible because of their inherent physiology and anatomy should remain a
consideration.  Likewise, a number of factors may modify normal lung clearance, including,
aging, gender, and disease.  It should be noted that the results of Mauderly et al. (1989) discussed
in Chapter 5 indicated that rats with diseased lungs (emphysematous) were no more susceptible
than rats with normal lungs to the effects of DE exposure. Although the exact role of these
factors is not resolved, all would influence the particle dose to the lung tissue from inhalation
exposure. Activity patterns related to occupation and habitation in the proximity of major
roadways are certain to be contributory for some subgroups in receiving higher DPM exposures
(Chapter 2).  In the absence of DE-specific data, this assessment relies on a default UF value of
10 to  account for possible interindividual human variability (U.S. EPA, 1994; Renwick and
Lazarus,  1998).
       Application of an animal-to-human extrapolation or interspecies uncertainty factor to an
assessment may be modified via a number of circumstances. When the assessment is based on
human data, no such UF is necessary. When the assessment is based on animal data, as is the
case with DPM, a default UF of 10 typically is applied to the animal  effect level. This latter
                                          6-14

-------
action implies that the effect observed in the animal study would occur in humans at a 10-fold
lower concentration, ostensibly from some combination of pharmacokinetic and
pharmacodynamic factors that would reflect greater dose (PK consideration) to the human target
or greater sensitivity (PK consideration) of the human tissue.
       The circumstances with DPM warrant modification away from application of the default
UF for animal-to-human extrapolation.  The first circumstance is the extensive effort in this
assessment to address the pharmacokinetic component of the UF.  The point of employing state-
of-the-art lung dosimetry models with specific parameterization for DPM in conversion of
animal exposures to human-equivalent exposures is to derive an estimate of interspecies
pharmacokinetics; to know this aspect of interspecies difference with some degree of certainty.
Having made this informed effort addresses a major portion of the PK component. It is
acknowledged, however, that uncertainties about the model employed here (or any other model)
persist. Although the model comparison shown in Chapter 3 indicates relatively minor
variability in output among the various human models examined (see Table 3-3 and Figure 3-9)
other sources of uncertainty and variability remain.  These include, but are not limited to, matters
such as the estimates if the model were applied to the general population or variability from the
animal portion of the model(s).  A second circumstance involves the pharmacodynamic or PD
component of the interspecies UF, especially the aspect as to whether the experimental animal
species used in the assessment is more or less sensitive than humans. In the consensus report of
ILSI (2000) a specific recommendation is made concerning the PD aspect of the interspecies
uncertainty factor for poorly soluble particles such as DPM. Because the pulmonary responses
from DPM in the principal experimental species, the rat, are present under exposure conditions
that do not appear to elicit any response in humans, the experimental species is considered more
sensitive  than humans. Accordingly, the report suggested that no  accommodation be made for
uncertainty concerning the pharmacodynamic component of the interspecies UF for DPM and
presumably for any other PSP, as the rat appeared to be a sensitive species, more so even than the
human. However, other information currently available on DPM  suggests that, at least with
regard to inflammatory effects, humans may indeed be as sensitive or even more so than rats.
Section 5.1.1.1.3 discusses several studies where humans were exposed to airborne DPM and
either precursors (Salvi et al., 2000; Nordenhall et al., 2000) or markers (Nightingale et al., 2000;
Salvi et al.,1999) of inflammation were detected.  These indicators of inflammation were in
response  to DPM levels of only 200-300 |ag/m3 of l-2h duration. Note that in Table 6-2,
NOAEL concentrations to which rats actually were exposed were only 100-400 pg/m3, clearly
within the range of the aforementioned human exposure levels.  Thus, adverse effects
(inflammation) have been shown to occur in humans at equivalent or possibly even lower levels
                                         6-15

-------
of DPM than observed in rats, indicating that humans may indeed be at least as sensitive if not
more so than rats.
       The sum of these considerations on the animal-to-human UF is that, although major
portions of uncertainty have been addressed, degrees of uncertainty persist in both the
pharmacodynamic and pharmacokinetic components of the factor. In considering both this
residual uncertainty and the information discussed above, it would be prudent to acknowledge
partial degrees of uncertainty in both these areas with a partial uncertainty factor, i.e.,  1005 vice
101, such that a factor of 3 would be applied for interspecies extrapolation.
       In summary, the application of UFs for the two areas discussed above, interhuman and
animal to human, would result in a composite uncertainty factor of 30, 10 for interhuman x 3  for
animal to human. Use of other UFs,  as discussed in the RfC methodology (U.S. EPA, 1994) for
deficiencies in database or for duration extrapolation, is not considered necessary.  It should be
noted that, given the emerging research on DE-induced immunological effects,  it may be
necessary at a later date to reconsider the basis for selection of the critical effect and UFs  and
thus the entire derivation of the DE RfC.

6.3.5. Derivation of the RfC for Diesel Exhaust
       On the basis of the above analysis, the value of 0.144 mg/m3 DPM was  selected as the
point of departure for the RfC evaluation. This value was derived from concentrations in rat
chronic studies that were modeled to obtain HECs. The pulmonary effects, histopathology and
inflammation,  were determined to be the critical noncancer effects. Response data on
inflammation also were suggested by a specific scientific working group as a satisfactory
surrogate for fibrogenic responses in assessing the pulmonary responses of poorly soluble
particles  such as DPM (TLSI, 2000).  Sufficient documentation from other studies showed no
effect in the portal-of-entry tissues, the extrathoracic (nasopharyngeal) region of the respiratory
system, or in other organs at the  lowest levels that produce pulmonary effects in chronic
exposures.  Application of the dosimetric model of Yu et al. (1991) to the exposure value from
Ishinishi  et al.  (1988) of 0.46 mg/m3  16 hr/day, 6 days/wk, aNOAEL, yielded a NOABL^ of
0.144 mg/m3.  Application of the composite UF yields the RfC:
                                    •*• UF = RfC
                    0.144 mg/m3 •*•  30 = O.C048mg/m3  = S^g/m3.
                                          6-16

-------
6.4.  EPIDEMIOLOGICAL EVIDENCE AND NAAQS FOR FINE PM
       Historically, EPA has established primary NAAQS to protect sensitive human population
groups against adverse health effects associated with ambient exposures to certain widespread air
pollutants, including PM, ozone (O3), carbon monoxide (CO), sulfur dioxide (SO2), nitrogen
dioxide (NO2), and lead (Pb).  The U.S. Clean Air Act (the Act) requires that EPA periodically
review and revise as appropriate the criteria (scientific bases) and standards for each pollutant or
class of pollutants (e.g., PM) for which NAAQS have been established. The primary, health-
based NAAQS must be based on the latest scientific information useful in indicating the kind and
extent of all effects on public health expected from the presence of the pollutant in the ambient
air, which is evaluated in a "Criteria Document" (CD).  The NAAQS are then set at levels that, in
the judgment of the EPA Administrator, protect public health (as contrasted with the health of
any individual) with an adequate margin of safety.  In determining the degree of protection that
will  satisfy this mandate, EPA considers the nature and severity of the effects, the types of health
evidence available, the kind and degree of scientific uncertainty that effects would in fact occur
at any particular level of pollution, and the size and nature of sensitive populations at risk  of
experiencing exposures of concern. The EPA develops a staff paper to bridge the gap between
the scientific criteria and the public health policy considerations the Administrator must take into
account in reaching a final judgment. The EPA also must consider the recommendations of the
Clean Air Scientific Advisory Committee (CASAC), an independent committee established by
the Act specifically to advise the Administrator on air quality criteria and NAAQS.  In contrast to
an RfC, the NAAQS are not intended to identify a concentration that is protective against  a
hypothetical continuous lifetime exposure to a given level, but rather take into account expected
actual exposure conditions of U.S. populations.
       The original PM NAAQS were set in 1971 in terms of total suspended particulate matter
(TSP) and included both inhalable and noninhalable particles, ranging in size up to 25-50 |_im.
A later periodic review of the PM criteria and NAAQS led to the setting in 1987 of PM10
NAAQS  (150 |-ig/m3, 24-h average; 50 pg/m3, annual average) aimed at protecting against health
effects associated with those inhalable particles capable of penetrating to lower (thoracic)  regions
of the human respiratory tract and depositing in tracheobronchial and alveolar tissue of the lung
(< 10.0 jam) (52 FR 24634, July 1, 1987). The most recently completed PM NAAQS review was
based on an assessment of the latest available scientific information characterized in the EPA PM
CD (U.S. EPA, 1996a) and additional staff assessments contained in an associated PM Staff
Paper (U.S. EPA, 1996b).  In 1997, on the basis of this information and taking into account
CASAC recommendations and extensive public comments, EPA established new PM25 NAAQS
(15 |-ig/m3,  annual average; 65 pg/m3, 24-h average) to protect against adverse health effects
associated with exposures to fine PM. At the same time, EPA retained, in modified form, the
                                          6-17

-------
PM10 NAAQS originally set in 1987 to protect against effects associated with coarse fraction PM
(62 FR 38652, July 18, 1997).3
       The 1997 PM NAAQS decisions were based, in part, on important distinctions already
highlighted by information present in the PM CD between the fine and coarse fractions of PM10
with regard to size, chemical composition, sources, and transport.  Also of key importance were
the assessment and interpretation of new epidemiological findings on health effects associated
with ambient PM. The epidemiological evidence and basis for the NAAQS for fine PM are
summarized below, followed by a discussion of the relevance of this information for noncancer
assessment of DE.

6.4.1.  Epidemiological Evidence for Fine PM
       The PM CD (U.S. EPA, 1996a) and Staff Paper (U.S. EPA, 1996b) highlighted more than
80 newly published community epidemiologic studies, of which more than 60 found significant
associations between increased mortality and/or morbidity risks and various ambient PM
indicators.  The main findings of concern were community epidemiology results showing
ambient PM exposures to be statistically associated with increased mortality (especially among
people over 65 years of age and those with preexisting cardiopulmonary conditions) and
morbidity (indexed by increased hospital admissions, respiratory symptom rates, and decrements
in lung function).
       Time-series mortality studies reviewed in the 1996 PM CD (U.S. EPA, 1996a) provide
strong evidence that ambient PM air pollution is associated with increases in daily human
mortality and morbidity (e.g., increased hospital admissions and respiratory symptoms). These
studies provided evidence that such effects occur at routine ambient PM levels, extending to 24-h
concentrations below the  150 |_ig/m3 level of the PM10 NAAQS set in 1987.  Overall, as shown in
Table 6-3, the PM10 effects estimates derived from the recent PM10 total mortality studies suggest
that an increase of 50 i-ig/m3 in 24-h average PM10 is significantly associated with an increase in
total mortality, with an RR on the order of 1.025 to 1.05 in the  general population. Table 6-3
also shows higher relative risks for increased hospital admissions for the elderly and for those
with preexisting respiratory conditions, both of which represent subpopulations at special risk for
mortality implications of acute exposures to air pollution, including PM; higher relative risks are
also shown for increased respiratory symptoms and decreased lung function in children. Results
are very similar over a range of statistical models used in the analyses, and are not artifacts of the
methods by which the data were analyzed. Further, these studies suggest a possible linear,
       3At present, the 1997 PM2 5 standards are the subject of ongoing litigation, although they legally remain in
effect, as do the 1987 PM10 standards.

                                          6-18

-------
Table 6-3. Effect estimates per 50 ng/m3 increase in 24-h PM10 concentrations from U.S.
and Canadian studies
Study location
RR (± CI)
only PM
in model
RR (± CI) Reported
other pollutants PM10 levels
in model mean (min/max)t
Increased total acute mortality
Six Cities3
Portage, WI
Boston, MA
Topeka, KS
St. Louis, MO
Kingston/Knoxville, TN
Steubenville, OH
St. Louis, MOC
Kingston, TNC
Chicago, ILh
Chicago, ILg
Utah Valley, UTb
Birmingham, ALd
Los Angeles, CAf

1.04 (0.98, 1.09)
1.06 (1.04, 1.09)
0.98 (0.90, 1.05)
1.03 (1.00, 1.05)
1.05 (1.00, 1.09)
1.05 (1.00, 1.08)
1.08(1.01, 1.12)
1.09 (0.94, 1.25)
1.04(1.00, 1.08)
1.03 (1.02, 1.04)
1.08(1.05, 1.11)
1.05(1.01, 1.10)
1.03 (1.00, 1.055)
—
— 18 (±11.7)
— 24 (±12.8)
— 27 (±16.1)
— 31 (±16.2)
— 32 (±14.5)
— 46 (±32.3)
1.06(0.98,1.15) 28(1/97)
1.09 (0.94, 1.26 30 (4/67)
— 37 (4/365)
1.02(1.01,1.04) 38(NR/128)
1.19(0.96,1.47) 47(11/297)
— 48 (21, 80)
1.02 (0.99, 1.036) 58( 15/177)
Increased hospital admissions (for elderly > 65 yrs.)
Respiratory Disease
Toronto, CAN1
Tacoma, WA
New Haven, CTJ
Cleveland, OHk
Spokane, WA1
COPD
Minneapolis, MNn
Birmingham, ALm
Spokane, WA1
Detroit, MI°
Pneumonia
Minneapolis, MNn
Birmingham, ALm

1.23 (1.02, 1.43)*
1.10(1.03, 1.17)
1.06(1.00, 1.13)
1.06(1.00, 1.11)
1.08(1.04, 1.14)

1.25(1.10, 1.44)
1.13(1.04, 1.22)
1.17(1.08, 1.27)
1.10(1.02, 1.17)

1.08(1.01, 1.15)
1.09(1.03, 1.15)

1.12(0.88,1.36)* 30-39*
1.11(1.02,1.20) 37(14,67)
1.07(1.01,1.14) 41(19,67)
— 43 (19, 72)
— 46 (16, 83)

— 36 (18, 58)
— 45 (19, 77)
— 46 (16, 83)
— 48 (22, 82)

— 36(18,58)
— 45 (19, 77)
                                        6-19

-------
Table 6-3.  Effect estimates per 50 ng/m3 increase in 24-h PM10 concentrations from U.S.
and Canadian studies (continued)
Study location
Spokane, WA1
Detroit, MF
Ischemic HP
Detroit, MP
RR (± CI)
only PM
in model
1.06(0.98, 1.13)
—

1.02(1.01, 1.03)
RR (± CI) Reported
other pollutants PM10 levels
in model mean (min/max)t
— 46 (16, 83)
1.06(1.02,1.10) 48(22,82)

1.02 (1.00, 1.03) 48 (22, 82)
Increased respiratory symptoms
Lower Respiratory
Six Cities'1
Utah Valley, UTr

Utah Valley, IIP
Cough
Denver, COX
Six Cities5
Utah Valley, IIP
Decrease in Lung Function
Utah Valley, UP
Utah Valley, UP
Utah Valley, UTW

2.03 (1.36, 3.04)
1.28 (1.06, 1.56)T
1.01 (0.81, l.Tiy
1.27 (1.08, 1.49)

1.09(0.57,2.10)
1.51(1.12,2.05)
1.29(1.12, 1.48)

55 (24, 86)"
30 (10, 50)"
29(7,51)"*

Similar RR 30(13,53)
— 46(11/195)

— 76(7/251)

— 22 (0.5/73)
Similar RR 30(13,53)
— 76(7/251)

— 46(11/195)
— 76(7/251)
— 55(1,181)
References:
 "Schwartz etal.(1996a).
 bPope et al. (1992, 1994)/O3.
 'Dockery et al. (1992)/O3.
 "Schwartz (1993).
 fKinney et al. (1995)/O3, CO.
 8Ito and Thurston (1996)/O3.
 hStyeretal. (1995).
 'Thurston et al. (1994)/O3.
 Schwartz (1995)/SO2.
 kSchwartz et al. (1996b).
'Schwartz (1996).
"Schwartz (1994e).
"Schwartz (1994f).
"Schwartz (1994d).
"Schwartz and Morris (1995)/O3, CO, SO2.
"Schwartz etal. (1994).
Tope etal. (1991).
Tope and Dockery (1992).
'Schwartz (1994g)
"Pope and Kanner (1993).
"Ostro etal. (1991)
fMin/Max 24-h PM10 in parentheses unless noted
 otherwise as standard deviation (± S.D), 10 and
 90 percentile (10, 90). NR = not reported.
"Children.
"Asthmatic children and adults.
'Means of several cities.
"PEFR decrease in ml/sec.
"FEV! decrease.
*RR refers to total population, not just>65 years.
Source: Adapted from U.S. EPA, 1996b, Tables V-3, V-6, and V-7. See U.S. EPA (1996a,b) for all reference
citations.
                                                      6-20

-------
non-threshold PM/mortality relationship, but the data do not rule out the existence of an
underlying nonlinear, threshold relationship (U.S. EPA, 1996a, 12-310-311; 1996b, VI-16).
Figure 6-2 illustrates the consistency and coherence of the PM10 epidemiology findings for
increased total and cause-specific mortality and morbidity risks in adults and children.  In
addition, Table 6-4 summarizes results from a wide array of U.S. and Canadian studies that
showed increased risks of mortality and morbidity to be related to changes in short-term (24-h)
fine PM (indexed by PM2 5 and other fine particle indicators).
       As summarized below, long-term exposure studies reviewed in the 1996 PM CD (U.S.
EPA, 1996a) also provide evidence of associations between indicators of PM,  including fine
particle indicators, and chronic mortality and morbidity. Table 6-5 shows the direct comparisons
of two key prospective studies of long-term PM mortality: the Harvard Six Cities Study
(Dockery et al., 1993) and the American Cancer Society (ACS) Study (Pope et al., 1995). These
two studies agree in their findings of strong associations between fine particles and increased
mortality.  The RR estimates for total mortality are large and highly significant in the Six Cities
study.  With their 95% confidence intervals, the RR estimate for a 50 |_ig/m3 increase in PM15/10 is
1.42(1.16, 2.01), the RR estimate for a 25 |_ig/m3 increase in PM25 is 1.31 (1.11, 1.68), and the
RR estimate for a 15 i-ig/m3 increase in SO4 is 1.46 (1.16,  2.16).  The ACS study estimates for
total mortality are smaller, but also more precise:  RR =1.17 (1.09, 1.26) for a 25 |ag/m3 increase
in PM25, and RR =1.10 (1.06, 1.16) for a 15 |_ig/m3 increase in SO4.  Both studies used Cox
regression models and were adjusted for similar sets of individual covariates.  In each case,
however, caution must be applied in use of the stated quantitative risk estimates, given that the
lifelong cumulative exposures of the study cohorts (especially in the dirtiest cities) included
distinctly higher past PM exposures than those indexed by the more current PM measurements
used to estimate  long-term PM exposures in the study. Thus, somewhat lower relative risk
estimates than the published ones may well apply.  A third study by Abbey et al. (1991, 1995)
reported no association between long-term PM exposure (indexed by TSP and other estimated
PM indices) after 10 years, although the PM CD (U.S. EPA, 1996a) noted TSP may have been an
inadequate index for exposure to inhalable particles and that additional follow-up might still
reveal chronic effects.
       An additional line of evidence concerning long-term effects may be seen in comparing
cause-specific deaths in the Six Cities and ACS studies.  The relative risks for the most versus
the least polluted cities in the two studies are very similar for mortality from cardiopulmonary
causes (U.S. EPA, 1996b, V-17). These two long-term exposure studies, taken together, suggest
that there may be increases in mortality for specific disease categories that are  consistent with
long-term exposure to ambient fine particles. Moreover, at least some fraction of these deaths is

                                          6-21

-------
 3.
 s
 0)
 o>


2.0-


n ^





]





Legend
X Adults
n All Children


• Symptomatc and/or
Asthmatic Children
£ Range represents 95%
confidence level.


LI
:i_x__* 	 _:







:]si jiixjt






?Ii.





.
-?











-ft*

Adults
II











1

[

i
*





i ~i
i
t J


i


(C.I. = 3.04)

1


i






1
ti


i -p .p
00

Children

           12345   12345     12345
             Total      Respiratory  Cardiovascular
           Mortality     Mortality      Mortality
          678 910    11128  1314    151617     15161718   15161718
          Respiratory   COPD or IHD   Cough      Lower      Upper
            Hospital       Hospital               Respiratory Respiratory
          Admissions    Admissions             Symptoms  Symptoms
            1  Pope etal. (1992) Utah Valley, UT
            2  Schwartz (1993) Birmingham, AL
            3  Styeretal. (1995) Chicago, IL
            4  Ostro etal. (1996) Santiago, Chile
            5  Ito and Thurston (1996) Chicago, IL
            6  Schwartz (1995) New Haven, CT
 7 Schwartz (1995) Tacoma, WA
 8 Schwartz (1996) Spokane, WA
 9 Thurston etal. (1994)Toronto, Canada
10 Schwartz et al. (1996b) Cleveland, OH
11 Schwartz (1994f) Minneapolis, MN
12 Schwartz (1994c) Birmingham, AL
13  Schwartz (1994d) Detroit, Ml
14  Schwartz and Morris (1995) Detroit, Ml
15  Hoekand Brunekreef (1993) The Netherlands
16  Schwartz et al. (1994) Six Cities
17  Pope and Dockery (1992) Utah Valley, UT
18  Pope etal. (1991) Utah Valley, VT
Figure 6-2.  Relative risk (RR) estimates for increased mortality and morbidity endpoints
               associated with 50 ug/m3 increments in PM10 concentrations as derived from
               studies cited by numbers listed above each given type of health endpoint.

Note:   Notice the consistency of RR elevations across studies for given endpoint and coherence of RR estimates
        across endpoints, e.g., higher RR values for symptoms versus hospital admissions and cause-specific
        mortality.
Source: PM Staff Paper (see U.S. EPA, 1996b for full reference citations for each study identified in figure.)
                                                    6-22

-------
Table 6-4. Effect estimates per variable increments in 24-h concentrations of fine particle
indicators (PM2 5, SO^, H+) from U.S. and Canadian studies
Acute mortality
Six Cities3
Portage, WI
Topeka, KS
Boston, MA
St. Louis, MO
Kingston/Knoxville,
TN
Steubenville, OH
Indicator

PM25
PM25
PM25
PM25
PM25
PM25
RR (± CI) per 25 jig/m3
PM increase

1.030 (0.993, 1.071)
1.020 (0.951, 1.092)
1.056(1.038, 1.0711)
1.028 (1.010, 1.043)
1.035 (1.005, 1.066)
1.025 (0.998, 1.053)
Reported PM
levels
mean (min/max)t

11.2 (±7.8)
12.2 (±7.4)
15.7 (±9.2)
18.7 (±10.5)
20.8 (±9.6)
29.6 (±21.9)
Increased hospitalization
Ontario, CANb
Ontario, CANC
NYC/Buffalo, NYd
Toronto*
so;
so;
03
so;
H+ (Nmol/m3)
so;
PM,,
1.03 (1.02, 1.04)
1.03 (1.02, 1.04)
1.03 (1.02, 1.05)
1.05(1.01, 1.10)
1.16(1.03, 1.30)*
1.12(1.00, 1.24)
1.15(1.02, 1.78)
R= 3. 1-8.2
R = 2.0-7.7
NR
28.8 (NR/391)
7.6 (NR, 48.7)
18.6 (NR, 66.0)
Increased respiratory symptoms
Southern California6
Six Citiesf
(Cough)
Six Citiesf
(Lower Resp. Symp.)
so;
PM25
PM2 5 Sulfur
H+
PM25
PM2 5 Sulfur
H+
1.48(1.14, 1.91)
.19(1.01, 1.42)"
.23 (0.95, 1.59)"
.06 (0.87, 1.29)"
.44(1.15-1.82)**
.82 (1.28-2.59)**
.05 (0.25-1.30)**
R = 2-37
18.0 (7.2, 37)"*
2.5(3.1,61)*"
18.1(0.8,5.9)*"
18.0 (7.2, 37)"*
2.5 (0.8, 5.9)"*
18.1(3.1,61)"*
Decreased lung function
Uniontown, PAg
PM25
PEFR23.1 (-0.3, 36.9) (per 25 ng/m3)
25/88 (NR/88)
References:
"Schwartz etal. (1996a)
bBurnettetal. (1994)
cBurnettetal. (1995) O3
dThurston et al. (1992,  1994)
eOstro etal (1993)
fSchwartzetal. (1994)
gNeasetal. (1995)
       x 24-h PM indicator level shown in parentheses unless
 otherwise noted as (± S.D.), 10 and 90 percentile (10,90)
 or R = range of values from min-max, no mean value reported.
'Change per 100 nmoles/m3.
"Change per 20 ng/m3 for PM2 5; per 5 ng/m3 for
 PM2 5 sulfur; per 25 nmoles/m3 for H+.
***50th percentile value (10,90 percentile).
Source: Adapted from U.S. EPA, 1996b, Table V-12.  See U.S. EPA (1996a,b) for all reference citations.
                                               6-23

-------
Table 6-5. Effect estimates per increments" in annual average levels of fine particle
indicators from U.S. and Canadian studies
 Type of health
 effect and location
Indicator
 Change in health indicator
    per increment in PMa
Range of city
  PM levels
mean (ng/m3)
 Increased total chronic mortality in adults
                           Relative risk (95% CI)
 Six Cityb
PM15/10
PM25
so;
       1.42(1.16-2.01)

       1.31 (1.11-1.68)

       1.46(1.16-2.16)
    18-47

    11-30

    5-13
 ACS Study0
 (151 U.S. SMSA)
PM,
       1.17(1.09-1.26)
    9-34

so:
Increased bronchitis in children
Six Cityd
Six City6
24 Cityf
24 Cityf
24 Cityf
24 Cityf
Southern California8
Decreased lung function
Six Cityd'h
Six City6
PM15/10
TSP
H+
SOJ
PM21
PM10
so:
in children
PM15/10
TSP
1.10(1.06-1.16)
Odds ratio (95% CI)
3.26(1.13, 10.28)
2.80(1.17,7.03)
2.65 (1.22, 5.74)
3.02 (1.28, 7.03)
1.97(0.85,4.51)
3.29(0.81, 13.62)
1.39(0.99, 1.92)

NS Changes
NS Changes
4-24

20-59
39-114
6.2-41.0
18.1-67.3
9.1-17.3
22.0-28.6
—

20-59
39-114
 24 Citylj
 24 City1
 24 City1
 24 City1
H+ (52 nmoles/m3)

PM,, (15 ^g/m3)

SO: (7
PM,n (17
-3.45% (-4.87, -2.01) FVC

-3.21% (-4.98,-1.41) FVC

-3.06% (-4.50,-1.60) FVC

-2.42% (-4.30,-.0.51) FVC
"Estimates calculated annual-average PM increments assume: a 100 ng/m3 increase for TSP; a 50 ng/m3
 increase for PM10 and PM15; a 25 ng/m3 increase for PM2 5; and a 15 ng/m3 increase for SO:, except where
 noted otherwise; a 100 nmole/m3 increase forH+.
bDockery et al. (1993).                     gAbbey et al. (1995a,b,c).
Tope et al. (1995).                        hNS Changes = No significant changes.
dDockery et al. (1989).                     'Raizenne et al. (1996).
6Ware et al. (1986).                        JPollutant data same as for Dockery et al. (1996).
TJockery  et al. (1996).
*Range of annual median values for subset of 50 cities.

Source: Adapted from U.S. EPA, 1996a, Table 12-6 and U.S. EPA, 1996b, Table V-8. See U.S. EPA (1996a,b) for
all reference citations.
                                                6-24

-------
likely to be a consequence of cumulative, long-term exposure effects. These effects extend
beyond the additive impacts of short-term exposure episodes, in terms of producing marked
increases above the expected number of daily deaths among especially susceptible groups, such
as the elderly and those with pulmonary disease.
       The PM CD (U.S. EPA, 1996a) also highlighted a growing body of evidence directly
comparing fine and coarse fraction PM effects that suggests that fine particles are more strongly
related than coarse fraction particles to increased mortality and morbidity in both short- and long-
term exposure studies.  Such evidence notably includes the results of analyses of the type
illustrated in Figure 6-3 through 6-5.  More specifically, Figure 6-3 shows a stronger relationship
between changes in short-term (24-h) concentrations of fine particles (indexed by PM25) and
increased mortality risks than for changes in short-term concentrations of coarse fraction particles
(indexed by PM15.2 5). Similarly, a stronger relationship is seen between chronic mortality and
long-term exposure to fine particles (including both the sulfate and nonsulfate components) than
exposure to coarse fraction particles (Figure 6-4), and a much stronger relationship between lung
function decrements and long-term exposure to fine particles than to coarse fraction particles
(Figure 6-5).

6.4.2.  NAAQS for Fine PM
       The health effects evidence discussed above is relevant to this current HAD, as both this
document (Chapter 2) and the PM CD present information that clearly shows DPM to be a
constituent of ambient fine particles.  Therefore, it is reasonable to conclude that DPM  is
associated, but to an undetermined degree, with the health effects described above.  Whereas
broader public health factors are taken into account in setting NAAQS than are relevant for this
noncancer assessment of lifetime exposure for DPM, the annual PM2 5 NAAQS based primarily
on this evidence is of interest in considering the extent to which the RfC for DE (as derived
above in Section 6.3) is concordant with the information on fine particles.
       As presented in the Federal Register final rule notice (62 FR 38652, July 18, 1997), EPA
drew upon the quantitative epidemiology information concisely summarized above to derive a
rationale for selection of an annual-average PM2 5 standard.4  First, to appropriately reflect the
       4As an initial matter, EPA concluded that the existing PM10 standards were not adequate to protect public
health, that fine and coarse fraction particles should be considered separately, that PM2 5 was the appropriate
indicator to use for fine particles, and that an annual PM2 5 standard could provide the requisite reduction in risk
associated with both annual and 24-h averaging times in most areas of the United States. This annual standard,
together with a 24-h standard, could provide supplemental protection against extreme peak fine particle levels that

                                            6-25

-------
                                           Relative Risk for 50 ^g/m3 PM15
                                              in  Six City Acute  Study
                          o
Topeka
Portage
Steubenville
St. Louis
Harriman
Boston
1 A

1 	




1

	 0 	 1
. 	 o — .
K^H
. 	 o 	 ,
•-0—1
                                          0.9
                                                                       1.1
Relative Risk
A
/R.
Topeka
Portage
Steubenville
—
St. Louis
Harriman
Boston
N
slative Risk for 25 Ljg/m3 Fine Particles Relative Risk for 25 Ljg/m3 Coarse Particles
(PM25 ) in Six City Acute Study (Pm15- PM25 ) in Six City Acute Study


h






-O 	 1

I-OH
— 0—
MM
Topeka
Portage
^, Steubenville
0
St. Louis
Harriman
Boston
I 	 0 	
I 	

"—
I —
I — '
H
O 	 1
i 	 0 	

^H
^> 	 1
^H
1




                0.9
                               1             1.1
                           Relative Risk
                                                                   0.9
                                                                                                1.1
Relative Risk
Figure 6-3.  Relative risks of acute mortality in Harvard Six Cities Study, for inhalable
              thoracic particles (PM15/PM10), fine particles (PM2 5), and coarse fraction
              particles (PM15-PM2 5).

Note: The coarse fraction effects are smaller and statistically nonsignificant (i.e., lower 95% confidence intervals do
not exceed relative risk of 1.0), except in Steubenville where there is high correlation between fine and coarse
particles (R2 = 0.69).

Source:  PM CD (U.S. EPA, 1996a) graphical depiction of results from Schwartz et al. (1996).
might occur in some localized situations or in areas with distinct variations in seasonal fine particle levels (62 FR
38652).
                                                 6-26

-------
                                                                    Total Particles
                                                                                          H

                                                                                          W
                                                                                       40  50 60 70  80 90 100
                                                                                        Total Particles, ug/m3
                                            Total Particles
                                  Divided into Inhalable and
                                     Non-lnhalable Particles  I


1.2


1.1


1.0

s

H
L

W

P T


1.2

1
" 1.1
»
Q:
1.0
0.9
S

H
L

W

P T

                      15  20 25 30 35 40 45 50
                         Inhalable Particles, ug/m3
                                                                                     10   20   30   40   50
                                                                                      Non-lnhalable Particles, MQ'ms
                   Inhalable Particles
                     Divided into Fine
                 and Coarse Particles


1.2


1.1

1.0

S

H
L

W
PT


1.2


* 1.1
I
1.0
09
S

H
L

W
P T

                                        10   15   20   25   30
                                            Fine Particles, ug/m3
                      68  10  12  14  16
                         Coarse Particles, |jg/m'
 Fine Particles
  Divided into
  Sulfate and f 1.1
  Non-Sulfate
     Particles
17

1.2


1.1
1.0

S

H
L
W
TP


12

1
I"
1.0
09
S

H
L
W
P T

                  4   6   8   10   12
                    Sulfate Particles, Mg'm!
5  7  9  11  13 15 17
 Non-Sulfate Fine Particles, ug/m'
Figure 6-4.   Adjusted relative risks for mortality are plotted against each of seven long-
               term average particle indices in the Harvard Six Cities Study, from largest
               range (total suspended particles,  upper right) through sulfate and nonsulfate
               fine particle concentrations (lower left).

Note: A relatively strong linear relationship is seen for fine particles, and for sulfate and nonsulfate components.
Topeka, which has a substantial coarse particle component of inhalable (thoracic) particle mass, stands apart from
the linear relationship between relative risk and inhalable particle concentration.
Source: U.S. EPA (1996a) replotting of results fromDockery et al. (1993).
                                                  6-27

-------
                  22 City Fine Mass vs. % Children <85% FVC
oo
V
O
    8-

    7-

    6-

    5-

    4-

    3-

    2-

    1-

    0
            % FVC .85
            Linear (% FVC .85)
 UJ
 60
 V

 >
 LL
    9

    8

    7



    5-




    3-

    2-

    1-
           2     4    6     8    10    12    14    16    18   20    22    24
                                 PM 2.1 (|Jg/m3)


           22 City Coarse Fraction Mass vs. % Children <85% FVC
           % FVC .85
           Linear (% FVC .85)
                                  6         8
                              PM 10-2.1 (|jg/m3)
                                                    10
                                                             12
                                                                       14
Figure 6-5. Percent of children with <85% normal FVC versus annual-
            average fine (PM21) particle concentrations and coarse
            fraction (PM10_21) levels for 22 North American cities.

Note: A much stronger connection appears between fine particles and lung function
decrements (top panel) than for coarse fraction particles (bottom panel).
Source: PM Staff Paper (1996b) graphical depiction of results from Razienne etal. (1996).
                                  6-28

-------
weight of evidence as a whole, EPA concluded that it was appropriate to limit annual PM2 5
concentrations to somewhat below those where the body of epidemiological evidence is most
consistent and coherent, recognizing both the strengths and limitations of the full range of
information on the health effects of PM, as well as associated uncertainties. In accordance with
EPA staff and CAS AC views on the relative strengths of the epidemiologic studies, major
reliance was placed on several short-term (24-h) exposure studies showing significantly
increased risks of daily mortality (Schwartz et al., 1996) and morbidity indexed by hospital
admissions (Thurston et al., 1994) and respiratory symptoms/lung function decrements in
children (Schwartz et al., 1994; Neas et al.,  1995) in relationship to increased fine particle (PM25)
concentrations. Whereas it was recognized that health effects may occur over the full range of
concentrations observed in these studies, it was concluded that the strongest evidence for short-
term PM2 5 effects occurs at concentrations near the long-term (e.g., annual) average. More
specifically, the strength of the evidence of effects increases for concentrations of PM25 that are
at or above the long-term mean levels reported for these studies.  Given the serious nature of the
potential effects, EPA judged that it was both prudent and appropriate to select a level for an
annual standard at or below such concentrations. More specifically, statistically significant
increases in relative risks for daily mortality or morbidity were most clearly observed in these
studies to be associated with 24-h  fine particle concentrations in cities with long-term mean fine
particle concentrations ranging from about 16 to about 21 pg/m3, leading to the judgment that an
annual standard level of 15 pg/m3  would be appropriate.
       Before reaching a final conclusion, the epidemiologic studies of long-term exposures to
fine particles were also considered, which may reflect the accumulation of daily effects over time
as well as potential effects uniquely associated with long-term exposures. Even subject to
additional uncertainties, these studies were judged to provide important insights with respect to
the overall protection afforded by an annual standard. In particular, the annual mean PM2 5
concentrations for the multiple cities included in the two key long-term exposure mortality
studies (Dockery et  al., 1993; Pope et al., 1995) were 18 |ag/m3 and about 21-22 |ag/m3,
respectively, with most of the 50 cities in the Pope,  et al. (1995) having mean PM25
concentrations above 15 i-ig/m3. Taken together with other long-term exposure studies and
considering other factors discussed in the final rule (62 FR 38676, July  18, 1997), EPA
concluded that the concordance of evidence for PM effects and associated levels provides clear
support for an annual standard set at 15 i-ig/m3.
                                           6-29

-------
6.4.3.  DPM as a Component of Fine PM
       Chapter 2 of this document, as well as the PM CD (U.S. EPA, 1996a), report the extent to
which DPM may contribute to ambient PM2 5 concentrations.  In some urban situations, the
annual average fraction of PM2 5 attributable to DPM (according to mass concentrations) is about
35% on the high end, although the proportion appears to be more typically in the range of about
10% (see Chapter 2, Table 2-23 and Section 2.4.2.1).
       An approach to considering the relationship of toxicity between DPM and PM2 5 would be
simply to assume that, as DPM is contributory to the content of ambient PM2 5, so too would it be
contributory to toxicity of PM25. This approach is qualitative only because no firm basis
currently exists for apportioning toxicity among the various components of PM2 5.  Nevertheless,
some qualitative information from laboratory animal studies does exist, showing that DPM is no
more potent at eliciting pulmonary pathology than other poorly soluble particles such as talc,
titanium dioxide, or carbon black in rats, or talc or titanium dioxide in mice. No data suggest
that DPM is any more potent in eliciting pulmonary pathology than any other poorly soluble
particle that typically may be present in ambient PM2 5.  It may be reasonable to suggest, then,
that DPM is no more likely to be lexicologically potent than any other fine particle constituents
that typically make up ambient PM2 5.
       Based on the foregoing aspects of such an approach, a conclusion could be drawn that as
long as DPM constituted its current approximate proportion to PM2 5, the annual PM2 5 standard
would also be expected to provide a measure of protection for DPM.  Even if a basis did exist to
apportion toxicity among the various components of ambient PM2 5, such as DPM, use of such
information in an approach to derive a safe air level for DPM would result in only a generalized,
nonspecific estimate limited by a variety of factors including the accuracy of the apportionment
of DPM from PM2 5.  The RfC derived in Section 6.3 was based on an approach that utilized
toxicological information from actual DPM exposures,  a more direct approach that would result
in a more specific estimate not limited by any apportionment scheme.

6.5.  CHARACTERIZATION OF THE NONCANCER ASSESSMENT FOR DIESEL
     EXHAUST
       Adverse health effects from short-term acute (high-level) exposures to DE such as
occupational reports of decreases in lung function, wheezing, chest tightness, increases in airway
resistance, and reports in laboratory animals of inflammatory airway changes and lung function
changes are acknowledged but are not assessed quantitatively.  The focus of this dose-response
assessment is on the adverse noncancer health consequences of a lifetime, low-level, continuous
air exposure by humans to DE.

                                          6-30

-------
       This assessment uses the whole particle, termed DPM, as the key index or measure of DE
dose. DPM includes any and all adsorbed organics, among which are a large number of PAHs,
heterocyclic compounds, and their derivatives (Chapter 2), as well as the carbon core.  It is not
possible to separate the carbon core of DPM from the adsorbed organics to compare the toxicity
in exposures other than with limited in-vitro-type scenarios.  The dosimetric model used in the
derivation of the RfC (Yu et al., 1991) is consistent with this designation, as it considers DPM as
well as the adsorbed organics as two types, slow-cleared and fast-cleared.  Studies with diesel do
occasionally report levels of accompanying gaseous components of DE (e.g., NOX, CO), but
nearly all report particle concentration and characteristics.
       Adverse responses occurring in the rat lung have been used in this assessment as the basis
for characterizing nonneoplastic human lung responses, yet use of these data in hazard evaluation
for cancer is not considered relevant to humans.  The basis for this use of these noncancer
pulmonary effects in rats for derivation of an RfC includes the fact that humans and rats exhibit
similar responses to other poorly soluble particles and also that similar noncancer effects are seen
in other species (ILSI, 2000; Freedman and Robinson, 1988). Thus,  when viewed across species
(including humans), the nonneoplastic pulmonary effects of inflammation and fibrosis used in
this assessment are dissociable from the cancer response and are of likely relevance to humans.
       As a part of the RfC methodology (U.S. EPA, 1994), dose-response assessments are
assigned levels of confidence that are intended to reflect the strengths and limitations of an
assessment as well as to indicate the likelihood of the assessment changing with any additional
information. Confidence levels of either low, medium, or high are assigned both to the study (or
studies) used in the assessment to characterize the critical effects and to the overall toxicological
database of the substance.  An overall confidence level also is assigned to the entire assessment.
Usually, it is the  same, or in any case no higher than the level assigned to the database.
       Compared with the databases of most other toxicants, the basic toxicological database for
DE is substantial. The critical effects are characterized using not one but multiple long-term
chronic studies conducted independently of one another (Tables 6-1  and 6-2). The exhaustive
manner in which these studies were conducted and reported also imparts a high degree of
confidence. Both developmental and reproductive areas are addressed.  Also, ancillary studies
that address mechanistic aspects of DE toxicity, either as the whole particle with adsorbed
organics, or segregated as a poorly soluble particle and extracted organics, are available and used
in this assessment. Although only limited human data are available,  extensive consideration has
been given to the relevancy of the animal studies to the human condition.  On the other hand,
data from related toxicants such as general ambient PM indicate effects in endpoints (e.g.,
cardiovascular measures) that have not been addressed in the DPM database. A major point to

                                          6-31

-------
consider in assigning confidence in this assessment, and a reason that the value of the RfC may
change in the future, is the emerging issue of allergenicity caused or exacerbated by DE.
Although information to evaluate allergenicity in parallel to the present effects (pulmonary
inflammation and histopathology) is currently lacking, future efforts to elucidate and characterize
this effect may well be a driver to make a reevaluation of the noncancer RfC derivation for DE
appropriate. With respect to the current RfC for DE, the confidence level is medium, both for the
database and overall.  The level reflects the relevance of (and information lacking on)
allergenicity effects associated with DE in humans, and the possibility that the current RfC could
change as a consequence of this information becoming available from the scientific community.
       In the introductory portion of this chapter, DPM is acknowledged as a constituent of
ambient PM (U.S. EPA, 1996a,b). A discussion of the quantitative epidemiology, particularly
regarding fine PM, indicated that public health effects, including premature mortality, increased
hospital admissions, respiratory symptoms, and decreased lung function, were observed in
populations living in areas with long-term mean PM25 levels generally ranging above 15 |_ig/m3.
Application of the RfC method, which involved critical consideration of the entirety of the
disparate DE database with many chronic studies from several different species, evaluation of a
myriad of possible DE-specific toxicological endpoints, and use of extrapolation models,
produced a value of 5  i-ig/m3.  As the accuracy of the RfC is stated in the definition ( "...within an
order of magnitude ...  "), this dose-response estimate could be considered to be not different from
the level of 15 pg/m3,  the lower end of the range identified for PM2 5. It is acknowledged here
again that the levels of the PM25 NAAQS should not be considered as indicative of the  same
degree of health protection for DE as intended by the RfC. Nevertheless, the congruence of these
estimates tends to enhance the overall confidence that this range of levels is near or inclusive of
those that would be expected to be protective of the human population  against the health effects
ofDE.

6.6.  SUMMARY
       Table 6-6 summarizes the key data and factors used in the dose-response analysis leading
to the derivation of the RfC for DE.  The DE RfC of 5 |_ig DPM/m3 is a chronic exposure likely to
be without  an appreciable risk of adverse human health effects.
       The link between ambient fine PM and DPM with respect to origin, content, and possible
health effects has been presented and discussed in this chapter, and the general congruence
between the DE RfC and the level of the annual NAAQS for fine particles has been noted.
Although these values should not be compared directly, it is reasonable to observe that the annual
PM25 standard would be expected to provide a measure of protection for DPM, reflecting DPM's

                                          6-32

-------
       Table 6-6. Decision summary for the quantitative noncancer RfC assessment for
Quantitative assessment for noncancer effects from
lifetime exposure to DPM
                                                                       5 |ig/m3
Critical effect

Principal study
Designated basis for quantitation (exposures in rats)

NOAELnEc (HEC)
Adjustments for uncertainty factors (interspecies
variability  and intraspecies extrapolation)
               = RfC
                                                          Pulmonary inflammation and
                                                          histopathology in rats
                                                          Array of four chronic rat studies
                                                          0.46 mg DPM /m3, 16 hr/day, 6
                                                          d/wk, 130wks;  aNOAEL
                                                          0. 144 mg DPM / m3
                                                          30
                                                           0.144 mg/m3 / 30 = 5
current approximate proportion to PM2 5.
       The estimated air concentration of 5 |_ig/m3 (the RfC, a lifetime exposure to DE measured
as DPM) is above the ambient air levels reported in most rural areas but could be below those
levels reported under short-term conditions in some urban scenarios, such as at busy intersections
or bus stops (see Chapter 2, Table 2-23). The RfC is intended to address lifetime chronic
exposures and aspects of time-averaging for less than lifetime scenarios, such as, for example,
acute exposures at busy intersections or bus stops, which are not addressed in this particular
assessment.

                            REFERENCES FOR CHAPTER 6
Abbey, DE; Mills, PK; Petersen, FF; et al. (1991) Long-term ambient concentrations of total suspended particulates
and oxidants as related to incidence of chronic disease in California Seventh-Day Adventists. Environ Health
Perspect 94:43-50.
Abbey, DE; Hwang, BL; Burchette, RJ; et al. (1995) Estimated long-term ambient concentrations of PM10 and
development of respiratory symptoms in a nonsmoking population. Arch Environ Health 50:139-152.
Barnes, DG; Dourson, M. (1988) Reference dose (RfD): description and use in health risk assessments. Regul
Toxicol Pharmacol 8:471-486.
Barnes, DG; Daston, GP; Evans, JS; et al.  (1995) Benchmark dose workshop: criteria for use of a benchmark dose to
estimate a reference dose. Regul Toxicol Pharmacol 21:296-306.
                                            6-33

-------
Barnhart, MI; Chen, S-T; Salley, SO; et al. (1981) Ultrastructure and morphometry of the alveolar lung of guinea
pigs chronically exposed to diesel engine exhaust: six months' experience. J Appl Toxicol 1:88-103.

Barnhart, MI; Salley, SO; Chen, S-T; et al. (1982) Morphometric ultrastructural analysis of alveolar lungs of guinea
pigs chronically exposed by inhalation to DE (DE). In: Toxicological effects of emissions from diesel engines:
proceedings of the Environmental Protection Agency diesel emissions symposium; Lewtas, J., ed. October, 1981;
Raleigh, NC. New York: Elsevier Biomedical; pp. 183-200. (Developments in toxicology and environmental
science: v. 10).

Bhatnagar, RS; Hussain, MZ; Sorensen, KR; et al. (1980) Biochemical alterations in lung connective tissue in rats
and mice exposed to diesel emissions. In: Health effects of diesel engine emissions: proceedings of an international
symposium. Pepelko, WE; Banner, RM; Clarke, NA., eds., v. 1; December 1979; Cincinnati, OH. Cincinnati, OH:
U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp. 557-570; EPA report no.
EPA-600/9-80-057a.  Available from: NTIS, Springfield, VA; PB81-173809.

Dockery, DW; Pope,  CA, III; Xu, X; etal. (1993) An association between air pollution and mortality in six U.S.
cities. N Engl J Med 329:1753-1759.

Freedman, AP; Robinson, SE. (1988) Noninvasive magnetopneumographic studies of lung dust retention and
clearance in coal miners. In: Respirable dust in the mineral industries: health effects, characterization, and control:
proceedings of the international symposium on respirable dust in the mineral industries; October 1986; University
Park, PA. Frantz, RL; Ramani, RV., eds. University Park, PA:  Pennsylvania State University Press; pp. 181-186.

Heinrich, U; Peters, L; Funcke, W; et al. (1982) Investigation of toxic and carcinogenic effects of DE in long-term
inhalation exposure of rodents. In: Toxicological effects of emissions from diesel engines: proceedings of the
Environmental Protection Agency diesel emissions symposium; October 1981; Raleigh, NC. Lewtas, J., ed. New
York: Elsevier Biomedical; pp. 225-242. (Developments in toxicology and environmental science: v. 10).

Heinrich, U; Muhle, H; Takenaka, S; et al. (1986) Chronic effects on the respiratory tract of hamsters, mice, and rats
after long-term inhalation of high concentrations of filtered and unfiltered diesel engine emissions. J Appl Toxicol
6:383-395.

Heinrich, U; Fuhst, R; Rittinghausen, S; et al. (1995) Chronic inhalation exposure of Wistar rats and two different
strains of mice to diesel engine exhaust, carbon black, and titantium dioxide. Inhal Toxicol 7:533-556.

Henderson, RF; Pickrell, JA; Jones, RK; et al. (1988) Response of rodents to inhaled diluted DE: biochemical and
cytological changes in bronchoalveolar lavage fluid and in lung tissue. Fundam Appl Toxicol 11:546-567.

Hyde, DM; Plopper, CG; Weir, AJ; et al. (1985) Peribronchiolar fibrosis in lungs of cats chronically exposed to DE.
Lab Invest 52:195-206.

International Life Sciences Institute Risk Science Institute Workshop Participants (ILSI). (2000) The relevance of
the rat lung response to particle overload for human risk assessment: a workshop consensus report. In: ILSI Risk
Science Institute Workshop: the relevance  of the rat lung response to particle overload for human risk assessment;
March 1998. Gardner, DE., ed. Inhal Toxicol 12:1-17.

Ishinishi, N; Kuwabara, N; Nagase, S; et al. (1986) Long-term inhalation studies on effects of exhaust from heavy
and light duty diesel engines onF344 rats.  In: Carcinogenic and mutagenic effects of diesel engine exhaust:
proceedings of the international satellite symposium on lexicological effects of emissions from diesel engines; July;
Tsukuba Science City, Japan. Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. Amsterdam, Holland: Elsevier
Science Publishers  B.V.; pp. 329-348. (Developments in toxicology and environmental science: v. 13).
                                                  6-34

-------
Ishinishi, N; Kuwabara, N; Takaki, Y; et al. (1988) Long-term inhalation experiments on DE. In: DE and health
risks: results of the HERP studies. Tsukuba, Ibaraki, Japan: Japan Automobile Research Institute, Inc., Research
Committee for HERP Studies; pp. 11-84.

Iwai, K; Udagawa, T; Yamagishi, M; et al. (1986) Long-term inhalation studies of DE on F344 SPF rats. Incidence
of lung cancer and lymphoma. In: Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the
international satellite symposium on lexicological effects of emissions from diesel engines; July; Tsukuba Science
City, Japan. Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. Amsterdam, Holland: Elsevier Science Publishers
B. V; pp. 349-360. (Developments in toxicology and environmental science: v. 13).

Johnston, CJ;  Finkelstein, JN; Mercer, P, et al. (2000) Pulmonary effects induced by ultrafine PTFE particles.
Toxicol Appl  Pharmacol 168:208-215.

Lewis, TR; Green, FHY; Moorman, WJ; et al. (1989) A chronic inhalation toxicity study of diesel engine emissions
and coal dust, alone and combined. J Am. Coll Toxicol 8:345-375.

Mauderly, JL; Jones, RK; Griffith, WC; et al.  (1987a) DE is a pulmonary carcinogen in rats exposed chronically by
inhalation. Fundam. Appl Toxicol 9:208-221.

Mauderly, JL; Bice, DE; Carpenter, RL; et al.  (1987b) Effects of inhaled nitrogen dioxide and DE on developing
lung. Cambridge, MA: Health Effects Institute; research report no. 8.

Mauderly, JL; Bice, DE; Cheng, YS; et al. (1989) Influence of experimental pulmonary emphysema on lexicological
effects from inhaled nitrogen dioxide and diesel exhaust Cambridge, MA: Heallh Effecls Inslilule; report no. HEI-
RR-89/30. Available from: NTIS, Springfield, VA; PB90-247347.

National Research Council (NRC). (1983) Risk assessmenl in Ihe federal government managing Ihe process.
Washington, DC: National Academy Press.

Neas, LM; Dockery, DW; Koulrakis, P; el al.  (1995) The association of ambienl air pollution wilh twice daily peak
expiratory flow rale measuremenls in children. Am. JEpidemiol 141:111-122.

Nightingale, JA; Maggs, S;  Cullinan, P; el al.  (2000) Airway inflammation after conlrolled exposure to diesel
exhausl particulales.  AmJ Respir Cril Care Med  162(1):161-166.

Nikula, KJ; Snipes, MB; Barr, EB; el al. (1995) Comparative pulmonary loxicities and carcinogenicilies of
chronically inhaled DE and carbon black in F344 rals. Fundam Appl Toxicol 25:80-94.

Nordenhall, C; Pourazar,  J; Blomberg,  A;  (2000) Airway inflammation following exposure to diesel exhausl: a sludy
of time kinetics using induced spulum. Eur Respir J 15(6):1046-1051.

Oberdorsler, G; Finkelslein, JN; Johnston, CJ; (2000) Ullrafine particles as inducers of acule lung injury;
mechanisms and correlation wilh disease and age. Heallh Effecls Inslilule 96:1-88.

Pepelko, WE. (1982a) Effecls of 28 days exposure to diesel engine emissions in rals. EnvironRes 27:16-23.

Pepelko, WE. (1982b) EPA sludies on Ihe lexicological effecls of inhaled diesel engine emissions. In: Toxicological
effecls of emissions from diesel engines: proceedings of Ihe Environmental Protection Agency diesel emissions
symposium; October 1981;  Raleigh, NC. Lewlas, J., ed. New York: Elsevier Biomedical; pp. 121-142.
(Development in toxicology and environmental science: v. 10).

Plopper, CG;  Hyde, DM; Weir, AJ. (1983) Cenlriacinar alterations in lungs of cals chronically exposed to DE. Lab
Invesl 49:391-399.
                                                 6-35

-------
Pope, CA, III; Thun, MZ; Namboodiri, MM; et al. (1995) Paniculate air pollution as a predictor of mortality in a
prospective study of U.S. adults. Am J Respir Crit Care Med 151:669-674.

Raizenne, M; Neas, LM; Damokosh, AI; et al. (1996) Health effects of acid aerosols on North American children:
pulmonary function. Environ Health Perspect 104:506-514.

Renwick, AG; Lazarus, NR. (1998) Human variability and noncancer risk assessment—an analysis of the default
uncertainty factor. Regul Toxicol Pharmacol 27:3-20.

Salvi, S; Blomberg,A; Rudell, B; et al. (1999) Acute inflammatory responses in the airways and peripheral blood
after short-term exposure to diesel exhaust in healthy human volunteers. Am J Respir Crit Care Med 159 (3):702-
709.

Salvi, SS; Nordenhall,C; Blomberg, A; et al. (2000) Acute exposure to diesel exhaust increases IL-8 and GRO-alpha
production in healthy human airways. Am  J Respir Crit Care Med 161:550-557.

Schwartz, J; Dockery, DW;  Neas, LM; Wypij, D; Ware,  JH; Spengler, JD; Koutrakis, P; Speizer, FE; Ferris, BG, Jr.
(1994) Acute effects of summer air pollution on respiratory symptom reporting in children. Am. J. Respir. Crit. Care
Med. 150: 1234-1242.

Schwartz, J; Dockery, DW; Neas, LM. (1996) Is daily mortality associated specifically with fine particles? J Air
Waste Manage Assoc 46:927-939.

Thurston, GD; Ito, K; Hayes, CG; et al. (1994) Respiratory hospital admissions and summertime haze air pollution in
Toronto,  Ontario: consideration of the role of acid aerosols. EnvironRes 65:271-290.

U.S. Environmental Protection Agency (U.S. EPA). (1994) Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry [draft final]. Research Triangle Park, NC: Office of Health
and Environmental Assessment, Environmental Criteria and Assessment Office; report no. EPA/600/8-88/066F.

U.S. EPA. (1995) The use of the benchmark dose approach in health risk assessment. Washington, DC: U.S. EPA,
Risk Assessment Forum; report no. EPA/630/R-94/007. Available from: NTIS, Springfield, VA;
PB95-213765/XAB.

U.S. EPA. (1996a) Air quality criteria for paniculate matter. Research Triangle Park, NC: National Center for
Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.

U.S. EPA. (1996b) Review of the national ambient air quality standards for participate matter: policy assessment of
scientific and technical information. OAQPS staff paper. Research Triangle Park, NC:  Office of Air Quality
Planning and Standards; report no. EPA/452/R-96-013. Available from: NTIS, Springfield, VA; PB97-115406REB.

Vostal, JJ; Chan, TL; Garg, BD; et al. (1981) Lymphatic transport of inhaled diesel particles in the lungs of rats and
guinea pigs exposed to diluted DE. Environ Int 5:339-347.

Yu, CP; Yoon, KJ. (1990) Retention modeling of DE particles in rats and humans. Amherst, NY: State University of
New York at Buffalo (Health Effects Institute research report no. 40).

Yu, CP; Yoon, KJ; Chen, YK. (1991) Retention modeling of DE particles in rats and humans. J Aerosol Med
4:79-115.
                                                 6-36

-------
                    7. CARCINOGENICITY OF DIESEL EXHAUST

7.1.  INTRODUCTION
       Initial health hazard concerns regarding the potential carcinogenicity of diesel engine
exhaust (DE) were based on the reported induction of skin papillomas by diesel particle extracts
(Kotin et al., 1955), evidence for mutagenicity of extracts (Huisingh et al.,  1978), evidence that
components of diesel extract act as weak tumor promoters (Zamora et al., 1983), and the
knowledge that diesel particles and their associated organics are respirable. During the 1980s,
both human epidemiologic studies and long-term animal cancer bioassays were initiated.  In
1981, Waller published the first epidemiologic investigation, a retrospective mortality  study of
London transport workers.  Since then a large number of retrospective cohort and case-control
studies have been carried out with railroad workers, dockworkers, truck drivers, construction
workers, miners, and bus garage employees. During 1986 and 1987, several chronic animal
cancer bioassays were published. These studies and numerous laboratory investigations carried
out since then have been directed toward assessing the carcinogenic potential of whole exhaust,
evaluating the importance of various exhaust components in the induction of cancer, and
understanding the mode of action and implications of deposition, retention, and clearance of DE
particles.

7.1.1.  Overview
       This chapter evaluates the carcinogenic potential of DE in both humans (Section 7.2) and
animals (Section 7.3), discusses mode(s) of action (Section 7.4), and provides an overall weight-
of-evidence evaluation (Section 7.5) for carcinogenicity in humans.  This chapter also
summarizes evaluations of DE conducted by other organizations (Section 7.6) and the  final
conclusions (Section  7.7) identify major uncertainties for which additional research is  needed.
This assessment focuses on DE, although it should be noted that diesel particles make  up a
portion of ambient particulate matter (PM)  (Chapter 2,  Section 2.2.3; Chapter 6,  Section 6.4.3),
and thus, the ambient PM data may have some relevance.

7.1.2. Ambient PM-Lung Cancer Relationships
       A brief overview of the data regarding exposure to ambient PM and lung cancer is
provided as background information and is  based on analyses contained in the 1996 Air Quality
Criteria for PM (PM  CD) (U.S. EPA, 1996a).J With DE being part of ambient PM, the question
       JAs noted in Chapter 6, a new PM CD is now being prepared to reflect the latest scientific studies on
ambient PM available since the last document was completed.

                                           7-1

-------
of what is seen in the ambient PM data is of interest, as epidemiologic evidence for an effect of
ambient PM on lung cancer mortality or incidence could possibly contribute to evaluation of
DE-specific epidemiologic data.
       Chapters 2 and 5 noted that DPM, consisting mostly of fine particles (<1.0 mm
diameter), represents a lexicologically important component of typical ambient fine particle
mixes. As discussed in Chapter 6, several large-scale prospective studies (Harvard Six Cities
Study; American Cancer Society (ACS) Study; Adventist Health Study of Smog (AHSMOG)
provide important evidence regarding associations between chronic exposures to ambient fine
particles and increased risks of noncancer mortality/morbidity effects (e.g., cardiorespiratory-
related deaths or hospital admissions) (U.S. EPA, 1996a).  As summarized below, these same
studies also evaluated relationships between chronic PM exposures and lung cancer mortality
and/or incidence.
       As an initial matter, both the Harvard  Six Cities Study (Dockery et al., 1993), of
approximately 8,000 adults in six cities comprising a transect across the northcentral and
northeastern United States, and the ACS Study (Pope et al., 1995), of 550,000 adults in 151
cities across all U.S. geographic regions, found markedly increased relative risks (RR) of lung
cancer mortality associated with smoking. More specifically, the Six City Study reported
increased risks of smoking for current (RR =  8.00, 95% CI = 2.97-21.6) and former (RR = 2.54,
CI = 0.90-7.18) smokers, with the ACS Study reporting striking similar increased risks for
current smokers (RR = 9.73, 95%  CI = 5.96-15.9).
        After  controlling for smoking and other risk factors, both the Six Cities  Study and the
ACS Study (using a subset of 50 of the 151 cities) evaluated relationships between long-term
exposure to fine PM (indexed by PM25), from the least to the most polluted of the cities in each
study, and lung cancer mortality.  In both studies, lung cancer mortality risks were not
statistically significantly associated with ambient PM2 5 concentrations in combined analyses of
data for both males and females (RR = 1.37, 95% CI = 0.81-2.31, in the Six Cities Study; RR =
1.03, 95% CI = 0.80-1.33, in the ACS Study). Also, lung cancer mortality risks were not
statistically significantly associated with ambient PM2 5 concentrations in the ACS Study for
smaller sample size subgroups broken out by  sex and smoking status. In addition, analyses of
data from the AHSMOG series of studies, of 6,338 nonsmoking long-term California adult
residents, found no statistically significant associations between PM2 5 (estimated from visibility
data) and lung cancer mortality or total mortality (Abbey et al., 1995); further, no such
associations were reported for PM10 (estimated from total suspended particulate matter [TSP]
data) in the same study.  Earlier AHSMOG analyses (Abbey et al., 1991) reported no statistically
significant associations between TSP  (which includes not only fine PM but also larger coarse-
                                           7-2

-------
mode particles ranging up to 25-50 |lm) and respiratory cancer for either sex (only respiratory
symptoms and any-site female cancers were reported to be associated with TSP in this study).
       The ACS Study and the later AHSMOG analyses (Abbey et al., 1995) also evaluated
relationships between long-term exposures to sulfates (SO4) (which are predominantly but not
exclusively found in fine-mode particles, and can be considered an index for ambient fine
particles) and lung cancer mortality.  The ACS Study reported somewhat elevated and
statistically significant lung cancer risk (RR = 1.36,95%CI = 1.11-1.66) across 151 cities in
combined analyses of data for both males and females.  However, in further analyses of
subgroups broken out by sex and smoking status (and thus having smaller sample sizes in each
than for the above overall combined analyses), only the lung cancer mortality risks for male
"ever-smokers" (RR = 1.44, 95% CI = 1.14-1.83) were statistically significant; no statistically
significant relationships were reported  for male "never-smokers" (RR = 1.36, 95% CI = 0.40 -
4.66), for female "ever-smokers" (RR =1.10, 95% CI = 0.72-1.68), or for female "never-
smokers" (RR =1.61; 95% CI = 0.66 -3.92). In the later AHSMOG analyses, Abbey et al.
(1995) found no statistically significant associations between  sulfates and lung cancer or total
mortality.
       In summary, the three key prospective cohort studies summarized above, and discussed
in more detail in the 1996 PM CD (U.S. EPA, 1996a), provide an equivocal array of results with
regard to possible associations between chronic exposures to ambient PM and lung cancer
mortality and/or incidence. None of the analyses of fine particles (as indexed by PM25) in these
three studies reported statistically significant relationships between long-term PM2 5
concentrations and lung cancer mortality. Only the ACS Study found a statistically significant
association of increased risk of lung cancer with one indicator of ambient fine particles
(sulfates).  Overall, then, these studies support a conclusion that there continues to be little
epidemiologic evidence for an effect of ambient PM on lung cancer mortality or incidence. It is
recognized, however, that subsequent AHSMOG analyses and other studies, published since
completion of the 1996 PM CD, have further analyzed relationships between ambient PM and
lung cancer. Results from these more recent studies are now being evaluated as part of the
integrated assessment of ambient PM that will be part of the new PM CD targeted for
completion in 2002.

7.2.  EPIDEMIOLOGIC STUDIES OF THE CARCINOGENICITY OF
     EXPOSURE TO DIESEL EXHAUST
     An increased risk from malignancies  of the lung, bladder, and lymphatic tissue has been
reported in populations potentially exposed to higher levels of DE than typically seen in the
environment.  A few authors have reported other malignancies, including testicular cancer
                                          7-3

-------
(Garland et al., 1988), gastrointestinal cancer (Balarajan and McDowall, 1988; Guberan et al.,
1992), and prostate cancer (Aronsen et al., 1996).  A detailed review of 22 lung cancer studies is
presented in this section; a few more studies exist, but these 22 are judged to be the key ones. A
detailed review of other health effect studies is not presented because findings are equivocal.
     Excess risk of bladder cancer has been reported in several studies (Howe et al., 1980;
Wynder et al., 1985; Hoar and Hoover et al., 1985; Silverman et al., 1983; Vineis and Magnani
1985; Silverman et al., 1986; Jensen et al., 1987;  Steenland et al., 1987; Isocovich et al., 1987;
Risch et al.,  1988; Iyer et al., 1990;  Steineck et al., 1990; Cordier et al., 1993; Notani et al.,
1993). Very few studies found significant excesses after adjustment for cigarette smoking.
Most studies failed to show  any association between exposure to DE and occurrence of bladder
cancer.  Some authors have  reported excess mortality from lymphohematopoietic system cancers
in people potentially exposed to diesel fumes.  Rushton and Alderson (1983) and Howe and
Lindsay (1983) found increased mortality from lymphatic neoplasms.  Balarajan and McDowall
(1983) found raised mortality for malignant lymphomas. Flodin et al. (1987) observed increased
risk for multiple myeloma, and Bender et al. (1989) reported excess mortality from leukemia.
Because evidence for bladder cancer and lymphohematopoietic cancer was found to be
equivocal, detailed reviews  of these studies are not presented here.
     The potential for elevated DE  exposure in the occupational setting generally includes
miners, railroad workers,  truckers, bus and taxi drivers, heavy equipment operators, farm tractor
drivers, and those involved with heavy duty marine engines. Regarding the mining industry
some assert that excess lung cancer  should be observed in the miners if exposure to DE is
causally associated with the occurrence of lung cancer since DE is allegedly present in the
mines. Our review of the mining industry data dose not support this assertion for the following
reasons.  In the United States, the introduction of diesel engines into metal mines dates from the
early to the mid 1960s.  Currently, there are approximately 265 underground metal/nonmetal
mines in the United States.  Virtually all of these mines use diesel powered equipment for
various tasks, such as haulage, roof bolting etc. (Department of Labor, Mine Safety and Health
Administration, 2001). Introduction of diesel equipment into coal mines was even later.  Of 910
existing underground coal mines in  the United States, only 145 currently use diesel-powered
equipment. Of these 145  mines, 32 mines are currently using diesel equipment for face coal
haulage.  The remaining mines use diesel equipment for transportation, materials handling, and
other support operations (Department  of Labor, Mine Safety and Health Administration, 2001).
It should be noted that there is a paucity of epidemiologic studies in miners where exposure to
DE and health effects are explored.  Furthermore,  the majority of epidemiologic studies in
miners do not mention exposure to diesel equipment use.  Thus, it is impossible to know how
many miners were exposed  to DE and for how long and at what concentrations in a given study,
                                          7-4

-------
if any. Hence the studies in miners (coal, metal, and nonmetal with the exception of potash
miners) are not reviewed in this chapter, because the available studies are uninformative relative
toDE.
     In this section, various mortality and morbidity studies of lung cancer from potential
exposure to diesel engine emissions are reviewed. Although an attempt was made to cover all
the relevant studies, a number of studies are not included for several reasons. In the United
States the change from steam to diesel engines in locomotives began after World War II.  By
1946 about 10% of the locomotives in service were diesel, by  1952 55%  were diesel, and
dieselization was about 95% complete by 1959 (Garshick et al.,  1988). Therefore, exposure to
DE was less common, and the follow-up period for studies conducted prior to 1960 (Raffle,
1957; Commins et al., 1957; Kaplan, 1959) was not long enough to cover the long latency
period of lung cancer. The usefulness of these studies in evaluating the carcinogenicity of DE is
greatly reduced; thus, they are not considered here.
     On the other hand, the trucking industry changed to diesel trucks by the 1960s.  In the
1960s sales of diesel-powered Class 8 trucks (long-haul trucks) were 48% of the market, and by
the 1970s sales had risen to 85%.  Thus, studies conducted among truck drivers prior to the
1970s may reflect exposures to gasoline exhaust as well  as DE. Hence, studies with ambiguous
exposures or studies that examined several occupational  risk factors were excluded because they
would have contributed little to the evaluation of the carcinogenicity of DE (Waxweiler et al.,
1973; Williams et al., 1977; Ahlberg et al., 1981; Stern et al.,  1981; Buiatti et al.,  1985;
Gustafsson et al., 1986; Siemiatycki et al., 1988). A study by  Coggon et al. (1984) was excluded
because occupational information  abstracted from death  certificates had not been validated; this
would have resulted in limited information.
     Several types of studies of the health effects of exposure to diesel engine emissions are
reviewed in this chapter, such as cohort studies, case-control studies, and studies that conducted
meta-analysis. In the cohort studies, cohorts of heavy  construction equipment operators, railroad
and locomotive workers, bus garage employees,  and miners were studied retrospectively to
determine increased mortality and morbidity resulting  from exposures to varying levels of diesel
emissions in the workplace. The evaluation of each study presents the study population,
methodology used for the study, i.e., data collection and  verification, analysis, results, and a
critique of the study. There are some methodologic limitations that are common to studies with
similar design. The total evidence, including limitations, is discussed at the end of the chapter in
the summary and discussion section.
                                           7-5

-------
7.2.1. Cohort Studies
7.2.1.1.   Waller (1981): Trends in Lung Cancer in London in Relation to Exposure
         to Diesel Fumes
      A retrospective mortality study of a cohort of London transport workers was conducted
to determine if there was an excess of deaths from lung cancer that could be attributed to DE
exposure.  From nearly 20,000 male employees in the early years, those aged 45 to 64 were
followed for the 25-year period between 1950 and 1974 (the actual number of employees is not
given in the paper), constituting a total of 420,700 man-years at risk.  These workers were
distributed among five job categories: drivers, garage engineers, conductors, motormen or
guards, and engineers  (works).  Lung cancer were ascertained from death certificates of
individuals who died while still employed, or if retired, following diagnosis.  Expected death
rates were calculated by applying greater London death rates to the population at risk within
each job category. Data were calculated in 5-year periods and 5-year age ranges, and the results
were combined to obtain the total expected deaths in the required age range for the calendar
period. A total  of 667 cases of lung cancer was reported, compared with 849 expected, to give a
cancer mortality ratio of 79%. In each of the five job categories, the observed numbers were
below those expected. Engineers in garages had the highest mortality ratio,  90%, motormen and
guards had a mortality ratio of 87%, and both the bus drivers and conductors had mortality ratios
of 75%. The engineers in the central works had a mortality ratio of 66%. These mortality ratios
did not differ significantly from each other. Environmental sampling was done at one garage, on
one day in 1979, for benzo[a]pyrene (B[a]P) concentrations and was compared with
corresponding values recorded in 1957.  Concentrations of B[a]P recorded in 1957 were at least
10 times greater than those measured in 1979.
      This study failed to find any association between DE and occurrence of lung cancer,
which may be due to several methodologic limitations. The lung cancer  deaths were ascertained
while the workers were employed (the worker either died of lung cancer  or retired after lung
cancer was diagnosed). Although man-years at risk were based on the entire cohort, no attempt
was made to trace or evaluate the individuals who had resigned from the  London transport
company for any other reason.  Hence, information  on resignees who may have had significant
exposure to DE, and on lung cancer deaths among them, was not available for analysis. This
may have led to a dilution effect, resulting in underascertainment of observed lung cancer deaths
and underestimation of mortality ratios.  Eligibility  criteria for inclusion  in the cohort, such as
starting date and length of service with the company, were not specified.  Therefore, there may
not have been sufficient latency for the development of lung cancer. Use of greater London
population death rates to obtain expected number of deaths may have resulted in a deficit in
mortality ratios reflecting the "healthy worker effect." Investigators did  not categorize the five
                                          7-6

-------
job categories either by qualitative or quantitative levels of DE exposure; neither did they use an
internal comparison group to derive risk estimates.
       The age range considered for this study was limited (45 to 64 years of age) for the period
between 1950 and 1974. It is not clear whether this age range was applied to calendar year 1950
or 1974, or at the midpoint of the 25-year follow-up period. No analyses were presented either
by latency or by duration of employment (surrogate for exposure).  The environmental survey
based on B[a]P concentrations suggests that the cohort in its earlier years was exposed to much
higher concentrations of environmental contaminants than currently exist. It is not clear when
the reduction in B[a]P concentration occurred, because there are no environmental readings
available between 1957 and 1979.  It is also important to note that the concentrations of B[a]P
inside the garage in 1957 were not very different from those outside the garage, thus indicating
that exposure for garage workers was not much different from that of the general population.
Thus, this study fails to provide either positive or negative association between the DE exposure
and the occurrence of lung cancer.

7.2.1.2. Howeetal. (1983): Cancer Mortality (1965 to 1977) in Relation to Diesel Fumes and
        Coal Exposure in a Cohort of Retired Railroad Workers
       This is a retrospective cohort study of the mortality experience of 43,826 male pensioners
of the Canadian National Railroad (CNR) between 1965 and 1977.  Members of this cohort
consisted of male CNR pensioners who had retired before 1965 and who were known to be alive
at the start of that year, as well as those who retired between 1965 and 1977. The records were
obtained from a computer file that is regularly updated and used by the company for payment of
pensions. To receive a pension,  each pensioner must provide, on a yearly basis, evidence that he
is alive. Specific cause of death  among members of this cohort was ascertained by linking these
records to the Canadian Mortality Data Base, which contains records of all deaths registered in
Canada since 1950.  Of the 17,838 deaths among members of the cohort between  1965 and
1977, 16,812 (94.4%) were successfully linked to a record in the mortality file.  A random
sample manual check on unlinked data revealed that failure to link was due mainly to some
missing information on the death records.
       Occupation at time of retirement was used by the Department of Industrial Relations to
classify workers into three diesel fume and coal dust exposure categories:  (1) nonexposed, (2)
possibly exposed, and (3) probably exposed. Person-years of observation were calculated and
classified by age at observation in  5-year age groups (35 to 39, 40 to 44, . . . , 80 to 84,  and >85
years).  The observed deaths were classified by age at death for different cancers,  for all cancers
combined, and for all causes of death combined. Standard mortality ratios (SMRs) were then
calculated using rates of the Canadian population for the period between 1965 and 1977. The
                                          7-7

-------
relative risks were calculated using the three exposure categories: nonexposed, possibly exposed,
and probably exposed.
       Both total mortality (SMR = 95,  /K0.001) and all cancer deaths (SMR = 99, p>0.05)
were close to that expected for the entire cohort. Analysis by exposure to diesel fume levels in
the three categories (nonexposed, possibly exposed, and probably exposed) revealed an increased
relative risk for lung cancer among workers with increasing exposure to diesel fumes. The
relative risk for nonexposed workers was presumed to be 1.0; for those possibly exposed, the
relative risk was significantly elevated to 1.2 (p=0.013); and for those probably exposed, it was
significantly elevated to 1.35 (p=0.001).  The corresponding rates for exposure to varying levels
of coal dust were very similar at 1.00, 1.21 (p=0.012), and 1.35 (p=0.001), respectively. The
trend tests were highly significant for both exposures (p<0.001). Analysis performed after the
exclusion of individuals who worked in the maintenance of steam engines, and hence were
exposed to high levels of asbestos, yielded a risk of lung cancer of 1.00, 1.21, and 1.33 for those
nonexposed, possibly exposed, and probably exposed to DE, respectively, with a highly
significant trend (/X0.001).
       An analysis done on individuals who retired prior to 1950 showed the relative risk of
lung cancer among nonexposed, possibly exposed, and probably exposed to be 1.00, 0.70, and
0.44, respectively, based on fewer than 15 deaths in each category.  A similar analysis of
individuals who retired after  1950 found the results in the same categories to be 1.00, 1.23, and
1.40, respectively.  Although retirement prior to 1950 indicated exposure to coal combustion
fumes alone, retirement after 1950 shows the results of mixed exposure to coal combustion
fumes and diesel fumes. As there was considerable overlap between occupations involving
probable exposure to diesel fumes and probable exposure to coal, and as most members of the
cohort were employed during the years in which the transition from coal to diesel occurred, it
was difficult to distinguish whether lung cancer was associated with exposure to coal
combustion fumes or diesel fumes or a mixture of both.
       Although this study showed a highly significant dose-response relationship between
diesel fumes and lung cancer, it has  some methodological limitations.  There were concurrent
exposures to both diesel fumes and coal  combustion fumes during the transition period;
therefore, misclassification of exposure may have occurred, because only occupation at
retirement was available for analysis.  It is possible that the elevated response observed for lung
cancer was  due to the combined effects of exposure to both coal dust/coal combustion products
and diesel fumes and not just one or the  other.  However, deaths due to lung cancer were not
elevated among workers who retired prior to the 1950s and thus would have been primarily
exposed to coal dust/coal combustion products.  Furthermore, it should be noted that so far coal
dust has not been demonstrated to be a pulmonary carcinogen in studies of coal miners. This
                                          7-8

-------
study was restricted to deaths among retired workers; therefore, it is unclear if a worker who
developed lung cancer when actively employed and filed for a disability claim instead of
retirement claim would be included in the study or not.  Thus, it is possible that workers with
heavy exposure might have been excluded from the study. Neither information on duration of
employment in diesel work, nor coal dust-related jobs other than those held at retirement, nor
details of how the exposure categories were created was provided.  Therefore, it was not possible
to evaluate whether this omission would have led to an under- or overestimate of the true
relative risk.  Although information on potential confounders such as smoking is lacking, the use
of an internal comparison group to compute the relative risks minimizes the potential for
confounding by smoking, as there is no reason to assume different smoking patterns among
individuals exposed to DE versus those not exposed. Despite these limitations, this study
provides suggestive evidence toward a causal association between exposure to DE and excess
lung cancer.

7.2.1.3. Rushtonetal.  (1983): Epidemiological Survey of Maintenance Workers in the
       London Transport Executive Bus Garages and Chiswick Works
       This is a retrospective mortality cohort study of male maintenance workers employed for
at least 1 continuous year between January 1, 1967, and December 31, 1975,  at 71 London
transport bus  (also known as rolling stock) garages and at Chiswick Works.  The following
information was obtained from computer listings:  surname with initials, date of birth, date of
joining company,  last or present job, and location of work.  For those individuals who left their
job, date of and reason for leaving were also obtained.  For those who died in service or after
retirement, and for men who had resigned, full name and last known address  were obtained from
an alphabetical card index in the personnel department.  Additional tracing of individuals who
had left was carried out through social security records. The area of residence was assumed to
be close to their work; therefore place of work was coded as residence.  One hundred different
job titles were coded into 20 broader groups. These 20 groups were not ranked for DE exposure,
however.  The reason for leaving was coded  as died in service, retired, or other.  The underlying
cause of death was coded using the eighth revision of the International Classification of Diseases
(ICD). Person-years were calculated from date of birth and dates of entry to  and exit from the
study using the man-years computer language program. The workers were then subdivided into
5-year age and calendar period groups. The expected number of deaths was calculated by
applying the 5-year age and calendar period death rates of the comparison population with the
person-years  of corresponding groups.  The mortality experience of the male population in
England and Wales was used as the comparison population.  Significance values were calculated
for the difference between the observed and expected deaths, assuming a Poisson distribution.
                                          7-9

-------
       The person-years of observation totaled 50,008 and were contributed by 8,490
individuals in the study, with a mean follow-up of 5.9 years. Only 2.2% (194) of the men were
not traced.  Observed deaths from all causes were significantly lower than expected (O = 495,
/X0.001). Observed deaths from all neoplasms and cancer of the lung were approximately the
same as those expected. The only significant excess observed, for cancer of the liver and gall
bladder at Chiswick Works, was based on four deaths (p<0.05). A few job groups showed a
significant excess of risks for various cancers. All the excess deaths observed for the various job
groups, except for the general hand category, were based on very small numbers (usually fewer
than five) and merited cautious interpretation. Only a notable excess in the general hand
category for lung cancer was based on as many as 48 cases (SMR = 133,/><0.03).
       This mortality study did not demonstrate any cancer excess. Details of work history
were not obtained to permit any analysis by DE exposure. The study's limitations, including
small sample size, short duration of follow-up (average of only 6 years),  and lack of sufficient
latency period, make it  inadequate to draw any conclusions.

7.2.1.4.  Wong et al. (1985):  Mortality Among Members of a Heavy Construction Equipment
        Operators Union With Potential Exposure to DE Emissions
       This retrospective mortality study was conducted on a cohort of 34,156 male members of
a heavy construction equipment operators union with potential exposure to DE emissions.  Study
cohort members were identified from records maintained at Operating Engineers' Local Union
No. 3-3 A in San Francisco, CA.  This union has maintained both work and death records on all
its members since 1964. Individuals with at least 1 year of membership in this union between
January 1, 1964, and December 31, 1978, were included in the study. Work histories of the
cohort were obtained from job dispatch computer tapes. The study follow-up period was
January 1964 to December 1978. Death information was obtained from a trust fund, which
provided information on retirement dates, vital status, and date of death for those who were
entitled to retirement and death benefits. Approximately 50% of the cohort had been union
members for less than 15 years, whereas the other 50% had been union members for 15 years or
more.  The average duration of membership was 15 years. As of December 31, 1978, 29,046
(85%) cohort members  were alive, 3,345 (9.8%) were dead, and 1,765 (5.2%) remained
untraced. Vital status of 10,505  members who had left the union as of December 31, 1978, was
ascertained from the Social Security Administration. Death certificates were obtained from
appropriate State health departments. Altogether, 3,243 deaths (for whom death certificates
were available) in the cohort were coded using the seventh revision of the ICD.  For 102
individuals, death certificates could not be obtained, only the date of death; these individuals
were included in the calculation  of the SMR for all causes of death but were deleted from the
                                          7-10

-------
cause-specific SMR analyses. Expected deaths and SMRs were calculated using the U.S.
national age-sex-race cause-specific mortality rates for 5-year time periods between 1964 and
1978. The entire cohort population contributed to 372,525.6 person-years in this 15-year study
period.
       A total of 3,345 deaths was observed, compared with 4,109 expected.  The corresponding
SMR for all causes was 81 (p=0.01), which is consistent with the "healthy worker effect."  A
total of 817 deaths was attributed to malignant neoplasms, slightly fewer than the 878 expected
based on U.S. white male cancer mortality rates (SMR = 93,/>=0.05). Mostly there were SMR
deficits for cause-specific cancers, including lung cancer for the entire cohort (SMR = 99, O =
309). The only significant excess SMR was observed for cancer of the liver (SMR = 167, O =
23,/K0.05).
       Analysis by length of union membership as a surrogate of duration for potential exposure
showed statistically significant increases in SMRs of cancer of the liver (SMR = 424,/X0.01) in
the 10- to 14-year membership group and of the stomach (SMR = 248,/><0.05)  in the 5- to 9-
year membership group. No cancer excesses were observed in the 15- to 19-year and 20+-year
membership groups.  Although the SMR for cancer of the lung had a statistically significant
deficit in the less-than-5-year duration group, it showed a positive trend with increasing length
of membership, which leveled off after 10 to 14 years.
       Cause-specific mortality  analysis by latency period showed a positive trend for SMRs of
all causes of death, although all of them were statistically significant deficits, reflecting the
diminishing "healthy worker effect." This analysis also demonstrated a statistically significant
SMR excess for cancer of the liver (10- to 19-year group, SMR = 258).  The SMR for cancer of
the lung showed a statistically significant deficit for a <10-year latency but showed a definite
positive trend with increasing latency.
       In addition to these analyses of the entire cohort, similar analyses were carried out in
various subcohorts.  Analyses of retirees, 6,678 individuals contributing to 32,670 person-years,
showed statistically significant increases (p<0.01) in SMRs for all cancers; all causes of death;
cancers of the digestive system, large intestine, respiratory system, and lung; emphysema; and
cirrhosis of the liver. The other two significant excesses (/X0.01) were for lymphosarcoma and
reticulosarcoma and nonmalignant respiratory  diseases. Further analysis of the 4,075 retirees
(18,678 person-years) who retired at age 65 or who retired earlier but had reached the age of 65
revealed statistically significant SMR increases (p<0.05) for all cancers, cancer  of the lung, and
lymphosarcoma and reticulosarcoma.
       To analyze cause-specific mortality by  job held (potential exposure to DE emissions), 20
functional job titles were used, which were further grouped into three potential categories:  high
exposure, low exposure, and unknown exposure. A person was classified in a job title if he ever
                                          7-11

-------
worked on that job.  Based on this classification system, if a person had ever worked in a high-
exposure job title he was included in that group, even though he may have worked for a longer
time in a low-exposure group or in an unknown exposure group.  Information on length of work
in any particular job, hence indirect information on potential length of exposure, was not
available either.
       For the high-exposure group a statistically significant excess was observed for cancer of
the lung among bulldozer operators who had 15 to 19 years of membership and 20+ years of
follow-up (SMR = 343,/><0.05). This excess was based on 5 out of 495 deaths observed in this
group of 6,712 individuals, who contributed 80,328 person-years of observation.
       The cause-specific mortality analysis in the low-exposure group revealed statistically
significant SMR excesses in individuals who had ever worked as engineers. These excesses
were for  cancer of the large intestine (SMR = 807, O = 3,/><0.05) among those with 15 to 19
years of membership and length of follow-up of at least 20 years, and cancer of the liver (SMR =
872, O = 3,/><0.05) among those with 10 to 14 years of membership and length of follow-up of
10 to 19 years.  There  were 7,032 individuals who contributed to 78,403 person-years of
observation in the low-exposure group.
       For the unknown exposure group, a statistically significant SMR was observed for motor
vehicle accidents only (SMR = 174, O = 21,/><0.05). There were 3,656 individuals who
contributed to 33,388 person-years of observation in this category.
       No work histories were available for those who started their jobs before 1967 and for
those who held the same job prior to and after 1967. This group comprised 9,707 individuals
(28% of the cohort)  contributing to 104,448 person-years. Statistically significant SMR excesses
were observed for all cancers (SMR = 112, O = 339,/><0.05) and cancer of the lung (SMR =
119, O = 141,/><0.01). A significant SMR elevation was also observed for cancer of the
stomach (SMR= 199, O = 30,/K0.01).
       This study demonstrates a statistically significant excess for cancer of the liver but also
shows statistically significant deficits in cancers of the large intestine and rectum.  It may be, as
the authors suggested, that the liver cancer cases actually resulted from metastases from the large
intestine  and/or rectum, as tumors of these sites will frequently metastasize to the  liver. The
excess in liver cancer mortality and the deficits in mortality that are due to cancer of the large
intestine  and rectum could also, as the authors indicate, be due to misclassification.  Both
possibilities have been considered by the investigators in their discussion.
       Cancer of the lung showed a positive trend with length of membership as well as with
latency, although none of the SMRs were statistically significant except for workers without any
work histories.  The individuals without any work histories may have been the ones  who were in
their jobs for the longest period of time, because workers without job histories included those
                                          7-12

-------
who had the same job before and after 1967 and thus may have worked 12 to 14 years or longer.
If they had belonged to the category in which heavy exposure to DE emissions was very
common for this prolonged time, then the increase in lung cancer, as well as stomach cancer,
might be linked to DE. Further information on those without work histories should be obtained
if possible, because such information may be quite informative with regard to the evaluation of
the carcinogenicity of DE.
       The study design is adequate, covers about a 15-year observation period, has a large
enough population, and is appropriately analyzed; however, it has too many limitations to permit
any conclusions. First, no exposure histories are available; one has to make do with job
histories, which provide limited information on exposure level. Any person who ever worked at
the job, or any person working at the same job over any period of time, is included in the same
category; this would have a dilution effect, because extremely variable exposures were
considered in the study.  Second, the length of time worked in any particular job is not available.
Third, work histories were not available for 9,707 individuals, who contributed 104,448 person-
years, a large proportion of the study cohort (28%). These individuals happen to show the most
evidence of a carcinogenic effect. Confounding by alcohol consumption for cancer of the liver
and smoking for emphysema and cancer of the lung was not ruled out. Fourth,  15 years' follow-
up may not provide sufficient latency to observe excess lung cancer. Last, although 34,156
members were eligible for the study, the vital status of 1,765 individuals was unknown.
Nevertheless, they were still considered in the denominator of all the analyses.  The investigators
fail to mention how the person-year calculation for these individuals was handled. Also, some
of the person-years might have been overestimated, as people may have paid the dues for a
particular year and then left work. These two causes of overestimation of the denominator may
have resulted in some or all the SMRs being underestimated.

7.2.1.5. Edlingetal. (1987): Mortality Among Personnel Exposed to DE
       This retrospective cohort mortality study of bus company employees investigated a
possible increased mortality of cardiovascular diseases and cancers from DE exposure.  The
cohort comprised all males employed at five different bus companies in southeastern Sweden
between 1950 and 1959.  Based on information from personnel registers, individuals were
classified into one or more categories and could have contributed person-years at risk in more
than one exposure category. The study period was from 1951 to  1983; information was
collected from the National Death Registry, and copies of death certificates were obtained from
the National Bureau of Statistics. Workers who died after age 79 were excluded from the study
because diagnostic procedures were likely to be more uncertain at higher ages (according to
investigators). The cause-, sex-, and age-specific national death rates in Sweden were applied to
                                          7-13

-------
the 5-year age categories of person-years of observation to determine expected deaths for all
causes, malignant diseases, and cardiovascular diseases.  A Poisson distribution was used to
calculated-values and confidence limits for the ratio of observed to expected deaths. The total
cohort of 694 men (after loss of 5 men to follow-up) was divided into three exposure categories:
(1) clerks with lowest exposure, (2) bus drivers with moderate exposure, and (3) bus garage
workers with highest exposure.
       The 694 men provided 20,304 person-years of observation, with 195 deaths compared
with 237 expected. A deficit in cancer deaths largely accounted for this lower-than-expected
mortality in the total cohort.  Among subcohorts, no difference between observed and expected
deaths for total mortality, total cancers, or cardiovascular causes was observed for clerks (lowest
diesel exposure), bus drivers  (moderate diesel exposure), and garage workers (high diesel
exposure). The risk ratios for all three categories were less than 1 except for cardiovascular
diseases among bus drivers, which was 1.1.
       When the analysis was restricted to members who had at least a 10-year latency period
and either any exposure or an exposure exceeding 10 years, similar results were obtained, with
fewer neoplasms than expected, whereas cardiovascular diseases showed risk around or slightly
above unity.
       Five lung cancer deaths were observed among bus drivers who had moderate DE
exposure, whereas seven were expected.  The only other lung cancer death was observed among
bus garage workers who had  the highest DE exposure. This study's major limitations, including
small size and poor data on DE exposure, make it inadequate to draw any conclusions.

7.2.1.6. Boffetta and Stellman (1988): DE Exposure and Mortality Among Males in the
       American Cancer Society Prospective Study
       Boffetta and Stellman conducted a mortality analysis of 461,981 males with known
smoking history and vital status at the end of the first 2 years of follow-up. The analysis was
restricted to males aged 40 to 79 years in 1982 who enrolled in the American Cancer Society's
prospective mortality study of cancer. Mortality was analyzed in relation to exposure to DE and
to employment in selected occupations related to DE exposure. In 1982, more than 77,000
American Cancer Society volunteers enrolled more than 1.2 million men and women from all 50
States, the District of Columbia, and Puerto Rico in a long-term cohort study, the Cancer
Prevention Study II (CPS-II). Enrollees were usually friends, neighbors, or relatives of the
volunteers; enrollment was by family groups,  with at least one person in the household 45 years
of age or older.  Subjects were asked to fill  out a four-page confidential questionnaire and return
it in a sealed envelope.  The questionnaire included history of cancer and other diseases; use of
medications and vitamins; menstrual and reproductive history; occupational history; and
                                          7-14

-------
information on diet, drinking, smoking, and other habits. The questionnaire also included three
questions on occupation:  (1) current occupation, (2) last occupation, if retired, and (3) job held
for the longest period of time, if different from the other two. Occupations were coded to an ad
hoc two-digit classification in 70 categories. Exposures at work or in daily life to any of the 12
groups of substances were also ascertained. These included diesel engine exhausts, asbestos,
chemicals/acids/solvents,  dyes, formaldehyde, coal or stone dusts, and gasoline exhausts.
Volunteers checked whether their enrollees were alive or dead and recorded the date and place of
all deaths every other year during the study.  Death certificates were then obtained from State
health departments and coded by a trained nosologist according to a system based on the ninth
revision of the ICD.
       The data were analyzed to determine the mortality for all causes and lung cancer in
relation to DE exposure, mortality for all causes and lung cancer in relation to employment in
selected occupations with high DE exposure, and mortality from other causes in relation to DE
exposure.  The incidence-density ratio was used as a measure of association, and test-based
confidence limits were calculated by the Miettinen method. For stratified analysis, the Mantel-
Haenszel method was used for testing linear trends.  Although data on 476,648 subjects
comprising 939,817 person-years of risk were available for analysis,  3% of the subjects (14,667)
had not given any smoking history, and 20% (98,026) did not give information on DE exposure
and were therefore excluded from the main DE analysis. Among individuals who had provided
DE exposure history, 62,800 were exposed and 307,143 were not exposed. Comparison of the
population with known information on DE exposure with the excluded population with no
information on DE exposure showed that the mean ages were 54.7 and 57.7 years, the
nonsmokers were 72.4% and 73.2%, and the total mortality rates per 1,000 per year were 23.0%
and 28.8%, respectively.
       All-cause mortality was elevated among railroad workers (relative risk [RR] = 1.43, 95%
confidence interval [CI] = 1.2, 1.72), heavy equipment operators (RR = 1.7, 95% CI =1.19,
2.44), miners (RR= 1.34, 95% CI = 1.06,  1.68), and truck drivers (RR = 1.19,  95% CI = 1.07,
1.31).  The age-adjusted lung cancer relative risk was elevated significantly (RR = 1.41, 95% CI
= 1.19,1.66), which was slightly decreased to 1.31 (95% CI = 1.10, 1.54).  For lung cancer
mortality the age- and smoking-adjusted risks were significantly elevated for miners (RR = 2.67,
95% CI = 1.63,  4.37) and heavy equipment operators (RR = 2.60, 95% CI =1.12, 6.06).  Risks
were also elevated, but not significantly, for railroad workers (RR =  1.59, 95% CI =  0.94, 2.69)
and truck drivers (RR = 1.24, 95% CI = 0.93, 1.66).  These risks were calculated with the
Mantel-Haenszel method, controlling for age and smoking. Although the relative risk was
nonsignificant for truck drivers, a small dose-response effect was observed when duration of DE
exposure was examined. For drivers who worked for 1  to 15  years, the relative risk was 0.87,
                                          7-15

-------
whereas for drivers who worked for more than 16 years, the relative risk was 1.33 (95% CI =
0.64, 2.75). Relative risks for lung cancer were not presented for other occupations.  Mortality
analysis for other causes and DE exposure showed a significant excess of deaths (p<0.05) in the
following categories: cerebrovascular disease, arteriosclerosis, pneumonia, influenza, cirrhosis
of the liver, and accidents.
       The main strength of this study is detailed information on smoking.  The two main
methodologic concerns are the representativeness of the study population and the quality of
information on exposure.  The sample, though very large, was composed of volunteers.  Thus,
the cohort was healthier and less frequently exposed to important risk factors such as smoking
and alcohol. Self-administered questionnaires were used to obtain data on occupation and DE
exposure. None of this information was validated. Nearly 20% of the individuals had an
unknown exposure status to DE, and they experienced a higher mortality for all causes and lung
cancer than both the DE exposed and unexposed groups.  This could have introduced a
substantial bias in the estimate of the association. Given that all DE exposure occupations, such
as heavy equipment operators, truck drivers, and railroad workers, showed elevated lung cancer
risk, this study is suggestive of a causal association.  It should be noted that after adjusting for
smoking, the RR reduced  slightly from 1.41 to 1.31 and remained significant, indicating that
observed excess of lung cancer was associated mainly with DE exposure.

7.2.1.7. Garshick et al. (1988): A Retrospective Cohort Study of Lung Cancer and DE
       Exposure in Railroad Workers
       An earlier case-control study of lung cancer and DE exposure in U.S. railroad workers by
these investigators had demonstrated a relative odds of 1.41 (95% CI = 1.06, 1.88) for lung
cancer with 20 years of work in jobs with DE exposure. To confirm these results, a large
retrospective cohort mortality study was conducted by the same investigators.  Data sources for
the study were the work records of the U.S. Railroad Retirement Board (RRB).  The  cohort was
selected based on job titles in  1959, which was the year by which 95% of the locomotives in the
United States were diesel powered.  DE exposure was considered to be a dichotomous variable
depending on yearly job codes between 1959 and death or retirement through 1980.  Industrial
hygiene evaluations and descriptions of job  activities were used to classify jobs as exposed or
unexposed to diesel emissions. A questionnaire survey of 534 workers at one of the railroads
where workers were asked to indicate the amount of time spent in railroad locations,  either near
or away from sources of DE, was used to validate this classification.  Workers selected for this
survey were actively employed at the time of the  survey,  40 to 64 years of age, started work
between  1939 and 1949 in the job codes sampled in  1959, and  eligible for railroad benefits. To
qualify for benefits, a worker must have had 10 years or more of service with the railroad and
                                          7-16

-------
should not have worked for more than 2 years in a nonrailroad job after leaving railroad work.
Workers with recognized asbestos exposure, such as repair of asbestos-insulated steam
locomotive boilers, passenger cars, and steam pipes, or railroad building construction and
repairs, were excluded from the job categories selected for study. However, a few jobs with
some potential for asbestos exposure were included in the cohort, and the analysis was done both
ways, with  and without them.
       The death certificates for all subjects identified in 1959 and reported by the RRB to have
died through 1980 were searched.  Twenty-five percent of them were obtained from the RRB
and the remainder from the appropriate State departments of health.  Coding of cause of death
was done without knowledge of exposure history, according to the eighth revision of the ICD. If
the underlying cause of death was not lung cancer, but was mentioned on the death certificate, it
was assigned as a secondary cause  of death, so that the ascertainment of all cases was complete.
Workers not reported by the RRB to have died by December 31, 1980, were considered to be
alive. Deceased workers for whom death certificates had not been obtained or, if obtained, did
not indicate cause of death, were assumed to have died of unknown causes.
       Proportional hazard models were fitted that provided estimates of relative risk for death
caused by lung cancer using the partial likelihood method described by Cox, using the time
dimension being the time since first entry into the cohort. The model also controlled for the
birth year and the calendar time. The 95% confidence intervals were constructed using the
asymptotic  normality of the estimated regression coefficients of the proportional hazards model.
Exposure was analyzed by DE-exposed jobs in 1959 and by cumulative number of years of DE
exposure through 1980. Directly standardized rate ratios for deaths from lung cancer were
calculated for DE exposed compared with unexposed for each 5-year age group in 1959. The
standardized rates were based on the overall 5-year person-year time distribution of individuals
in each age group starting in 1959.  The only exception to this was between 1979 and 1980,
when a 2-year person-year distribution was used.  The Mantel-Haenszel analogue for person-
year data was used to calculate 95% confidence intervals for the standardized rate ratios.
       The cohort consisted of 55,407 workers, 19,396 of whom had died by the end of 1980.
Death certificates were not available for 11.7% of all deaths.  Of the 17,120 deaths for whom
death certificates were obtained, 48.4% were attributable to diseases of the circulatory system,
whereas 21% were attributable to all neoplasms. Of all neoplasms, 8.7% (1,694 deaths) were
due to lung cancer. A higher proportion of workers in the younger age groups, mainly brakemen
and conductors, were exposed to DE, while a higher proportion of workers in the older age
groups were potentially exposed to asbestos.  In a proportional hazards model, analyses by age in
1959 found a relative risk of 1.45 (95% CI = 1.11, 1.89) among the age group 40 to 44 years  and
a relative risk of 1.33 (95% CI = 1.03, 1.73) for the age group 45 to 49 years. Risk estimates in
                                         7-17

-------
the older age groups 50 to 54, 55 to 59, and 60 to 64 years were 1.2, 1.18, and 0.99, respectively,
and were not statistically significant. The two youngest age groups in 1959 had workers with
the highest prevalence and longest duration of DE exposure and lowest exposure to asbestos.
When potential asbestos exposure was considered as a confounding variable in a proportional
hazards model, the estimates of relative risk for asbestos exposure were all near null value and
not significant. Analysis of workers exposed to DE in 1959 (n = 42,535), excluding workers
with potential past exposure to asbestos, yielded relative risks of 1.57 (95% CI = 1.19, 2.06) and
1.34 (95% CI = 1.02, 1.76) in the 1959 age groups 40 to 44 years and 45 to 49 years.  Directly
standardized rate ratios were also calculated for each 1959 age group based on DE exposure in
1959. The results confirmed those obtained by using the proportional hazards model.
       Relative risk estimates were then obtained using duration of DE exposure as a surrogate
for dose. In a model that used years of exposure up to and including exposure in the year of
death, no exposure duration-response relationship was obtained. When analysis was done by
disregarding exposure in the year of death and 4 years prior to death, the risk of dying from lung
cancer increased with the  number of years worked in a diesel-exhaust-exposed job. In this
analysis, exposure to DE was analyzed by exposure duration groups and in a model entering age
in 1959 as a continuous variable. The workers with greater than 15 years of exposure had a
relative risk of lung cancer of 1.72 (95% CI = 1.27, 2.33).  The risk for 1 to 4  years of
cumulative exposure was  1.20 (95% CI = 1.01, 1.44); for 5 to 9 years of cumulative exposure, it
was 1.24 (95% CI = 1.06, 1.44); and for 10 to 14 years of cumulative exposure, it was 1.32
(95%CI= 1.13, 1.56).
       The results of this study, demonstrating a positive association between DE  exposure and
increased lung cancer, are consistent with the results of the case-control study  conducted by the
same investigators in railroad workers dying of lung cancer from March 1981  through February
1982. This cohort study has addressed many  of the weaknesses of the other epidemiologic
studies.  The large sample size (55,400) allowed sufficient power to detect small risks and also
permitted the exclusion of workers with potential past exposure to asbestos. The stability of job
career paths in the cohort  ensured that of the workers 40 to 44 years of  age in 1959 classified as
DE-exposed, 94% of the cases were still in DE-exposed jobs 20 years later.
       The main limitation of the study is the lack of quantitative data on exposure to DE in
either individual workers or overall job categories. This is one of the few studies in which
industrial hygiene measurements of DE were  done. These measurements were correlated with
job titles to divide the  cohort in dichotomous  exposure groups of exposed and nonexposed. This
may have led to an underestimation of the risk of lung  cancer because exposed groups included
individuals with low to high exposure.  The number of years exposed to DE was used as a
surrogate for dose. The dose, based on duration of employment, was inaccurate because
                                          7-18

-------
individuals were working on steam and diesel locomotives during the transition period.  It
should be noted that the investigators only included  exposures after 1959; the duration of
exposure prior to 1959 was not known. If the categories of exposure to DE had been set up as
no, low, moderate, and high exposure, the results would have been more meaningful, as would
the dose-response relationship. Another limitation of this study was its inability to examine the
effect of years of exposure prior to 1959 and latency. No  adjustment for smoking was made in
this study.  However, an earlier case-control study done in the same cohort (Garshick et al.,
1987) showed no significant difference in the risk estimate after adjusting for smoking.  Despite
these limitations, the results of this study indicate that occupational exposure to DE is associated
with a modest risk (1.5) of lung cancer.
       The data of this study were used by Crump et al. (1991) to explore the development of
dose-response-based quantitative estimates of lung cancer associated with DE exposure by using
diesel exposure estimate data from the industrial hygiene (IH) studies conducted by Hammond
(1998) and Woskie et al. (1988a,b). These  studies were conducted in conjunction with the
Garshick et al. (1988) study. The Woskie et al. (1988a,b) IH studies were conducted in four
small northern railroads where the workers were exposed to DE in the early 1980s, prior to the
Garshick et al. (1988) epidemiologic study. A total of 39 job titles were identified by Woskie et
al. (1988a,b), which were subsequently combined into 13 job groups and finally merged into 5
career exposure job codes as follows: brakers, conductors, and hostlers; clerks; engineers and
firers;  signal maintainers; and shop workers.  The average exposure estimates were assigned to
the cohort members by Crump et al. (1991) based on the job codes in 1959. Cumulative
exposures were calculated using these average exposures for each job code. The exposures in
the IH study by Hammond (1998) were defined as the concentrations of respirable-sized
particles (RSP), the adjusted respirable particles (ARP) concentrations, and the adjusted
extractable mass (AEM).  The concentrations of ARP were estimated in the IH study by
removing the particle contribution of environmental tobacco smoke (ETS). Crump et al. (1991)
also used another index called total extractable material (TEX), which was the extractable RSP
including the particle contribution of ETS.  Using these four exposure indices and the regional
climates for the United States, Crump et al. (1991) constructed various exposure metrics. They
conducted more than 50 analyses based on calendar year, age in 1959, attained age, and five job
codes identified in 1959: brakers, conductors, and hostlers; clerks; engineers and firers; signal
maintainers; and shop workers; using the exposure metrics.  Crump et al. (1991) used the U.S.
general population age- and year-specific death rates for comparison and found that the relative
risk can be positively or negatively related to the duration of exposure depending on how age
was controlled in a model. Their use of the U.S. general population rates instead of the  internal
unexposed group of railroad workers that was used by Garshick et al. (1988)  identified that the
                                          7-19

-------
death ascertainment between 1977 and 1980 as incomplete. The Crump et al. (1991) analysis,
limited to 1959 through 1976, found an excess lung cancer risk similar to the subsequent
Garshick analysis (letter from Garshick, Harvard Medical School, to Chao Chen, U.S. EPA,
dated August 15, 1991).
       Garshick conducted some additional analyses after confirming the underascertainment of
deaths by RRB identified by Crump et al. (1991). He reported that the relationship between
years of exposure, when adjusted for attained age and calendar year, was flat to negative
depending upon which model was used.  He also found that in the years 1977-1980 the death
ascertainment was incomplete; approximately 20% to 70% of deaths were missing depending
upon the calendar year.  Garshick's analysis, based on job titles in 1959 and limited to deaths
occurring through 1976, showed that even though the relative risk for all exposure groups was
elevated, the youngest workers still had the highest risk of dying of lung cancer.
       Crump (1999), on the other hand, reported that the negative dose-response continued to
be upheld in his latest analysis when age was controlled more carefully and years of exposure
quantified more accurately.  Crump (1999) asserted that the negative dose-response trends for
lung cancer observed either with the cumulative exposure or with duration of exposure may be
due to underascertainment of deaths in the last 4 years of follow-up of the Garshick et al. (1988)
study as well as incomplete follow-up in earlier years.
       California EPA's (Cal EPA, 1998) Office of Environmental Health Hazard Assessment
(OEHHA) used the same railroad worker data for its quantitative risk assessment. The five job
categories defined by Woskie et al. (1988a,b) and used by Crump et al. (1991) were combined
into three exposure categories: exposed (engineers and firers; brakers, conductors, and hostlers;
collectively known as "train workers"), unexposed (clerks and signalmen), and uncertain
exposure (shop workers). In its analysis, OEHHA found  a positive dose-response and a steadily
increasing risk of lung cancer with increasing duration of exposure by using age in 1959 but
allowing for an interaction term of age and calendar year in the model. This positive dose-
response finding was contradictory to the negative to flat dose-response findings of both Crump
et al. (1991) and Garshick (letter from Garshick, Harvard Medical School, to Chao Chen, U.S.
EPA, dated August 15, 1991).
       The Health Effects Institute (HEI, 1999) convened an expert panel specifically to
evaluate strengths and limitations of two epidemiologic studies that had some exposure data,  for
quantitative risk estimation and to resolve the discrepancies in the dose-response results reported
by Garshick et al. (1988), Crump et al. (1991), and OEHHA (Cal EPA, 1998).  In their
evaluation of the epidemiologic study  of railroad worker data for quantitative risk assessment,
the panel conducted their own analysis of the Garshick et al. (1988) data. They excluded the last
4 years of follow-up (1977-1980) because of underascertainment of deaths during these years.

                                          7-20

-------
The panel categorized the duration of exposure in 12 categories that were basically the duration
of employment. The exposure was assumed to be linearly increasing for 15 years prior to 1959.
Lags of 5 and 10 years were also considered in the analysis.  The job  categories based on job
held in 1959 were classified as clerks, signalmen, engineers and firers, conductors and brakers,
hostlers, and shop workers.  For final analysis these were collapsed into three groups:  clerks and
signalmen, train workers (engineers and firers, conductors and brakers, and hostlers), and shop
workers.  Seven different models were used. The panel's analysis revealed consistently elevated
lung cancer risk for train workers compared with clerks for each duration of employment (1-4,
5-9, 10-14, 15-17, 18+) in years and that shop workers had an intermediate risk of lung cancer.
Their analysis also revealed decreasing risk of lung cancer with increasing duration of
employment in all three job categories. These findings were similar to those of Garshick (letter
from Garshick, Harvard Medical School, to  Chao Chen, U.S. EPA, dated August 15, 1991) and
Crump etal. (1991).
       In addition to differences in adjusting the age (age in 1959 versus attained age) in their
respective analyses, these three investigators made different assumptions in estimating exposure
patterns in these railroad workers.  Garshick et al. (1988) assumed that there was no exposure to
DE prior to 1959 and that the exposure to DE was constant throughout the period of follow-up,
i. e., 1959 to 1980 (block exposure pattern). Crump et al. (1991) assumed that the exposure to
DE increased steadily from 1945 to 1959 to the same level as assumed in the block exposure
pattern by Garshick et al. (1988) and then remained constant from 1959 through 1980 (ramp
exposure pattern).  OEHHA assumed that the exposure increased steeply from 1945 to 1959.
The peak exposure attained in 1959 according to OEHHA was twice as high as assumed in the
block and ramp exposure patterns by Garshick et al. (1988) and Crump et al. (1991),
respectively. The exposures then declined steeply from 1959 to reach the levels assumed in the
block and ramp exposure patterns in 1980 (roof exposure pattern).  The roof exposure pattern
was constructed on the assumption that diesel engines were "smokier" in the past. A detailed
discussion of divergent results observed by  Crump and Cal EPA can be found in Chapter 8.
       The panel discussed various possibilities for the negative dose-response found among
train workers and to a lesser extent among shop workers.  They asserted that several types of
biases could affect the data, alone or in combination, and mask a true positive association.  The
biases enumerated by the panel were: unmeasured confounding by smoking, exposure to other
sources of pollution, previous occupational  exposures, exposure misclassification, use of
"duration of employment" as a surrogate measure for exposure, healthy worker survivor effect,
and differential or incomplete ascertainment of lung cancer deaths (for detailed discussion of
how an individual bias affects the results, please see HEI, 1999).  The panel concluded,
"However, despite the reason or reasons why the relative risks  in these data decrease with
                                          7-21

-------
duration of employment, the lack of a positive exposure-response association in the railroad
worker cohort substantially weakens that study's potential to provide a reliable quantitative
estimate of risk of exposure to diesel engine emissions." Thus, the panel recommended against
using the current railroad worker data as the basis for quantitative risk assessment in ambient
settings.
       The panel also reported that the Garshick et al. study (1987, 1988) had several strengths,
such as a large number of study subjects (55,407 subjects, including 1,694 lung cancer deaths in
the cohort study and 1,256 lung cancer cases for the case-control study). The workers were
employed in an industry where many of them were exposed to DE.  Confounding by asbestos
was handled by either excluding certain job categories from the analyses or controlling for it in
the analyses.  Confounding by smoking was controlled in the analyses of case-control study.
The panel concluded that the overall results of the Garshick studies were generally consistent
with findings of a weak association between exposure to DE and occurrence of lung cancer.
       Thus, it should be noted that although the railroad worker data are unsuitable for
quantitative risk assessment, they provide qualitative support for a positive association between
exposure to DE and occurrence of lung cancer.

7.2.1.8. Gustavsson et al.  (1990): Lung Cancer and Exposure to DE Among Bus
        Garage Workers
       A retrospective mortality study (from  1952 to 1986), cancer incidence study (from 1958
to 1984), and nested case-control study were conducted among a cohort  of 708 male workers
from five bus garages in Stockholm, Sweden, who had worked for at least 6 months between
1945 and 1970. Thirteen individuals were lost to follow-up, reducing the cohort  to 695.
       Information was available on location of workplace, job type, and beginning and ending
of work periods. Workers were traced through a computerized register of the living population,
death and burial books,  and data from the Stockholm city archives.
       For the cohort mortality analyses, death rates of the general population  of greater
Stockholm were used. Death rates of occupationally active individuals,  a subset of the general
population of greater Stockholm, were used as a second comparison group to reduce the bias
from "healthy worker effect." Mortality analysis was conducted using the "occupational
mortality analysis program" (OCMAP-PC). For cancer incidence analysis, the "epidemiology in
Linkoping" (EPILIN) program was used, with the incidence rates obtained from the cancer
registry.
       For the nested case-control study, both dead and incident primary lung cancers identified
in the register of cause of deaths and the cancer register were selected. Six  controls matched on
age ± 2 years, selected from the noncases at the time of the diagnosis of  cases, were drawn at
                                          7-22

-------
random without replacements. Matched analyses were done to calculate odds ratios using
conditional logistic regression.  The EGRET and Epilog programs were used for these analyses.
       DE and asbestos exposure assessments were performed by industrial hygienists based on
the intensity of exposure to DE and asbestos, specific for workplace, work task, and calendar
time period. A DE exposure assessment was based on (1) amount of emission (number of buses,
engine size, running time, and type of fuel), (2) ventilatory equipment and air volume of the
garages, and (3) job types and work practices. Based on detailed historical data and very few
actual measurements, relative exposures were estimated (these were not absolute exposure
levels).  The scale was set to 0 for unexposed and 1 for lowest exposure, with each additional
unit increase corresponding to a 50% increase in successive intensity (i.e., 1.5, 2.25, 3.38, and
5.06).
       Based on personal sampling of asbestos during 1987, exposures were estimated and time-
weighted annual mean exposures were classified on a scale of three degrees (0, 1, and 2).
Cumulative exposures for both DE and asbestos were calculated by multiplying the level of
exposure by the duration of every work period.  An exposure index was calculated by adding for
every individual contribution from all work periods for both DE and asbestos. Four DE index
classes were created: 0 to 10, 10 to 20, 20 to 30, and >30. The four asbestos index classes were
0 to 20, 20 to 40, 40 to 60, and >60. The cumulative exposure indices were used for the nested
case-control study.
       Excesses were observed for all cancers and some other site-specific cancers using both
comparison populations for the cohort mortality study, but none of them was statistically
significant. Based on 17 cases, SMRs for lung cancer were 122 and 115 using Stockholm
occupationally active and general population, respectively. No dose-response was observed with
increasing cumulative exposure in the mortality  study.  The cancer incidence study reportedly
confirmed the mortality results (results not given).
       The nested case-control study, on the other hand, showed increasing risk of lung cancer
with increasing  exposure. Using 0 to 10 DE exposure index as the comparison group yielded
RRsof 1.34(95%CI= 1.09 to  1.64), 1.81 (95% CI = 1.20 to 2.71), and 2.43 (95% CI = 1.32 to
4.47) for the DE indices  10 to 20, 20 to 30, and >30, respectively.  The study was based on  17
cases and 6 controls for each case matched on age ±  2 years.  Adjustment  for asbestos exposure
did not change the lung cancer risk for DE.
       The main strength of this study is the detailed exposure matrices constructed for both DE
and asbestos exposure, although they were based primarily on job tasks and very few actual
measurements.  There are a few methodological limitations to this study.  The cohort is small
and there were only 17 lung cancer deaths; thus  the power is low. Exposure or outcome may be
misclassified, although any resulting bias in the  relative risk  estimates is likely to be toward
                                          7-23

-------
unity, because exposure classification was done independently of the outcome. Although the
analysis by dose indices was done, no latency analysis was performed. Although data on
smoking were missing, it is unlikely to confound the results because this is a nested case-control
study; therefore, smoking is not likely to be different among the individuals irrespective of their
exposure status to DE. Overall, this study provides some support to the excess lung cancer
results found earlier among populations exposed to DE.

7.2.1.9.  Hansen (1993): A Follow-up Study on the Mortality of Truck Drivers
      This is a retrospective cohort mortality study of unskilled male laborers, ages  15 to 74
years, in Denmark, identified from a nationwide census file of November 9,  1970. The exposed
group included all truck drivers employed in the road delivery or long-haul business (14,225).
The unexposed group included all laborers in certain selected  occupational groups considered to
be unexposed to fossil fuel combustion products and to resemble truck drivers in terms of work-
related physical demands and various personal background characteristics (43,024).
      Through automatic record linkage between the 1970 census register (the Central
Population Register 1970 to 1980) and the Death Certificate Register (1970 to 1980), the
population was followed for cause-specific mortality or emigration up to November 9, 1980.
Expected number of deaths among truck drivers was calculated by using the 5-year age group
and 5-year time period death rates of the unexposed group and applying them to the person-years
accumulated by truck drivers. ICD Revision 8 was used to code the underlying cause of death.
Test-based CIs were calculated using Miettinen's method. A Poisson distribution was assumed
for the smaller numbers, and CI was calculated based on exact Poisson distribution (Ciba-
Geigy).  Total person-years accrued by truck drivers were 138,302, whereas  for the unexposed
population, they were 407,780.  There were 627 deaths among truck drivers  and 3,811 deaths in
the unexposed group. Statistically significant excesses were observed for all cancer mortality
(SMR =121, 95% CI = 104 to 140); cancer of respiratory organs (SMR = 160, 95% CI = 128 to
198), which was due mainly to cancer of bronchus and lung (SMR = 160, 95% CI = 126 to 200);
and multiple myeloma (SMR = 439, 95% CI = 142 to 1,024).  When lung cancer mortality was
further explored by age groups, excesses were observed in most age groups (30 to 39, 45 to 49,
50 to 54, 55 to 59, 60 to 64, and 65 to 74), but there were small numbers of deaths in each group
when stratified by age, and the excesses were statistically significant for the 55 to 59 (SMR =
229, O = 19, 95% CI = 138 to 358) and 60 to 64 (SMR = 227, O = 22, 95% CI = 142 to 344) age
groups only.
      As acknowledged by the author, the study has quite a few methodologic limitations.  The
exposure to DE is assumed in truck drivers based on use of diesel-powered trucks, but no
validation of qualitative or quantitative exposure is attempted. It is also not known whether any
                                         7-24

-------
of these truck drivers or any other laborers had changed jobs after the census of November 9,
1970, thus creating potential misclassification bias in exposure to DE. The truck drivers and the
unexposed laborers were from the same socioeconomic class and may have the same smoking
habits. Still, the lack of information on smoking data and a 36% rural population (usually
consuming less tobacco) in the unexposed group may potentially confound the lung cancer
results.  However, a population survey carried out in 1988 showed very little difference in
smoking habits of residents of rural areas and the total Danish male population.  The investigator
reports that diesel trucks were introduced in Denmark after World War II, and since the late
1940s the majority of the Danish fleet has been composed of diesel trucks. Consequently, even
though the follow-up period is relatively short, the truck drivers  may have had exposure to DE
for 20 to 30 years. Therefore, the finding of excess lung cancer  in this study is consistent with
the findings of other truck driver studies.

7.2.1.10. Saverin et al. (1999): DE and Lung Cancer Mortality in Potash Mining
       This is a cohort mortality study conducted in male potash miners in Germany. The
mines began using mobile diesel-powered vehicles in 1969 and 1970. Miners who had worked
underground for at least  1 year after 1969 to 1991, when the mines were closed, were followed
from 1970 to 1994. A total of 5,981 individuals were identified from the medical records  by a
team of medical personnel familiar with the mining technology.  A total of 5,536 were eligible
for follow-up after 5.5% were excluded due to implausible or incomplete work history and 1.9%
were lost to follow-up. A subcohort of 3,258 miners who had worked for at least 10 years
underground (80% had held a single job) was also identified. The miners' biannual medical
examination records were used to extract the information about personal data, smoking data, and
pre-mining occupation, and to reconstruct a chronology of workplaces occupied by the worker
since hire for each person.
       Exposure categories were defined as production, maintenance, and workshop, roughly
corresponding to high, medium, and low.  Concentrations of total carbon, including elemental
and organics, were measured in the airborne fine dust in 1992. A total of 255 samples covering
all workplaces was obtained. Most were personal dust samples;  some were area dust samples.
Cumulative exposure was calculated for each miner, for each year of observation, using the work
chronology and the work category. For the workshop category years of employment were
considered as exposure time; for production and maintenance years of employment was
weighted by a factor of 5/8, since these workers for an 8-hour shift worked for only 5 hours
underground. As neither the mining technology nor the type of machinery used had changed
substantially from 1970 to 1992, the exposure measurements were considered to represent the
exposures throughout the study period.  Accrued person-years were classified into cumulative
                                         7-25

-------
exposures and were expressed in intervals of 0.5 ymg/m3. Both the exposure data and the
smoking data obtained from the medical files were validated by personal interviews with 1,702
cohort members. Death certificates were obtained from local health centers for 94.4% of
deceased members. Autopsy  data were available for 13% of the deceased.  Internal comparison
was done between production and workshop categories.  Using East German general male
population rates, SMRs were  computed for the total cohort as well as the subcohort. Analyses
were done using Poisson and  Cox regression models.
       The concentrations of total carbon for production, maintenance, and workshop categories
were 0.39 mg/m3, 0.23 mg/m3, and 0.12 mg/m3, respectively.  The cumulative exposure ranged
from 0.25 ymg/m3 to 6.25 ymg/m3. The regression analysis showed that the cohort's smoking
habits were homogenous and  that smoking had an even distribution over cumulative exposure.
       A total of 424 deaths were observed for the entire cohort (SMR = 54). The all-cancer
deaths were 133, of which 38 were from lung cancer (SMR = 78). Analysis for the subcohort
using the internal comparison group of low exposure (workshop category, mean cumulative
exposure =  2.12 ymg/m3) RR of 2.17 (95% CI = 0.79, 5.99) was found for the production
category (mean  cumulative exposure = 4.38 ymg/m3).  The relative risks for lung cancer for 20
years of exposure in the production category (highest exposure = cumulative exposure of 4.9
ymg/m3) were calculated using Poisson and Cox regression methods. RRs of 1.16 and 1.68 were
observed for the total cohort, while RRs of 1.89 and 2.7 were observed for the subcohort by
Poisson and Cox regression methods respectively.
       The main strengths of the study are the information available on DE exposure and
smoking. Although these potash miners were exposed to salt dust and nitric gases, exposures to
other confounders such as heavy metals and radon were absent. Smoking does not seem to be a
confounder in this study but cannot be completely ruled out. Unfortunately, the age distribution
of the cohort is not available.  Since there were only 424 deaths in 25 years of follow-up in this
cohort of 5,536, it appears that the cohort is young. Although lung cancer risk was elevated by
twofold in the production category of the subcohort of miners who had worked  for at least  10
years underground at the same job for 80% of their time  and did not have more  than 3 jobs, it
was not statistically significant. The follow-up period for this study was 25 years, but the cohort
members could have entered the cohort any time between 1970 and  1990, as long as they
worked underground for a year, i.e., they could have worked in the mines for 1  year to 21 years.
Thus, the authors may not have had enough follow-up or latency to observe the lung cancer
excess.  Despite these limitations, the results of this study provide suggestive evidence for the
causal association between DE and excess lung cancer.
       Table 7-1 summarizes the above cohort studies.
                                         7-26

-------
      	Table 7-1. Epidemiologic studies of the health effects of exposure to DE: cohort mortality studies	
        Authors	Population studied	DE exposure assessment	Results	Limitations	
      Waller     Approximately 20,000 male      Five job categories used to  SMR = 79 for lung cancer for the Exposure measurement of B [a]P
      (1981)      ~   '             '                                          '   '                       	"
            London transportation workers

            Aged 45 to 64 years
define exposure
total cohort
                                                                                     showed very little difference between
                                                                                     inside and outside the garage
                               Environmental B[a]P       SMRs for all five job categories
                               concentrations measured in  were less than 100 for lung      Incomplete information on cohort
25 years follow-up (1950-1974)   1957 and 1979             cancer                       members

                                                                                     No adjustment for confounding such
                                                                                     as other exposures, cigarette
                                                                                     smoking, etc.

	No latency analysis	
to
Howe et al.  43,826 male pensioners of the
(1983)      Canadian National Railway
            Company

            Mortality between 1965 and
            1977 among these pensioners
            was compared with mortality
            of general Canadian population
Exposure groups
classified by a group
of experts based on
occupation at the time
of retirement

Three exposure groups:
Nonexposed
Possibly exposed
Probably exposed
RR= 1.2 (p=0.013) and
RR = 1.3 (p=0.001) for lung
cancer for possible and probable
exposure, respectively

A highly significant
dose-response relationship
demonstrated by trend
test(p<0.001)
                                                                                     Incomplete exposure assessment due
                                                                                     to lack of lifetime occupational
                                                                                     history

                                                                                     Mixed exposures to coal
                                                                                     dust/combustion products and DE

                                                                                     No validation of method was used to
                                                                                     categorize exposure

                                                                                     Lack of data on smoking but use of
                                                                                     internal comparison group to
                                                                                     compute RRs minimizes the potential
                                                                                     confounding by smoking
                                                                                                       No latency analysis

-------
            Table 7-1.   Epidemiologic studies of the health effects of exposure to DE: cohort mortality studies
                         (continued)
         Authors
Population studied
DE exposure assessment
           Results
Limitations
       Rushton     8,490 male London transport
       et al. (1983)  maintenance workers

                   Mortality of workers employed
                   for 1 continuous year between
                   January 1, 1967, and December
                   31, 1975, was compared with
                   mortality of general population
                   of England and Wales
                        100 different job titles were SMR = 133 (p<0.03) for lung
                        grouped in
                        20 broad categories

                        The categories were not
                        ranked for DE exposure
                        cancer in the general hand job
                        group

                        Several other job
                        categories showed SS increased
                        SMRs for several other sites
                        based on fewer than five cases
                              Ill-defined DE exposure without any
                              ranking

                              Average 6-year follow-up i.e., not
                              enough time for lung cancer latency

                              No adjustment for confounders
       Wong et al.  34,156 male heavy construction
       (1985)       equipment operators

                   Members of the local union for
                   at least 1 year between
                   January 1, 1964, and December 1.
                   1978
to
oo
                        20 functional job titles
                        grouped into three job
                        categories for potential
                        exposure

                        Exposure groups (high,
                        low, and unknown) based
                        on job description and
                        proximity to source of DE
                        emissions
                        SMR = 166 (p<0.05) for liver
                        cancer for total cohort
                              No validation of exposure categories,
                              which were based on surrogate
                              information
SMR = 343 (observed = 5,
/><0.05) for lung cancer for high- Incomplete employment records
exposure bulldozer operators
with 15-19 years of membership,
20+ years of follow-up
                                                      Employment history other than from
                                                      the union not available
                                                                          SMR =119 (observed =141,
                                                                          /><0.01) for workers with no
                                                                          work histories
                                                      15 year follow-up may not provide
                                                      sufficient time for lung cancer
                                                      latency

                                                      No data on confounders such as other
                                                      exposures, alcohol, smoking, etc.
       Edling et al.  694 male bus garage employees   Three exposure groups
       (1987)
                   Follow-up from 1951 through
                   1983

                   Mortality of these men was
                   compared with mortality of
                   general population of Sweden
                        based on job titles:
                        High exposure, bus
                         garage workers
                        Intermediate exposure,
                         bus drivers
                        Low exposure, clerks
                        No SS differences were observed Small sample size
                        between observed and expected
                        for any cancers by different      No validation of exposure
                        exposure groups
                                                      No data on confounders such as other
                                                      exposures, smoking, etc.

-------
             Table 7-1.  Epidemiologic studies of the health effects of exposure to DE: cohort mortality studies
                          (continued)
          Authors
Population studied
DE exposure assessment
Results
Limitations
        Boffetta and 46,981 male volunteers enrolled
        Stellman    in the American Cancer Society's
        (1988)      Prospective Mortality Study of
                    Cancer in 1982

                    Aged 40 to 79 years at enrollment

                    First 2-year follow-up
                        Self-reported occupations
                        were coded into 70 job
                        categories

                        Employment in high DE
                        exposure jobs were
                        compared with nonexposed
                        jobs
                        Total mortality (SS) elevated for Exposure information based on self-
                        railroad workers (RR=1.43),
                        heavy equipment operators
                        (RR=1.7), miners (RR=1.34),
                        and truck drivers (RR=1.19)

                        Lung cancer mortality (SS)
                        adjusted for age & smoking,
                        elevated for total cohort
                        (RR=1.31), miners (RR=2.67),
                        and heavy equipment operators
                        (RR=2.6)

                        Lung cancer mortality (SNS)
                        elevated among railroad workers
                        and truck drivers

                        Truck drivers also showed a
                        dose-response
                  reported occupation for which no
                  validation was done

                  Volunteer population, probably
                  healthy population
to

-------
     Table 7-1.  Epidemiologic studies of the health effects of exposure to DE: cohort mortality studies
                  (continued)
  Authors
Population studied
                                                   DE exposure assessment
Results
Limitations

-------
     Table 7-1.   Epidemiologic studies of the health effects of exposure to DE: cohort mortality studies
                  (continued)
  Authors
Population studied
 DE exposure assessment
Results
Limitations
Gustavsson  695 male workers from 5 bus
et al. (1990)  garages in Stockholm, Sweden,
            who had worked for 6 months
            between 1945 and 1970

            34 years follow-up (1952-1986)

            Nested case-control study
            17 cases, six controls for each
            case matched on age ± 2 years
                                                 SNS SMRs of 122 and 115 (OA
                                                 and GP), respectively
                        and duration of work
Four DE indices were
created:
0 to 10, 10 to 20, 20-30,
and >30 based on job tasks  Case-control study results
                         showed dose response:
                         RR= 1.34 (10 to 20)
                         RR= 1.81 (20 to 30)
                         RR = 2.43 (>30)

                         All SS with 0-10 as comparison
                         group
                  Exposure matrix based on job tasks
                  (not on actual measurements)

                  Small cohort, hence low power

                  Lack of smoking data is unlikely to
                  confound the results since it is a
                  nested case-control study
Hansen      Cohort of 57,249 unskilled       DE exposure assumed
(1993)      laborers, ages 15 to 74, in        based on diesel-powered
            Denmark (nationwide census file) trucks
            November 9, 1970

            Follow-up through November 9,
            1980
                                                 SS SMRs for lung cancer :
                                                 SMR =160 for total population
                                                 SMR = 229 for age 55-59 years
                                                 SMR = 227 for age 60-64 years
                                                       No actual exposure data available

                                                       Lack of smoking data but population
                                                       survey showed very little difference
                                                       between rural and urban smoking
                                                       habits

                                                       Job changes may have occurred from
                                                       laborer to driver

                                                       Short follow-up period
Saverin et   Cohort of 5,536 potash miners    DE exposure categories
al. (1999)    who had worked underground for defined as:
            at least 1 year after 1969

            Subcohort of 3,258 who had
            worked for at least 10 years
            underground

            Follow-up from 1970 to 1994
                        production (high)
                        maintenance (medium)
                        workshop (low)

                        225 air samples obtained:
                        for total carbon, organics,
                        & fine dust in 1992
                         SNS increased RRs adjusted for  Small, young cohort
                         smoking: 1.68 and 2.7 for total
                         cohort & subcohort, respectively  Few deaths

                                                       No latency analysis
   Abbreviations:  RR = relative risk; SMR = standardized mortality ratio; SNS = statistically nonsignificant; SS = statistically significant;
   O = occupationally active; GP = general population.

-------
7.2.2. Case-Control Studies of Lung Cancer
7.2.2.1.  Hall and Wynder (1984): A Case-Control Study ofDE Exposure and Lung Cancer
      Hall and Wynder (1984) conducted a case-control study of 502 male lung cancer cases
and 502 controls without tobacco-related diseases that examined an association between
occupational DE exposure and lung cancer. Histologically confirmed primary lung cancer
patients who were 20 to 80 years old were ascertained from 18 participating hospitals in 6 U.S.
cities 12 months prior to the interview.  Eligible controls, patients at the same hospitals without
tobacco-related diseases, were matched to cases by age (± 5 years), race, hospital, and hospital
room status.  The number of male lung cancer cases interviewed totaled 502, which was 64% of
those who met the study criteria for eligibility.  Of the remaining 36%, 8% refused, 21% were
too ill or had died, and 7% were unreliable. Seventy-five percent of eligible controls completed
interviews. Of these interviewed controls,  49.9% were from the all-cancers category, whereas
50.1% were from the all-noncancers category. All interviews were obtained in hospitals to
gather detailed information on smoking history, coffee consumption, artificial sweetener use,
residential history, and abbreviated medical history as well as standard demographic variables.
Occupational information was elicited by a question on the usual lifetime occupation and was
coded by the abbreviated list of the U.S. Bureau of Census Codes.  The odds ratios were
calculated to evaluate the association between DE exposure and risk of lung cancer incidence.
Summary odds ratios were computed by the Mantel-Haenszel method after adjusting for
potential confounding by age, smoking, and socioeconomic class. Two-sided, 95% confidence
intervals were computed by Woolf s method. Occupational exposure to DE was defined by two
criteria.  First, occupational titles were coded "probably high exposure"  as defined by the
industrial hygiene standards established for the various jobs. The job titles included under this
category were warehousemen, bus and truck drivers, railroad workers, and heavy equipment
operators and repairmen. The second method used the National Institute for Occupational Safety
and Health (NIOSH) criteria to analyze occupations by diesel exposure. In this method, the
estimated proportion of exposed workers was computed for each occupational category by using
the NIOSH estimates of the exposed population as the numerator and the estimates of
individuals employed in each occupational category from the 1970 census as the denominator.
Occupations estimated to have at least 20% of their employees exposed to DE were defined as
"high exposure," those with 10% to 19% of their employees exposed were defined as "moderate
exposure," and those with less than 10% of their employees exposed were defined as "low
exposure."
      Cases and controls were compared with respect to exposure.  The relative risk was 2.0
(95% CI = 1.2, 3.2) for those workers who were exposed to DE versus those who were not.  The
risk, however, decreased to a nonsignificant 1.4 when the data were adjusted for smoking.
                                         7-32

-------
Analysis by NIOSH criteria found a nonsignificant relative risk of 1.7 in the high- exposure
group. There were no significantly increased cancer risks by occupation either by the first
method or by the NIOSH method. To assess any possible synergism between DE exposure and
smoking, the lung cancer risks were calculated for different smoking categories.  The relative
risks were 1.46 among nonsmokers and ex-smokers, 0.82 among current smokers of <20
cigarettes/day, and 1.3 among current smokers of 20+ cigarettes/day, indicating a lack of
synergistic effects.
       The major strength of this study is the availability of a detailed smoking history for all
the study subjects.  However, this is offset by lack of DE exposure measurements, use of a poor
surrogate for exposure, and lack of consideration of latency period. Information was collected
on only one major lifetime occupation, and it is likely that those workers who had more than one
major job may not have reported the occupation with the heaviest DE exposures. Furthermore,
the exposure categories based on job titles were broad, and thus would have made a true effect of
DE difficult to detect.

7.2.2.2. Dumber andLarsson (1987): Occupation and Male Lung Cancer, a Case-Control
        Study in Northern Sweden
       A case-control study of lung cancer was conducted in northern Sweden to determine the
occupational risk factors that could explain the large geographic variations of lung cancer
incidence in that country. The study region comprised the three northernmost counties of
Sweden, with a total male population of about 390,000. The rural municipalities, with 15% to
20% of the total population, have forestry and agriculture as dominating industries, and the
urban areas have a variety of industrial activities (mines, smelters, steel factories, paper mills,
and mechanical workshops).  All male cases of lung  cancer reported to the Swedish Cancer
Registry during the 6-year period between 1972 and 1977 who had died before the start of the
study were selected.  Of 604 eligible cases, 5 did not have microscopic confirmation, and in
another 5 the diagnosis was doubtful, but these cases were included nevertheless. Cases were
classified as small-cell carcinomas, squamous cell  carcinomas, adenocarcinomas, and other
types.  For each case a dead control was drawn from the National Death Registry matched by
sex, year of death, age, and municipality.  Deaths in controls classified as lung cancer and
suicides were excluded.  A living control matched to the case by sex,  year of birth, and
municipality was also drawn from the National Population Registry. Postal questionnaires were
sent to close relatives of cases and dead controls, and to living controls themselves to collect data
on occupation, employment, and smoking habits. Replies were received from 589 cases (98%),
582 surrogates of dead controls (96%), and 453 living controls (97%).
                                          7-33

-------
       Occupational data were collected on occupations or employment held for at least 1 year
and included type of industry, company name, task, and duration of employment.
Supplementary telephone interviews were performed if occupational data were lacking for any
period between age 20 and time of diagnosis. Data analysis involved calculation of the odds
ratios by the exact method based on the hypergeometric distribution and the use of a linear
logistic regression model to adjust for the potential confounding effects of smoking. Separate
analyses were performed with dead and living controls, and on the whole there was good
agreement between the two control groups.  A person who had been active for at least 1 year in a
specific occupation was in the analysis assigned to that occupation.
       Using dead controls, the odds ratios adjusted for smoking were 1.0 (95% CI = 0.7, 1.5)
and 2.7 (95% CI = 1.0, 8.1) for professional drivers (> 1 year of employment) and underground
miners (> 1 year of employment), respectively.  For 20 or more years of employment in those
occupations, the odds ratios adjusted for smoking were 1.2 (95% CI = 0.9, 2.6) and 9.8  (95% CI
= 1.5, 414). These were the only two occupations listed with potential DE exposure. An excess
significant risk was detected for copper smelter workers, plumbers, electricians, and asbestos
workers, as well as concrete and asphalt workers. All the odds ratios were calculated by
adjusting for age, smoking, and municipality.  A comparison with the live controls resulted in
the odds ratios being lower than those observed with dead controls, and none were statistically
significant in this comparison.
       This study did not detect any excess risk of lung cancer for professional drivers, who,
among all the occupations listed, had the most potential for exposure to motor vehicle exhaust.
However, it is not known whether these drivers were exposed exclusively to gasoline exhaust,
DE, or varying degrees of both.  An excess risk was detected for underground miners, but it is
not known if this was due to diesel emissions from engines or from radon daughters in poorly
ventilated mines.  Although a high response rate (98%) was obtained by the postal
questionnaires, the use of surrogate respondents is known to lead to misclassification errors that
can bias the results in either direction.

7.2.2.3. Lerchenetal. (1987): Lung Cancer and Occupation in New Mexico
       This is a population-based case-control study conducted in New Mexico that examined
the association between occupation and occurrence of lung cancer in Hispanic and non-Hispanic
whites.  Cases involved residents of New Mexico, 25 through 84 years of age, and diagnosed
between January 1, 1980, and December 31, 1982, with primary lung cancer, excluding
bronchioalveolar carcinoma.  Cases were ascertained through the New Mexico Tumor Registry,
which is a member of the Surveillance Epidemiology and End Results  (SEER) Program of the
National Cancer Institute. Controls were chosen by randomly selecting residential telephone
                                         7-34

-------
numbers and, for those over 65 years of age, from the Health Care Financing Administration's
roster of Medicare participants.  They were frequency-matched to cases for sex, ethnicity, and
10-year age category with a ratio of 1.5 controls per case.  The 506 cases (333 males and 173
females) and 771 controls (499 males and 272 females) were interviewed, with a nonresponse
rate of 11% for cases. Next of kin provided interviews for 50% and 43% of male and female
cases, respectively. Among controls, only 2% of the interviews were provided by next of kin for
each sex. Data were collected by personal interviews conducted by bilingual interviewers in the
participants' homes.  A lifetime occupational history and a self-reported history of exposure to
specific agents were obtained for each job held for at least 6 months since age 12.  Questions
were asked about the title of the position, duties performed, location and nature of industry, and
time at each job title.  A detailed smoking history was also obtained.  The variables on
occupational exposures were coded according to the Standard Industrial Classification scheme
by a single person  and reviewed by another. To test the hypothesis about high-risk jobs for lung
cancer, the principal investigator created an a priori listing of suspected occupations and
industries by a two-step process involving a literature review for implicated industries and
occupations.  The principal investigator also determined the appropriate Standard Industrial
Classification  and  Standard Occupational Codes associated with job titles.  For four
agents—asbestos, wood dust, DE, and formaldehyde—the industries and occupations determined
to have exposure were identified, and linking of specific industries  and occupations was based
on literature review and consultation with local industrial hygienists.
       The relative odds were calculated for suspect occupations and industries, classifying
individuals as  ever employed for at least 1 year in an industry or occupation and defining the
reference group as those subjects never employed in that particular industry or occupation.
Multiple logistic regression models were used to control simultaneously for age, ethnicity, and
smoking status.  For occupations with potential DE exposure, the analysis showed no excess
risks for diesel engine mechanics and auto mechanics.  Similarly, when analyzed by exposure to
specific agents, the odds ratio (OR) adjusted for age, smoking, and  ethnicity was not elevated for
DE fumes (OR = 0.6, 95% CI = 0.2, 1.6). Significantly elevated ORs were found for uranium
miners (OR =  2.8), underground miners (OR = 2.4), construction workers,  and welders (OR =
4.3). No excess risks were  detected for the following industries: shipbuilding, petroleum
refining, printing, blast furnace, and steel mills.  No excess risks were detected for the following
occupations:  construction workers, painters, plumbers, paving equipment operators, roofers,
engineers and  firemen, woodworkers, and shipyard workers. Females were excluded from
detailed analysis because none of the Hispanic female controls had  been employed in high-risk
jobs; among the non-Hispanic white controls, employment in a high-risk job was recorded for at
                                          7-35

-------
least five controls for only two industries, construction and painting, for which the OR were not
significantly elevated. Therefore, the analyses were presented for males only.
       Among the many strengths of this study are its population-based design, high
participation rate, detailed smoking history, and the separate analysis done for two ethnic groups,
southwestern Hispanic and non-Hispanic white males. The major limitations pertain to the
occupational exposure data. Job titles obtained from occupational histories were used as proxy
for exposure status, but these were not validated. Further, for nearly half the cases, next of kin
provided occupational histories.  The authors acknowledge the above sources of bias but state
without substantiation that these biases would not strongly affect their results.  They also did not
use a job exposure matrix to link occupations to exposures and did not provide details on the
method they used to classify individuals as DE exposed based on reported occupations. The
observed absence of an association for exposure to  asbestos, a well-established lung carcinogen,
may be explained by the misclassification errors in  exposure status or by sample size constraints
(not enough power). Likewise, the association for DE reported by only 7 cases and 17 controls
also may have gone undetected because of low power. In conclusion,  there is insufficient
evidence from this study to confirm or refute an association between lung cancer and DE
exposure.

7.2.2.4. Garshick et al. (1987):  A Case-Control Study of Lung Cancer and DE Exposure in
         Railroad Workers
       An earlier pilot study of the mortality of railroad workers by the same investigators
(Schenker et al.,  1984) found a moderately high risk of lung cancer among workers exposed to
DE compared with those who were not.  Based on these findings the investigators conducted a
case-control study of lung cancer in the same population.  The population base for this case-
control study was approximately 650,000 active and retired male U.S. railroad workers with 10
years or more of railroad service who were born in  1900 or later.  The U.S. Railroad Retirement
Board (RRB), which operates the retirement system, is separate from the Social Security System,
and to qualify for the retirement or survivor benefits the workers had to acquire 10 years or more
of service.  Information on deaths that occurred between March 1,  1981, and February 28, 1982,
was obtained from the RRB.  For 75% of the deceased population, death certificates were
obtained from the RRB, and, for the remaining 25%, they were obtained from the appropriate
State departments of health. Cause of death  was coded according to the eighth revision of the
ICD. The cases were selected from deaths with primary lung cancer, which was the underlying
cause of death in most cases. Each case was matched to two deceased controls whose dates of
birth were within 2.5 years of the date of birth of the case and whose dates of death were within
31 days of the date of death noted in the case. Controls were selected  randomly from workers
                                          7-36

-------
who did not have cancer noted anywhere on their death certificates and who did not die of
suicide or of accidental or unknown causes.
       Each subject's work history was determined from a yearly job report filed by his
employer with the RRB from 1959 until death or retirement. The year 1959 was chosen as the
effective start of DE exposure for this study since by this time 95% of the locomotives in the
United States were diesel powered.  Investigators acknowledge that because the transition to
diesel-powered engines took place in the early 1950s, some workers had additional exposure
prior to 1959; however, if a worker had died or retired prior to  1959, he was considered
unexposed. Exposure to DE was considered to be dichotomous for this study, which was
assigned based on an industrial hygiene evaluation of jobs and work areas.  Selected jobs with
and without regular DE exposure were identified by a review of job title and duties. Personal
exposure was assessed in 39 job categories representative of workers with and without DE
exposure. Those jobs for which no personal sampling was done were considered exposed or
unexposed based on similarities in job activities and work locations and by degree of contact
with diesel equipment. Asbestos exposure was categorized based on jobs held in 1959, or on the
last job held if the subject retired before 1959.  Asbestos exposure in railroads occurred
primarily during the steam engine era and was related mostly to the repair of locomotive steam
boilers that were insulated with asbestos.  Smoking history information was obtained from the
next of kin.
       Death certificates were obtained for approximately 87% of the 15,059 deaths reported by
the RRB, from which 1,374 cases of lung cancer were identified.  Fifty-five cases of lung cancer
were excluded from the study for either incomplete data (20) or refusal by two States to use
information on death certificates to contact the next of kin.  Successful matching to at least one
control with work histories was achieved for 335 (96%) cases <64 years of age at death and 921
(95%) cases >65 years of age at death. In both age groups, 90% of the cases were matched with
two controls.  There were 2,385 controls in the study; 98% were matched within ±31 days of
the date of death, whereas the remaining 2% were matched within 100 days. Deaths from
diseases of the circulatory system predominated among controls.  Among the younger workers,
approximately 60% had exposure to DE, whereas among older workers, only 47% were exposed
toDE.
       Analysis by a regression model, in which years of DE exposure were the sum total of the
number of years in diesel-exposed jobs, used as a continuous exposure variable, yielded an odds
ratio of lung cancer of 1.39 (95% CI = 1.05, 1.83) for >20 years of DE exposure in the <64
years of age group. After adjustment for asbestos exposure and lifetime smoking (pack-years),
the odds ratio was 1.41 (95% CI = 1.06, 1.88).  Both crude odds ratio and asbestos exposure as
well as lifetime smoking-adjusted odds ratio for the >65 years of age group were not significant.
                                         7-37

-------
Increasing years of DE exposure, categorized as >20 diesel years and 5 to 19 diesel years, with 0
to 4 years as the referent group, showed significantly increased risk in the <64 years of age
group after adjusting for asbestos exposure and pack-year category of smoking. For individuals
who had >20 years of DE exposure, the odds ratio was 1.64 (95% CI = 1.18, 2.29), whereas
among individuals who had 5 to 19 years of DE exposure, the odds ratio was 1.02 (95% CI =
0.72, 1.45).  In the >65 years of age group, only 3% of the workers were exposed to DE for
more than 20 years.  Relative odds for 5 to 19 years and >20 years of diesel exposure were less
than 1 (p>0.01) after adjusting for smoking and asbestos exposure.
       Alternative models to explain past asbestos exposure were tested. These were variables
for regular and intermittent exposure groups and an estimate of years of exposure based on
estimated years worked prior to 1959. No differences in results were seen.  The interactions
between DE exposure  and the three pack-year categories (<50, >50, and missing pack-years)
were explored.  The cross-product terms were not significant.  A model was also tested that
excluded recent DE  exposure occurring within the 5 years before death and gave an odds ratio of
1.43 (95% CI = 1.06, 1.94), adjusted for cigarette smoking and asbestos exposure, for workers
with 15 years of cumulative exposure.  For workers with 5 to 14 years of cumulative exposure,
the OR were not significant.
       The many  strengths of the study are consideration of confounding factors such as
asbestos exposure and smoking; classification of DE exposures by job titles and industrial
hygiene sampling; exploration of interactions between smoking, asbestos exposure, and DE
exposure; and good  ascertainment (87%)  of death certificates from the 15,059 deaths reported by
theRRB.
       The investigators also recognized  and reported the following limitations: overestimation
of cigarette consumption by surrogate respondents, which may have exaggerated the
contribution of smoking to lung cancer risk, and use of the Interstate Commerce Commission
(ICC) job classification as a surrogate for exposure, which may have led to misclassification of
DE exposure jobs with low intensity and intermittent exposure, such as railroad police and bus
drivers, as unexposed.  These two limitations would result in underestimation of the lung cancer
risk. This source of error could have been avoided if DE exposures were categorized by a
specific dose range associated with a job title that could have been classified as heavy, medium,
low, and zero exposure instead of a dichotomous variable. The use of death certificates to
identify cases and controls may have resulted in misclassification.  Controls may have had
undiagnosed primary lung cancer, and lung cancer cases might have been secondary lesions
misdiagnosed as primary lung cancer. However, the investigators quote a third National Cancer
Survey report in which the death certificates for lung cancer were coded appropriately in 95% of
the cases.  Last, as in all previous studies, there is a lack of data on the contribution of unknown
                                          7-38

-------
occupational or environmental exposures and passive smoking. In conclusion, this study
provides strong evidence that occupational DE emission exposure increases the risk of lung
cancer.

7.2.2.5. Benhamou et al. (1988): Occupational Risk Factors of Lung Cancer in a French
        Case-Control Study
       This is a case-control study of 1,625 histologically confirmed cases of lung cancer and
3,091 matched controls, conducted in France between 1976 and 1980. This study was part of an
international study to investigate the role of smoking and lung cancer. Each case was matched
with one or two controls, whose diseases were not related, to tobacco use, sex, age at diagnosis
(± 5 years), hospital of admission, and interviewer. Information was obtained from  both cases
and controls on place of residence since birth, educational level, smoking, and drinking habits.
A complete lifetime occupational history was obtained by asking participants to give their
occupations from the most recent to the first. Women were excluded because most of them had
listed no occupation.  Men who smoked cigars and pipes were excluded because there were very
few in this category.  Thus, the study was restricted to nonsmokers and cigarette smokers.
Cigarette smoking exposure was defined by age at the first cigarette (nonsmokers, <20 years, or
>20 years), daily consumption of cigarettes (nonsmokers, <20 cigarettes a day,  and >20
cigarettes a day), and duration of cigarette smoking (nonsmokers, <35 years, and >35 years).
The data on occupations were coded by a panel of experts according to their own chemical or
physical exposure determinations. Occupations were recorded blindly using the International
Standard Classification of Occupations. Data on 1,260 cases and 2,084 controls were available
for analysis. The remaining 365 cases and 1,007 controls were excluded because they did not
satisfy the required smoking status criteria.
       A matched logistic regression analysis was performed to estimate the effect of each
occupational exposure after adjusting for cigarette status.  Matched relative risk ratios were
calculated for each occupation with the baseline category, which consisted of patients who had
never been engaged in that particular occupation.  The matched RR ratios,  adjusted for cigarette
smoking for the major groups of occupations, showed that the risks were significantly higher for
production and related workers, transport equipment operators, and laborers (RR = 1.24, 95% CI
= 1.04, 1.47).  On further analysis of this group, for occupations with potential diesel emission
exposure, significant excess risks were found for motor vehicle drivers (RR = 1.42,  95% CI =
1.07,  1.89) and transport equipment operators (RR = 1.35, 95% CI = 1.05,  1.75). No interaction
with smoking status was found in any of the occupations.  The only  other significant excess was
observed for miners and quarrymen (RR = 2.14, 95% CI = 1.07, 4.31). None of the significant
associations showed a dose-response relationship with duration of exposure.
                                          7-39

-------
       This study was designed primarily to investigate the relationship between smoking (not
occupations or environmental exposures) and lung cancer. Although an attempt was made to
obtain complete occupational histories, the authors did not clarify whether, in the logistic
regression analysis, they used the subjects' first occupation, predominant occupation, last
occupation, or ever worked in that occupation as the risk factor of interest.  The most important
limitation of this study is that the occupations were not coded into exposures for different
chemical and physical agents, thus precluding the calculation of relative risks for diesel
exposure. Using occupations as surrogate measures of diesel exposure, an excess significant risk
was obtained for motor vehicle drivers and transport equipment operators, but not for motor
mechanics. However, it is not known if subjects in these occupations worked with diesel
engines or nondiesel engines.

7.2.2.6. Hayes et al.  (1989):  Lung Cancer in Motor Exhaust-Related Occupations
       This study reports the findings from an analysis of pooled data from three lung cancer
case-control studies that examine in detail the association between employment in motor
exhaust-related (MER) occupations and lung cancer risk adjusted for confounding by smoking
and other risk factors. The three studies were carried out by the National Cancer Institute in
Florida (1976 to 1979), New Jersey (1980 to 1981), and Louisiana (1979 to 1983). These three
studies were selected because the combined group would provide a sufficient sample to detect a
risk of lung cancer in excess  of 50% among workers in MER occupations.  The analyses were
restricted to males who had given occupational history. The Florida study was hospital based,
with cases ascertained through death certificates.  Controls were randomly selected from hospital
records and death certificates, excluding psychiatric diseases, matched by age and county. The
New Jersey study was population based, with cases ascertained through hospital records, cancer
registry, and death certificates. Controls were selected from among the pool of New Jersey
licensed drivers and death certificates.  The Louisiana study was hospital based (it is not
specified how the cases were ascertained), and controls were randomly selected from hospital
patients, excluding those with lung diseases and tobacco-related cancers.
       A total of 2,291 cases of male lung cancers and 2,570 controls were eligible, and the data
on occupations were collected by next-of-kin interviews for all jobs held for 6 months or more,
including the industry, occupation, and number of years employed.  The proportion of next-of-
kin interviews varied by site  from 50% in Louisiana to 85% in Florida. The coding schemes
were reviewed to identify MER occupations, which included truck drivers and heavy equipment
operators (cranes, bulldozers, and graders); bus drivers, taxi drivers, chauffeurs, and other motor
vehicle drivers; and automobile and truck mechanics.  Truck drivers were classified as routemen
and delivery men and other truck drivers.  All jobs were also classified with respect to potential
                                          7-40

-------
exposure to known and suspected lung carcinogens.  ORs were calculated by the maximum
likelihood method, adjusting for age by birth year, usual amount smoked, and study area.
Logistic regression models were used to examine the interrelationship of multiple variables.
       A statistically significant excess risk was detected for employment of 10 years or more
for all MER occupations (except truck drivers) adjusted for birth cohort, usual daily cigarette
use, and study area. The odds ratio for lung cancer using data gathered by direct interviews was
1.4 (95% CI = 1.1, 2.0), allowing for multiple MER employment, and 2.0 (95% CI = 1.3, 3.0),
excluding individuals with multiple MER employment. ORs for all MER employment, except
truck drivers who were employed for less than 10 years, were 1.3 (95% CI = 1.0, 1.7) and 1.3
(95% CI = 0.9, 1.8) including and excluding multiple MER employment, respectively. ORs
were then derived for specific MER occupations and, to avoid the confounding effects of
multiple MERjob classifications, analyses were also done excluding subjects with multiple
MER job exposures.  Truck drivers employed for more than 10 years had an odds ratio of 1.5
(95% CI = 1.1, 1.9). A similar figure was obtained excluding subjects with multiple MER
employment. An excess risk was not detected for truck drivers employed less than 10 years.
The only other job category that showed a statistically  significant excess for lung cancer
included taxi drivers and chauffeurs who worked multiple MER jobs for less than 10 years (OR
= 2.5, 95% CI =  1.4, 4.8). For the same category, the risk for individuals working in that job for
more than 10 years was 1.2 (95% CI = 0.5,  2.6).  A statistically significant positive trend
(p<0.05) with increasing employment of <2 years, 2 to 9 years, 10 to 19 years, and 20+ years
was observed for truck drivers but not for other MER occupations. A statistically nonsignificant
excess risk was also observed for heavy equipment operators, bus drivers, taxi drivers and
chauffeurs, and mechanics employed for 10 years or more. All of the above-mentioned ORs
were derived, adjusted for birth cohort, usual daily cigarette use, and State of residence.
Exposure to  other occupational suspect lung carcinogens did not account for the excess risks
detected.
       Results of this large study provide evidence that workers  in MER jobs are at an excess
risk of lung cancer that is not explained by their smoking habits or exposures to other lung
carcinogens. Because no information on type of engine had been collected, it was not possible
to determine if the excess risk was due to exposure to DE or gasoline exhaust or a mixture of the
two.  Among the study's other limitations are a possible bias due to misclassification of jobs
reported by the large proportion of next-of-kin interviews.  Such a bias would make the effect of
DE harder to detect due to broad categorization of jobs and the problems in classifying
individuals into uniform occupational groups based on the pooled data in the three studies that
used different occupational classification schemes.
                                          7-41

-------
7.2.2.7. Steenland et al (1990): A Case-Control Study of Lung Cancer and Truck Driving in
        the Teamsters Union
       Steenland et al. conducted a case-control study of lung cancer deaths in the Teamsters
Union to determine the risk of lung cancer among different occupations. Death certificates were
obtained from the Teamsters Union files in the central States for 10,485 (98%) male decedents
who had filed claims for pension benefits and who had died in 1982 and 1983.  Individuals were
required to have 20 years' tenure in the union to be eligible to claim benefits.  Cases comprised
all deaths (n = 1,288) from lung cancer, coded as ICD 162 or 163 for underlying or contributory
cause on the death certificate.  The 1,452 controls comprised every sixth death from the entire
file, excluding deaths from lung cancer, bladder cancer, and motor vehicle accidents.  Detailed
information on work history and potential confounders such as smoking, diet, and asbestos
exposure was obtained by questionnaire.  Seventy-six percent of the interviews were provided by
spouses and the remainder by  some other next of kin. The response rate was 82% for  cases and
80% for controls. Using these interview data and the 1980  census occupation  and industry
codes, subjects were classified either as nonexposed  or as having held other jobs with  potential
DE exposure. Data on job categories were missing for 12% of the study subjects.  A second
work history file was also created based on the Teamsters Union pension application that lists
occupation, employer, and dates of employment. A three-digit U.S. census code for occupation
and industry was assigned to each job for each individual. This Teamsters Union work history
file did not have information on whether men drove diesel or gasoline trucks, and the  four
principal occupations were long-haul drivers, short-haul or  city drivers, truck mechanics, and
dockworkers. Subjects were assigned the job category in which they had worked the longest.
       The case-control  analysis was done using unconditional logistic regression.  Separate
analyses were conducted for work histories from the Teamsters Union pension file and from
next-of-kin interviews.  Covariate data  were  obtained from  next-of-kin interviews.  Analyses
were also performed for  two time periods: employment after 1959 and employment after 1964.
These two cut-off years reflect years of presumed dieselization: 1960 for most trucking
companies and 1965 for  independent driver and nontrucking firms.  Data for analysis  could be
obtained for 994 cases and 1,085 controls using Teamsters Union work history and for 872 cases
and 957 controls using next-of-kin work history. When exposure was considered as a
dichotomous variable, for both Teamsters Union and next-of-kin work history, no single job
category had an elevated risk.  From the next-of-kin  data, diesel truck drivers had an odds  ratio
of 1.42 (95% CI = 0.74, 2.47)  and diesel truck mechanics had an odds ratio of 1.35 (95% CI  =
0.74,2.47).  ORs  by duration  of employment as a categorical variable were then estimated.  For
the Teamsters Union work history data, when only employment after 1959 was considered, both
long-haul (p<0.04) and short-haul drivers (not significant) showed an increase in risk with
                                          7-42

-------
increased years of exposure.  The length-of-employment categories for which the trends were
analyzed were 1 to 11 years,  12 to 17 years, and 18 years or more. Using 1964 as the cutoff
date, long-haul drivers continued to show a significant positive trend (p=0.04), with an odds
ratio of 1.64 (95% CI = 1.05, 2.57) for those who worked for 13+ years, the highest category.
Short-haul drivers, however,  did not show a positive trend when 1964 was used as the cutoff
date. Similar trend analysis was done for most next-of-kin data.  A marginal increase in risk with
increasing duration of employment as a truck driver (p=0.12) was observed.  For truck drivers
who primarily drove  diesel trucks for 35 years or longer, the odds ratio for lung cancer was 1.89
(95% CI = 1.04, 3.42).  Similarly, the corresponding odds ratio was 1.34 (95% CI = 0.81, 2.22)
for both gasoline truck drivers and drivers who drove both types of trucks, and 1.09 (95% CI =
0.44, 2.66) for truck mechanics.
       No significant interactions between age and DE exposure or smoking and DE exposure
were observed. All the ORs  were adjusted for age, smoking, and asbestos in addition to various
exposure categories.
       This  is a well-designed and analyzed study. The main strengths of the study are the
availability of detailed records from the Teamsters Union, a relatively large sample size,
availability of smoking data , and measurements of exposures. The authors acknowledge some
limitations of this study, which include possible misclassifications of exposure and smoking
habits, as information was provided by next of kin; lack of sufficient latency  to observe lung
cancer excess; and a small nonexposed group (n = 120).  Also, they could not evaluate the
concordance between Teamsters Union and next-of-kin job categories easily because job
categories were defined differently in each data set. No data were available on levels of diesel
exposure for the different job categories.  Despite these limitations, the positive findings  of this
study, which are probably underestimated, provide a positive evidence toward causal association
between DE exposure and excess lung cancer.

7.2.2.8. Steenlandet al.  (1998): DE and Lung Cancer in the Trucking Industry:
        Exposure-Response Analyses and Risk Assessment
       Steenland et al. (1998) conducted an exposure-response analysis by supplementing the
data from their earlier case-control study of lung cancer and truck drivers in the Teamsters
Union (Steenland et al., 1990) with exposure estimates based on a 1990 industrial hygiene
survey of elemental carbon exposure, a surrogate for DE in the trucking  industry.
       Study subjects were long-term Teamsters enrolled in the pension system who died during
the period 1982-1983. Using death certificate information, the researchers identified 994 cases
of lung cancer for the study period, and 1,085 non-lung-cancer deaths served as controls.
Subjects were divided into job categories based on the job each held the  longest.  Most had held
                                          7-43

-------
only one type of job.  The job categories were short-haul driver, long-haul driver, mechanic,
dockworker, other jobs with potential diesel exposure, and jobs outside the trucking industry
without occupational  diesel exposure. Smoking histories were obtained from next of kin. ORs
were calculated for work in an exposed job category at any time and after 1959 (an estimated
date when the majority of heavy-duty trucks had converted to diesel) compared with work in
nonexposed jobs.  ORs were adjusted for age, smoking, and potential asbestos exposure.  Trends
in effect estimates for duration of work in an exposed job were also calculated.
       An industrial hygiene survey by Zaebst et al. (1991) of elemental carbon exposures in the
trucking industry provided exposure estimates for each job  category in 1990. The elemental
carbon measurements were generally consistent with the epidemiologic results, in that mechanics
were found to have the highest exposures and relative risk,  followed by long-haul and then
short-haul drivers, although dockworkers had the highest exposures and the lowest relative risks.
       Past exposures were estimated assuming that they were a function of (1) the number of
heavy-duty trucks on the road, (2) the particulate emissions (grams/mile) of diesel engines over
time, and (3) leaks from truck exhaust systems for long-haul drivers. Estimates of past exposure
to elemental carbon, as a marker for DE exposure, for subjects in the case-control study were
made by assuming that average 1990 levels for a job category could be assigned to all subjects in
that category, and that levels prior to 1990 were directly proportional to vehicle miles traveled
by heavy-duty trucks and the estimated emission levels of diesel engines.  A 1975 exposure level
of elemental carbon in terms of micrograms per cubic meter was estimated by the following
equation: 1975 level = 1990 level*(vehicle miles 1975/vehicle miles 1990) (emissions
1975/emissions 1990). Once estimates of exposure for each year of work history were derived
for  each subject, analyses were conducted by cumulative level of estimated carbon exposure.
       Estimates were made for long-haul drivers (n = 1,237), short-haul drivers (n = 297),
dockworkers (n = 164), mechanics (n = 88), and those outside the trucking industry (n = 150).
Logistic regression was used to estimate ORs adjusted for five categories of age, race, smoking
(never, former-quitting before 1963, former-quitting in  1963 or later, current-with <1  pack per
day, and current-with 1 or more packs per day), diet, and reported asbestos exposure.  A variety
of models for cumulative exposure were considered, including a log-linear model with
cumulative exposure, a model adding a quadratic term for cumulative exposure, a log transform
of cumulative exposure, dummy variables for quartile of cumulative exposure, and smoothing
splines of cumulative exposure.  The estimates of rate ratios from  logistic regression for specific
levels of exposure to elemental carbon were then used to derive excess risk estimates for lung
cancer after lifetime exposure to elemental carbon.
        The survey found that mechanics had the highest current levels of DE exposures and
dockworkers who mainly used propane-powered forklifts had the lowest exposure.  ORs of 1.69
                                          7-44

-------
and 0.93 were observed for the mechanics and dockworkers, respectively. The finding of the
highest lung cancer risk for mechanics and lowest for dockworkers is indicative of causal
association between the DE exposure and development of lung cancer.  The log of cumulative
exposure was found to be the best-fitting model and was a significant predictor (p = 0.01).
However, the risk among mechanics did not increase with increasing duration of employment.
       OR for quartile of cumulative exposure show a pattern of significantly increasing trends
in risk with increasing exposure, ranging between 1.08 and 1.72, depending on the exposure
level and lag structure used. The lifetime excess risk of lung cancer death (through age 75) for a
male truck driver was estimated to be in the range of 1.4%-2.3% (95% confidence limits ranged
from 0.3% to 4.6%) above the background risk, depending on the emissions scenarios assumed.
The authors found that current exposures indicated that truck drivers are exposed to DE at levels
about the same as ambient levels on the highways, which are about double the background levels
in urban air. They conclude that the data suggest a positive and significant increase in lung
cancer risk with increasing estimated cumulative exposure to DE among workers in the trucking
industry.  They assert that these estimates suggest that the lifetime excess risk for lung cancer is
10 times higher than the OSHA standards, but caution that the results should be viewed as
exploratory.
       The authors acknowledge that the increasing trend in risk with increasing estimates of
cumulative exposure is partly due to the fact that a component of cumulative dose is simple
duration of exposure, and that analyses by simple duration also exhibit a positive trend with
duration.  This analysis essentially weights the duration by contrived estimates of exposure
intensity, and the authors acknowledge that this weighting depends on very broad assumptions.
       This is not an analysis of new data that provides independent estimates of relative risk for
DE and lung cancer incidence. Instead, it is an attempt to convert the data from Steenland's
earlier study of lung cancer for the purpose of estimating a different risk metric, "lifetime excess
risk of lung cancer," by augmenting these data with limited industrial hygiene data and
rationalizations about plausible models for cumulative exposure.
       The Health Effects Institute (HEI, 1999) and others have raised some concerns about the
exposure estimations, selection of controls, and control for confounding variables, and hence,
this study's usefulness for quantitative risk assessment. EPA and NIOSH will address these
concerns  in the year 2001.  The HEI (1999) panel noted that some of the strengths of this study
include the relevance of exposure levels to the general population and the use of an exposure
marker for diesel engine emissions that was an improvement over the concentration of
respirable-size particles (RSP). The number of study subjects (996 lung cancer cases) is large.
Histories of exposures to asbestos and smoking were obtained, and confounding by these two
variables was controlled in the analysis.  Thus, it should be noted that these concerns are about
                                          7-45

-------
the use of these data for quantitative risk assessment, due to limitations of the exposure data. As
far as qualitative risk assessment is concerned, this study is still considered to be positive and
strong.

7.2.2.9.  Boffetta et al. (1990):  Case-Control Study on Occupational Exposure to DE and
         Lung Cancer Risk
       This is an ongoing (since 1969) case-control study of tobacco-related diseases in 18
hospitals (six U.S. cities). Cases comprise 2,584 males with histologically  confirmed primary
lung cancers.  Sixty-nine cases were matched to 1 control, whereas 2,515 were matched to 2
controls. Controls were individuals who were diagnosed with non-tobacco-related diseases. The
matching was done for sex,  age (±2 years), hospital, and year of interview.  The interviews were
conducted at the hospitals at the time of diagnosis. In 1985, the occupational section of the
questionnaire was modified to include the usual occupation and up to five other jobs as well as
duration (in years) worked in those jobs.  After 1985, information was also obtained on exposure
to 45 groups of chemicals, including DE at the workplace or during hobby  activities. A priori
aggregation of occupations was categorized into low probability of DE exposure (reference
group), possible  exposure (19 occupations), and probable exposure (13 occupations). Analysis
was conducted based on "usual occupation" on all study subjects, and any occupation with
sufficient cases was eligible for further analysis.  In addition, cases enrolled after 1985 for which
there were self-reported DE exposure and detailed work histories were also analyzed separately.
       Both matched and unmatched analyses were done by calculating the adjusted (for
smoking and education) relative odds using the Mantel-Haenzael method and calculating the
test-based 95% confidence interval using the Miettinen method.  Unconditional logistic
regression was used to adjust for potential confounders (the PROC LOGIST of SAS).  Linear
trends for risk were also tested  according to Mantel.
       Adjusted relative odds for possible and probable exposure groups as well as the truck
drivers were slightly below  unity, none being statistically significant for the entire study
population. Although slight excesses were observed for the self-reported DE exposure group
and the subset of post-1985  enrollees for highest duration of exposure (for  self-reported
exposure, occupations with  probable exposure, and truck drivers), none was statistically
significant. Trend tests for the risk of lung cancer among self-reported DE exposure, probable
exposure, and truck drivers  with increasing exposure (duration of exposure used as surrogate for
increasing dose) were nonsignificant too.  Statistically significant lung cancer excesses were
observed for cigarette smoking only.
       The major strength of this study is availability of detailed smoking history. Even though
detailed information was obtained for the usual and five other occupations  (1985), because it
                                          7-46

-------
was difficult to estimate or verify the actual exposure to DE, duration of employment was used
as a surrogate for dose instead.  The numbers of cases and controls were large; however, the
number of individuals exposed to DE was relatively few,  thus reducing the power of the study.
This study did not attempt latency analysis either. Due to these limitations, the findings of this
study are unable to provide either positive or negative evidence for a causal association between
DE and occurrence of lung cancer.

7.2.2.10. Emmelin et al. (1993): DE Exposure and Smoking:  A Case-Referent Study of
         Lung Cancer Among Swedish Dock Workers
       This case-control study of lung cancer was drawn  from a cohort defined as all male
workers who had been employed as dockworkers for at least 6 months between 1950 and 1974.
In the population of 6,573 from 20 ports, there were 90 lung cancer deaths (cases), identified
through Swedish death and cancer registers, during the period 1960 to 1982. Of these 90 deaths,
the 54 who were workers at the 15 ports for which exposure surrogate information was available
were chosen for the case-control study.  Four controls, matched on port and age, were chosen for
each case from the remaining cohort who had survived to the time of diagnosis of the case. Both
live and deceased controls were included. The final analyses were done on 50 cases and 154
controls who had complete information on employment dates  and smoking data.  The smoking
strata were created by classifying ex-smokers as nonsmokers if they had not smoked for at least
5 years prior to the date of diagnosis of the case; otherwise they were classified as smokers.
       Relative odds and regression coefficients were calculated using conditional logistic
regression models. Comparisons were made both with  and without smoking included as a
variable, and the possible interaction between smoking and DE was tested.  Both  the weighted
linear regressions of the adjusted relative odds and the regression coefficients were used to test
mortality trends with all three exposure variables.
       Exposure to DE was assessed indirectly by initially measuring:  (1) exposure intensity
based on exhaust emission, (2) characteristics of the environment in terms of ventilation, and (3)
measures of proportion of time in higher exposed jobs.  For exhaust emissions, annual diesel fuel
consumption at a port was used as the surrogate. For ventilation, the annual proportion of ships
with closed or semiclosed holds was used as the surrogate.  The proportion of time spent below
decks was used as the surrogate for more exposed jobs. Although data were collected for all
three measures, only the annual fuel consumption was used  for analysis. Because every man
was likely to rotate through the various  jobs, the authors thought using annual consumption of
diesel fuel was the appropriate measure of exposure.  Consequently, in a second analysis, the
annual fuel consumption was divided by the number of employees in the same port that year to
come up with the fuel-per-person measure, which was further used to create a second measure,
                                         7-47

-------
"exposed time."  The "annual fuel" and exposed-time data were entered in a calendar time-
exposure matrix for each port, from which individual exposure measures were created. A third
measure, "machine time" (years of employment from first exposure), was also used to compare
the results with other studies. All exposure measures were accumulated from the first year of
employment or first year of diesel machine use, whichever came later. The last year of exposure
was fixed at 1979.  All exposures up to 2 years before the date of lung cancer diagnosis were
omitted from both cases and matched controls. A priori classification into three categories of
low, medium, and high exposure was done for all three exposure variables: machine time, fuel,
and exposed time.
       Conditional logistic regression models, adjusting for smoking status and using low
exposures and/or nonsmokers as a comparison group, yielded positive trends for all exposure
measures, but no trend test results were reported,  and only the relative odds for the exposed-time
exposure measure in the high-exposure group (OR = 6.8, 90% CI = 1.3 to 34.9) was reported as
statistically significant. For smokers, adjusting for DE exposure level, the relative odds were
statistically significant and about equal for all three exposure variables: machine time, OR = 5.7
(90%  CI = 2.4 to 13.3); fuel, OR = 5.5 (90% CI = 2.4 to 12.7); and exposed time, OR = 6.2
(90%  CI = 2.6 to 14.6). Interaction between DE and smoking was tested by conditional logistic
regression in the  exposed-time variable.  Although there were positive trends for both smokers
and nonsmokers, the trend for smokers was much steeper: low, OR = 3.7 (90% CI = 0.9 to 14.6);
medium, OR = 10.7 (90% CI = 1.5 to 78.4);  and high, OR = 28.9 (90% CI = 3.5 to 240),
indicating more than additive interaction between these two variables.
       In the weighted linear regression model with the exposed-time variable, the results were
similar to those using the logistic regression model. The authors also explored the smoking
variable further in various analyses, some of which  suggested a strong interaction between DE
and smoking. However, with just six nonsmokers and no further categorization of smoking
amount or duration, these results are of limited value.
       The DE exposure matrices created using three different variables are intricate.  Analyses
by any of these variables yield essentially the same  positive results and positive trends, providing
consistent support for a real effect of DE exposure,  at least in smokers. However,
methodological limitations to this study prevent a more  definitive conclusion.  The numbers of
cases  and controls are small.  There are very  few nonsmokers; thus, testing the effects of DE
exposure in them is futile. Lack of information on asbestos exposure, to which dockworkers are
usually exposed,  may also confound the  results. Also, no latency analyses are presented.
Overall, despite these limitations, this study supports the earlier findings of excess lung cancer
mortality among  individuals exposed to DE.
                                          7-48

-------
7.2.2.11. Swanson et al. (1993): Diversity in the Association Between Occupation and Lung
         Cancer Among Black and White Men
       This population-based case-control study of lung cancer was conducted in metropolitan
Detroit. The cases and controls for this study were identified from the Occupational Cancer
Incidence Surveillance Study (OCISS). A total  of 3,792 incident lung cancer cases and 1,966
colon and rectal cancer cases used as controls, diagnosed between 1984 and 1987 among white
and black males aged 40 to 84 years, were selected for the study. Information was obtained by
telephone interview either with the individual or a surrogate about lifetime work history and
smoking history, as well as medical, demographic, and residential history.  Occupation and
industry data were  coded using the 1980 U.S. Census Bureau classification codes.  The
investigators selected certain occupations and industries as having little or no exposure to
carcinogens and defined them as an unexposed group. Analysis was done using logistic
regression method  and adjusting for age at diagnosis, pack-years of cigarette smoking, and race.
       The results  were presented by various occupations and industries; those with potential
exposures to DE were drivers of heavy trucks and light trucks, farmers, and railroad workers,
respectively. Among white males, increasing lung cancer risks were observed with increasing
duration of employment for drivers of heavy trucks, drivers  of light trucks, and farmers.
Although none of the individual  ORs were statistically significant, trend tests were significant
for all three occupations (p<0.05). On the other hand, among black males increasing  lung
cancer risks with increasing duration of employment were observed for farmers only,  with an
OR of 10.4 (95% CI = 1.4, 77.1) reaching significance for employment of 20+ years.  As for the
railroad industry, increasing lung cancer risks with increasing duration of employment were
observed for both white and black males. The trend test was significant for white males only,
with an OR of 2.4 (95% CI = 1.1, 5.1) reaching  significance for employment of 10+ years.
       The main strengths of the study are large sample size, availability of lifetime work
history and smoking history, and the population-based study format, precluding selection bias.
The major limitation, as in other studies, is lack of direct information on specific exposures. The
interesting result of this study is lung cancer excesses observed in farmers, mainly among crop
farmers, who have  potential exposure to DE from their tractors in addition to pesticides,
herbicides,  and other PM10 The authors point out that this is the first study to find excess lung
cancer in this occupation.

7.2.2.12. Hansen et al. (1998):  Increased Risk of Lung Cancer Among Different Types of
         Professional Drivers in Denmark
       This is a population-based case-control study of lung cancer, conducted in professional
drivers in Denmark. The cases first diagnosed as primary lung cancer between 1970 and 1989
                                          7-49

-------
among males born between 1897 and 1966 were identified from the Danish Cancer Registry.
The registry provided the information on diagnosis from ICD-7, name, sex, and unique personal
identification number (PIDN). Information about past employment was obtained by linkage
with the nationwide pension fund.  The fund keeps the records by name and PIDN about the date
of start and end of each job and unique company number of the employer. The records are kept
even after the employee has retired or died. Information about current employment was
obtained from the Danish Central Population Registry (CPR) by linkage with the PIDN.
       Of 37,597 cases identified from the Registry, 8,853 did not have any employment
records. Controls (1:1) for 28,744 lung cancer cases with employment histories were selected
randomly from CPR, matched with the case by year of birth and sex. Furthermore, these
controls had to be alive, cancer free, and employed prior to the diagnosis of lung cancer in the
corresponding case. Employment histories were obtained for the controls in the same fashion as
cases from the pension fund.  The employment record search resulted in a total of 1,640
lorry/bus drivers  and 426 taxi drivers. They were further divided into subgroups by their
duration of employment. Information about smoking in drivers was acquired from two national
surveys conducted in 1970-72 and 1983.  No direct information on smoking was available in
either cases or controls. A separate case-control study of mesothelioma indirectly looked at
asbestos exposure among professional drivers. OR, adjusting for socioeconomic status and 95%
CI, were computed using conditional logistic regression (PECAN procedure in the statistical
package EPICURE).
       Significant ORs for lung cancer were found for lorry/bus drivers (OR = 1.31, 95% CI =
1.17, 1.46), taxi drivers (OR = 1.64, 95% CI = 1.22, 2.19), and unspecified drivers (OR = 1.39,
95% CI = 1.30, 1.51).  Significant ORs were found for both lorry/bus drivers and taxi drivers by
duration of employment in 1-5 years and >5 years categories, with no lag time and with a 10-
year lag time. The  ORs remained the same for lorry/bus drivers in these employment categories
for no lag time and 10-year lag time. Among taxi drivers,  on the other hand, the OR of 2.2 in >5
year employment in no-lag-time analysis increased to 3.0 in the 10-year lag time analysis. The
authors asserted that the higher risk seen in the taxi drivers may be due to higher exposure
attributable due to longer time spent in traffic congestion.  The trend tests for increasing risk
with increasing duration of employment (surrogate for exposure) were statistically significant
(/X0.001) for both lorry/bus drivers and taxi drivers in no-lag-time and  10-year lag time
analysis.  All the ORs were adjusted for socioeconomic status.
       The main strengths of the study are the large sample size, availability of information on
socioeconomic status, and detailed employment records. The main limitation, however, is lack
of information on what type of fuel these vehicles used. It is probably safe to assume that the
lorry/buses  were  diesel powered, whereas the taxis could be either diesel or gasoline powered.  A
                                         7-50

-------
personal communication with Dr. Johnni Hansen confirmed that dieselization in Denmark was
completed in the late 1940s and lorries, buses, and taxis have been using diesel fuel since then.
Although direct adjustments were not done for smoking and exposure to asbestos, indirect
information on both these confounders indicates that they are unlikely to explain the observed
excesses and the increasing risk with increasing duration of employment.  Thus, the results of
this study are strongly supportive of DE being associated with increased lung cancer.

7.2.2.13.  Bruske-Hohlfeld et al (1999): Lung Cancer Risk in Male Workers Occupationatty
         Exposed to Diesel Motor Emissions in Germany
       This paper presents a pooled analysis of two case-control studies of lung cancer. The
first study, by Jockel et al. (1995, 1998), was conducted between 1988 and 1993 and had 1,004
cases and 1,004 controls matched for sex, age, and region of residence,  selected randomly from
the compulsory municipal registries. The inclusion criteria for cases were: they should have
been born in or after 1913, should have been of German nationality, and should have been
diagnosed with lung cancer within 3 months prior to the interview. The second study, by
Wichmann  et al. (1998), was ongoing when it was included in this study. The study span
covered the years 1990 to 1996.  By 1994 a total of 3,180 cases and 3,249 controls, randomly
selected from the compulsory population registries, were frequency matched on sex, age, and
region. The cases were less than 76 years old, were residents of the region and living in
Germany for more than 25 years, and had a diagnosis not more than 3 months old.  Of 4,184
pooled cases and 4,253 pooled controls,  the analysis was conducted on 3,498 male cases and
3,541 male controls.  A personal interview was conducted with each study participant.  Data
were collected on basic demographic information, detailed smoking history, and lifelong
occupational history about jobs held and industries worked in.  The job  titles and industries were
classified into 33 and 21 categories, respectively, using the German Statistical Office codes.
       Based on job codes with potential exposure to diesel motor emission (DME), four
exposure groups were constituted. Group A comprised professional drivers of trucks, buses,
taxis, etc. Group B comprised other traffic-related jobs such as switchmen, diesel locomotive
drivers, and diesel forklift truck drivers.  Group C comprised bulldozer operators, graders, and
excavators. Group D comprised full-time farm tractor drivers.  Validation of the jobs was done
by written evaluation of the job task descriptions, which also avoided misclassification.  The
following information was acquired for the construction of job task descriptions: (1) What
were your usual tasks at work and how often (in % of daily working hours) were they
performed?  (2)  What did you produce,  manufacture, or transport?  (3) Which material was
used? (4)  What kind of machine did you operate?  Some individuals had more than one job task
                                          7-51

-------
with DME exposure.  The exposure assessment was done without knowing the status of the
case/control.
       For each individual, cumulative exposure was calculated for the complete work history
by categorizing the duration of exposure as >0-3, >3-10, > 10-20, >20-30, >30 years, and
beginning and end of exposure. The first year of exposure was defined as <1945, 1946-1955,
and >1956 while the last year of exposure was defined as <1965, 1966-1975, and >1976. For
professional drivers, hours driven per day were accumulated and were classified as "driving
hours."
       A smoker was defined as any individual who had smoked regularly for at least 6 months.
Smoking information was acquired in series with the starting time, type of tobacco, amount
smoked, duration in years, and calender year of quitting. Asbestos exposure was estimated by
certain job-specific supplementary questions.
       The cases and controls were post-hoc stratified into 6 age and 17 region categories. ORs
adjusted for smoking and asbestos exposure were calculated by conditional logistic regression,
using "never exposed" workers as the reference group.  The adjustment for cigarette smoking
was done by using pack-years as a continuous variable; adjustment for other tobacco products
was done by considering them as a binary variable.  A total of 716 cases and 430 controls were
found to be ever exposed to DME.  The smoking- and asbestos-adjusted OR of 1.43 (95% CI =
1.23, 1.67) for all DME exposed was reduced from the crude OR of 1.91. For the entire group
the various analyses yielded statistically significant ORs ranging from 1.25 to 2.31, adjusted for
smoking and asbestos exposure (West Germany, >10-20 years and >20-30 years of exposure,
first year of exposure in 1946-1955 and 1956+, end  of exposure in 1966-1975 and  1976+,  and
for the job categories of Group A, B, and C). The risk increased with  increasing years of
exposure, and for both the first year of exposure (<1945, 1946-1955, and >1956) and end year of
exposure (<1965, 1966-1975, and >1976).
       Separate analyses by four job categories (all  the ORs were adjusted for smoking and
asbestos exposure)  showed that for professional drivers (Group A) the overall OR was 1.25
(95% CI = 1.05, 1.47). Significant ORs were found for various factors in West Germany only.
The factors were: >0-3 years and >10-20 years of exposure (OR = 1.69, 95% CI =1.13, 2.53,
and OR = 2.02, 95% CI = 1.32, 3.08, respectively), beginning of exposure in 1956+ and end of
exposure in 1976+(OR = 1.56, 95% CI = 1.21, 2.03, and OR = 1.5, 95% CI = 1.14, 1.98,
respectively), and 1,000-49,999 driving hours (OR = 1.54, 95% CI =1.15, 2.07). None of the
ORs were significant in East Germany in this group.
       For other traffic-related jobs (Group B) the overall OR was 1.53 (95% CI = 1.04, 2.24).
The ORs for beginning of exposure in 1956+ and end of exposure in 1976+ were OR = 1.71,
95% CI = 1.05, 2.78,  and OR = 2.68, 95% CI = 1.47, 4.90, respectively. The risk increased with
                                         7-52

-------
increasing duration of exposure and was statistically significant for >10-20 years (OR = 2.49)
and more than 20 years (OR = 2.88). No separate analyses for West Germany and East
Germany were presented in this category.
       For heavy equipment operators (Group C) the overall OR of 2.31 (95% CI = 1.44, 3.7)
was highest among all the job categories.  Significant ORs were observed for beginning
exposure in 1946-1955 (OR = 2.83, 95% CI = 1.10, 7.23) and end exposure in 1966-1975 (OR =
3.74, 95% CI = 1.20, 11.64).  The risk increased with increasing duration of exposure and was
statistically significant for more than 20 years of exposure (OR = 4.3). Although no separate
analyses for West Germany and East Germany were presented, investigators mentioned that for
this job group hardly any difference was seen between West Germany and East Germany.
       For drivers of the farming tractors (Group  D) the overall OR of 1.29 was not significant.
Risk increased with increasing duration of exposure and was significant for exposure of more
than 30 years (OR  = 6.81, 95% CI = 1.17, 39.51).  No separate analyses for West Germany and
East Germany were presented in this category.
        The professional drivers and the other traffic-related job categories probably have mixed
exposures to gasoline exhaust in general traffic. On the other hand, it should be noted that
exposure to DME among heavy equipment and farm tractor drivers is much higher and not as
mixed as in professional drivers.  The heavy equipment drivers usually drive repeatedly through
their own equipment's exhaust.  Therefore, the observed highest risk for lung cancer in this job
category establishes a direct link with the DME. The only other study that found significantly
higher risk for heavy equipment operators (RR = 2.6) was conducted by Boffeta et al. (1988).
Although the only  significant excess was observed for farming tractor operators among
individuals with more than 30 years of exposure, a steady increase in risk was observed for this
job category with increasing exposure. The investigators stated that the working conditions and
the DME of tractors remained fairly constant over the years.  This increase may be due mainly to
exposure to DME and, in addition, PM10
       This is a well-designed, well-conducted, and well-analyzed study. Its main strengths are
large sample size, resulting  in good statistical power; inclusion of incident cases that were
diagnosed not more than 3 months prior to the interview; use of only personal interviews,
reducing recall bias; diagnosis ascertained by cytology  or histology; and availability of lifelong
detailed occupational and smoking history. Exposure estimation for each individual was based
on job codes and industry codes, which were validated by written job descriptions to avoid
misclassification. The main limitation of the study is lack of data on actual exposure to DME.
The cumulative quantitative exposures were calculated based on time spent in each job with
potential exposure  to DME  and the type of equipment used. Thus, this study provides strong
evidence for a causal association between exposure to DE and occurrence of lung cancer.
   Table 7-2 summarizes the above lung cancer case-control  studies.
                                          7-53

-------
     Table 7-2.  Epidemiologic studies of the health effects of exposure to DE: case-control studies of lung cancer
  Authors         Population studied        DE exposure assessment              Results                        Limitations
Hall and     502 histologically confirmed
Wynder     lung cancers
(1984)      Cases diagnosed 12 mo prior to
            interviews

            502 matched hospital controls
            without tobacco-related diseases,
            matched for age, sex, race, and
            geographical area

            Population from 18 hospitals in
            controls
Based on previous
Industrial Hygiene
Standards for a
particular occupation,
usual lifetime occupation
coded as "probably high
exposure" and "no
exposure"

NIOSH standards used
to classify exposures:
High
Moderate
Low
SNS excess risk after adjustment
for smoking for lung cancer:
RR= 1.4 (1st criteria)
and
RR = 1.7 (NIOSH criteria)
Complete lifetime employment
history not available

Serf-reported occupation history not
validated

No analysis by dose, latency, or
duration of exposure

No information on nonoccupational
diesel exposure
Damber and 589 lung cancer cases who had
Larsson     died prior to 1979 reported to
(1987)      Swedish registry between 1972
            and 1977

            582 matched dead controls (sex,
            age, year of death, municipality)
            drawn from National Registry
            of Cause of Death

            453 matched living controls
            (sex, year of birth, municipality)
            drawn from National
            Population Registry
Occupations held for at    For underground miners:  SSOR=  Uncertain DE exposure
least 1 year or more
2.7 (> 1 year of employment)
A 5-digit code was used to SS OR = 9.8 (>20 years of
classify the occupations
according to Nordic
Classification of
Occupations
employment)

For professional drivers:  SNS OR
= 1.2 (>20 years of employment)
with dead controls

All ORs adjusted for smoking
No validation of exposure done

Underground miners data not
adjusted for other confounders such
as radon, etc.

-------
Table 7-2. Epidemiologic studies of the health effects of exposure to DE: case-control studies of lung cancer (continued)
  Authors         Population studied       DE exposure assessment              Results                        Limitations
Lerchen    506 lung cancer cases from
et al. (1987) New Mexico tumor registry
            (333 males and 173 females)

            Aged 25-84 years

            Diagnosed between January 1,
            1980, and  December 31, 1982

            771 (499 males and 272 females)
            frequency  matched with cases,
	selected from telephone directory
Lifetime occupational
history and self-reported
exposure history were
obtained

Coded according to
Standard Industrial
Classification Scheme
No excess of relative odds were
observed for DE exposure
Exposure based on occupational
history and serf-report, which was not
validated

50% occupational history provided
by next of kin

Absence of lung cancer association
with asbestos suggests
misclassification of exposure
Garshick    1,319 lung cancer cases who died  Personal exposure
etal. (1987) between March 1, 1981,          assessed for 39 job
            and February 28, 1982           categories
                        SS OR = 1.41 (<64 year age group) Probable misclassification of DE
                                                        exposure jobs
                        SS OR = 1.64 (<64 year age group)
            2,385 matched controls (two each,
            age and date of death)

            Both cases and controls drawn
            from railroad worker cohort
            who had worked for 10 or
            more years
This was corrected with
job titles to dichotomize
the exposure into:
Exposed
Not exposed

Industrial hygiene
sampling done	
for >20 years DE exposure group
when compared to 0- to 4-year
exposure group

All ORs adjusted for lifetime
smoking and asbestos exposure
Years of exposure used as surrogate
for dose

13% of death certificates not
ascertained

Overestimation of smoking history
Benhamou  1,260 histologically confirmed
etal. (1988) lung cancer cases

            2,084 non-tobacco-related
            disease matched controls
            (sex, age at diagnosis,
            hospital admission, and
            interviewer)

            Occurring between 1976 and
            1980 in France
Based on exposures
determined by panel of
experts

The occupations were
recorded blindly using
International Standard
Classification of
Occupations as chemical
or physical exposures
Significant excess risks were found Exposure based on occupational
in motor vehicle drivers
(RR= 1.42) and
transport equipment operators
(RR =1.35) (smoking adjusted)
histories not validated

Exposures classified as chemical and
physical exposures, not specific to
DE

-------
Table 7-2. Epidemiologic studies of the health effects of exposure to DE:  case-control studies of lung cancer (continued)
  Authors         Population studied        DE exposure assessment             Results                        Limitations
Hayes et al. Pooled data from three different
(1989)      studies consisting of 2,291 male
            lung cancer cases

            2,570 controls
Occupational information
from next of kin for all
jobs held

Jobs classified with
respect to potential
exposure to known and
suspected pulmonary
carcinogens	
SS OR = 1.5 for truck drivers (>10
years of employment)

SS positive trend with increasing
employment as truck driver

Adjusted for age, smoking, & study
area
Exposure data based on job
description given by next of kin,
which was not validated

Could have been mixed exposure to
both diesel and gasoline exhausts

Job description could have led to
misclassification
 Steenland   1,058 male lung cancer deaths
 et al. (1990) between 1982 and 1983

            1,160, every sixth death from
            entire mortality file, sorted by
            Social Security number
            (excluding lung cancer,
            bladder cancer, and motor
            vehicle accidents)

            Cases and controls were from
            Central State Teamsters who
            had filed claims (requiring
	20-year tenure)	
Longest job held:  diesel
truck driver, gasoline
truck driver, both types
of trucks, truck
mechanic, and
dockworkers
As 1964 cut-off point:

SS OR = 1.64 for long-haul drivers
with 13+years of employment

Positive trend test for long-haul
drivers (p=0.04)

SS OR = 1.89 for diesel truck
drivers of 35+ years of
employment

Adjusted for age, smoking, &
asbestos
Exposure based on job titles not
validated

Possible misclassification of exposure
and smoking, based on next-of-kin
information

Lack of sufficient latency
 Steenland et Exposure-response analyses of
 al. (1998)   their 1990 case-control study
Industrial hygiene data of  For mechanics: OR = 1.69 (had
elemental carbon in       the highest DE exposure)
trucking industry collected
by Zaebst et al. (1991)     Lowest DE exposure and lowest
used to estimate individual OR = 0.93 observed for
exposures                dockworkers
                                           Cumulative exposures
                                           calculated based on
                                           estimated lifetime
                                           exposures
                         Increasing risk of lung cancer with
                         increasing exposure

                         Adjusted for age & smoking

-------
Table 7-2. Epidemiologic studies of the health effects of exposure to DE:  case-control studies of lung cancer (continued)
  Authors         Population studied       DE exposure assessment             Results                        Limitations
Boffetta et  From 18 hospitals (since 1969),
al. (1990)   2,584 male lung cancer cases
            matched to either one control (69)
            or two controls (2,515) were
            drawn. Matched on age, hospital,
            and year of interview
A priori aggregation of
occupations categorized
into low probability,
possible exposure (19
occupations), and
probable exposure (13
occupations) to DE
OR slightly below unity SNS

Adjusted for smoking
No verification of exposure

Duration of employment used as
surrogate for dose

Number of individuals exposed to DE
was small
Emmelin et 50 male lung cancer cases from
al. (1993)   15 ports (worked for at least
            6 months between 1950 and
            1974), 154 controls matched on
            age and port
Indirect DE exposure
assessment done based on
(1) exposure intensity, (2)
characteristics of
ventilation, (3) measure of
proportion of time in
higher exposure jobs
SS OR for high-exposure group =
6.8

Positive trend for DE observed
(trend much steeper for smokers
than nonsmokers)

Adjusted for smoking
Numbers of cases and controls are
small

Very few nonsmokers

Lack of exposure information on
asbestos

No latency analysis
 Swanson et  Population based case-control
 al. (1993)   study in metropolitan Detroit
Telephone interviews with SS excess ORs observed for
the individual or surrogate - black farmers OR= 10.4 for 20+
                                Lack of direct information on
                                specific exposures
                                           about lifetime work
            3,792 lung cancer cases and 1,966 history
            colon cancer (cases) controls,
            diagnosed between 1984 and 1987 Occupation and industry
            in white and black males (aged    data coded per 1980 U.S.
            between 40-84)                  Census Bureau
                                           classification codes
                        years employment
                        - white railroad industry workers   No latency analysis
                        OR= 2.4 for 10+ years employment

                        Among white trend tests were SS
                        for
                        -drivers of heavy duty trucks
                        - drivers of light duty trucks
                        - farmers
                                           Certain occupations and
                                           industries were selected as - railroad workers
                                           unexposed to carcinogens
                                                                   Among blacks trend test was SS for
                                                                   farmers only

                                                                   All the ORs were adjusted for age
                                                                   at diagnosis, pack-years of
                                                                   cigarette smoking and race

-------
        Table 7-2.  Epidemiologic studies of the health effects of exposure to DE:  case-control studies of lung cancer (continued)
          Authors         Population studied       DE exposure assessment             Results                        Limitations
        Hansen et  Population-based case-control
        al. (1998)   study of professional drivers in
                   Denmark
                               Information about past
                               employment obtained by
                               linkage with nationwide
                               pension fund
Male lung cancer cases diagnosed
between 1970-1989, controls      Employment as lorry/bus
matched by year of birth and sex   drivers (n= 1,640) and taxi
                               drivers (n=426) was used
                               as surrogate for exposure
                               toDE
For lorry /bus drivers: SS OR =1.31 Lack of information on the type of
                                fuel (personal communication with
                                the principal investigator confirmed
                                that diesel fuel is used for the
                                lorry /buses and taxis since early
                                1960s)
For taxi drivers: SS OR = 1.64,
which increased to 2.2 in > 5-year
employment with no lag time &
3.0 in > 5 year employment with
10-year lag time
                                                                          SS trend test for increasing risk
                                                                          with increasing employment for
                                                                          both lorry /bus drivers & taxi
                                                                          drivers (pO.OOl)

                                                                          All ORs adjusted for
                                                                          socioeconomic status
                                Even though direct adjustment was
                                not done for smoking/asbestos,
                                indirect methods indicate that the
                                results are not likely to be
                                confounded by these factors

-------
Table 7-2.  Epidemiologic studies of the health effects of exposure to DE:  case-control studies of lung cancer (continued)
  Authors         Population studied       DE exposure assessment              Results                        Limitations
 Briiske-    Pooled analysis of two case-
 Hohlfeld et control studies (3,498 cases &
 al. (1999)   3,541 controls)

            Controls frequency matched on
            sex, age, & region, randomly
            selected from the compulsory
            population registry

            Inclusion criteria: (1) born in or
            after 1913/less than 75 years old,
            (2) German nationality/resident of
            the region - lived in Germany for
            more than 25 years, & (3) lung
            cancer diagnosis should be 3
            months prior to the study

            Information obtained by personal
            interview on:
Lifetime detailed
occupational & smoking
histories obtained from
each individual in a
personal interview

Based on job codes (33
job titles & 21 industries)
potential DE exposure
classified in 4 categories:
A- professional drivers of
trucks, buses, & taxis; B-
other traffic related i.e.,
switchman, locomotive, &
forklift drivers; C-
bulldozer operators,
graders,& excavators; D-
farm tractor drivers

Cumulative DE exposures
and pack-years (smoking)
calculated for each
individual
SS higher risk adjusted for smoking Lack of data on actual exposure to
observed for all 4 categories:

A- ORs ranged from 1.25 to 2.53
B- ORs ranged from 1.53 to 2.88
C- ORs ranged from 2.31 to 4.3
D- 6.81 (exposure < 30 years)

Risk increased with increasing
exposure
diesel exaust
Abbreviations: OR = odds ratio; RR = relative risk; SNS = statistically nonsignificant; SS = statistically significant.

-------
7.2.3. Summaries of Studies and Meta-Analyses of Lung Cancer
7.2.3.1. Cohen and Higgins (1995): Health Effects of DE: Epidemiology
       The Health Effects Institute (HEI) reviewed all published epidemiologic studies on the
health effects of exposure to DE available through June 1993, identified by a MEDLINE search
and by reviewing the reference sections of published research and earlier reviews. HEI
identified 35 reports of epidemiologic studies (16 cohort and 19 case-control) of the relation of
occupational exposure to diesel emissions and lung cancer published between 1957 and 1993.
HEI reviewed the 35  reports for epidemiologic evidence of health effects of exposure to DE for
lung cancer, other cancers, and nonmalignant respiratory disease. They found that the data were
strongest for lung cancer. The evidence suggested that occupational exposure to DE from
diverse sources increases the rate of lung cancer by 20% to 40% in exposed workers generally,
and to a greater extent among workers with prolonged exposure. They also found that the results
are not explicable by confounding caused by cigarette smoking or other known sources of bias.
       Control for smoking was identified in 15  studies.  Six studies (17%) reported relative risk
estimates less than 1; 29 studies (83%) reported at least one relative risk greater than one
indicating positive association. Twelve studies indicating a relative risk greater than 1 had 95%
confidence intervals,  which excluded unity.
       The authors conclude that epidemiologic data consistently show weak associations
between exposure to DE and lung cancer. They find that the evidence suggests that long-term
exposure to DE in a variety of occupational circumstances is associated with a 1.2- to 1.5-fold
increase in the relative risk of lung cancer compared with workers classified as unexposed. Most
of the studies that controlled for smoking found that the association between increased risk of
lung cancer and exposure to DE persisted after such controls were applied, although in some
cases the excess risk was lower. None of the studies measured exposure to diesel emissions or
characterized the actual emissions from the source of exposure for the time period most relevant
to the development of lung cancer. Most investigators classified exposure based on work
histories reported by  subjects or their next of kin, or by retirement records. Although these data
provide relative rankings of exposure, the absence of concurrent exposure information is the key
factor that limits interpretation of the epidemiologic findings and subsequently their utility in
making quantitative estimates of cancer risks.
       This is a comprehensive and thorough narrative review of studies of the health effects of
DE. It does not undertake formal estimation of summary measures of effect or evaluation of
heterogeneity in the results. The conclusion drawn about the consistency of the  results is based
on the author's assessment of the failure of potential biases and alternative explanations for the
increase in risk to account for the observed consistency.  In many if not most studies, the quality
of the data used to control confounding was relatively crude. Although the studies do include
qualitative assessment of whether control for smoking is taken into account, careful scrutiny of
                                          7-60

-------
the quality of the control or adjustment for smoking among the studies is absent. This leaves
open the possibility that prevalent residual confounding by inadequate control for smoking in
many studies may account for the consistent associations seen.

7.2.3.2. Bhatiaetal (1998): DE Exposure and Lung Cancer
       Bhatia et al. (1998) report a meta-analysis of 29 published2 cohort and case-control
studies of the relation between occupational exposure to DE and lung cancer.  A search of the
epidemiologic literature was conducted for all studies concerning lung cancer and DE exposure.
Occupational studies involving mining were excluded because of concern about the possible
influence of radon  and silica exposures.  Studies in which the minimum interval from time of
first exposure to end of follow-up was less than 10 years, and studies in which work with diesel
equipment or engines could not be confirmed or reliably inferred, were excluded. When studies
presented  risk estimates for more than one specific occupational category of DE-exposed
workers, the subgroup risk estimates were used in the meta-analysis.  Smoking-adjusted effect
measures were used when present.
       Of 29 studies 23 met the criteria for inclusion in the meta-analysis.  The observed
relative risk estimates were greater than 1 in 21 of these studies; this result is unlikely to be due
to chance. The pooled relative risk weighted by study precision was 1.33 (95% CI = 1.24, 1.44),
indicating increased relative risk for lung cancer from occupational exposure to DE.
Subanalyses by study design (case-control and cohort studies) and by control for smoking
produced results that did not differ from those of the overall pooled analysis.  Cohort studies
using internal comparisons showed higher relative risks than those using external comparisons
(see Figure 7-1).
       Bhatia and  colleagues conclude that the analysis shows a small but consistent increase in
the risk for lung cancer among workers with exposure to DE.  The authors evaluate the
dependence  of the relative risk estimate on the presence of control for smoking among studies,
and provide a table that allows assessment of whether the quality of the data contributing to
control for smoking is related to the relative risk estimates (albeit in a limited number of
studies). Bhatia et al. assert that residual confounding is not affecting the summary estimates or
conclusions  for the following reasons: (1) the pooled relative risks for studies adjusted for
smoking were the same as those for studies not adjusting for smoking; (2) in those studies giving
risk estimates adjusted for smoking and risk estimates not adjusted for smoking, there was only a
small reduction in the pooled relative risk from DE exposure; and (3) in studies with internal
        Of 35 studies identified in the literature search, 6 pairs of studies represented analyses of the same study
population, reducing the number of studies to 29.
                                          7-61

-------
                     0.5
                                       RR estimates & 95% Cl
                                     1              1.5
               All Studies
         CaseControl Studies
            Cohort Studies
         internal Comparison
           Population
        External Comparison
           Population
          Smoking Adjusted
        Smoking Not Adjusted
            Sub-analysis by
             Occupation
           Railroad Workers
        Equipment Operators
             Truck Drivers
             Bus Workers
       I	n-
i	n-
                                                I	D
Figure 7-1. Pooled relative risk estimates and heterogeneity-adjusted 95% confidence
intervals for all studies and subgroups of studies included in the meta-analysis.

Source: Bhatiaetal., 1998.
comparison populations, in which confounding is less likely, the pooled relative risk estimate
was 1.43.
       The validity of this assessment depends on the adequacy of control for smoking in the
individual studies.  If inadequate adjustment for smoking is employed and residual confounding
by cigarette smoking pertains in the result of the individual studies, then the comparisons and
contrasts of the pooled estimates the authors cite as reasons for dismissing the effect of residual
confounding by smoking will remain contaminated by residual confounding in the individual
studies. In fact, Bhatia et al. erroneously identify the treatment of the smoking data in the main
                                            7-62

-------
analysis for the 1987 report by Garshick et al. as a continuous variable representing pack-years
of smoking, whereas the analysis actually dichotomized the pack-years data into two crude dose
categories (above and below the 50 pack-years level).  This clearly reduced the quality of the
adjustment for smoking, which already suffered from the fact that information on cumulative
cigarette consumption was missing for more than 20% of the lung cancer cases. In this instance,
the consistency between the adjusted and unadjusted estimates of the relative risk for DE
exposure may be attributable to failure of adjustment rather than lack of confounding by
cigarette smoking. A similar problem exists for the Bhatia et al. representation of the control for
confounding in the study by Boffetta and  Stellman (1988).
       An evaluation of the potential for publication bias is presented that provides reassurance
that the magnitude of published effects is  not a function of the precision or study power;
however, this assessment cannot rule out the possibility of publication bias.

7.2.3.3. Lipsett and Campleman (1999): Occupational Exposure to DE and Lung Cancer: A
       Meta-Analysis
       Lipsett and Campleman (1999) conducted electronic searches to identify epidemiologic
studies published between 1975 and 1995 of the relationship of occupational exposure to DE and
lung cancer.  Studies were selected based  on the following criteria: (1) Estimates of relative risks
and their standard errors must be reported or derivable from the information presented. (2)
Studies must  have allowed for a latency period of 10 or more years for development of lung
cancer after onset of exposure. (3) No obvious bias resulted from incomplete case ascertainment
in follow-up studies. (4) Studies must be independent: that is, a single representative study
selected from any set of multiple analyses of data from the same population.  Studies focusing
on occupations involving mining were excluded because of potential confounding by radon,
arsenic, and silica, as well as possible interactions between cigarette smoking and exposure to
these substances in lung cancer induction.
       Thirty of the 47 studies initially identified as relevant met the specified inclusion criteria.
Several risk estimates were extracted from six studies reporting results from multiple mutually
exclusive diesel-related occupational subgroups.   If a study reported effects associated with
several levels or durations of exposure, the effect reported for the highest level  or longest
duration of exposure was used. If estimates for several occupational subsets were reported, the
most diesel-specific occupation or exposure was  selected.  Adjusted risk estimates were used
when available.
       Thirty-nine independent estimates of relative risk and standard errors were extracted.
Pooled estimates of relative risk were calculated  using a random-effects model. Among study
                                          7-63

-------
populations most likely to have had substantial exposure to DE, the pooled smoking- adjusted
relative risk was 1.47 (95% CI = 1.29, 1.67) (see Figure 7-2).
            IB
          R
          0
          05 1 6
            1.4
£
E
•D
£
0
£
            12
                                                      ^'
                                Categories of Epidemiological Studies Included
    Note. CI = confidence interval; HWE = healthy worker effect.
Figure 7-2. Pooled estimates of relative risk of lung cancer in epidemiologic studies
involving occupational exposure to DE (random-effects models).

Source: Lipsett and Campleman, 1999.
       The between-study variance of the relative risks indicated the presence of significant
heterogeneity in the individual estimates.  The authors evaluated the potential sources of
 heterogeneity by subset analysis and linear meta-regressions. Major sources of heterogeneity
included control for confounding by smoking, selection bias (a healthy worker effect), and
exposure patterns characteristic of different occupational categories. A modestly higher, pooled
relative risk was derived for the subset of case-control studies, which, unlike the cohort studies,
showed little evidence of heterogeneity.
       This meta-analysis also evaluated the potential for publication bias, which provides
reassurance that the magnitude of published effects is not a function of the precision or study
                                            7-64

-------
power. Again, as stated in the Bhatia et al. (1998) review, this assessment cannot rule out the
possibility of publication bias.
       Although a relatively technical approach was used in deriving summary estimates of
relative risk and the evaluation of possible sources of variation in the relative risks in this meta-
analysis, this approach should not be confused with rigorous evaluation of the potential
weaknesses among the studies included in the analysis.  The heterogeneity attributable to
statistical adjustment for smoking was evaluated based on a dichotomous assessment of whether
control for  smoking could be identified in the studies considered. This does not reflect the
adequacy of the adjustment for smoking employed in the individual studies considered.

7.2.4. Summary and Discussion
       Certain extracts of DE have been demonstrated as both mutagenic and carcinogenic in
animals and in humans. Animal data suggest that DE is a pulmonary carcinogen among rodents
exposed by inhalation to high doses over long periods of time. While rat lung cancer response to
DE is not suitable for dose-response extrapolation to humans, the positive lung cancer response
doses imply a hazard for humans.  Because large working populations are currently exposed to
DE and because nonoccupational ambient exposures currently are of concern as well, the
possibility that exposure to this complex mixture may be carcinogenic to humans has become an
important public health issue.
       Because diesel emissions become diluted in the ambient air, it is  difficult to study the
health effects in the general population.  Nonoccupational exposure to DE is worldwide in urban
areas. Thus, "unexposed" reference populations used in occupational cohort studies are likely to
contain a substantial number of individuals who are nonoccupationally exposed to DE.
Furthermore, the "exposed" group in these studies is based on job titles,  which in most instances
are not verified or correlated with environmental hygiene measurement.  The issue of health
effect measurement is further complicated by the fact that occupational cohorts tend to be
healthy and have below-average mortality, usually referred to as the "healthy worker effect."
Hence, the usual standard mortality ratios observed in cohort mortality studies are  likely to be
underestimations of true risk.
       A major difficulty with the occupational studies considered here was measurement of
actual DE exposure. Because all the cohort mortality studies were retrospective, assessment of
health effects from exposure to DE was naturally indirect.  In these occupational settings, no
systematic quantitative records of ambient air were available. Most studies compared men in job
categories with presumably some exposure to DE with either standard populations (presumably
no exposure to DE) or men in other job categories from industries with little or no potential for
DE exposure. A few studies have included measurements of diesel fumes, but there is no
                                          7-65

-------
standard method for the measurement.  No attempt is made to correlate these exposures with the
cancers observed in any of these studies, nor is it clear exactly which extract should have been
measured to assess the occupational exposure to DE.  All studies have relied on the job
categories or self-report of exposure to DE. Gustavsson et al. (1990), Emmelin et al. (1993),
and Briiske-Hohlfeld et al. (1999) estimated exposure levels by getting detailed histories of job
tasks/categories and computing cumulative exposures, which unfortunately were not verifiable
due to of the lack of industrial hygiene data. In the studies by Garshick et al. (1987, 1988), the
diesel-exhaust-exposed job categories were verified based on an industrial hygiene survey done
by Woskie et al.  (1988a,b).  The investigators found that in most cases the job titles were good
surrogates for DE exposure. Also, in the railroad industry, where only persons who had at least
10 years of work experience were included in the study, the workers tended not to change job
categories over the years. Thus, a job known only at one point in time was a reasonable marker
of past DE exposure. Unfortunately, the exposure was only qualitatively verified. Quantitative
use of this information would have been much more meaningful.  Zaebst et al.  (1991) conducted
an industrial hygiene survey of elemental carbon exposure in the trucking industry by job
categories. Using these exposure measurements, Steenland et al. (1998) conducted an exposure-
response analysis of their earlier lung cancer case-control  study (Steenland et al., 1990). These
exposure data are currently being verified and will be used for quantitative risk assessment in the
near future.
       Occupations involving potential exposure to DE are miners, truck drivers, transportation
workers, railroad workers, and heavy equipment  operators. No known studies in metal  miners
have assessed whether DE is associated with lung cancer.  Currently, there are about 265
underground metal/nonmetal mines in the United States (Department of Labor, Mine Safety and
Health Administration, 2001). Approximately 20,000 miners are employed, but not all  of them
are currently working in the mines.  Diesel engines were introduced in metal mines in the  United
States in the early to mid-1960s. Although all these mines use diesel equipment, it is difficult to
estimate how many of these miners were actually exposed to diesel fumes.
       Diesel engines were introduced in coal mines at an even later date in the United  States,
and their use is still quite limited. There are 910  underground coal mines in the United  States, of
which only 145 currently use diesel powered equipment (Department of Labor, Mine Safety and
Health Administration, 2001). Even if it were possible to estimate how many miners (metal and
coal) were exposed to DE, it would be very difficult to separate out the confounding effects of
other potential pulmonary carcinogens, such as radon decay products or heavy  metals (e.g.,
arsenic, chromium). Furthermore, the relatively short latency period limits the usefulness of
these cohorts of miners.
                                          7-66

-------
       Both metal and coal mines in Europe and Australia, on the other hand, have been using
diesel equipment for more than 50 years.  The epidemiologic studies of coal miners conducted in
these countries discuss only exposures to coal dust. In most of the coal miner studies, DE
exposures are not even mentioned by the investigators as confounding exposures. Therefore, it
is not known how many miners, if any, were exposed to DE, for how long, and at what
concentrations.  Although studies of coal miners reviewed by IARC (1997) generally found
lower than expected lung cancer mortality (with some exceptions where some excess of lung
cancer was observed), without knowing the concentrations, duration of exposure, and number of
miners exposed to DE, it is inappropriate to conclude that the reported lung cancer mortality
deficit  in these studies provides a proof positive of absence of causal association between DE
exposure and occurrence of lung cancer.

7.2.4.1. Summary of the Cohort Mortality Studies
       The cohort studies mainly demonstrated an increase in lung cancer. Studies of bus
company workers by Waller (1981), Rushton et al. (1983), and Edling et al. (1987) failed to
demonstrate any statistically significant excess risk of lung cancer, but these studies have certain
methodological problems, such as small sample sizes, short follow-up periods (just 6 years in the
Rushton et al. study), lack of information on confounding variables, and lack of analysis by
duration of exposure, duration of employment, or latency that preclude their use in determining
the carcinogenicity of DE.  Although the Waller (1981) study had a 25-year follow-up period,
the cohort was restricted to employees (ages 45 to 64) currently in service. Employees who left
the job earlier, as well as those who were still employed after age 64 and who may have died
from cancer, were excluded.
       Wong et al. (1985) conducted a mortality study of heavy equipment operators that
demonstrated a nonsignificant positive trend for cancer of the lung with length of membership
and latency.  Analysis of deceased retirees showed a significant excess of lung cancer.
Individuals without work histories who started work prior to 1967, when records were not kept,
may have been in the same jobs for the longest period of time. Workers without job histories
included those who had the same job before and after 1967 and thus may have worked about 12
to  14 years longer; these workers exhibited significant excess risks of lung cancer and stomach
cancer. If this assumption about duration of jobs is correct, then these site-specific causes can be
linked  to DE exposure.  One of the methodologic limitations of this study is that most of these
men worked outdoors; thus, this cohort might have had relatively low exposure to DE.  The
authors did not present any environmental measurement data either.  Because of the absence of
detailed work histories for 30% of the cohort and the availability of only partial work histories
for the remaining 70%, jobs were classified and ranked according to presumed diesel exposure.
                                         7-67

-------
Information is lacking regarding duration of employment in the job categories (used for
surrogate of exposure) and other confounding factors (alcohol consumption, cigarette smoking,
etc.).  Thus, this study cannot be used to support or refute a causal association between exposure
to DE and lung cancer.
       A 2-year mortality analysis by Boffetta and Stellman (1988) of the American Cancer
Society's prospective study, after controlling for age and smoking, demonstrated an excess risk
of lung cancer in certain occupations with potential exposure to DE. These excesses were
statistically significant among miners (RR = 2.67, 95% CI = 1.63, 4.37) and heavy equipment
operators (RR = 2.6, 95% CI =1.12, 6.06).  Recently Briiske-Hohlfeld et al. (1999) also have
observed significantly higher risk for lung cancer, in the  range of 2.31 to 4.3, for heavy
equipment operators. The elevated risks were nonsignificant in railroad workers (RR = 1.59)
and truck drivers (RR = 1.24). A dose response was also observed for truck drivers. With the
exception of miners, exposure to DE occurred in the three other occupations showing an increase
in the risk of lung cancer. Despite methodologic limitations,  such as the lack of representiveness
of the study population (composed of volunteers only,  who were probably healthier than the
general population), leading to an underestimation of the risk, and the questionable reliability of
exposure data based on self-administered questionnaires that were not validated, this study is
suggestive of a causal association between exposure to DE and excess risk of lung cancer.
       Two mortality studies were conducted by Gustavsson et al. (1990) and Hansen (1993)
among bus garage workers (Stockholm, Sweden) and truck drivers, respectively. An  SMR of
122 was found among bus garage workers, based on 17 cases. A nested case-control study was
also conducted in this cohort.  Detailed exposure matrices based on job tasks were assembled for
both DE and asbestos exposures.  Statistically significant increasing lung cancer relative risks of
1.34, 1.81, and 2.43 were observed for DE indices of 10 to 20, 20 to 30, and >30, respectively,
using 0 to 10  as a comparison  group. Adjustment for asbestos exposure did not change the
results. The main strength of this study is the detailed  exposure matrices; some of the limitations
are low power (small cohort) and lack of smoking histories.  But smoking is not likely to be
different among study individuals irrespective of their  exposure status to DE.
       Hansen (1993), on the  other hand, found statistically significant SMR of 160 from cancer
of bronchus and lung.  No dose response was observed, although the excesses were observed in
most of the age groups (30 to 39, 45  to 49, 50 to 54, 55 to 59, 60 to 64, and 65 to 74).  There are
quite a few methodologic limitations to this study.  Exposure to DE was assumed in truck drivers
for diesel-powered trucks, but no validation of exposure was attempted.  Follow-up period was
short, no latency  analysis was  done, and smoking data  were lacking. However, a population
survey carried out in 1988 showed very little difference in smoking habits of residents of rural
area and the total Danish male population, thus, smoking is unlikely to confound the finding of
                                          7-68

-------
excess lung cancer.  The findings of both these studies are consistent with the findings of other
truck driver studies and are supportive of causal association.
        Two mortality studies of railroad workers were conducted by Howe et al. (1983) and
Garshick et al. (1988). The Howe et al. study, which was conducted in Canada, found relative
risks of 1.2 (p<0.01) and 1.35 (/X0.001) among "possibly" and "probably" exposed groups,
respectively.  The trend test showed a highly significant dose-response relationship with
exposure to DE and the risk of lung cancer. The main limitation of the study was the inability to
separate overlapping exposures of coal dust/combustion fumes and DE fumes. Information on
jobs was available at retirement only. There also was insufficient detail on the classification of
jobs by DE exposure.  The exposures could have been nonconcurrent or concurrent, but because
the data are lacking, it is possible that the observed excess could be due to the effect of both coal
dust/combustion fumes and DE fumes and not just one or the other.  It should be noted that, so
far, coal dust has not been demonstrated to be a pulmonary carcinogen in studies of coal miners.
However, lack of data on confounders such as asbestos and smoking (though use of the internal
comparison group to compute relative risks minimizes confounding by smoking) makes
interpretation of this study difficult.  When three DE exposure categories were examined for
smoking-related diseases such as emphysema, laryngeal cancer, esophageal cancer, and buccal
cancer, positive trends were observed, raising a possibility that the dose response demonstrated
for diesel exposure may  have been due to smoking. The findings of this study are at best
suggestive of DE being a lung carcinogen.
       The strong evidence for linking DE exposure to lung cancer comes from the Garshick  et
al. (1988) railroad worker study conducted in the United States. Relative risks of 1.57 (95% CI
= 1.19, 2.06) and 1.34 (95% CI = 1.02, 1.76) were found for ages 40 to 44 and 45 to 49,
respectively, after the exclusion of workers exposed to asbestos.  The investigators reported that
the risk of lung cancer increased with increasing duration of employment.  As this was a large
cohort study with a lengthy follow-up and adequate analysis, including dose response (based on
duration of employment as a surrogate) as well as adjustment for other confounding factors such
as asbestos, the observed association between increased lung cancer and exposure to DE is more
meaningful. Even though the reanalysis of these data by Crump et al. (1991) found that the
relative risk could be positively or negatively related to duration of exposure depending on how
age was controlled, additional analysis by Garshick et al. (letter from Garshick, Harvard Medical
School, to Chao Chen, U.S. EPA, dated August 15, 1991) found that the relationship between
years exposed when adjusted for the attained age and calendar years was flat to negative,
depending on the choice of the model. They also found that deaths were underreported by
approximately 20% to 70% between 1977 and 1980, and their analysis based on job titles,
limited to 1959-1976, showed that the youngest workers still had the highest risk of dying of
                                          7-69

-------
lung cancer. On the other hand, an analysis of the same data by California EPA (CalEPA, 1998)
yielded a positive dose response set using age at 1959 and adding an interaction term of age and
calendar year in the model.  However, Crump (1999) reported that the negative dose-response
continued to be upheld in his latest analysis when age was controlled more carefully and years of
exposure quantified more accurately. Crump (1999) asserted that the negative dose-response
trends for lung cancer observed with either the cumulative exposure or duration of exposure may
be due to underascertainment of deaths in the last 4 years of follow-up of the Garshick et al.
(1988) study, as well as incomplete follow-up in earlier years. The HEI (1999) special panel
conducted its own analyses using Garshick et al. (1988) data to evaluate their usefulness for
quantitative risk assessment and found results similar to those of Crump et al. (1991) and
Garshick (letter from Garshick, Harvard Medical School, to Chao Chen, U.S. EPA, dated August
15, 1991).  The HEI panel reported consistently elevated risk of lung cancer for train workers
compared with clerks for each duration of employment, and that shop workers had an
intermediate risk of lung cancer. But they found decreasing risk of lung cancer with increasing
duration of employment. The panel discussed various possibilities (different types of biases) for
the negative dose-response and advised against using the Garshick et al. (1988) data for
quantitative risk assessment. The panel also reported the strengths of the Garshick et al. (1988)
study such as large population, control for asbestos, and smoking, and concluded that the study
was generally consistent with findings of weak association between exposure to DE and
occurrence of lung cancer.  Hence, the divergent results of these recent analyses do not negate
the positive evidence this study provides for the qualitative evaluation.  The observance of dose-
response would have strengthened the causal association, but an absence of a dose-response does
not negate it.
       Suggestive  evidence is provided by a recent study of potash miners in Germany. The
information on the  exposure (including elemental carbon and organics), work chronology, and
work category was used by the investigators to calculate cumulative exposures for each worker.
Furthermore, information on smoking habits indicated homogeneity in the cohort.
A statistically nonsignificant twofold increase in lung cancer was observed in the production
workers as  compared to workshop workers.  The lack of significance for this finding could be
due to short follow-up, not enough latency, and relatively young age of the cohort.

7.2.4.2. Summary of the Case-Control Studies of Lung Cancer
       Among the  11 lung cancer case-control studies reviewed in this chapter, only 2 studies
did not find any increased risk of lung cancer. Lerchen et al. (1987) did not find any excess risk
of lung cancer, after adjusting for age and  smoking, for diesel fume exposure. The major
limitation of this study was a lack of adequate exposure data derived from the job titles obtained
                                          7-70

-------
from occupational histories. Next of kin provided the occupational histories for 50% of the
cases that were not validated.  The power of the study was small (analysis done on males only,
333 cases). Similarly, Boffeta et al. (1990) did not find any excess of lung cancer after adjusting
for smoking and education.  This study had a few methodological limitations.  The lung cancer
cases and controls were drawn from the ongoing study of tobacco-related diseases. It is
interesting to note that the leading risk factor for lung cancer is cigarette smoking. The exposure
was not measured.  Instead,  occupations were used as surrogates for exposure. Furthermore,
there were very few individuals in the study who were exposed to DE.  On the other hand,
statistically nonsignificant excess risks were observed for DE exposure by Hall and Wynder
(1984) in workers who were exposed to DE versus those who were not (OR =1.4 and 1.7 with
two different criteria) and by Damber and Larsson (1987) in professional drivers (OR = 1.2).
These rates were adjusted for age and smoking.  Hall and Wynder (1984) had a high
nonparticipation rate of 36%.  Therefore,  the positive results found in this study  are
underestimated at best. In addition, the self-reported exposures used in the study by Hall and
Wynder (1984) were not validated.  This  study also had low power to detect excess risk of lung
cancer for specific occupations.
       The study by Benhamou  et al.  (1988),  after adjusting for smoking, found significantly
increased risks of lung cancer  among French motor vehicle drivers (RR = 1.42) and transport
equipment operators (RR =  1.35). The main limitation of the study was the inability to separate
exposures to DE from those to gasoline exhaust because both motor vehicle drivers and transport
equipment operators probably  were  exposed to the exhausts of both types of vehicles.
       Hayes et al. (1989) combined data from three studies (conducted in three different states)
to increase the power to detect an association between lung cancer and occupations with a high
potential for exposure to DE.  They  found that truck drivers employed for more than 10 years
had
a significantly increased risk of lung cancer (OR = 1.5, 95% CI = 1.1, 1.9).  This study also
found a significant trend of increasing risk of lung cancer with increasing duration of
employment among truck drivers. The relative odds were computed by adjusting for birth
cohort, smoking, and State of residence.  The main limitation of this study is again the mixed
exposures to diesel and gasoline exhausts, because information on type of engine was lacking.
Also, potential bias may have been introduced because the way in which the cause of death was
ascertained for the  selection of cases varied in the three studies. Furthermore, the methods used
in these studies to classify occupational categories were different, probably leading to
incompatibility of occupational categories.
       Emmelin et al. (1993),  in their Swedish dockworkers from 15 ports, found increased
relative odds of 6.8 (90% CI = 1.3 to 34.9). A strong interaction between smoking and DE was
                                          7-71

-------
observed in this study.  Of 50 cases and 154 controls, only 6 individuals were nonsmokers.
Although intricate exposure matrices were created using three different variables,  no direct
exposure measurement was done. Despite the limitations of small number of cases and controls;
lack of data on asbestos exposure, which is fairly common in dockworkers; and very few
nonsmokers; this study provides consistent support for a real  effect of DE exposure and
occurrence of lung cancer, at least in smokers.
       The most convincing evidence comes from the case-control studies among railroad
workers by Garshick et al. (1987); among truck drivers of the Teamsters Union by Steenland et
al. (1990, 1998); among truck drivers, railroad workers, and farmers in a population-based study
by Swanson et al. (1993); among different professional drivers in Denmark by Hansen et al.
(1998); and among male workers occupationally exposed to diesel motor emissions in Germany
by Briiske-Hohlfeld et al.  (1999). Garshick et al. (1987) found that after adjustment for asbestos
and smoking, the relative  odds for continuous exposure were 1.39 (95% CI = 1.05, 1.83).
Among the younger workers with longer DE exposure, the risk of lung cancer increased with
duration of exposure after adjusting for asbestos and smoking. Even after the exclusion of
recent DE exposure (5 years before death), the relative odds increased to 1.43 (95% CI = 1.06,
1.94). This appears to be  a well-conducted and well-analyzed study with reasonably good
power. Potential confounders were controlled adequately, and interactions between DE and
other lung cancer risk factors were tested.  Some of the limitations of this study are
misclassification of exposure because ICC job classification was used as surrogate for exposure
and use of death certificates for identification of cases and controls.
       Steenland et al. (1990), on the other hand, created two separate work history files, one
from Teamsters Union pension files and the other from next-of-kin interviews. Using duration
of employment as a categorical variable and considering employment after 1959 (when
presumed dieselization occurred) for long-haul drivers, the risk of lung cancer increased with
increasing years of exposure. Using 1964  as the cutoff, a similar trend was observed for long-
haul drivers. For short-haul drivers, the trend was positive with a 1959 cutoff, but not when
1964 was used as the cutoff.  For truck drivers who primarily drove diesel trucks and worked for
35 years, the relative odds were  1.89.  The main strengths of the  study are availability of detailed
records from the Teamsters Union, a relatively large sample size, availability of smoking data,
and measurements of exposure.  The limitations of this study include possible misclassifications
of exposure and smoking, lack of levels of diesel exposure, a smaller nonexposed group, and an
insufficient latency period. Recently Steenland et al.  (1998) conducted an exposure-response
analysis on these cases and controls, using the industrial hygiene survey results of Zaebst et al.
(1991).  The estimates were made for long-haul drivers, short-haul drivers, dockworkers,
mechanics, and those outside the trucking industry. The survey found that mechanics had the
                                          7-72

-------
highest current levels of DE exposures and dockworkers who mainly used propane- powered
forklifts had the lowest exposure. The finding of the highest lung cancer risk for mechanics and
lowest for dock workers is indicative of a causal association between the DE exposure and
development of lung cancer.  However, the risk among mechanics did not increase with
increasing duration of employment. The ORs for quartile cumulative exposures, computed by
using logistic regression adjusted for age, race, smoking, diet, and asbestos exposure, showed a
pattern of increasing trends in risk with increasing exposure, between 1.08 and 1.72 depending
upon exposure level and lag structure used.
       In a population-based lung cancer case-control study Swanson et al. (1993) found
statistically significant excess risks adjusted for age at diagnosis, smoking, and race, among
white male drivers of heavy trucks employed for >20 years and railroad workers employed for
>10 years (OR = 2.5, 95% CI = 1.1, 4.4, and OR = 2.4, 95 % CI = 1.1, 5.1, respectively), and
among black farmers employed for >20 years (OR = 10.4, 95% CI = 1.4, 77.1).  Although
individual ORs were not significant for various occupations with potential exposure to DE,
statistically significant trends were observed for drivers of heavy trucks, light trucks, farmers,
and railroad industry workers among whites, and among black farmers (p<0.05).  The main
strengths of the study are availability of data on lifetime work history and smoking history; the
main limitation is absence of actual specific exposure data.  This is the first study that found
increased lung cancer risk for farmers, who are exposed to DE of their  farm tractors.
       Hansen et al. (1998), in their study of professional drivers in Denmark, found statistically
significant ORs (adjusted for socioeconomic status) of 1.31, 1.64, and 1.39 for lorry/bus drivers,
taxi drivers, and unspecified drivers, respectively. The lag time analyses for duration of
employment were unchanged for lorry/bus drivers but increased to OR = 3 from 2.2 in taxi
drivers with a lag time of 10 years and duration of employment of > 5 years. The authors
asserted that the higher risk seen in the taxi drivers may be due to higher exposure to these
drivers because of longer time spent in traffic congestion. Furthermore, the trend tests for
increasing risk of lung cancer with increasing duration of employment were statistically
significant for both lorry/bus drivers and taxi drivers in both 10-year lag time and no lag time.
The main strengths of the study are the large sample size, availability of detailed employment
records, and information on socioeconomic status.  The main limitations are absence of
individual data on smoking habits and asbestos exposure, and information about the type of fuel
used for the vehicles driven by these professional drivers. A personal communication with the
main investigator revealed that the lorries/buses and taxis have been using diesel fuel since the
late 1940s. Moreover, indirect information about smoking and asbestos exposure indicated that
these two confounders are unlikely to explain the observed excesses or the trends, resulting in
strong support of earlier positive studies.
                                          7-73

-------
       Briiske-Hohlfeld et al. (1999) recently conducted a pooled analysis of two case-control
studies among male workers occupationally exposed to DME in Germany. The investigators
collected data on demographic information, detailed smoking, and occupational history. Job
titles and industries were classified in 33 and 21 categories respectively.  Job descriptions were
written and verified to avoid misclassification of estimated exposure to diesel emissions.
Individual cumulative DME exposures and smoking pack-years were calculated. Asbestos
exposures were estimated by certain job-specific supplementary questions. Analysis of 3,498
lung cancer cases and 3,541 controls yielded statistically significant ORs ranging from 1.25 to
2.31 adjusted for smoking and asbestos exposure. The risk increased with increasing years of
exposure for both the first year of exposure and the  end year of exposure. These investigators
presented analyses by various job categories, by years of exposure, first and end years of
exposure and, when possible, separately for West and East Germany.  Significantly higher risks
were found among all four job categories. For professional drivers (of trucks, buses, and taxis)
ORs ranged from 1.25 to 2.53. For other traffic-related jobs (switchmen, diesel locomotive
drivers, diesel forklift truck drivers), ORs ranged from 1.53 to 2.88. For heavy equipment
operators (bulldozers, graders, and excavators), ORs ranged from 2.31 to 4.3, and for drivers of
farming equipment the only significant excess (OR  = 6.81) was for exposure for <30 years.
       This study shows increased risk for all the DME-exposed job categories. The
professional drivers and the other traffic-related jobs also have some mixed exposures to
gasoline exhaust in general  traffic.  On the other hand, it should be noted that exposure to DME
among heavy equipment and farm tractor drivers is  much higher  and not as mixed as in
professional drivers.  The heavy equipment drivers usually drive  repeatedly through their own
equipment's exhaust. Therefore, the observed highest risk for lung cancer in this job category
establishes a strong link with the DME. The only other study that found significantly higher risk
for heavy equipment operators (RR = 2.6) was conducted by Boffeta et al. (1988).  Although the
only significant excess in the group was observed for farming tractor operators with more than
30 years of exposure, a steady increase in risk was observed for this job category with increasing
exposure. The investigators stated that the working conditions and the DME of tractors
remained fairly constant over the years. This increase may be due mainly to  exposure to DME
and PM10
       The main strengths of the study are large sample size, resulting in good statistical  power;
inclusion of incident cases diagnosed not more than 3 months prior to the interview; use of only
personal interviews, reducing recall bias; diagnoses ascertained by cytology or histology; and
availability of lifelong detailed occupational and smoking history. Exposure estimation done for
each individual was based on job codes and industry codes, which were validated by written job
descriptions to avoid misclassification.
                                          7-74

-------
       The main limitation of the study is lack of data on actual exposure to DME. The
cumulative quantitative exposures were calculated based on time spent in each job with potential
exposure to DME and the type of equipment used.  Thus, this study provides strong evidence for
causal association between exposure to DE and occurrence of lung cancer.

7.2.4.3. Summary of the Reviews and Meta-Analyses of Lung Cancer
       Three summaries of studies concerned with the relationship of DE exposure and lung
cancer risk are reviewed.  The HEI report is a narrative study of 35 epidemiologic studies (16
cohort and 19 case-control) of occupational exposure to diesel emissions published between
1957 and 1993. Control for smoking was identified in 15 studies.  Six of the studies (17%)
reported relative risk estimates less than 1, whereas 29 (83%) reported at least 1 excess relative
risk, indicating a positive association.  Twelve studies indicating a relative risk greater than 1
had 95% confidence intervals that excluded unity. These studies found that the evidence
suggests that occupational exposure to DE from diverse sources increases the rate of lung cancer
by 20% to 40% in exposed workers generally, and to a greater extent among workers with
prolonged exposure.  They also found that the results are not explicable by confounding due to
cigarette smoking or other known sources of bias.
       Bhatia et al. (1998) identified 23 studies that met criteria for inclusion in the meta-
analysis.  The observed relative risk estimates were greater than 1 in 21 of these studies. The
pooled relative risk weighted by study precision was 1.33 (95% CI= 1.24, 1.44), which indicated
increased relative risk for lung cancer from occupational exposure to DE. Subanalyses by study
design (case-control and cohort studies) and by control for smoking produced results that did not
differ from those of the overall pooled analysis.  Cohort studies using internal comparisons
showed higher relative risks than those using external comparisons.
       Lipsett and Campleman (1999) identify 39 independent estimates of relative risk among
30 eligible studies of DE and lung cancer published between 1975 and 1995. Pooled relative
risks for all studies and for study subsets were estimated using a random effect model.
Interstudy heterogeneity was also modeled and evaluated. A pooled smoking-adjusted relative
risk was 1.47 (95% CI = 1.29, 1.67). Substantial heterogeneity was found  in the pooled-risk
estimates. Adjustment for confounding by smoking, having a lower likelihood of selection bias,
and increased study power were all found to contribute to lower heterogeneity and increased
pooled estimates of relative risk.
       There is some variability in the  conclusions of these summaries of the association of DE
and lung cancer. The three analyses find that smoking is unlikely to account for the observed
effects, and all conclude that the data support a causal association between lung cancer and DE
exposure.  On the  other hand, Stober and Abel (1996), Muscat and Wynder (1995), and Cox
                                          7-75

-------
(1997) call into question the assertions by Cohen and Higgins (1995), Bhatia et al. (1998), and
Lipsett and Campleman (1999) that the associations seen for DE and lung cancer are unlikely to
be due to bias.  They argue that methodologic problems are prevalent among the studies,
especially in evaluation of diesel engine exposure and control of confounding by cigarette
smoking, and thus, the observed association between exposure to DE and excess risk of lung
cancer is more likely to be due to bias.  The conclusions of the two meta-analyses are based on
magnitude of pooled relative risk estimates and evaluation of potential sources of heterogeneity
in the estimates. Despite the statistical sophistication of the meta-analyses, the statistical models
used cannot compensate for  deficiencies in the original studies and will remain biased to the
extent that bias exists in the  original studies.

7.2.4.4. Discussion of Relevant Methodologic Issues
       A persistent association of risk for lung cancer and DE exposure has been observed in
more than 30 epidemiologic studies published in the literature over the past 40 years. Evaluation
of whether this association can be attributed to a causal relation between DE exposure and lung
cancer requires careful consideration of whether chance, bias, or confounding might be likely
alternative explanations.
       A total of 10 cohort and 12 case-control studies are reviewed in this chapter. An
increased lung cancer risk was observed in 8 cohort and 10 case-control studies, even though the
results were not always statistically significant.  There is a consistent tendency for point
estimates of relative risk to be greater than one in studies that adjusted (either directly or
indirectly) for smoking, had a long enough follow-up, and sufficient statistical power among
truck drivers, railroad workers, dock workers, and heavy equipment workers. If this elevated
risk was due to chance one would expect almost equal distribution of these point estimates to be
above and below one. Many of the studies provide confidence intervals for their estimates of
excess risk or statistical tests, which indicate that it is unlikely that the individual study findings
were due to random variation. The persistence of this association between DE and lung cancer
risk in so many studies indicates that the possibility is remote that the observed association in
aggregate is due to chance.  It is unlikely that chance alone accounts for the observed relation
between DE and lung cancer.
       The excess risk is observed in both cohort and case-control designs, which contradicts the
concern that a methodologic bias specifically characteristic of either design (e.g., recall bias)
might account for the observed effect.  Selection bias is certainly present in some of the
occupational cohort studies that use external  population data in estimating relative risks, but this
form of selection bias (a healthy worker effect) would only obscure, rather than spuriously
produce, an association between DE and lung cancer.  Several occupational epidemiologic
                                           7-76

-------
studies that use more appropriate data for their estimates are available.  Selection biases may be
operating in some case-control studies, but it is not obvious how such a bias could be sufficiently
uniform in effect, prevalent, and strong enough to lead to the consistent association seen in the
aggregate data. Given the variety of designs used in studying the DE and lung cancer
association and the number of studies in different populations, it is unlikely that routinely
studying noncomparable groups is an explanation for the consistent association seen. Exposure
information bias is certainly a problem for almost all of the studies concerned. Detailed and
reliable individual-level data on DE exposure for the period of time relevant to the induction of
lung cancer are not available and are difficult to obtain. Generally, the only information from
which diesel exposure can be inferred is occupational data, which is a poor surrogate for the true
underlying exposure distribution.  The variability in actual lifetime exposure to DE in an
occupational cohort may not be reflected in differences in job title, and there might be
considerable variability in actual exposure despite similar job titles. Study endpoints are
frequently mortality data taken from death certificate information, which is frequently inaccurate
and often does not fully  characterize the lung cancer incidence experience of the population in
question.  Using inaccurate surrogates for lung cancer incidence and for diesel exposure can lead
to substantial bias, and these shortcomings are endemic in the field. In most cases these
shortcomings will lead to misclassification of exposure and of outcome, which is nondifferential.
Nondifferential misclassification of exposure and/or outcome can bias estimates of a DE-lung
cancer association, if one exists, toward the null; but it is unlikely that such misclassification
would produce a spurious estimate in any one study. It is even more unlikely that it would bias a
sufficient number of studies in a uniform direction to account for the consistent aggregate
association observed.
       Moreover, throughout this chapter, various methodologic limitations of individual studies
have been discussed, such as small sample size,  short follow-up period, lack of data  on
confounding variables, use of death certificates to identify the lung cancer cases, and lack of
latency analysis. The studies with small sample sizes (i.e., not enough power) and short follow-
up periods (i.e., not enough latent period) have been difficult to interpret due to these limitations.
       The most important confounding variable is smoking which is a strong risk factor for
lung cancer. All the studies considered for this report are  either cohort retrospective mortality or
case-control studies where history of exposures in the past is elicited.  Smoking history is usually
difficult to obtain in such instances. The smoking histories obtained  from surrogates (next of
kin, either spouse or offspring) were found to be accurate by Lerchen and Samet (1986) and
McLaughlin et al. (1987).  Lerchen and Samet did not detect any consistent bias in the report of
cigarette consumption. In contrast, overreporting of cigarette smoking by surrogates was
observed by Rogot and Reid (1975), Kolonel et  al. (1977), and Humble et al. (1984). Kolonel et
                                          7-77

-------
al. found that the age at which an individual started smoking was reported within 4 years of
actual age 84% of the time. These studies indicate that surrogates were able to provide fairly
credible information on the smoking habits of the study subjects. If the surrogates of the cases
were more likely to overreport cigarette smoking compared with the controls, then it might be
harder to find an effect of DE because most of the increase in lung cancer would be attributed to
smoking rather than to exposure to DE.
       Some studies  do not adjust for tobacco smoke exposure.  Even though smoking is a
strong risk for lung cancer, it is only a confounder if there are differential smoking habits among
individuals exposed to DE versus individuals who are not exposed.  Most of the occupational
cohorts include workers from the same socioeconomic background or used an internal
comparison group; hence, it is unlikely that confounding by cigarette smoking is substantial in
these studies.  Some studies have adjusted for socioeconomic status and some studies have
compared the cigarette smoking habits by conducting rural and urban general  population
surveys.  Besides, in studies with long enough latency, adjustment for cigarette smoking did not
alter substantially the observed higher risk.
       Another methodologic concern in these studies is use of death certificates to determine
cause of death. Death certificates were used by all of the cohort mortality studies and some of
the case-control studies of lung cancer to determine cause of death.  Use of death certificates
could lead to misclassification bias because of overdiagnosis. Studies of autopsies done between
1960 and 1971 demonstrated that lung cancer was overdiagnosed when compared with hospital
discharge, with no  incidental cases found at autopsy (Rosenblatt et al., 1971).  Schottenfeld et al.
(1982) also found an  overdiagnosis  of lung cancer among autopsies conducted in 1977 and 1978.
On the other hand, Percy et al. (1981) noted 95% concordance when comparing 10,000 lung
cancer deaths observed in the Third National Cancer Survey from 1969 to 1971 (more  than 90%
were confirmed histologically) to death-certificate-coded cause of death. These more recent
findings suggest that the diagnosis of lung cancer on death  certificates is better than anticipated.
In reality, lung cancer is one cause of death that has been found to be generally reliably reported
on the death certificate. Thus, the misclassification bias probably is minimal in the studies
described in this chapter.
       Finally, several investigators have not conducted latency analysis in their studies. The
latent period for lung cancer development is from 20 to 30  years or more. Considering the fact
that dieselization was not complete till almost 1959 for locomotives and the 1970s for the
trucking industry in the United States, most of the cohort studies conducted in the U.S.
population do not have a long enough follow-up period to allow for latency of 20 to 30+ years.
In addition, the study inclusion criteria for most of the studies are individuals who worked in the
industry for at least 6 months /I year from the beginning of the follow-up period to the end of
                                          7-78

-------
the follow-up period. Hence, the later the individual enters the cohort, the shorter the follow-up
period; thus, the latent period is insufficient for the occurrence of lung cancer in these late
entrants. Therefore, the observed slight to moderate increase in risk of lung cancer could be due
to insufficient latency.  On the other hand, in certain case-control studies the elapsed period
between the identification of the lung cancer cases and exposure to DE is long enough to allow
for the 30+ years latency needed for the development of lung cancer (Hansen et al., 1998;
Briiske-Hohlfeld et al.,  1999).  These investigators identified lung cancer cases in the early to
mid-1990s and found significant excess risks for lung cancer among the individuals exposed to
DE. It should be noted that the use of diesel fuel for trucks, buses, and taxis had started in their
countries (Denmark and Germany, respectively) in the late 1940s.

7.2.4.5.  Evaluation of Causal Association
       In most situations, epidemiologic data are used to delineate the causality of certain health
effects.  Several cancers have been causally associated with exposure to  agents for which there is
no direct biological evidence. Insufficient knowledge about the biological basis for diseases in
humans makes it difficult to identify exposure to an agent as causal, particularly for malignant
diseases when the exposure was in the distant past. Consequently, epidemiologists and biologists
have used the original or modified version of a set of criteria provided by Hill  (1965)3 that
define a causal relationship between exposure and the health outcome. A causal interpretation is
enhanced for studies that meet these criteria. None of these criteria actually proves causality;
actual proof is rarely attainable when dealing with environmental carcinogens.  None of these
criteria should be considered either necessary (except temporality of exposure) or sufficient in
itself.  The absence of any one or even several of these criteria does not prevent a causal
interpretation.  However,  if more criteria apply, this provides more credible evidence for
causality.
       Thus, applying the Hill criteria (1965) of causal inference, as modified by Rothman
(1986), to the studies reviewed here resulted in the following:

       •      Strength of association.  This phrase  refers to the  magnitude of the ratio of
              incidence or mortality (RRs or ORs).  Several studies found statistically
              significant RRs and ORs that ranged  from 1.2 to 2.6 (Howe et al., 1983; Rushton
        Hill in his address to the Royal Society of Medicine in 1965 on "The environment and disease:
association or causation" explored several aspects of association between exposure and occurrence of an event
before deciding that the most likely interpretation of it is causation. He provided nine different aspects of
association that he characterized as his viewpoints before interpreting the association being causal. The
epidemiologic community universally adopted these (aspects/viewpoints) later as criteria for causality/causal
association.
                                            7-79

-------
et al., 1983; Wong et al., 1985; Gustavsson et al., 1990; Emmlin et al., 1993;
Hansen, 1993; Hansen et al., 1998) and, after adjustment for smoking and/or
asbestos, RRs and ORs remained statistically significant and in the same range in
certain studies (Dambar and Larson 1987; Garshick et al., 1987, 1988; Benhamou
et al., 1988; Boffetta and Stellman, 1988; Hays et al., 1989; Steenland et al.,
1990; Swanson et al., 1993; Briisk-Hohlfeld et al., 1999). In addition, two meta-
analyses demonstrated that not only did excess in lung cancer remain the same
after stratification/adjustment for smoking and occupation, but in several
instances the pooled RRs showed modest increases, with little evidence of
heterogeneity. Overall, the studies in epidemiologic terms show relatively
modest to weak association between DE and occurrence of lung cancer. Even
though strong associations are more likely to be causal than modest-to-weak
associations, the fact that association is relatively modest or weak does not rule
out the causal link.

Consistency.  Increased lung cancer risk has been observed in several  cohort and
case-control studies, conducted in several industries and occupations in which
workers were potentially exposed to DE. However, not all the excesses were
statistically significant. Statistically significant lung cancer excesses adjusted for
smoking were observed in truck drivers (Hayes et al., 1989; Hansen, 1993;
Swanson et al., 1993; Briiske-Hohlfeld et al., 1999), professional drivers
(Benhamou et al., 1988;  Briiske-Hohlfeld et al., 1999), railroad workers
(Garshick et al., 1987; Swanson et al., 1993), heavy equipment drivers (Boffetta
and Stellman, 1988; Briiske-Hohlfeld et al., 1999), and farm tractor drivers
(Swanson et al., 1993; Briiske-Hohlfeld et al., 1999). Furthermore, the two
recent meta-analyses by Bhatia et al. (1998) and Lipsett and Campleman (1999)
found that even though a substantial heterogeneity existed in their initial pooled
estimates, stratification on several factors demonstrated a relationship between
exposure to DE and excess lung cancer that remained positive throughout various
analyses.

Specificity.  This criterion requires that a single cause lead to a single  effect. With
respect to exposure to DE, excess for lung cancer is the only effect that is found
to be consistently elevated and statistically significant in several studies. Quite a
few studies have examined DE for other effects such as bladder cancer, leukemia,
                            7-80

-------
gastrointestinal cancers, prostate cancer etc. The evidence for these effects is
inadequate.

Temporality.  The only necessary, but not sufficient, criterion described by Hill
for causality inference is that exposure to a causal agent precedes the effect in
time.  This criterion is clearly satisfied in the studies reviewed here. Temporality
can be explored further in addressing the latency issue.  A certain period is
necessary for development of an effect after exposure to a causal agent has
occurred. For instance, in cancer-causing agents a latent period can vary from 5
years (childhood leukemia) to >30 years (mesothelioma). Most of the studies
reviewed here did not conduct the latency analysis.  Some studies had a short
follow-up period that did not allow enough time for the latency period (Waller,
1981; Howe et al., 1983; Rushton et al., 1983; Wong et al.,  1985, Hansen, 1993)
while several studies clearly allowed for an adequate latency period (Garshick et
al., 1987; Gustavsson et al., 1990; Steenland et al., 1990; Swanson et al., 1993;
Briiske-Hohlfeld et al., 1999).  Both type of studies showed mixed results.

Biological gradient.  This criterion refers to the dose-response curve. Due to the
lack of quantitative data on DE exposure in most studies reviewed here, analyzing
the dose-response curve directly was not possible.  In very few studies, exposure
to DE was addressed specifically.  Most investigators have used job
titles/categories and duration of employment as surrogates for exposure and thus
have presented response in relation to duration of employment.  Significant dose-
response (using duration of employment as a surrogate) was observed in various
studies for railroad workers (Howe et al., 1983; Garshick et al., 1987; Garshick et
al., 1988; Swanson et al., 1993; Cal EPA, 1998), truck drivers (Boffetta and
Stellman, 1988; Hayes et al., 1989; Steenland et al., 1990; Swanson et al., 1993;
Hansen et al.,  1998; Briiske-Hohlfeld et al.,  1999), transportation/heavy
equipment operators (Wong et al., 1985; Gustavsson et al., 1990; Briiske-
Hohlfeld  et al., 1999), farmers/farm tractor users (Swanson et al., 1993; Briiske-
Hohlfeld et al., 1999), and dockworkers (Emmelin et al., 1993).

Biological plausibility. This criterion refers to the biologic plausibility of the
hypothesis, an important concern that may be  difficult to judge. The hypothesis
considered for this review is that occupational exposure to DE is causally
associated with the occurrence of lung cancer and is supported by the following:
                             7-81

-------
             First, DE has been shown to cause lung and other cancers in animals (Heinrich et
             al., 1986b; Iwai et al., 1986b; Mauderly et al., 1987; Pott et al., 1990; Mauderly,
             1994).  Second, it contains highly mutagenic substances such as poly cyclic
             aromatic hydrocarbons as well as nitroaromatic compounds (Claxton, 1983; Ball
             et al., 1990; Gallagher et al., 1993; Sera et al., 1994; Nielsen et al., 1996a) that
             are recognized human pulmonary carcinogens (IARC, 1989).  Third, DE consists
             of carbon core particles with surface layers of organics and gases; the tumorigenic
             activity may reside in one, some, or all of these components. As explained in
             Chapter 4, there is clear evidence that the mixture of organic constituents, both in
             particles and vapor phases, have the capacity to interact with DNA and give rise
             to mutations, chromosomal aberrations, and cell transformations, all well-
             established steps in the process of carcinogenesis. Further, increased levels of
             peripheral blood cell DNA adducts associated with occupational exposure to DE
             have been observed in humans (Nielsen et al., 1996a,b). Thus, the above
             evidence makes a convincing case that occupational exposures to DE are causally
             associated with the occurrence of lung cancer is highly plausible biologically.

       In conclusion, the epidemiologic studies of exposure to DE and occurrence of lung
cancer furnish evidence that is consistent with a causal association. This association observed in
several studies is unlikely to be due to chance or bias. Although many studies did not have
information on smoking, significant confounding by smoking is unlikely in these studies because
the comparison population was from the same socioeconomic class. The  strength of association
(i.e., RRs/ORs between 1.2 and 2.6) was weak to modest by epidemiologic standards, with dose-
response relationships observed in several  studies. Last, but not least, there is highly plausible
biological evidence that exposure to DE could result in excess risk of lung cancer in humans.

7.3.  CARCINOGENICITY OF DIESEL EXHAUST IN LABORATORY ANIMALS
       This chapter summarizes studies that assess the carcinogenic potential of DE in
laboratory animals.  The first portion of this chapter summarizes results of inhalation studies.
Experimental protocols for the inhalation studies typically consisted of exposure (usually
chronic) to diluted exhaust in whole-body exposure chambers using rats, mice, and hamsters as
model species. Some of these studies used both filtered (free of particulate matter) DE and
unfiltered (whole) DE to differentiate gaseous-phase effects from effects induced by diesel PM
(DPM) and its adsorbed components.  Other studies were designed to evaluate the relative
importance of the carbon core of the diesel particle versus that of particle-adsorbed compounds.
                                          7-82

-------
Finally, a number of exposures were carried out to determine the combined effect of inhaled DE
and tumor initiators, tumor promoters, or cocarcinogens.
       Particulate matter concentrations in the DE used in these studies ranged from 0.1 to 12
mg DPM /m3.  In this chapter, any mention of statistical significance implies thatp<0.05 was
reported in the reviewed publications. A summary of the animal inhalation carcinogenicity
studies and their results is presented in Table 7-3.
       Results of lung implantation and intratracheal instillation studies of whole diesel
particles, extracted diesel particles, and particle extracts are reported in Section 7.3.3 and in
Tables 7-4 and 7-5.  Studies destined to assess the carcinogenic effects of DPM as well as
solvent extracts of DPM following subcutaneous (s.c.) injection, intraperitoneal (i.p.) injection,
or intratracheal (itr.) instillation in rodents are summarized in Section  7.3.5.  Individual
chemicals present in the gaseous phase or adsorbed to the particle surface were not included in
this review because assessments of those of likely concern (i.e., formaldehyde, acetaldehyde,
benzene, polycyclic aromatic hydrocarbons [PAHs]) have been published elsewhere (U.S. EPA,
1993).

7.3.1. Inhalation Studies (Whole Diesel Exhaust)
7.3.1.1. Rat Studies
       The potential carcinogenicity of inhaled DE was first evaluated by Karagianes et al.
(1981).  Male Wistar rats (40 per group) were exposed to room air or diesel engine exhaust
diluted to a DPM  concentration of 8.3 (± 2.0) mg/m3, 6 hr/day, 5 days/week for up to 20 months.
The animals were exposed in 3,000 L plexiglass chambers. Airflow was equal to 50 liters per
minute. Chamber temperatures were maintained between 25 °C and 26.5 °C. Relative humidity
ranged from 45%  to 80%. Exposures were carried out during the daytime.  The connected to an
electric generator  and operated at varying loads and speeds to simulate operating conditions in
an occupational situation. To control the CO concentration at 50 ppm, the exhaust was diluted
35:1 with clean air.  Six rats per group were sacrificed after 4, 8, 16, and 20 months exposure for
gross  necropsy and histopathological examination.
       The only tumor detected was a bronchiolar adenoma in the group exposed over 16
months to DE. No lung tumors were reported in controls. The equivocal response may have
been caused by the relatively short exposure durations (20 months) and small numbers of
animals examined. In more recent studies, for  example, Mauderly et al. (1987), most of the
tumors were detected in rats exposed for more than 24 months.
                                          7-83

-------
             Table 7-3.  Summary of animal inhalation carcinogenicity studies
oo
Species/
Study strain

Karagianes Rat/Wistar
etal. (1981)






Kaplan et al. Rat/F344
(1983)
White et al.
(1983)





Heinrich et Rat/
al. (1986a,b) Wistar
Mohr et al.
(1986)






Iwai et al. Rat/F344
(1986a,b)



Sex/total
number

M, 40
M, 40






M, 30
M, 30
M, 30
M, 30





F, 96
F, 92

F, 95






F,24
F,24

F,24

Exposure
atmosphere

Clean air
Whole
exhaust





Clean air
Whole
exhaust
Whole
exhaust
Whole
exhaust


Clean air
Filtered
exhaust
Whole
exhaust





Clean air
Filtered
exhaust
Whole
exhaust
Particle
concentration Other
(mg/m3) treatment

8.3 None
None






0 None
0.25 None
0.75 None
1.5 None





4 None
None

None






4.9 None
None

None

Exposure
protocol

6 hr/day,
5 days/
week,
for up to
20 mo



20 hr/day,
7 days/
week,
for up to
15 mo




19 hr/day,
5 days/
week
for up to
35 mo





8 hr/day,
7 days/
week,
for 24 mo

Post-
exposure
observation Tumor type and incidence (%)a Comments
Adenomas
NA 0/6 (0)
1/6(16.6)



Bronchoalveolar carcinoma
0/30 (0)
1/30(3.3)
8 mo 3/30 (10.0)
8mo 1/30(3.3)
8 mo
8 mo



Squamous
Adenomas Carcinomas cell tumors All tumors
NA 0/96(0) 0/96(0) 0/96(0) 0/96(0)
0/92(0) 0/92(0) 0/92(0) 0/92(0)

8/95(8.4) 0/95(0) 9/95(9.4) 17/95
(17.8)c
Large cell
Adenocarcinoma and
and squamous
adenosquamous cell
Adenomas carcinoma carcinomas All tumors
NA 1/22(4.5) 0/22(0) 0/22(0) l/22(4.5)f
0/16(0) 0/16(0) 0/16(0) 0/16(0)

3/19(0) 3/19(15.8) 2/19(10.5) 8/19
(42.1)c'g

-------
              Table 7-3.  Summary of animal inhalation carcinogenicity studies (continued)

-------
              Table 7-3. Summary of animal inhalation carcinogenicity studies (continued)

-------
              Table 7-3. Summary of animal inhalation carcinogenicity studies (continued)
oo


Study

Brightwell
etal. (1989)















Henrich et
al. (1989a)








Lewis et al.
(1989)



Species/ Sex/total Exposure
strain number atmosphere

Rat/344 M + F, 260 Clean air
M + F, 144 Filtered
exhaust
(medium
exposure)
M + F, 143 Filtered
exhaust
(high
M + F, 143 exposure)
M + F, 144 Whole
M + F, 143 exhaust
Whole
exhaust
Whole
exhaust


Rat/Wistar F, NS Clean air
F, NS Whole
F, NS exhaust
Filtered
F, NS exhaust
F, NS Clean air
F, NS Whole
exhaust
Filtered
exhaust
Rat/F344 M + F, Clean air
288n Whole
exhaust

Particle
concentration
(mg/m3)

0
0



0


0.7
2.2
6.6






0
4.2
0

0
4.2
0



2




Other
treatment

None
None



None


None
None
None






DPNd
DPNd
DPNd

DPNe
DPNe
DPNe



None
None



Exposure
protocol

16 hr/day,
5 days/
week,
for 24 mo













19 hr/day,
5 days/
week
for 24 to
30 mo





7 hr/day,
5 days/
week,
24 mo
Post-
exposure
observation Tumor type and incidence (%)a
Primary lung tumors
NA 3/260(1.2)
0/144 (0)



0/143 (0)


1/143 (0.7)
14/144 (9.7)c
55/143 (38.5)c




Squamous All lung
cell tumors
NA carcinoma (84.8)
(4.4) (83.0)
(46.8)c (67.4)
(4.4)
(93.8)
(16.7) (89.6)
(31.3)c (89.6)
(14.6)


NA No tumors 0/192(0)
0/192(0)




Comments

Tumor
incidence for
all rats dying
or sacrificed



? 24/25
(96%) after 24
mo
a* 12/27
(44%)
after 24 mo



















-------
                  Table 7-3.  Summary of animal inhalation carcinogenicity studies (continued)
               Study
             Species/  Sex/total   Exposure
             strain    number  atmosphere
                         Particle
                      concentration
                         (mg/m3)
           Other
         treatment
        Exposure
         protocol
              Post-
            exposure
           observation
         Tumor type and incidence (%)a
                                    Comments
Takaki et al. Rat/F344  M + F, 123  Clean air
             (1989)
             Light-duty
             engine
                     M + F, 123
                     M + F, 125
                     M + F, 123
                     M + F, 124
oo
oo
            Nikulaetal. Rat/F344
            (1995)
        Whole
        exhaust
        Whole
        exhaust
        Whole
        exhaust
        Whole
        exhaust
Heinrich et  Rat/Wistar
al.
(1995)
F, 220   Clean air
F, 200   Whole
F, 200   exhaust
F, 100   Whole
F, 100   exhaust
F, 100   Whole
        exhaust
        Carbon black
        TiO,
                     M + F,     Clean air
                     214b       Whole
                     M + F, 210 exhaust
                     M + F, 212 Whole
                     M + F, 213 exhaust
                     M + F, 211 Carbon black
                                Carbon black
 0
 0.1
 0.4
 1.1
 2.3
 0
 0.8
 2.5
 7.0
11.6
10.0
                             0
                             2.5
                             6.5
                             2.5
                             6.5
None
None
None
None
None
                                                                           None
                                                                           None
           None
           None
           None
           None
           None
16hr/day,
6 days/
week, for
up to
30 mo
                                                                           None   18hr/day,
        5 days/
        week,
                                                                           None   for up to
                                                                           None   24 mo
                                                                           None
        16hr/day,
        5 days/
        week for
        up to
        24 mo
                                                                                                 NA
                      6 mo
             6 weeks
                                                                                                           Adeno-
                                                                                                          squamous   Squamous cell
                                                                                                         carcinomas   carcinomas    All tumors
1/23 (0.8)
1/23 (0.8)
1/25 (0.8)
0/23 (0)
1/24(8.1)
2/123(1.6)
1/23 (0.8)
0/125 (0)
5/123(4.1)
2/124(1.6)
1/23 (0.8)
1/23 (0.8)
0/125 (0)
0/123 (0)
0/124 (0)
4/123 (3.3)
3/123 (2.4)
1/125 (0.8)
5/123(4.1)
3/124 (2.4)
                                                                                                                                                Benign
                                                                                                                                    Squamous  squamous
                                                                                                                    Adenocarcinoma     cell        cell
Adenomas
0/217(0)
0/198(0)
2/200(1)
4/100 (4)
13/100
(13)
4/100 (4)


Adenomas
1/214 (<1)
7/210(3)
23/212
(11)
3/213(1)
13/211(6)
s
1/217 (<1)
0/198(0)
1/200 (<1)
4/100 (4)
13/100(13)
13/100(13)


Adenocarcinoma
s
1/214 (<1)
4/210(2)
22/212(10)
7/213(3)
21/211(10)

carcinomas
0/217(0)
0/198(0)
0/200 (0)
2/100 (2)
4/100 (4)
3/100(3)

Squamous
cell
carcinoma
1/214 (<1)
3/210(1)
3/212(1)
0/213(0)
3/211(1)

tumors
0/217(0)
0/198(0)
7/200(3.5)
14/100 (14) Tumor
20/100 (20) incidences
20/100 (20) after 30 mo

Adeno-
squamous Other
carcinoma neoplasms
0/214(0) 0/214(0)
0/210(0) 0/210(0)
1/212 (<1) 0/212(0)
0/213(0) 1/213 (<1)
2/211 (<1) 0/211(0)


-------
              Table 7-3. Summary of animal inhalation carcinogenicity studies (continued)
Species/
Study strain


Iwai et al. F/344
(1997)



Orthoefer et Mouse/
al. (1981) Strong A
(Pepelko
and Peirano,
1983)



Particle
Sex/total Exposure concentration
number atmosphere (mg/m3)

Clean air
121, F Filtered air
108,F Whole
153, F exhaust


M, 25 Clean air


Whole
exhaust

Whole
exhaust


0
0
3.2-9.4


0


6.4

6.4


Other
treatment


None
None
None


None


None

UV
irradiated

Exposure
protocol
NA
48-56
hr/day
48-56
hr/day


20 hr/day,
7 days/
week,
for 7
weeks



Post-
exposure
observation Tumor type and incidence (%)a


NA 5/121 (4%) type not stated
6 mo 2/1 08(4%) type not stated
6 mo 53/153(35%) 61.3% adenoma, 25.8% adenocarcinoma,
2.2% benign squamous cell tumor, 7.5% squamous cell
carcinoma, 3.2% adenosquamous carcinoma
3/22(13.6)


26 weeks 7/19(36.8)

26 weeks 6/22 (27.3)


Comments


Cumulative
exposure dose
ranged from
154-274
mg/m3
0.13 tumors/
mouse

0.63 tumors/
mouse
0.27 tumors/
mouse

oo
VO

-------
              Table 7-3. Summary of animal inhalation carcinogenicity studies (continued)

-------
Table 7-3.  Summary of animal inhalation carcinogenicity studies (continued)
Species/ Sex/total Exposure
Study strain number atmosphere
Pepelko and Mouse/
Peirano Sencar
(1983)






M + F,260 Clean an-
dean air
Clean air
Whole
exhaust
Whole
exhaust
Whole
exhaust
Particle Post-
concentration Other Exposure exposure
(mg/m3) treatment protocol observation Tumor type and
121212 None Continuou
BHT1 s for 15
Urethank mo
None
BHT1
Urethan1



NA Adenomas
(5.1)
(12.2)
(8.1)
(10.2)c
(5.4)
(8.7)


Carcinomas
(0.5)
(1.7)
(0.9)
(1.0)
(2.7)
(2.6)


incidence (%)a Comments
All tumors
(5.6)
(2.8)
(9.0)
(11.2)c
(8.1)
(11.2)



-------
             Table 7-3. Summary of animal inhalation carcinogenicity studies (continued)
to
Species/
Study strain
Pepelko and Mouse/
Peirano Strain A
(1983)






Heinrich et Mouse/
al. (1986a,b)NMRI

Takemoto et Mouse/
al. IRC
(1986)
Mouse/
C57BL

Sex/total
number
M + F, 90







M + F, 84
M + F, 93
M + F, 76
M + F, 45
M + F, 69
M + F, 12
M + F, 38

Exposure
atmosphere
Clean air
Clean air
Whole
exhaust
Whole
exhaust
Clean air
Whole
exhaust

Clean air
Filtered
exhaust
Whole
exhaust
Clean air
Whole
exhaust
Clean air
Whole
exhaust

Particle
concentration Other Exposure
(mg/m3) treatment protocol
1212012 None
Exposure
(darkness)
Exposure
(darkness)
Urethan111
Urethan111



4 None 19hr/day,
None 5 days/
week
None for up to
30 mo
0 None 4 hr/day,
2-4 None 4 days/
week, for
19-28 mo
0 None 4 hr/day,
2-4 None 4 days/
week for
19-28 mo
Post-
exposure
observation Tumor type and incidence (%)a
All tumors
NA 21/87(24)
59/237 (24.9)
10/80(12.5)
22/250(0.10)
66/75 (88)
42/75 (0.95)


Squamous
cell
Adenomas Adenocarcinoma tumors All tumors
NA 9/84(11) 2/84(2) — 11/84(13)
11/93(12) 18/93(19)° — 29/93(31)°
11/76(15) 13/76(17)° — 24/76(32)°
NA
Adenoma Adenocarcinoma
NA 3/45 (6.7) 1/45 (2.2)
6/69 (8.7) 3/69 (4.3)

Comments
0.29 tumors/
mouse
0.27 tumors/
mouse
0.14
0.10
2.80
0.95









-------
Table 7-3.  Summary of animal inhalation carcinogenicity studies (continued)
Species/ Sex/total
Study
Heinrich et
al.
(1995)

















Mauderly et
al. (1996)









Heinrich et
al.
(1986a,b)


strain number
Mouse/ F,
C57BL/6N
F,

F,


Mouse/ F,
NMRI F,



Mouse/ F,
NMRI F,
F,





Mouse/ M +
CD-I 157b
M +
M +
M +







Hamster/ M +
Syrian M +

M +
120

120

120


120
120



120
120
120





F,

F, 171
F, 155
F, 186







F, 96
F, 96

F, 96
Particle
Exposure concentration Other
atmosphere
Clean air

Whole
exhaust

Particle-free
exhaust
Clean air
Whole
exhaust
Carbon black
Ti02
Clean air
Whole
exhaust
Particle-free
exhaust



Clean air
Whole
exhaust
Whole
exhaust
Whole
exhaust



Clean air
Filtered
exhaust

Whole
exhaust
Exposure
Post-
exposure



(mg/m3) treatment protocol observation Tumor type and incidence (%)a
4.5 None

None

None


0 None
4.5 None
11.6 None
10 None

4.5 None
None
None





0 None
0.35 None
3.5 None
7.1 None








None
None

4 None
18hr/day,
5 days/
week,
for up to
21 mo


18 hr/day,
5 days/
week for
up to
13.5 mo
18hr/day,
5 days/
week,
23 mo




7 hr/day, 5
days/week,
for up to
24 mo







19 hr/day
5
days/week
for up to
30 mo
6 mo






9.5 mo




None





Multiple
adenomas
None 1/157(0.6)
2/171(1.2)
0/155(0)
0/186(0)





Adenomas


0/96(0)
0/96(0)

NA 0/96(0)
1/12(8.3)
8/38(21.1)





Adenomas
(25)
(21.8)
(11.3)
(11.3)
(25)
(18.3)
(31.7)



Multiple
carcinomas
2/157(1.3)
1/171 (0.6)
1/155(0.6)
0/186(0)





Adenocarcinoma


0/96(0)
0/96(0)

0/96(0)
0/12 (0)
3/38 (7.9)













Comments
5.1% tumor
rate
8. 5% tumor
rate
3. 5% tumor
rate

Adenocarcinomas
(15.4)
(15.4)
(10)
(2.5)
(8.8)
(5.0)
(15)



Adenomas/
carcinoma
1/157(0.6)
1/171 (0.6)
0/155(0)
0/186(0)



Squamous
cell
tumors


0/96
0/96

0/96









Alveolar/
bronchiolar
adenoma
10/157
(6.4)
16/171
(9.4)
8/155 (5.2)
10/186
(5.4)


All tumors


0/96(0)
0/96(0)

0/96(0)









Alveolar/
bronchiolar
carcinoma
7/157(4.5)
5/171 (2.9)
6/155(3.9)
4/186(2.2)













-------
      Table 7-3.  Summary of animal  inhalation carcinogenicity studies (continued)
Study


Brightwell
etal.
(1989)
















Particle
Species/ Sex/total Exposure concentration
strain number atmosphere (mg/m3)


Hamster/ M + F, Clean air
Syrian M + F, 202 Clean air
Golden M + F, 104 Filtered
exhaust
(medium
dose)
M + F, 104 Filtered
exhaust
(high dose)
M + F, 101 Whole
M + F, 102 exhaust
M + F, 101 Whole
M + F, 204 exhaust
Whole
exhaust
M + F, 203 Filtered
exhaust
(high dose)
Whole


0
0
0



0


0.7
2.2
6.6
0


6.6



Post-
Other Exposure exposure
treatment protocol observation


None 16hr/day, NA
DEN 5 days/
DEN week, for
24 mo


DEN


DEN
DEN
DEN
None


None



Tumor type and incidence (%)a Comments
Primary lung
tumors
7/202(3.5)
4/104(3.8)
9/104 (8.7)



2/101 (2.0)


6/102(5.9)
4/101 (3.9)
1/204 (0.5)
0/203 (0)








Respiratory
tract tumors
not related to
exhaust
exposure for
any of the
groups












"Table values indicate number with tumors/number examined (% animals with tumors).
bNumber of animals examined for tumors.
"Significantly different from clean  air controls.
dDipentylnitrosamine; 6.25 mg/kg/week s.c. during first 25 weeks of exposure.
"Dipentylnitrasamine; 12.5 mg/kg/week s.c. during first 25 weeks of exposure.
fSplenic lymphomas also detected  in controls (8.3%), filtered exhaust group (37.5%) and whole
exhaust group (25%).
S5.3% incidence of large cell carcinomas.
hl g/kg, i.p. I/week for 3 weeks starting 1 mo into exposure.
'Includes adenomas, squamous cell carcinomas, adenocarcinomas, adenosquamous cell carcinoma,
and mesotheliomas.
J4.5 mg/diethylnitrosamine (DEN)/kg, s.c., 3 days prior to start of inhalation exposure.
'Single i.p. dose 1 mg/kg at start of exposure.
'Butylated hydroxytoluene 300 mg/kg, i.p. for week 1,  83 mg/kg for week 2, and 150 mg/kg for
weeks 3 to 52.
m!2 mg/m3from 12 weeks of age to termination of exposure. Prior exposure (in utero) and of parents
was 6 mg/m3.
"120-121 males and 71-72 females examined histologically.
"Not all animals were exposed for full term, at least 10  males were killed at 3, 6, and 12 mo of
exposure.
NS = Not specified.
NA = Not applicable.

-------
Table 7-4. Tumor incidences in rats following intratracheal instillation of DE particles
(DPM), extracted DPM, carbon black (CB), benzo[a]pyrene (B[a]P), or particles plus
B[a]P
Experimental group
Control
DPM (original)
DPM (extracted)
DPM (extracted)
CB (printex)
CB (lampblack)
B[o]P
B[«]P
DEP + B[a]P
CB (printex) + B[a]P
Number of
animals
47
48
48
48
48
48
47
48
48
48
Total dose
4.5 mL
15 mg
30 mg
15 mg
15 mg
14 mg
30 mg
15 mg
15 mg + 170 |_ig
B[«]P
15 mg + 443 |ag
B[«]P
Animals with
tumors (percent)
0 (0)
8 (17)
10 (21)
2 (4)
10 (21)
4 (8)
43 (90)
12 (25)
4 (8)
13 (27)
Statistical
significance3
-
<0.01
< 0.001
NS
< 0.001
NS
< 0.001
< 0.001
NS
< 0.001
Table 7-5. Tumorigenic effects of dermal application of acetone extracts of DPM
Number
of
animals
52


50


25




Strain/sex
C57BL/40 F
C57BL/12 M

Strain A/M


Strain A/F



Sample Time to first
material tumor (mo)
Extract of DPM 13
obtained during
warmup
Extract of DPM 15
obtained during
full load
Extract of DPM 13
obtained during
full load
Survivors at
time of first
tumor
33


8


20


Duration of
Total experiment
tumors (mo)
2 22


4 23


17 17


Source: Kotinetal., 1955.
                                         7-95

-------
       General Motors Research Laboratories sponsored chronic inhalation studies at the
Southwest Research Institute using male Fischer 344 rats, 30 per group, exposed to DPM
concentrations of 0.25, 0.75, or 1.5 mg/m3 (Kaplan et al., 1983; White et al., 1983).  The animals
were exposed in  12.6 m3 exposure chambers. Airflow was adjusted to provide 13 changes per
hour.  Temperature was maintained at 22 ± 2 °C.  The exposure protocol was 20 hr/day, 7
days/week for 9 to 15 months. Exposures were halted during normal working hours for
servicing. Some animals were sacrificed following completion of exposure, while others were
returned to clean air atmospheres  for an additional 8 months.  Control animals received clean air.
Exhaust was generated by 5.7-L Oldsmobile engines (four different engines used throughout the
experiment) operated at a steady speed and load simulating a 40-mph driving speed of a full-size
passenger car.
       Although five instances of bronchoalveolar carcinoma were observed in 90 rats exposed
to DE for 15 months and held an additional 8 months  in clean air,  compared with none among
controls, statistical significance was not achieved in any of the exposure groups. These included
one tumor in the 0.25 mg/m3 group, three in the 0.75 mg/m3 group, and one in the 1.5 mg/m3
group.  Rats kept in  clean-air chambers for 23 months did not exhibit any carcinomas. No
tumors were observed in any of the 180 rats exposed to DE for 9 or 15 months without a
recovery period,  or in the respective controls for these groups. Equivocal results may again have
been due to less-than-lifetime duration of the study as well as insufficient exposure
concentrations.  Although the increases in tumor incidences in the groups exposed for 15 months
and held an additional 8 months in clean air were not  statistically significant, relative to controls,
they were slightly greater than the historic background incidence of 3.7% for this specific lesion
in this strain of rat (Ward, 1983).  The first definitive  studies linking inhaled DE to induction of
lung cancer in rats were reported by researchers in Germany, Switzerland, Japan, and the United
States in the mid-to-late 1980s. In a study conducted  at the Fraunhofer Institute exhaust-
generating system and exposure atmosphere characteristics are presented in Appendix A.  The
type of engine used (3-cylinder, 43 bhp diesel) is normally used in mining situations and was of
Toxicology and Aerosol Research, female Wistar rats were exposed for 19  hr/day, 5 days/week
to both filtered and unfiltered (total) DE at an average particulate matter concentration of 4.24
mg/m3.  Animals were exposed for a maximum of 2.5 years. The exposure system as described
by Heinrich et al. (1986a) used a 40 kilowatt 1.6-L diesel engine operated continuously under
the U.S. 72 FTP driving cycle. The engines used European Reference Fuel with a sulfur
content of 0.36%. Filtered exhaust was obtained by passing engine exhaust through a Luwa FP-
65 HT 610 particle filter heated to 80 °C and a secondary series of filters (Luwa FP-85, Luwa
NS-30, and Drager CH 63302) at  room temperature.  The filtered and unfiltered exhausts were
diluted 1:17 with filtered air and passed through respective 12m3 exposure chambers. Mass
                                          7-96

-------
median aerodynamic diameter of DPM was 0.35 ± 0.10 |im (mean ± SD).  The gas-phase
components of the DE atmospheres are presented in Appendix A.
       The effects of exposure to either filtered or unfiltered exhaust were described by
Heinrich et al. (1986b) and Stober (1986). Exposure to unfiltered exhaust resulted in 8
bronchoalveolar adenomas and 9 squamous cell tumors in 15 of 95 female Wistar rats examined,
for a 15.8% tumor incidence. Although statistical analysis was not provided, the increase
appears to be highly significant. In addition to the bronchioalveolar adenomas and squamous
cell tumors, there was a high incidence of bronchioalveolar hyperplasia (99%) and metaplasia of
the bronchioalveolar epithelium (65%). No tumors were reported among rats exposed to filtered
exhaust (n = 92) or clean air (n = 96).
      Mohr et al. (1986) provided a more detailed description of the lung lesions and tumors
identified by Heinrich et al. (1986a,b) and Stober (1986).  Substantial alveolar deposition of
carbonaceous particles was noted for rats exposed to unfiltered DE. Squamous metaplasia was
observed in 65.3% of the rats breathing unfiltered DE, but not in the control rats.  Of nine
squamous  cell tumors, one was characterized as a Grade I carcinoma (borderline atypia, few to
moderate mitoses, and slight evidence  of stromal invasion), and the remaining eight were
classified as benign keratinizing cystic tumors.
      Iwai et al. (1986b) examined the long-term effects of DE inhalation on female F344 rats.
The exhaust was generated by a 2.4-L displacement truck engine. The exhaust was diluted 10:1
with clean air at 20 °C to 25 °C and 50% relative humidity.  The engines were operated at 1,000
rpm with an 80% engine load. These operating conditions were found to produce exhaust with
the highest particle concentration and lowest NO2 and SO2 content.  For those chambers using
filtered exhaust, proximally installed high-efficiency particulate air (FIEPA) filters were used.
Three groups of 24 rats  each were exposed to unfiltered DE, filtered DE, or filtered room air for
8 hr/day, 7 days/week for 24 months. Particle concentration was 4.9 mg/m3 for unfiltered
exhaust. Concentrations of gas-phase exhaust components were 30.9 ppm NOX, 1.8 ppm NO2,
13.1  ppm SO2, and 7.0 ppm CO.
      No lung tumors were found in the 2-year control (filtered room air) rats, although one
adenoma was noted in a 30-months control rat, providing a spontaneous tumor incidence of
4.5%. No lung tumors were observed in rats exposed to filtered DE. Nineteen of the 24
exposed to unfiltered exhaust survived for 2 years. Of these, 14 were randomly selected for
sacrifice at this time.  Four of the rats developed lung tumors; two of these were malignant.  Five
rats of this 2-year exposure group were subsequently placed in clean room  air for 3 to 6 months
and four eventually (time not specified) exhibited lung tumors (three malignancies).  Thus, the
lung tumor incidence for total tumors was 42.1% (8/19) and 26.3% (5/19) for malignant tumors
in rats exposed to whole DE. The tumor types identified were adenoma (3/19), adenocarcinoma
                                         7-97

-------
(1/19), adenosquamous carcinoma (2/19), squamous carcinoma (1/19), and large-cell carcinoma
(1/19). The lung tumor incidence in rats exposed to whole DE was significantly greater than
that of controls (p<0.01).  Tumor data are summarized in Table 7-3.  Malignant splenic
lymphomas were detected in 37.5% of the rats in the filtered exhaust group and in 25.0% of the
rats in the unfiltered exhaust group; these values were significantly (p<0.05) greater than the
8.2% incidence noted in the control rats. The study demonstrates production of lung cancer in
rats following 2-year exposure to unfiltered DE. In addition, splenic malignant lymphomas
occurred during exposure to both filtered and unfiltered DE. This is the only report to date of
tumor induction at an extrarespiratory site by inhaled DE in animals.
       A chronic (up to 24 months) inhalation exposure study was conducted by Takemoto et al.
(1986), in which female Fischer 344 rats were exposed to DE generated by a 269-cc YANMAR-
40CE NSA engine operated at an idle state (1,600 rpm).  Exposures were 4 hours/day, 4
days/week. The animals were exposed in a 376-L exposure chamber. Air flow was maintained
at 120 L/min. Exhaust was diluted to produce a particle concentration of 2-4 mg/m3. When not
exposed the animals were maintained in an air-conditioned room at a temperature of 24 ± 2°C
and a relative humidity of 55 ± 5% with 12 hr of light and darkness.  Temperature and humidity
in the exposure chambers was not noted. The particle concentration of the DE  in the exposure
chamber was 2 to 4 mg/m3. B[a]P and 1-nitropyrene concentrations were 0.85  and 93  |lg/g of
particles, respectively. No lung tumors were reported in the diesel-exposed animals. It was also
noted that the diesel engine employed in this  study was originally used as an electrical generator
and that its operating characteristics (not specified) were different from those of a diesel-
powered automobile. However, the investigators deemed it suitable for assessing the effects of
diesel emissions.
       Mauderly et al. (1987) provided data affirming the carcinogenicity of automotive diesel
engine exhaust in F344/Crl rats following chronic inhalation exposure. Male and female rats
were exposed to diesel engine exhaust at nominal DPM concentrations of 0.35  (n = 366), 3.5
(n = 367), or 7.1 (n = 364) mg/m3 for 7 hr/day, 5 days/week for up to 30 mo. Sham-exposed
(n = 365) controls breathed filtered room air. A total of 230, 223, 221, and 227 of these rats
(sham-exposed, low-, medium-, and high-exposure groups, respectively) were examined for lung
tumors. These numbers include those animals that died or were euthanized during exposure and
those that were terminated following  30 months of exposure. The exhaust was generated by
1980 model 5.7-L Oldsmobile V-8 engines operated through continuously repeating U.S.
Federal Test Procedure (FTP) urban certification cycles.  The engines were equipped with
automatic transmissions connected to eddy-current dynamometers and flywheels simulating
resistive and inertial loads of a midsize passenger car. The D-2 diesel control fuel (Phillips
Chemical Co.) met U.S. EPA certification standards and contained approximately 30% aromatic
                                         7-98

-------
hydrocarbons and 0.3% sulfur.  Following passage through a standard automotive muffler and
tailpipe, the exhaust was diluted 10:1 with filtered air in a dilution tunnel and serially diluted to
the final concentrations.  The primary dilution process was such that particle coagulation was
retarded. Mokler et al. (1984) provided a detailed description of the exposure system. No
exposure-related changes in body weight or lifespan were noted for any of the exposed animals,
nor were there any signs of overt toxicity. Collective lung tumor incidence was greater (z
statistic,p<0.05) in the high (7.1 mg/m3) and medium (3.5 mg/m3) exposure groups (12.8% and
3.6%, respectively) versus the control and low (0.35 mg/m3) exposure groups (0.9% and  1.3%,
respectively). In the high-dose group the incidences of tumor types reported were adenoma
(0.4%), adenocarcinomas plus squamous cell carcinomas (7.5%), and squamous cysts (4.9%).
In the medium-dose group adenomas were reported in 2.3% of animals, adenocarcinomas plus
squamous cell carcinomas in 0.5%, and squamous cysts in 0.9%. In the low-exposure group
adenocarcinomas plus squamous cell carcinomas were detected in 1.3% of the rats.  Using the
same statistical analysis of specific tumor types, adenocarcinoma plus squamous cell carcinoma
and squamous cyst incidence was significantly greater in the high-exposure group, and the
incidence of adenomas was significantly greater in the medium-exposure group. A significant
(p<0.001) exposure-response relationship was obtained for tumor incidence relative to exposure
concentration and lung burden of DPM. These data  are summarized in Table 7-3.  A logistic
regression model estimating tumor prevalence as a function of time, dose (lung burden of DPM),
and sex indicated a sharp increase in tumor prevalence for the high dose level at about 800 days
after the commencement of exposure. A less pronounced, but definite, increase in prevalence
with time was predicted for the medium-dose level. Significant effects were not detected at the
low concentration. DPM (mg per lung) of rats exposed to 0.35, 3.5, or 7.1 mg of DPM/m3 for
24 months were 0.6, 11.5, and 20.8, respectively, and affirmed the greater-than-predicted
accumulation that was the result of decreased particle clearance following high-exposure
conditions.
       In summary, this study demonstrated the pulmonary carcinogenicity of high
concentrations of whole, diluted DE in rats following chronic inhalation exposure.  In addition,
increasing lung particle burden resulting from this high-level exposure and decreased clearance
was demonstrated. A logistic regression model presented by Mauderly et al. (1987) indicated
that both lung DPM burden and exposure concentration may be useful for expressing exposure-
effect relationships.
       A long-term inhalation study (Ishinishi et al.,  1988a; Takaki et al., 1989) examined the
effects of emissions from a light-duty (LD) and a heavy-duty (HD) diesel engine on male and
female Fischer 344/Jcl rats. The LD engines were 1.8-L, 4-cylinder, swirl-chamber-type power
plants, and the HD engines were 11-L, 6-cylinder, direct-injection-type power plants. The
                                         7-99

-------
engines were connected to eddy-current dynamometers and operated at 1,200 rpm (LD engines)
and 1,700 rpm (HD engines). Nippon Oil Co. JIS No. 1 or No. 2 diesel fuel was used. The 30-
months whole-body exposure protocol (16 h/day, 6 days/week) used DPM concentrations of 0,
0.5, 1,  1.8, or 3.7 mg/m3 from HD engines and 0, 0.1, 0.4, 1.1, or 2.3 mg/m3 from LD engines.
The animals inhaled the exhaust emissions from 1700 to 0900 h.  Sixty-four male rats and 59 to
61 female rats from each exposure group were evaluated for carcinogenicity.
       For the experiments using the LD series engines, the highest incidence of hyperplastic
lesions plus tumors (72.6%) was seen in the highest exposure (2.3 mg/m3) group. However, this
high value was the result of the 70% incidence of hyperplastic lesions; the incidence of
adenomas was only 0.8% and that of carcinomas 1.6%. Hyperplastic lesion incidence was
considerably lower for the lower exposure groups (9.7%, 4.8%, 3.3%, and 3.3% for the 1.1, 0.4,
and 0.1 mg/m3 and control groups, respectively). The incidence of adenomas and carcinomas,
combining males and females, was not significantly different among exposure groups (2.4%,
4.0%, 0.8%, 2.4%, and 3.3% for the 2.3, 1.1, 0.4, and 0.1 mg/m3 groups and the controls,
respectively).
       For the experiments using the HD series engines, the total incidence of hyperplastic
lesions, adenomas, and carcinomas was highest (26.6%) in the 3.7 mg/m3 exposure  group. The
incidence of adenomas plus carcinomas for males and females combined equaled 6.5%, 3.3%,
0%, 0.8%, and 0.8% at 3.7, 1.8, 1, and 0.4 mg/m3 and for controls, respectively. A statistically
significant difference was reported between the 3.7 mg/m3 and the control groups for the HD
series engines. The carcinomas were identified as adenomas, adenosquamous carcinomas, and
squamous cell carcinomas.  Although the number of each was not reported, it was noted that the
majority were squamous cell carcinomas. A progressive dose-response relationship was not
demonstrated. Tumor incidence data for this experiment are presented in Table 7-3.
       The Ishinishi et al. (1988a) study also included recovery tests in which rats exposed to
whole DE (DPM concentration of 0.1 or 1.1 mg/m3 for the LD engine and 0.5 or 1.8 mg/m3 for
the HD engine) for 12 months were examined for lung tumors following 6-, 12-, or 18-month
recovery  periods in clean air. The incidences of neoplastic lesions were low, and pulmonary
DPM burden was lower than for animals continuously exposed to whole DE and not provided a
recovery  period.  The only carcinoma observed was in a rat examined 12 months following
exposure to exhaust (1.8 mg/m3) from the HD engine.
       Brightwell et al. (1986, 1989) studied the effects of DE on male and female  F344 rats.
The DE was generated by a 1.5-L Volkswagen engine that was computer-operated according to
the U.S. 72  FTP driving cycle. The engine was replaced after 15 mo. The engine emissions
were diluted by conditioned air delivered at 800 m3/h to produce the high-exposure  (6.6 mg/m3)
DE atmosphere. Further dilutions of 1:3 and 1:9 produced the medium- (2.2 mg/m3) and low-
                                        7-100

-------
(0.7 mg/m3) exposure atmospheres. The CO and NOX concentrations (mean ± SD) were 32 ± 11
ppm and 8 ± 1 ppm in the high-exposure concentration chamber. The inhalation exposures were
conducted overnight to provide five 16-h periods per week for 2 years; surviving animals were
maintained for an additional 6 mo.
       For males and females combined, a 1.2% (3/260), 0.7% (1/144), 9.7% (14/144), and
38.5% (55/143) incidence of primary lung tumors occurred in F344 rats following exposure to
clean air or 0.7, 2.2, and 6.6 mg of DPM/m3, respectively (Table 7-3). DE-induced tumor
incidence in rats was dose-related and higher in females than in males (Table 7-3). These data
included animals sacrificed at the interim periods (6, 12, 18, and 24 mo); therefore, the tumor
incidence does not accurately reflect the effects of long-term exposure to the DE atmospheres.
When tumor incidence is expressed relative to the specific intervals, a lung tumor incidence of
96% (24/25), 76% (19/25) of which were malignant, was reported for female rats in the high-
dose group exposed for 24 months and held in clean air for the remainder of their lives. For
male rats in the same group, the tumor incidence equaled 44% (12/27), of which 37% (10/27)
were malignant.  It was also noted that many of the animals  exhibiting tumors had more than one
tumor, often representing multiple histological types. The numbers and types of tumors
identified in the rats exposed to DE included adenomas (40), squamous cell carcinomas (35),
adenocarcinomas (19), mixed adenoma/adenocarcinomas (9), and mesothelioma (1).  It should
be noted that exposure during darkness (when increased activity would result in greater
respiratory exchange and greater inhaled dose) could account, in part, for the high response
reported for the rats.
       Lewis et al.  (1989) also examined the effects of inhalation exposure of DE and/or coal
dust on tumorigenesis on F344 rats. Groups of 216 male and 72 female rats were exposed to
clean air, whole DE (2 mg soot/m3), coal dust (2 mg/m3 respirable concentration; 5 to 6 mg/m3
total concentration), or DE plus coal dust (1 mg/m3 of each respirable concentration; 3.2 mg/m3
total concentration) for 7 h/day, 5 days/week during daylight hours for up to 24 mo.  Groups of
10 or more males were sacrificed at intermediate intervals (3, 6, and 12 mo). The DE was
produced by a 7.0-L, 4-cycle, water-cooled Caterpillar Model 3304 engine using No. 2 diesel
fuel (<0.5% sulfur by mass).  The exhaust was passed through a Wagner water scrubber, which
lowered the exhaust temperature and quenched engine backfire. The animals were exposed  in
100-cubic-foot chambers.  Temperature was controlled at 22 ± 2 °C and relative humidity at
50%±10%. The exhaust was diluted 27-fold with chemically and biologically filtered clean air
to achieve the desired particle concentration.
       Histological examination was performed on 120 to 121 male and 71 to 72 female rats
terminated after 24  months of exposure. The exhaust exposure did not significantly  affect the
tumor incidence beyond what would be expected for aging F344 rats. There was no
                                        7-101

-------
postexposure period, which may explain, in part, the lack of significant tumor induction. The
particulate matter concentration was also less than the effective dose in several other studies.
        In a more recent study reported by Heinrich et al. (1995), female Wistar rats were
exposed to whole DE (0.8, 2.5, or 7.0 mg/m3) 18 h/day, 5 days/week for up to 24 mo, then held
in clean air an additional 6 mo. The animals were exposed in either 6 or 12 m3 exposure
chambers. Temperature and relative humidity were maintained at 23-25 °C and 50%-70%,
respectively. DE was generated by two 40-kw 1.6-L diesel engines (Volkswagen). One of them
was operated according to the U.S. 72 cycle. The other was operated under constant load
conditions.  The first engine did not supply sufficient exhaust, which was filled by the second
engine.  Cumulative exposures for the rats in the various treatment groups were 61.7, 21.8, and
7.4 g/m3 x h for the high, medium, and low whole-exhaust exposures.  Significant increases in
tumor incidences were observed in the high (22/100;/7<0.001) and mid (11/200;/?<0.01)
exposure groups  relative to clean-air controls (Table 7-3). Only one tumor (1/217), an
adenocarcinoma, was observed in clean-air controls. Relative to clean-air controls, significantly
increased incidences were observed in the high-exposure rats for benign squamous cell tumors
(14/100; p<0.001), adenomas (4/100; p<0.01), and adenocarcinomas (5/100; p<0.05). Only the
incidence of benign squamous cell tumors (7/200; p<0.01) was significantly increased in the
mid-exposure group relative to the clean-air controls.
       Particle lung burden and alveolar clearance also were determined in the Heinrich et al.
(1995) study. Relative to clean air controls, alveolar clearance was significantly compromised
by exposure to mid and high DE.  For the high-diesel-exhaust group, 3-mo recovery time in
clean air failed to reverse the compromised alveolar clearance.
       In a  study conducted at the Inhalation Toxicology Research Institute (Nikula et  al.,
1995) F344 rats (114-115 per sex per group) were exposed 16 hr/day, 5 days/week during
daylight hours to DE diluted to achieve particle concentrations of 2.5 or 6.5 mg/m3 for up to 24
mo.  Controls (118 males, 114 females) were exposed to clean air. Surviving rats were
maintained an additional 6 weeks in clean air, at which time mortality reached 90%. DE was
generated with two 1988 Model LH6 General Motors  6.2-L V-8 engines burning D-2 fuel that
met EPA certification standards.  Chamber air flow was sufficient to provide about 15
exchanges per hour. Relative humidity was 40% to 70%  and temperature ranged from 23 to 25
°C.
       Following low and  high DE exposure, the lung burdens were 36.7 and 80.7 mg,
respectively, for females and 45.1 and 90.1 mg, respectively, for males. The percentages of
susceptible rats (males and females combined) with malignant neoplasms were 0.9 (control),  3.3
(low DE), and 12.3 (high DE). The percentages of rats (males and females combined) with
malignant or benign neoplasms were  1.4 (control), 6.2 (low DE), and 17.9 (high DE). All
                                         7-102

-------
primary neoplasms were associated with the parenchyma rather than the conducting airways of
the lungs.  The first lung neoplasm was observed at 15 mo. Among 212 males and females
examined in the high-dose group, adenomas were detected in 23 animals, adenocarcinomas in 22
animals, squamous cell carcinomas in 3 animals, and an adenosquamous carcinoma in 1 animal.
For further details see Table 7-3. Analysis of the histopathologic data suggested a progressive
process from alveolar epithelial hyperplasia to adenomas and adenocarcinomas.
       Iwai et al. (1997) carried out a series of exposures to both filtered and whole exhaust
using a light-duty (2,369 mL) diesel engine. The protocol for engine operation was not stated.
Groups of female SPF F344 Fischer rats were exposed for 2 years for 8 hr/day, 7 days/week, 8
hr/day, 6 days/week, or  18 hr/day, 3 days/week to either filtered exhaust or exhaust diluted to a
particle concentration of 9.4, 3.2, and 5.1 mg/m3, respectively.  Cumulative exposure (mg/m3 x
hrs of exposure) equaled 274.4, 153.6, and 258.1 mg/m3.  The animals were then held for an
additional  6 months in clean air. Lung tumors were reported in 5/121 (4%) of controls, 4/108
(4%) of those exposed to filtered exhaust, and 50/153 (35%) among those exposed to whole
exhaust. Among rats exposed to whole DE the following number of tumors were detected; 57
adenomas, 24 adenocarcinomas, 2 benign squamous cell tumors, 7 squamous cell carcinomas,
and 3 adenosquamous carcinomas.  The authors stated that benign squamous cell tumors
probably corresponded to squamous cysts in another classification.

7.3.1.2. Mouse Studies
       A series of inhalation studies using strain A mice was conducted by Orthoefer et al.
(1981). Strain A mice are usually given a series of intraperitoneal injections with the test agent;
they are then sacrificed at about 9 months and examined for lung tumors.  In the  present series,
inhalation  exposure was substituted. DE was provided by one of two Nissan CN6-33 diesel
engines having a displacement of 3244 cc and run on a Federal  Short Cycle. Flow through the
exposure chambers was  sufficient to provide 15 air changes per hour. Temperature was
maintained at 24 °C and relative humidity at 75%.  In the first study,  groups of 25 male Strong
A strain (A/S) mice were exposed to irradiated DE (to simulate  chemical reactions induced by
sunlight) or nonirradiated DE (6 mg/m3) for 20 h/day, 7 days/week.  Additional groups of 40
Jackson A strain (S/J) mice (20 of each sex) were exposed similarly to either clean air or DE,
then held in clean air until sacrificed at 9 months of age.  No tumorigenic effects were detected
at 9 months of age. Further studies were conducted in which male A/S  mice were exposed 8
hr/day, 7 days/week until sacrifice (approximately 300 at 9 months of age and approximately
100 at  12 months of age). With the exception of those treated with urethan, the number of
tumors per mouse did not exceed historical control levels in any of the studies. Exposure to DE,
                                        7-103

-------
however, significantly inhibited the tumorigenic effects of the 5-mg urethan treatment. Results
are listed in Table 7-3.
       Kaplan et al. (1982) also reported the effects of diesel exposure in strain A mice. Groups
of male strain A/J mice were exposed for 20 h/day, 7 days/week for 90 days and held until 9
months of age. Briefly, the animals were exposed in inhalation chambers to DE generated by a
5.7-L Oldsmobile engine operated continuously at 40 mph at DPM concentrations of 0, 0.25,
0.75, or 1.5 mg/m3.  Controls were exposed to clean air.  Temperature was maintained at 22 ± 2
°C and relative humidity  at 50% ± 10% within the chambers.  Among 458 controls and 485
exposed animals, tumors  were detected in 31.4% of those breathing clean air versus 34.2% of
those exposed to DE. The mean number of tumors per mouse also failed to show significant
differences.
       In a follow-up study, strain A mice were exposed to DE for 8 months (Kaplan et al.,
1983; White et al., 1983). After exposure to the highest  exhaust concentration (1.5 mg/m3), the
percentage of mice with pulmonary adenomas and the mean number of tumors per mouse were
significantly less (p<0.05) than those for controls (25.0% vs. 33.5% and 0.30 ± 0.02 [S.E.] vs.
0.42 ± 0.03 [S.E.]) (Table 7-3).
       Pepelko and Peirano (1983) summarized a series  of studies on the health effects of diesel
emissions in mice. Exhaust was provided by two Nissan CN 6-33, 6-cylinder, 3.24-L diesel
engines coupled to a Chrysler A-272 automatic transmission and Eaton model 758-DG
dynamometer.  Sixty-day pilot studies were conducted at a 1:14 dilution, providing DPM
concentrations of 6 mg/m3  The engines were operated using the Modified California Cycle.
These 20-hr/day, 7-days/week pilot studies using rats, cats, guinea pigs, and mice produced
decreases in weight gain  and food consumption. Therefore,  at the beginning of the long-term
studies, exposure time was reduced to 8 h/day, 7 days/week at an exhaust DPM concentration of
6 mg/m3. During the final 12 months of exposure, however, the DPM concentration was
increased to 12 mg/m3. For the chronic studies, the engines were operated using the Federal
Short Cycle. Chamber temperature was maintained at 24 °C and relative humidity at 50%.
Airflow was sufficient for 15 changes per hour.
       Pepelko and Peirano (1983) described a two-generation study using Sencar mice exposed
to DE. Male and female  parent-generation mice were exposed to DE at a DPM concentration of
6 mg/m3 prior to (from weaning to sexual  maturity) and throughout mating. The dams continued
exposure through gestation, birth, and weaning. Groups  of offspring (130 males and 130
females) were exposed to either DE or clean air. The exhaust exposure was increased to a DPM
concentration of 12 mg/m3 when the offspring were  12 weeks of age and was maintained until
termination of the experiment when the mice were 15 months old.
                                        7-104

-------
       The incidence of pulmonary adenomas (16.3%) was significantly increased in the mice
exposed to DE compared with 6.3% in clean-air controls. The incidence in males and females
combined was 10.2% in 205 animals examined compared with 5.1% in 205 clean-air controls.
This difference was also significant. The incidence of carcinomas was not affected by exhaust
exposure in either sex. These results provided the earliest evidence for cancer induction
following inhalation exposure to DE.  The increase in the sensitivity of the study, allowing
detection of tumors at 15 mo, may have been the result of exposure from conception. It is likely
that Sencar mice are sensitive to induction of lung tumors because they are also sensitive to
induction of skin tumors. These data are summarized in Table 7-3.
       Takemoto  et al. (1986) reported the effects of inhaled DE (2 to 4 mg/m3, 4 h/day, 4
days/week, for up to 28 mo) in ICR and C57BL mice exposed from birth. Details of the
exposure conditions are presented in Section 7.3.2.1. All numbers reported are for males and
females combined. Four adenomas and 1 adenocarcinoma were detected in 34 DE-exposed ICR
mice autopsied at  13 to 18 mo, compared with 3 adenomas among 38 controls. Six adenomas
and 3 adenocarcinomas were reported in 22 diesel-exposed ICR mice autopsied at 19 to 28 mo,
compared with 3 adenomas and 1 adenocarcinoma in 22 controls. Four adenomas and 2
adenocarcinomas were detected in 79 C57BL mice autopsied at  13 to 18 mo, compared with
none in 19 unexposed animals.  Among males and females autopsied at 19 to 28 mo, 8 adenomas
and 3 adenocarcinomas were detected in 71 exposed animals, compared with 1 adenoma among
32 controls. No significant increases in adenoma or adenocarcinoma were reported for either
strain of exposed mice.  However, the significance of the increase in the combined incidence of
adenomas and carcinomas was not evaluated statistically. A statistical analysis by Pott and
Heinrich (1990a) indicated that the difference in combined benign and malignant tumors
between whole DE-exposed C57BL/6N mice and corresponding controls was significant at
p<05. See Table  7-3 for details of tumor incidence.
       Heinrich et al. (1986b) and Stober (1986), as part of a larger study, also evaluated the
effects of DE in mice. Details of the exposure conditions reported by Heinrich et al.  (1986a) are
given in Section 7.3.1.1 and Appendix A.  Following lifetime (19 h/day, 5 days/week, for a
maximum of 120 weeks) exposure to DE diluted to achieve a particle concentration of 4.2
mg/m3, 76 female NMRI mice exhibited a total lung tumor incidence of adenomas and
adenocarcinomas combined of 32%. Tumor incidences reported for control mice (n = 84)
equaled 11% for adenomas and adenocarcinomas combined. While the incidence of adenomas
showed little change, adenocarcinomas increased significantly from 2.4% for controls to 17%
for exhaust-exposed mice. In a follow-up study, however, Heinrich et al. (1995) reported a lack
of tumorigenic response in either female NMRI or C57BL/6N mice exposed 17 h/day, 5
                                        7-105

-------
days/week for 13.5 to 23 months to whole DE diluted to produce a particle concentration of 4.5
mg/m3.  These data are summarized in Table 7-3.
       The lack of a carcinogenic response in mice was reported by Mauderly et al. (1996). In
this study, groups of 540 to 600 CD-I male and female mice were exposed to whole DE (7.1,
3.5, or 0.35 mg DPM/m3) for 7 hr/day, 5 days/week for up to 24 mo. Controls were exposed to
filtered air. DE was provided by 5.7-L Oldsmobile V-8 engines operated continuously on the
U.S. Federal Test Procedure urban certification cycle.  The chambers were maintained at 25 °C-
28 °C, relative humidity at 40%-60%, and a flow rate sufficient for 15 air exchanges per hour.
Animals were exposed during the light cycle, which ran from 6:00 AM to 6:00 PM. DPM
accumulation in the lungs of exposed mice was assessed at 6, 12, and 18 months of exposure and
was shown to be progressive; DPM burdens were 0.2 ± 0.02, 3.7 ± 0.16, and 5.6 ± 0.39 mg for
the low-, medium-, and high-exposure groups, respectively.  The lung burdens in both the
medium- and high-exposure groups exceeded that predicted by exposure concentration ratio for
the low-exposure group.  Contrary to what was observed in rats (Heinrich et al., 1986b; Stober,
1986; Nikula et al., 1995; Mauderly et al., 1987), an exposure-related increase in primary lung
neoplasms was not observed in the CD-I mice,  supporting the contention of a species difference
in the pulmonary carcinogenic response to poorly soluble particles.  The percentage incidence of
mice (males and females combined) with one or more malignant or benign  neoplasms was 13.4,
14.6, 9.7, and 7.5 for controls and low-, medium-, and high-exposure groups, respectively.
       Although earlier studies provided some evidence for tumorigenic responses in diesel-
exposed mice, no increases were reported in the two most recent studies by Mauderly et al.
(1996) and Heinrich et al. (1995), which utilized large group sizes and were well designed and
conducted.  Overall, the results in mice must therefore be considered to be equivocal.

7.3.1.3. Hamster Studies
       Heinrich et al. (1982) examined the effects of DE exposure on tumor frequency in female
Syrian golden hamsters.  Groups of 48 to 72 animals were exposed to clean air or whole DE at a
mean DPM concentration of 3.9 mg/m3.  Inhalation exposures were conducted 7 to 8 hr/day, 5
days/week for 2 years. The exhaust was produced by a 2.4-L Daimler-Benz engine operated
under a constant load and a constant speed of 2,400 rpm. Flow rate was sufficient for about 20
exchanges per hour in the 250-L chambers. No lung tumors were reported in either exposure
group.
       In a subsequent study, Syrian hamsters were exposed 19 hr/day, 5 days/week for a
lifetime to DE diluted to a DPM concentration of 4.24 mg/m3 (Heinrich et al., 1986b; Stober,
1986). Details of the  exposure conditions are reported in Appendix A. Ninety-six animals per
                                        7-106

-------
group were exposed to clean air or exhaust. No lung tumors were seen in either the clean-air
group or in the DE-exposed group.
       In a third study (Heinrich et al., 1989b), hamsters were exposed to exhaust from a
Daimler-Benz 2.4-L engine operated at a constant load of about 15 kW and at a uniform speed
of 2,000 rpm. The exhaust was diluted to an exhaust-clean air ratio of about 1:13, resulting in a
mean particle concentration of 3.75 mg/m3. Exposures were conducted in chambers maintained
at 22 to 24 °C and 40% to 60% relative humidity  for up to 18 mo. Surviving hamsters were
maintained in clean air for up to an additional 6 mo.  The animals were exposed 19 hr/day, 5
days/week beginning at noon each day, under a 12-hr light cycle starting at 7 AM.  Forty
animals per group were exposed to whole DE or clean air.  No lung tumors were detected in
either the clean-air or diesel-exposed hamsters.
       Brightwell et al. (1986, 1989) studied the  effects of DE on male and female Syrian
golden hamsters. Groups of 52 males and 52 females, 6 to 8 weeks old, were exposed to DE at
DPM concentrations of 0.7, 2.2, or 6.6 mg/m3. They were exposed 16 hr/day, 5 days/week for a
total of 2 years and then sacrificed. Exposure conditions are described in Section 7.3.1.1. No
statistically significant (t test) relationship between tumor incidence and exhaust exposure was
reported.
       In summary,  DE alone did not induce an increase in lung tumors in hamsters of either sex
in several studies of chronic duration at high exposure concentrations.

7.3.1.4. Monkey Studies
       Fifteen male  cynomolgus monkeys were exposed to DE (2 mg/m3) for 7 hr/day, 5
days/week for 24 months (Lewis et al.,  1989).  The same numbers of animals were also exposed
to coal dust (2 mg/m3 respirable concentration; 5  to 6 mg/m3 total concentration), DE plus coal
dust (1 mg/m3 respirable concentration for each component; 3.2 mg/m3 total concentration), or
filtered air. Details of exposure conditions were listed previously in the description of the Lewis
et al. (1989) study with rats (Section 7.3.1.1) and  are listed in Appendix A.
       None of the monkeys exposed to DE exhibited a significantly increased incidence of
preneoplastic or neoplastic lesions. It should be noted, however, that the 24-mo time frame
employed  in this study may not have allowed the  manifestation of tumors in primates, because
this duration is only  a small fraction of the monkeys' expected lifespan.  In fact, there have been
no near-lifetime exposure studies in nonrodent species.

7.3.2. Inhalation Studies (filtered DE)
       Several studies have been conducted in which animals were exposed to DE filtered  to
remove PM. As these studies also included groups exposed to whole exhaust, details can be
found in Sections 7.3.1.1 for rats, 7.3.1.2 for mice, and 7.3.1.3 for hamsters.  Heinrich et al.

                                         7-107

-------
(1986b) and Stober (1986) reported negative results for lung tumor induction in female Wistar
rats exposed to filtered exhaust diluted to produce an unfiltered particle concentration of 4.24
mg/m3.  Negative results were also reported in female Fischer 344 rats exposed to filtered
exhaust diluted to produce an unfiltered particle concentration of 4.9 mg/m3 (Iwai et al., 1986a),
in Fischer 344 rats of either sex exposed to filtered exhaust diluted to produce an unfiltered
particle concentration of 6.6 mg/m3 (Brightwell et al., 1989), in female Wistar rats exposed to
filtered exhaust diluted to produce an unfiltered particle concentration of 7.0 mg/m3 (Heinrich et
al.,  1995), and in female Fischer 344 rats exposed to filtered exhaust diluted to produce
unfiltered particle concentrations of 5.1, 3.2, or 9.4 mg/m3 (Iwai et al., 1997). In the Iwai et al.
(1986a) study, splenic lymphomas were detected in 37.5% of the exposed rats compared with
8.2% in controls.
       In the study reported by Heinrich at al. (1986a) and Stober (1986), primary lung tumors
were seen in 29/93 NMRI mice (males and females combined) exposed to filtered exhaust,
compared with 11/84 in clean-air controls, a statistically significant increase. In a repeat study
by Heinrich et al. (1995), however,  significant lung tumor increases were not detected in either
female NMRI or C57BL/6N mice exposed to  filtered exhaust diluted to produce an unfiltered
particle concentration of 4.5 mg/m3.
        Filtered exhaust also failed to induce lung tumor induction in Syrian Golden hamsters
(Heinrich et al., 1986a; Brightwell et al., 1989).
        Although lung tumor increases were reported in one study and lymphomas in another,
these results could not be confirmed in subsequent investigations. It is therefore concluded that
little direct evidence exists for carcinogenicity of the vapor phase of DE in laboratory animals at
concentrations tested.

7.3.3. Inhalation Studies (DE plus Cocarcinogens)
       Details of the studies reported here have been described earlier and in Table 7-3.  Tumor
initiation with urethan (1 mg/kg body weight i.p. at the start of exposure) or promotion with
butylated hydroxytolulene (300 mg/kg body weight i.p. week 1, 83 mg/kg week 2, and 150
mg/kg for weeks 3-52) did not influence tumorigenic responses in Sencar mice of both sexes
exposed to concentrations of DE up to 12 mg/m3 (Pepelko and Peirano, 1983).
       Heinrich et al. (1986b) exposed Syrian hamsters of both sexes to DE diluted to a particle
concentration of 4 mg/m3.  See Section 7.3.1.1 for details of the exposure conditions.  At the
start of exposure the hamsters received either  one dose of 4.5 mg diethylnitrosamine (DEN)
subcutaneously per kg body weight or 20 weekly intratracheal instillations of 250 |lgB[a]P.
Female NMRI mice received weekly intratracheal instillations of 50 or 100 |lgB[a]P for 10 or
20 weeks, respectively, or 50 \Jig dibenz[ah]anthracene (DBA) for 10 weeks. Additional groups
of 96 newborn mice received one s.c. injection of 5 or 10 [Lg DBA between 24 and 48 hr after
birth. Female Wistar rats received weekly subcutaneous injections of dipentylnitrosamine

                                         7-108

-------
(DPN) at doses of 500 and 250 mg/kg body weight, respectively, during the first 25 weeks of
exhaust inhalation exposure. Neither DEN, DBA, or DPN treatment enhanced any tumorigenic
responses to DE. Response to B[a]P did not differ from that of BaP alone in hamsters, but
results were inconsistent in mice.  Although 20 B[a]P instillations induced a 71% tumor
incidence in mice, concomitant diesel exposure resulted in only a 41% incidence. However,
neither 10 B[a]P instillations nor DBA instillations induced significant effects.
       Takemoto et al. (1986) exposed Fischer 344 rats for 2 years to DE at particle
concentrations of 2 to 4 mg/m3.  One month after start of inhalation exposure one group of rats
received di-isopropyl-nitrosamine (DIPN) administered i.p. at 1 mg/kg weekly for 3 weeks.
Among injected animals autopsied at 18 to 24 mo, 10 adenomas and 4 adenocarcinomas were
reported in 21 animals exposed to clean air, compared with 12 adenomas and 7 adenocarcinomas
in 18 diesel-exposed rats. According to the authors,  the incidence of adenocarcinomas was not
significantly increased by exposure to DE.
       Brightwell et al. (1989) investigated the concomitant effects of DE and DEN in Syrian
hamsters exposed to DE diluted to produce particle concentrations of 0.7, 2.2, or 6.6 mg/m3 for 2
years.  The  animals received a single dose of 4.5 mg DEN s.c. 3 days prior to start of inhalation
exposure. DEN did not affect the lack of responsiveness to DE alone. Heinrich et al. (1989b)
also exposed Syrian hamsters of both sexes to DE diluted to a particle  concentration of 3.75
mg/m3 for up to 18 mo.  After 2 weeks of exposure, groups were treated with either 3 or 6 mg
DEN/kg body weight, respectively. Again, DEN did not significantly  influence the lack of
tumorigenic responses to DE.
       Heinrich et al. (1989a) investigated the effects of DPN in female Wistar rats exposed to
DE diluted to achieve a particle concentration of 4.24 mg/m3 for 2-2.5 years. DPN at doses of
250 and 500 mg/kg body weight was injected subcutaneously once a week for the first 25 weeks
of exposure. The tumorigenic responses to DPN were not affected by exposure to DE.  For
details of exposure conditions of the hamster studies see Section 7.3.1.3.
       Heinrich et al. (1986a) and Mohr et al. (1986) compared the effects of exposure to
particles having only a minimal carbon core but a much greater concentration of PAHs than
DPM does.  The desired exposure conditions were achieved by mixing coal oven flue gas with
pyrolyzed pitch. The concentration of B[a]P and other PAHs per milligram  of DPM was about
three orders of magnitude greater than that of DE. Female rats were exposed to the flue gas-
pyrolyzed pitch for 16 hr/day, 5 days/week at particle concentrations of 3 to 7 mg/m3 for 22 mo,
then held in clean air for up to an additional 12 mo.  Among 116 animals exposed, 22 tumors
were reported in 21 animals, for an incidence of 18.1%. One was a bronchioloalveolar
adenoma, one was a bronchioloalveolar carcinoma, and 20 were squamous cell tumors. Among
the latter, 16 were classified as benign keratinizing cystic tumors and 4 were classified as
carcinomas. No tumors were reported in 115 controls. The tumor incidence in this study was
comparable to that reported previously for the DE-exposed animals.

                                        7-109

-------
       In analyzing the studies of Heinrich et al. (1986a,b), Heinrich (1990b), Mohr et al.
(1986), and Stober (1986), it must be noted that the incidence of lung tumors occurring
following exposure to whole DE, coal oven flue gas, or carbon black (15.8%, 18.1%, and 8% to
17%, respectively) was very similar. This occurred despite the fact that the PAH content of the
PAH-enriched pyrolyzed pitch was more than three orders of magnitude greater than that of DE;
carbon black, on the other hand, had only traces of PAHs. Based on these findings, particle-
associated effects appear to be the primary cause of diesel-exhaust-induced lung cancer in rats
exposed at high concentrations.  This issue is discussed further in Chapter 7.

7.3.4.  Lung Implantation or Intratracheal Instillation Studies
7.3.4.1. Rat Studies
       Grimmer et al. (1987), using female Osborne Mendel rats (35 per treatment group),
provided evidence that PAHs in DE that consist of four or more rings have carcinogenic
potential. Condensate was obtained from the whole exhaust of a 3.0-L passenger-car diesel
engine connected to a dynamometer operated under simulated city traffic driving conditions.
This condensate was separated by liquid-liquid distribution into hydrophilic and hydrophobic
fractions representing 25% and 75% of the total condensate, respectively. The hydrophilic,
hydrophobic, or reconstituted hydrophobic fractions were surgically implanted into the lungs of
the rats. Untreated controls, vehicle (beeswax/trioctanoin) controls,  and positive (B[a]P)
controls were also included in the protocol (Table  7-6). Fraction lib (made up of PAHs with
four to seven rings), which accounted for only 0.8% of the total weight of DPM condensate,
produced the highest incidence of carcinomas following implantation into rat lungs. A
carcinoma incidence of 17.1% was observed following implantation of 0.21 mg lib/rat, whereas
the nitro-PAH fraction (lid) at 0.18 mg/rat accounted for only a 2.8% carcinoma incidence.
Hydrophilic fractions of the DPM extracts, vehicle (beeswax/trioctanoin) controls, and untreated
controls failed to exhibit carcinoma formation. Administration of all hydrophobic fractions (Ila-
d) produced a carcinoma incidence (20%) similar to the summed incidence of fraction lib
(17.1%) and lid (2.8%).  The B[a]P positive controls (0.03, 0.1, 0.3 mg/rat) yielded a carcinoma
incidence of 8.6%, 31.4%, and 77.1%, respectively.  The study showed that the tumorigenic
agents were primarily four- to seven-ring PAHs and, to a lesser extent, nitroaromatics.
However, these studies demonstrated that simultaneous administration of various PAH
compounds resulted in a varying of the tumorigenic effect, thereby implying that the
tumorigenic potency of PAH mixtures may not depend on any one individual PAH. This study
did not provide any information regarding the bioavailability of the particle-associated PAHs
that might be responsible for carcinogenicity.
       Kawabata et al. (1986) compared the effects of activated carbon and DE on lung tumor
formation.  One group of 59 F344 rats was intratracheally instilled with DPM (1 mg/week for 10
                                         7-110

-------
          Table 7-6.  Tumor incidence and survival time of rats treated by surgical lung implantation with

          fractions from DE condensate (35 rats/group)
Material portion by weight (%)
Hydrophilic fraction (I) (25)
Hydrophobic fraction (II) (75)
Nonaromatics +
PACC 2 + 3 rings (Ha) (72)
PAHd 4 to 7 rings (lib) (0.8)
Polar PAC (He) (1.1)
Nitro-PAH (lid) (0.7)
Reconstituted hydrophobics
(Ia,b,c,d)(74.5)
Control, unrelated
Control (beeswax/trioctanoin)
B[a]P


Dose (mg)
6.7
20.00

19.22
0.21
0.29
0.19
19.91



0.3
0.1
0.03
Median
survival time
in weeks Number of
(range) carcinomas"
97 (24-139) 0
99 (50-139) 50601

103 (25-140)
102(50-140)
97 (44-138)
106(32-135)
93(46-136) 70027113

110(23-138)
103(51-136)
69(41-135)
98 (22-134)
97(32-135)
Number of Carcinoma
adenomas15 incidence (%)
1 0
1000 14.2

0
17.1
0
2.8
101000 20.0

0
0
77.1
31.4
8.6
aSquamous cell carcinoma.

bBronchiolar/alveolar adenoma.
CPAC = polycyclic aromatic compounds.

dPAH = polycyclic aromatic hydrocarbons.


Source: Adapted from Grimmer et al., 1987.

-------
weeks). A second group of 31 rats was instilled with activated carbon using the same dosing
regime. Twenty-seven rats received only the solvent (buffered saline with 0.05% Tween 80),
and 53 rats were uninjected. Rats dying after 18 months were autopsied. All animals surviving
30 months or more postinstillation were sacrificed and evaluated for histopathology.  Among 42
animals exposed to DPM surviving 18 months or more, tumors were reported in 31, including 20
malignancies.  In the subgroup surviving for 30 mo, tumors were detected in 19 of 20 animals,
including 10 malignancies. Among the rats exposed to activated carbon, the incidence of lung
tumors equaled 11 of 23  autopsied, with 7 cases of malignancy. Data for those dying between
18 and 30 months and  those sacrificed at 30 months were not reported separately. Statistical
analysis indicated that  activated carbon induced a significant increase in lung tumor incidence
compared with no tumors in 50 uninjected controls and 1 tumor in 23 solvent-injected controls.
The tumor incidence was significantly greater in the DPM-instilled group and was significantly
greater than the increase  in the carbon-instilled group.
       A study reported by Rittinghausen et al. (1997) suggested that organic constituents of
diesel particles play a role in the induction of lung tumors in rats. An incidence of 16.7%
pulmonary cystic keratinizing squamous cell lesions was noted in rats intratracheally instilled
with 15 mg whole DE  particles, compared with 2.1% in rats instilled with 15 mg particles
extracted to remove all organic constituents, and none among controls. Instillation of 30 mg of
extracted particles induced a 14.6% incidence of squamous lesions, indicating the greater
effectiveness of particles alone as lung particle overload increased.
       Iwai et al. (1997) instilled 2, 4, 8, and 10 mg of whole diesel particles over a 2- to 10-
week period into female  F/344 rats, 50 or more per group. Tumors were reported in  6%, 20%,
43%, and 74% of the rats, with incidence of malignant tumors equal to 2%, 13%, 34%, and
48%, respectively.  In  a second experiment comparing whole with extracted diesel particles,
tumor incidence equaled 1/48 (2%) in uninjected controls, 3/55 (5%) in solvent controls, 12/56
(21%) in extracted diesel particles, and 13/106 (12%)  in animals injected with unextracted
particles. Although the extracted particles appeared to be more potent, when converted to a lung
burden basis (mg/100 mg dry lung) the incidence was only 14% among those exposed to
extracted exhaust compared with 31% in those exposed to whole particles.
       Dasenbrock et  al. (1996)  conducted a study to  determine the relative importance  of the
organic constituents of diesel particles and particle surface area in the induction of lung cancer in
rats.  Fifty-two female Wistar rats were intratracheally instilled with 16-17 doses of DPM,
extracted DPM, printex carbon black (PR), lampblack (LB), B[a]P, DPM + B[a]P, or PR +
B[a]P. The animals were held for a lifetime or sacrificed when moribund.  The lungs were
necropsied and examined for tumors. Diesel particles were collected from a Volkswagen 1.6-L
engine operating on a US FTP-72 driving cycle. The mass median aerodynamic diameter
(MMAD) of the diesel particles was 0.25 |im and the specific surface area was 12 m2/gm.
                                         7-112

-------
Following extraction with toluene, specific surface area increased to 138 m2/gm.  The MMAD
for extracted PR was equal to 14 nm, while the specific surface area equaled 271  m2/gm. The
MMAD for extracted lampblack was equal to 95 nm, with a specific surface area equal to 20
m2/gm.  The B[a]P content of the treated particles was 11.3 mg per gm diesel particles and 29.5
mg B[a]P per gm PR.  Significant increases in lung tumors were detected in rats instilled with 15
mg unextracted DPM and 30 mg extracted DPM, but not 15 mg extracted DPM.  Printex CB was
more potent than lampblack CB for induction of lung tumors, whereas B[a]P was effective only
at high doses.  Total dose and tumor responses are shown in Table 7-4.
       A number of conclusions can be drawn from these results.  First of all, particles devoid
of organics are capable of inducing lung tumor formation, as indicated by positive results in the
groups treated with high-dose extracted diesel particles and printex. Nevertheless, toluene
extraction of organics from diesel particles results in a decrease in potency, indicating that the
organic fraction does play a role in cancer induction. A relationship between cancer potency and
particle surface area was also suggested by the finding that printex with a large specific surface
area  was more potent than either extracted DPM or lampblack, which  have smaller specific
areas. Finally, while very large doses of B[a]P are very effective in the induction of lung
tumors, smaller doses adsorbed to particle surfaces had little detectable effect, suggesting that
other organic components of DE may be of greater importance in the induction of lung tumors at
low doses pf B[a]P (0.2-0.4 mg).

7.3.4.2. Syrian Hamster Studies
       Kunitake et al. (1986) and Ishinishi et al. (1988b) conducted a study in which total doses
of 1.5, 7.5, or 15 mg of a dichloromethane extract of DPM were instilled intratracheally over 15
weeks into male Syrian hamsters that were then held for their lifetimes.  The tumor incidences of
2.3% (1/44), 0% (0/56), and  1.7% (1/59) for the high-, medium-, and  low-dose groups,
respectively, did not differ significantly from the 1.7% (1/56) reported for controls.  Addition of
7.5 mg of B[a]P to a DPM extract dose of 1.5 mg resulted in a total tumor incidence of 91.2%
and malignant tumor incidence of 88%. B[a]P (7.5 mg over 15 weeks) alone produced a tumor
incidence rate of 88.2% (85% of these being malignant), which was not significantly different
from the DPM extract + B[a]P group. Intratracheal  administration of 0.03 |lg B[a]P, the
equivalent content in 15 mg of DPM extract, failed to cause a significant increase in tumors  in
rats.  This study demonstrated a lack of detectable interaction between DPM extract and B[a]P,
the failure of DPM extract to induce carcinogenesis, and the propensity for respiratory tract
carcinogenesis following intratracheal instillation of high doses of B[a]P. For studies using  the
DPM extract, some concern must be registered regarding the known differences in chemical
composition between DPM extract and DPM. As with all intratracheal instillation protocols,
DPM extract lacks the complement of volatile chemicals found in whole DE.
                                         7-113

-------
       The effects on hamsters of intratracheally instilled DPM suspension, DPM with Fe2O3, or
DPM extract with Fe2O3 as the carrier were studied by Shefner et al. (1982). The DPM
component in each of the treatments was administered at concentrations of 1.25, 2.5, or 5.0
mg/week for 15 weeks to groups of 50 male Syrian golden hamsters.  The total volume instilled
was 3.0 mL (0.2 mL/week for 15 weeks). The DPM and dichloromethane extracts were
suspended in physiological saline with gelatin (0.5% w/v), gum arable (0.5% w/v), and
propylene glycol (10% by volume). The Fe2O3 concentration, when used, was 1.25 mg/0.2 mL
of suspension. Controls received vehicle and, where appropriate, carrier particles (Fe2O3)
without the DPM component.  Two replicates of the experiments were performed.
Adenomatous hyperplasia was reported to be most severe in those animals treated with DPM or
DPM plus Fe2O3 particles and least severe in those animals receiving DPM plus Fe2O3.  Of the
two lung adenomas detected microscopically,  one was in an animal treated with a high dose of
DPM and the other was in an animal receiving a high dose of DPM extract. Although lung
damage was increased by instillation of DPM, there was no evidence of tumorigenicity.

7.3.4.3. Mouse Studies
       Ichinose et al. (1997a) intratracheally instilled 36 four-week-old male ICR mice per
group weekly for 10 weeks with sterile saline or 0.05, 0.1, or 0.2 mg DPM. Particles were
collected from a 2.74-L four-cylinder Isuzu engine run at a steady speed of 1,500 rpm under a
load of 10 torque (kg/m).  Twenty-four hours after the last instillation, six animals per group
were sacrificed for measurement of lung 8-hydroxydeoxyguanosine (8-OHdG).  The remaining
animals were sacrificed after 12 months for histopathological analysis. Lung tumor incidence
varied from 4/30 (13.3%) for controls to 9/30 (30%), 9/29 (31%), and 7/29 (24.1%)  for mice
instilled with 0.05, 0.1, and 0.2 mg/week, respectively. The increase in animals with lung
tumors compared with controls was statistically significant for the 0.1 mg dose group, the only
group analyzed statistically. Increases in 8-OHdG, an indicator of oxidative DNA damage,
correlated well with the increase in tumor incidence in the 0.05 mg  dose group, although less so
with the other two. The correlation coefficients r = 0.916, 0.765, and 0.677 for the 0.05, 0.10,
and 0.20 mg DPM groups, respectively.
       In a similar study, 33 four-week-old male ICR mice per group were intratracheally
instilled weekly for 10 weeks with sterile saline, 0.1 mg DPM, or 0.1 mg DPM from which the
organic constituents were extracted with hexane (Ichinose et al., 1997b). Exhaust was collected
from a 2.74-L four-cylinder Izuzu engine run at a steady speed of 2,000 rpm under a load of 6
torque  (kg/m). Twenty-four hours after the last instillation, six animals per group were
sacrificed for measurement of 8-OHdG.  Surviving animals were sacrificed after 12 mo.  The
incidence of lung tumors increased from 3/27 (11.1%) among controls to 7/27 (25.9%) among
those instilled with extracted diesel particles and 9/26 (34.6%) among those instilled with
                                         7-114

-------
unextracted particles. The increase in number of tumor-bearing animals was statistically
significant compared with controls (p<0.05) for the group treated with unextracted particles.
The increase in 8-OHdG was highly correlated with lung tumor incidence, r = 0.99.

7.3.5. Subcutaneous and Intraperitoneal Injection Studies
7.3.5.1. Mouse Studies
       In addition to inhalation studies, Orthoefer et al. (1981) also tested the effects of i.p.
injections of DPM on male (A/S) strain mice. Three groups of 30 mice were injected with 0.1
mL of a suspension (particles in distilled water) containing 47, 117, or 235 |lg of DPM collected
from Fluoropore  filters in the inhalation exposure chambers.  The exposure system and exposure
atmosphere are described in Appendix A. Vehicle controls received injections of particle
suspension made up of particulate matter from control exposure filters, positive controls
received 20 mg of urethan, and negative controls received no injections.  Injections were made
three times weekly for 8 weeks, resulting in a total DPM dose of 1.1, 2.8, and 5.6 mg for the
low-, medium-, and high-dose groups and 20 mg of urethan for the positive control group.
These animals were sacrificed after 26 weeks and examined for lung tumors. For the low-,
medium-, and high-dose DPM groups, the tumor incidence was 2/30, 10/30, and 8/30,
respectively. The incidence among urethan-treated animals (positive controls) was 100%
(29/29), with multiple tumors per animal. The tumor incidence for the DPM-treated animals did
not differ significantly from that of vehicle  controls (8/30) or negative controls (7/28).  The
number of tumors per mouse was also unaffected by treatment.
       In further studies conducted by Orthoefer et al.  (1981), an attempt was made to compare
the potency of DPM with that of other environmental pollutants. Male and female Strain A mice
were injected i.p. three times weekly for 8 weeks with DPM, DPM extracts, or various
environmental mixtures of known carcinogenicity, including cigarette smoke condensate, coke
oven emissions, and roofing tar emissions.  Injection of urethan or dimethylsulfoxide (DMSO)
served as positive or vehicle controls, respectively.  In addition to DPM from the Nissan diesel
previously described, an eight-cylinder Oldsmobile engine operated at the equivalent of 40 mph
was also used to compare emission effects from different makes  and models of diesel engine.
The mice were sacrificed  at 9 months of age and their lungs examined for histopathological
changes. The only significant findings, other than for positive controls, were small increases in
numbers of lung adenomas per mouse in male mice injected with Nissan DPM and in female
mice injected with coke oven extract.  Furthermore, the increase in the extract-treated mice was
significant only in comparison with uninjected controls (not injected ones) and did not occur
when the experiment was repeated. Despite the use of a strain of mouse known to be sensitive to
tumor induction,  the overall findings of this study were  negative. The authors provided several
possible explanations for these findings, the most likely of which were (1) the carcinogens that
                                         7-115

-------
were present were very weak, or (2) the concentrations of the active components reaching the
lungs were insufficient to produce positive results.
       Kunitake et al.  (1986) conducted studies using DPM extract obtained from a 1983 HD
MMC—6D22P 11-L V-6 engine.  Five s.c. injections of DPM extract (500 mg/kg per injection)
resulted in a significant (p<0.01) increase in subcutaneous tumors for female C57BL mice (5/22
[22.7%] vs. 0/38 among controls).  Five s.c. doses of DPM extract of 10, 25, 30, 100, or 200
mg/kg  failed to produce a significant increase in tumor incidence. One of 12 female ICR mice
(8.3%) and 4 of 12 male ICR mice (33.3%) developed malignant lymphomas following neonatal
s.c. administration of 10 mg of DPM extract per mouse. The  increase in malignant lymphoma
incidence for the male  mice was statistically significant at/><0.05 compared with an incidence of
2/14(14.3%) among controls.  Treatment of either sex with 2.5  or 5 mg of DPM extract per
mouse did not result in statistically significant increases in tumor incidence.
       Additional studies using DPM extract from LD (1.8-L, 4-cylinder) as well  as FID engines
with female ICR and nude mice (BALB/c/cA/JCL-nu) were also reported (Kunitake et al.,
1988).  Groups of 30 ICR and nude mice each were given a single s.c. injection of 10 mg HD
extract, 10 mg HD + 50 |ig 12-O-tetradecanoylphorbol  13-acetate (TPA), 10 mg LD extract +
50 |lg TPA, or 50 |lg TPA.  No malignant tumors or papillomas were observed. One
papillomatous lesion was observed in an ICR mouse receiving LD extract + TPA,  and acanthosis
was observed in one nude mouse receiving only TPA.
       In what appears to be an extension of the Kunitake et al. (1986) s.c. injection studies,
Takemoto et al. (1988) presented additional data for subcutaneously administered  DPM extract
from HD and LD diesel engines. In this report, the extracts were administered to 5-week-old
and neonatal (<24 hr old) C57BL mice of both sexes. DPM extract from HD or LD engines was
administered weekly to the 5-week-old mice for 5 weeks at doses of 10, 25, 50, 100, 200, or 500
mg/kg, with group sizes ranging from  15 to 54 animals. After 20 weeks, comparison with a
control group indicated a significant increase in the incidence of subcutaneous tumors for the
500 mg/kg HD group (5 of 22 mice [22.7%],/X0.01), the 100 mg/kg LD group (6 of 32
[18.8%],
/X0.01), and the 500 mg/kg LD group (7 of 32 [21.9%],/K0.01) in the adult mouse
experiments. The tumors were characterized as malignant fibrous histiocytomas.  No tumors
were observed  in other organs.  The neonates were given single doses of 2.5, 5, or 10 mg DPM
extract subcutaneously within 24 hr of birth. There was a significantly higher incidence of
malignant lymphomas  in males receiving 10 mg of HD extract and of lung tumors for males
given 2.5 mg HD extract and for males given 5 mg and females given 10 mg LD extract. A
dose-related trend that  was not significant was observed for the  incidences of liver tumors for
both the HD extract- and LD extract-treated neonatal mice. The incidence of mammary tumors
in female mice and multiple-organ tumors in male mice was also greater for some extract-treated
                                        7-116

-------
mice, but was not dose related.  The report concluded that LD DPM extract showed greater
carcinogenicity than did HD DPM extract.

7.3.6. Dermal Studies
7.3.6.1. Mouse Studies
       In one of the earliest studies of diesel emissions, the effects of dermal application of
extract from DPM were examined by Kotin et al. (1955).  Acetone extracts were prepared from
the DPM of a diesel engine (type and size not provided) operated at warmup mode and under
load. These extracts were applied dermally three times weekly to male and female C57BL and
strain A mice.  Results of these experiments are summarized in Table 7-5.  In the initial
experiments using 52 (12 male,  40 female) C57BL mice treated with DPM extract from an
engine operated in warmup mode, two papillomas were detected after 13 mo. Four tumors were
detected 16 months after the start of treatment in 8  surviving of 50 exposed male strain A mice
treated with DPM extract from an engine operated under full load. Among female strain A mice
treated with DPM extract from an engine operated under full load, 17 tumors were detected in
20 of 25 mice surviving longer than 13 mo. This provided a significantly increased tumor
incidence of 85%. Carcinomas as well as papillomas were seen,  but the numbers were not
reported.
       Depass et al. (1982) examined the potential of DPM and dichloromethane extracts of
DPM to act as complete carcinogens,  carcinogen initiators, or carcinogen promoters.  In skin-
painting studies, the DPM was obtained from an Oldsmobile 5.7-L diesel engine operated under
constant load at 65 km/h.  The DPM was collected  at a temperature of 100°C.  Groups of 40
C3H/HeJ mice were used because of their low spontaneous tumor incidence.  For the complete
carcinogenesis experiments, DPM was applied as a 5% or 10% suspension  in acetone.
Dichloromethane extract was applied as 5%, 10%, 25%, or 50%  suspensions. Negative controls
received acetone, and positive controls received 0.2% B[a]P.  For tumor-promotion experiments,
a single application of 1.5% B[a]P was followed by repeated applications of 10% DPM
suspension, 50% DPM extract, acetone only (vehicle control), 0.0001% phorbol 12-myristate
13-acetate (PMA) as a positive promoter control, or no treatment (negative control).  For the
tumor-initiation studies, a single initiating dose of 10% diesel particle suspension, 50% diesel
particle extract, acetone, or PMA was followed by repeated applications of 0.0001% PMA.
Following 8 months of treatment, the PMA dose in the initiation and promotion studies was
increased to 0.01%.  Animals were treated three times per week in the complete carcinogenesis
and initiation experiments and five times per week in promotion  experiments.  All test
compounds were applied to a shaved  area on the back of the mouse.
       In the complete carcinogenesis experiments, one mouse receiving the high-dose (50%)
suspension of extract developed a squamous cell carcinoma after 714 days of treatment.  Tumor
                                        7-117

-------
incidence in the B[a]P group was 100%, and no tumors were observed in any of the other
groups.  For the promotion studies, squamous cell carcinomas with pulmonary metastases were
identified in one mouse of the 50% DPM extract group and in one in the 25% extract group.
Another mouse in the 25% extract group developed a grossly diagnosed papilloma. Nineteen
positive control mice had tumors (11 papillomas, 8 carcinomas). No tumors were observed for
any of the other treatment groups. For the initiation studies, three tumors (two papillomas and
one carcinoma) were identified in the group receiving DPM suspension and three tumors (two
papillomas and one fibrosarcoma) were found in the DPM extract group. These findings were
reported to be statistically insignificant using the Breslow and Mantel-Cox tests.
       Although these findings were not consistent with those of Kotin et al. (1955), the
occurrence of a single carcinoma in a strain known to have an extremely low spontaneous tumor
incidence may be of importance.  Furthermore, a comparison between studies employing
different strains of mice with varying spontaneous tumor incidences may result in erroneous
assumptions.
      Nesnow et al. (1982) studied the formation of dermal papillomas and carcinomas
following dermal application of dichloromethane extracts from coke oven emissions, roofing tar,
DPM, and gasoline engine exhaust. DPM from five different engines, including a preproduction
Nissan 220C, a 5.7-L Oldsmobile, a prototype Volkswagen Turbo Rabbit, a Mercedes 300D, and
a HD Caterpillar 3304, was used for various phases of the study. Male and female Sencar mice
(40 per group) were used for tumor initiation, tumor promotion, and complete carcinogenesis
studies.  For the tumor-initiation experiments, the DPM extracts were topically  applied in single
doses of 100, 500, 1,000, or 2,000 |ig/mouse. The high dose (10,000 |ig/mouse) was applied in
five daily doses of 2,000 |ig. One week later, 2 |ig of the tumor promoter TPA was applied
topically twice weekly. The tumor-promotion experiments used mice treated with 50.5 |lg of
B[a]P followed by weekly (twice weekly for high dose) topical applications (at the
aforementioned doses) of the extracts. For the complete carcinogenesis experiments, the test
extracts were applied weekly (twice weekly for the high doses) for  50 to 52 weeks. Only
extracts from the Nissan, Oldsmobile, and Caterpillar engines were used in the complete
carcinogenesis experiments.
      In the tumor-initiation studies, both B[a]P alone and the Nissan engine DPM extract
followed by TPA treatment produced a significant increase in tumor (dermal papillomas)
incidence at 7 to 8 weeks postapplication.  By 15 weeks, the tumor  incidence was greater than
90% for both groups. No  significant carcinoma formation was noted for mice in the tumor-
initiation experiments following exposure to DPM extracts of the other diesel engines, although
the Oldsmobile engine DPM extract at 2.0 mg/mouse did produce a 40% papilloma incidence in
male mice at 6 mo.  This effect, however, was not dose dependent.
                                        7-118

-------
       B[a]P (50.5 |ig/week), coke oven extract (at 1.0, 2.0, or 4.0 mg/week), and the highest
dose of roofing tar extract (4.0 mg/week) all tested positive for complete carcinogenesis activity.
DPM extracts from only the Nissan, Oldsmobile, and Caterpillar engines were tested for
complete carcinogenic potential, and all three proved to be negative using the Sencar mouse
assay.
       The results of the dermal application experiments by Nesnow et al. (1982) are presented
in Table 7-7.  The tumor initiation-promotion assay was considered positive if a dose-dependent
response was obtained and if at least two doses provided a papilloma-per-mouse value that was
three times or greater than that of the background value. Based on these criteria, only emissions
from the Nissan were considered positive.  Tumor initiation and complete carcinogenesis assays
required that at least one dose produce a tumor incidence of at least 20%. None of the DPM
samples yielded positive results based on this criterion.
       Kunitake et al.  (1986, 1988) evaluated the effects of a dichloromethane extract of DPM
obtained from a 1983 MMC M-6D22P  11-L V-6 engine.  An  acetone solution was applied in 10
doses every other day,  followed by promotion with 2.5 |lg of TPA three times weekly for 25
weeks. Exposure groups received a total dose of 0.5, 5, 15, or 45 mg of extract. Papillomas
were reported in 2 of 50 animals examined in the 45 mg exposure group and in 1 of 48 in the 15
mg group compared with 0 of 50 among controls. Differences, however, were not statistically
significant.

7.3.7. Summary and  Conclusions of Laboratory Animal Carcinogenicity Studies
       As early as 1955, Kotin et al.  (1955) provided evidence for tumorigenicity and
carcinogenicity of acetone extracts of DPM following dermal application and also provided data
suggesting a difference in this potential depending on engine operating mode.  Until the early
1980s, no chronic studies assessing inhalation of DE, the relevant mode for human exposure,
had been reported. Since then long-term inhalation bioassays with DE have been carried out in
the United States, Germany, Switzerland, and Japan, testing responses of rats,  mice, and Syrian
hamsters, and to a limited extent cats and monkeys.
                                         7-119

-------
               Table 7-7. Dermal tumorigenic and carcinogenic effects of various emission extracts
to
o
Tumor initiation
Sample Papillomasa Carcinomas'5
Benzo[a]pyrene +/+c +/+
Topside coke oven +/+ -/+
Coke oven main +/+ +/+
Roofing tar +/+ +/+
Nissan +/+ +/+
Oldsmobile +/+ -/-
VW Rabbit +/+ -/-
Mercedes +/- -/-
Caterpillar -/- -/-
Residential furnace -/- -/-
Mustang +/+ -/+
Complete
carcinogenesis
Carcinomas'5
+/+
NDd
+/+
+/+
-/-
-/-
F
ND
-/-
ND
ND
Tumor promotion
Papillomas"
+/+
ND
+/+
+/+
ND
ND
ND
ND
ND
ND
ND
          aScored at 6 mo.
          bCumulative score at 1 year.
          cMale/female.
          dND = Not determined.
          el = Incomplete.
          Source:  Nesnow et al., 1982.

-------
       It can be reasonably concluded that with adequate exposure, inhalation of DE is capable
of inducing lung cancer in rats. Responses best fit cumulative exposure (concentration x daily
exposure duration x days of exposure). Examination of rat data shown in Table 7-8 indicates a
trend of increasing tumor incidence at exposures exceeding 1 x 104 mg'hr/m3. Exposures greater
than approximately this value result in lung particle overload, characterized by slowed particle
clearance and lung pathology, as discussed in Chapters 3 and 5, respectively.  Tumor induction
at high doses may therefore be primarily the result of lung particle overload with associated
inflammatory responses.  Although tumorigenic responses could not be detected under non-
particle-overload conditions, the animal experiments lack sensitivity to determine if a threshold
exists.  However, studies such as those reported by Driscoll et al. (1996) support the existence of
a threshold if it is assumed that inflammation is a prerequisite for lung tumor induction. If low-
dose effects  do occur, it can be hypothesized that the organic constituents are playing a role. See
Chapter 7 for a discussion of this issue.
       Although rats develop adenomas, adenocarcinomas, and adenosquamous cell carcinomas,
they also develop squamous keratinizing lesions. This latter spectrum appears for the most part
to be peculiar to the rat.  In a recent workshop aimed at classifying these tumors (Boorman et al.,
1996), it was concluded that when these lesions occur in rats as part of a carcinogenicity study,
they must be evaluated on a case-by-case basis and regarded as a part of the total biologic profile
of the test article. If the  only evidence of tumorigenicity is the presence of cystic keratinizing
epitheliomas, it may not have relevance to human safety evaluation of a substance or particle.
Their use in  quantifying  cancer potency is even more questionable.
       The evidence for response of common strains of laboratory mice exposed under standard
inhalation protocols is equivocal. Inhalation of DE induced significant increases in lung tumors
in female NMRI mice (Heinrich et al., 1986b; Stober, 1986) and in female Sencar mice (Pepelko
and Peirano, 1983). An apparent increase was also seen in female C57BL mice (Takemoto et
al.,  1986). However, in a repeat of their earlier study, Heinrich et al. (1995) failed to detect lung
tumor induction in either NMRI or C57BL/6N mice.  No increases in lung tumor rates were
reported in a series of inhalation studies using strain A mice (Orthoefer et al., 1981;  Kaplan et
al.,  1982, 1983; White et al., 1983). Finally, Mauderly et al. (1996) reported no tumorigenic
responses in CD-I  mice exposed under conditions resulting in positive responses in rats. The
successful induction of lung tumors in mice  by Ichinose et al. (1997a,b) via intratracheal
instillation may have been the result of focal deposition of larger doses.  Positive effects in
Sencar mice may be due to use of a strain sensitive to tumor induction in epidermal tissue by
organic agents, as well as exposure from conception, although proof for such a hypothesis is
lacking.
                                         7-121

-------
             Table 7-8. Cumulative (concentration x time) exposure data for rats exposed to whole DE
to
to
Study
Mauderlyetal. (1987)


Nikulaetal. (1995)

Heinrich et al. (1986a)
Heinrichetal. (1995)


Ishinishi et al. (1988a)
(Light-duty engine)

(Heavy-duty engine)




Exposure
rate/duration Total exposure
(hr/week, mo) time (hr)
35, 30 4.20042004e+15
35,30
35,30
35,30
80, 23 736073607360
80,23
80,23
95,35 1330013300
95,35
90,24 8.64086409e+15
90,24
90,24
90,24
96,30 1.15201152e+49
96,30
96,30
96,30
96,30
96,30
96,30
96,30
96,30
96,30
Particle
concentration
(mg/m3)
0
0.35
3.5
7.1
0
2.5
6.5
4.24
0
0.8
2.5
7.0
0
0.1
0.4
1.1
2.3
0
0.5
1.0
1.8
3.7
Cumulative exposure
(mg-hr/m3)
Per week Total
0 14701470029820
12.25
122.5
248.5
0 1840047840
200.0
520.0
402.8 56392
0 74002180061700
72.0
225.0
630.0
0 1.1524
9.6 60813e+37
38.4
105.6
220.8
0
48.0
96.0
172.8
355.2
Tumor incidence
(%)"
0.9
1.3
3.6
12.8
1.0
7.0
18.0
17.8
0
0
5.5
22.0
3.3
2.4
0.8
4.1
2.4
0.8
0.8
0
3.3
6.5

-------
to
             Table 7-8.  Cumulative (concentration x time) exposure data for rats exposed to whole DE (continued)



                                                                                 Cumulative exposure
Study
Brightwell et al. (1989)


Kaplan etal. (1983)


Iwaietal. (1986b)
Takemoto et al. (1986)
Karagianes et al.
(1981)
Iwaietal. (1997)

iJjAIUJSltl V-
rate/duration
(hr/week, mo)
80,24
80,24
80,24
80,24
140, 15
140, 15
140, 15
140, 15
56,24
56,24
16, 18-24
16, 18-24
30,20
30,20
56,24
48,24
54,24
Total exposure
time (hr)
7.6807681e+15


8.4008401e+15


53765376
1,152-1,536
1,152-1,536
24002400
537649925616

_•_ til IL\,L\, —
concentration
(mg/m3)
0
0.7
2.2
6.6
0
0.25
0.75
1.5
4.9
0
2-4
8.3
9.4
3.2
5.1
Per week
0
56.0
176.0
528.0
0
35.0
105.0
210.0
274.4
0
32-64
249
526154275

Tumor incidence
Total (%)a
53761689650688 1.2
0.7
9.7
38.5
2100630012600 0
3.3
10.0
3.3
26342 36.8
0 0
3,456-4,608
19920 16.6
5.47041597e+14 421242


-------
       Attempts to induce significant increases in lung tumors in Syrian hamsters by inhalation
of whole DE were unsuccessful (Heinrich et al., 1982, 1986b, 1989b; Brightwell et al., 1986).
However, hamsters are considered to be relatively insensitive to lung tumor induction. For
example, while cigarette smoke, a known human carcinogen, was shown to induce laryngeal
cancer in hamsters, the lungs were relatively unaffected (Dontenwill et al., 1973).
       Neither cats (Pepelko and Peirano, 1983 [see Chapter 7]) nor monkeys (Lewis et al.,
1989) developed tumors following 2-year exposure to DE. The duration of these exposures,
however, was likely to be inadequate for these two longer-lived species, and group sizes were
quite small. Exposure levels were also below the maximum tolerated dose (MTD) in the
monkey studies and, in fact, only borderline for detection of lung tumor increases in rats.
       Long-term exposure to DE filtered to remove particulate matter failed to induce lung
tumors in rats (Heinrich  et al., 1986b; Iwai et al., 1986b; Brightwell et al., 1989), or in Syrian
hamsters (Heinrich et al., 1986b; Brightwell, 1989).  A significant increase in lung carcinomas
was reported by Heinrich et al. (1986b) in NMRI mice exposed to filtered exhaust.  However, in
a more recent study the authors were unable to confirm earlier results in either NMRI or
C57BL/6N mice (Heinrich et al.,  1995). Although filtered exhaust appeared to potentiate the
carcinogenic effects of DEN (Heinrich et al., 1982), because of the lack of positive data in rats
and equivocal or negative data in mice it can be concluded that filtered exhaust is either not
carcinogenic or has a low cancer potency.
       Kawabata et al. (1986) demonstrated the induction of lung tumors in Fischer 344 rats
following intratracheal instillation of DPM. Rittinghausen et al. (1997) reported an increase in
cystic keratinizing epitheliomas following intratracheal instillation of rats with either original
DPM or DPM extracted to remove the organic fraction, with the unextracted particles inducing a
slightly greater effect.  Grimmer et al. (1987) showed not only that an extract of DPM was
carcinogenic when instilled in the lungs of rats, but also that most of the carcinogenicity resided
in the portion containing PAHs with four to seven rings.  Intratracheal instillation did not induce
lung tumors in Syrian  hamsters (Kunitake et al., 1986; Ishinishi et al., 1988b).
       Dermal exposure and s.c. injection in mice provided additional evidence for tumorigenic
effects of DPM.  Particle extracts applied dermally to mice have been shown to induce
significant skin tumor  increases in two studies (Kotin et al., 1955; Nesnow et al., 1982).
Kunitake et al. (1986)  also reported a marginally significant increase in skin papillomas in ICR
mice treated with an organic extract from an HD  diesel engine.  Negative results were reported
by Depass et al. (1982) for skin-painting studies using mice and acetone extracts of DPM
suspensions. However, in this study the exhaust particles were collected at temperatures of 100
°C, which would minimize the condensation of vapor-phase organics and, therefore, reduce the
availability of potentially carcinogenic compounds that might normally be present on DE
particles. A significant increase in the incidence  of sarcomas in female C57B1 mice was
reported by Kunitake et al. (1986) following s.c. administration of LD DPM extract at doses of

                                         7-124

-------
500 mg/kg.  Takemoto et al. (1988) provided additional data for this study and reported an
increased tumor incidence in the mice following injection of LD engine DPM extract at doses of
100 and 500 mg/kg. Results of i.p. injection of DPM or DPM extracts in strain A mice were
generally negative (Orthoefer et al., 1981; Pepelko and Peirano, 1983), suggesting that the strain
A mouse may not be a good model for testing diesel emissions.
       Results of experiments using tumor initiators such as DEN, B[a]P, DPN, or DBA
(Brightwell et al., 1986; Heinrich et al.,  1986b;  Takemoto et al., 1986) were generally
inconclusive regarding the tumor-promoting potential  of either filtered or whole DE. A report
by Heinrich et al. (1982), however, indicated that filtered exhaust may promote the tumor-
initiating effects of DEN in hamsters.
       Several reports (Wong et al., 1985; Bond et al., 1990) affirm observations of the
potential carcinogenicity of DE by providing evidence for DNA damage in rats.  These findings
are discussed in more detail in Chapter 3,  Section 3.6.  Evidence for the mutagenicity of organic
agents present in diesel engine emissions is also provided in Chapter 4.
       Evidence for the importance of the carbon core was initially provided by studies of
Kawabata et al.  (1986), which showed induction of lung tumors following intratracheal
instillation of carbon black that contained no more than traces of organics, and studies of
Heinrich (1990b) that indicated that exposure via inhalation to carbon black (Printex 90)
particles induced lung tumors at concentrations  similar to those effective in DPM studies.
Additional studies by Heinrich et al. (1995) and Nikula et al. (1995) confirmed the capability of
carbon particles to induce lung tumors.  Induction of lung tumors by other particles of low
solubility, such  as titanium dioxide (Lee et al., 1986), confirmed the capability of particles to
induce lung tumors. Pyrolyzed pitch, on the other hand, essentially lacking a carbon core but
having much higher PAH concentrations than DPM, also was effective in tumor induction
(Heinrich  et al., 1986a, 1994).
       The relative importance of the adsorbed organics, however, remains to be elucidated and
is of some concern because of the known carcinogenic capacity of some of these chemicals.
These include polycyclic aromatics as well as nitroaromatics, as described in Chapter 2. Organic
extracts of particles also have been shown to induce tumors in a variety of injection, intratracheal
instillation, and skin-painting studies, and Grimmer et al. (1987) have, in fact, shown that the
great majority of the carcinogenic potential following instillation resided  in the fraction
containing four- to seven-ring PAHs.
       In  summary, based on positive inhalation studies in rats exposed to high concentrations,
intratracheal instillation studies in rats and mice exposed to high doses, and supported by
positive mutagenicity studies, the evidence for carcinogenicity of DE is considered to be
adequate in animals.  The contribution of the various fractions of DE to the carcinogenic
response is less  certain. Exposure to filtered exhaust generally failed to induce lung tumors.
The presence of known carcinogens adsorbed to diesel particles and the demonstrated

                                          7-125

-------
tumorigenicity of particle extracts in a variety of injection, instillation, and skin-painting studies
indicate a carcinogenic potential for the organic fraction.  Studies showing that long-term
exposure at high concentrations of poorly soluble particles (e.g., carbon black, TiO2) can also
induce tumors, on the other hand, have provided definitive evidence that the carbon core of the
diesel particle is primarily instrumental in the carcinogenic response observed in rats under
sufficient exposure conditions. The ability of DE to induce lung tumors at non-particle-overload
conditions, and the relative contribution of the particles' core versus the particle-associated
organics (if effects do occur at low  doses) remains to be determined.

7.4.  MODE OF ACTION OF DIESEL EXHAUST-INDUCED CARCINOGENESIS
       As noted in Chapter 2, DE is a complex mixture that includes a vapor phase and a
particle phase.  The particle phase consists of an insoluble carbon core with a large number of
organic compounds, as well as inorganic compounds such as sulfates, adsorbed to the particle
surface. Some of the semivolatile and particle-associated compounds, in particular PAHs, nitro-
PAHs, oxy-PAHs, and  oxy-nitro-PAHs (Scheepers and Bos, 1992), are considered likely to be
carcinogenic in humans. The vapor phase also contains a large number of organic compounds,
including several known or probable carcinogens  such as benzene and 1,3-butadiene. Because
exposure to the vapor phase alone, even at high concentrations, failed to induce lung cancer in
laboratory animals (Heinrich et al.,  1986b), the mode-of-action discussion will focus on the
particulate matter phase. Additive or synergistic effects of vapor-phase components, however,
cannot be ruled out, as  chronic inhalation bioassays involving exposure to diesel particles alone
have not been carried out.
       Several hypotheses regarding the primary  mode of action of DE have been proposed.
Initially it was generally believed that cancer was induced by particle-associated organics acting
via a genotoxic mechanism. By the late 1980s, however,  studies indicated that carbon particles
virtually devoid of organics could also induce lung cancer at sufficient inhaled concentrations
(Heinrich, 1990b).  This finding provided support for a hypothesis originally proposed by Vostal
(1986) that induction of lung tumors arising in rats exposed to high concentrations of DE is
related to overloading of normal lung clearance mechanisms, accumulation of particles, and cell
damage followed by regenerative cell proliferation.  The action of particles is therefore mediated
by epigenetic mechanisms that can be characterized more by promotional than initiation stages
of the carcinogenic process. More recently several studies have focused upon the production of
reactive oxygen species generated from particle-associated organics, which may induce oxidative
DNA damage at exposure concentrations lower than those required to produce lung particle
overload. Because it is likely that more than one of these factors is involved in the carcinogenic
process, a key consideration is their likely relative contribution at different exposure levels. The
following discussion  will therefore  consider the possible relationship of the organic components
of exhaust, inflammatory responses associated with lung particle overload, reactive oxygen

                                          7-126

-------
species, and physical characteristics of diesel particles to cancer induction, followed by a
hypothesized mode of action, taking into account the likely contribution of the factors discussed.

7.4.1.  Potential Role of Organic Exhaust Components in Lung Cancer Induction
       More than 100 carcinogenic or potentially carcinogenic components have been
specifically identified in diesel emissions, including various PAHs and nitroarenes such as
1-nitropyrene (1-NP) and dinitropyrenes (DNPs). The majority of these compounds are
adsorbed to the carbon core of the particulate phase of the exhaust and, if desorbed, may become
available for biological processes such as metabolic activation to mutagens.  Among such
compounds identified from DE are B[a]P, dibenz[a,h]anthracene, pyrene, chrysene, and
nitroarenes such as 1-NP, 1,3-DNP, 1,6-DNP, and 1,8-DNP, all of which are mutagenic,
carcinogenic, or implicated as procarcinogens or cocarcinogens (Stenback et al., 1976;
Weinstein and Troll, 1977; Thyssen et al., 1981; Pott and Stober, 1983; Howard et al.,  1983;
Hirose et al., 1984; Nesnow et al., 1984; El-Bayoumy et al., 1988). More recently Enya et al.
(1997) reported isolation of 3-nitrobenzanthrone, one of the most powerful direct-acting
mutagens known to date, from the organic extracts of DE.
       Grimmer et al. (1987) separated DE particle extract into a water- and a lipid-soluble
fraction, and the latter was further separated into a PAH-free, a PAH-containing, and a polar
fraction by column chromatography.  These fractions were then tested in Osborne-Mendel rats
by pulmonary implantation at doses corresponding to the composition of the original DE. The
water-soluble fraction did not induce tumors; the incidences induced by the lipid-soluble
fractions were  0% with the PAH-free fraction, 25% with the PAH and nitro-PAH- containing
fractions, and 0% with the polar fraction. The PAH and nitro-PAH-containing fraction,
comprising only  1% by weight of the total extract, was thus shown to be responsible for most, if
not all, of the carcinogenic activity.
       Exposure of rats by inhalation to 2.6 mg/m3 of an aerosol of tar-pitch condensate with no
carbon core but containing 50 |lg/m3 B[a]P along with other PAHs for 10 months induced lung
tumors in 39% of the animals. The same amount of tar-pitch vapor condensed onto the surface
of carbon black particles at 2 and 6 mg/m3 resulted in tumor rates that were roughly two times
higher (89% and 72%).  Because exposure to 6 mg/m3 carbon black almost devoid of extractable
organic material induced a lung tumor rate of 18%, the combination of PAHs and particles
increases their effectiveness (Heinrich et al., 1994). Although this study shows the tumor-
inducing capability of PAHs resulting from combustion, it should be noted that the B[a]P
content in the coal-tar pitch was about three orders of magnitude greater than in diesel soot.
Moreover, because organics are present on diesel particles in a thinner layer and the particles are
quite convoluted, they may be more tightly bound and less bioavailable. Nevertheless, these
studies provide evidence supporting the involvement of organic constituents of diesel particles in
the carcinogenic process.

                                         7-127

-------
       Exposure of humans to related combustion emissions provides some evidence for the
involvement of organic components. Mumford et al. (1989) reported greatly increased human
lung cancer mortality in Chinese communes burning so-called smoky coal, but not wood, in
unvented open-pit fires used for heating and cooking. Although particle concentrations were
similar, PAH levels were five to six times greater in the air of communes burning smoky coal.
Coke oven emissions, containing high concentrations of PAHs but lacking an insoluble carbon
core, have also been shown to be carcinogenic in humans (Lloyd, 1971).
       Adsorption of PAHs to a carrier particle such as hematite, CB, aluminum, or titanium
dioxide enhances their carcinogenic potency (Farrell and Davis, 1974).  As already noted,
adsorption to carbon particles greatly enhanced the tumorigenicity of pyrolyzed pitch condensate
containing B[a]P and other aromatic carcinogens (Heinrich et al., 1995). The increased
effectiveness can be partly explained by more efficient transport to the deep lung. Slow  release
also enhances residence time in the lungs and prevents  overwhelming of activating pathways.
As discussed in Chapter 3, free organics are likely to be rapidly absorbed into the bloodstream,
which may explain why the vapor-phase component of exhaust is relatively ineffective in the
induction of pathologic or carcinogenic effects.
       Even though the organic constituents may be tightly bound to the particle surface,
significant elution is still likely because particle clearance half-times are nearly  1 year in humans
(Bohning et al., 1982).  Furthermore, Gerde et al. (1991) presented a model demonstrating that
large aggregates of inert dust containing crystalline PAHs are unlikely to form at doses typical of
human exposure. This allows the particles to deposit and react with the surrounding lung
medium, without interference from other particles.  Particle-associated PAHs can then be
expected to be released more rapidly from the particles. Bond et al. (1984) provided evidence
that alveolar macrophages from beagle dogs metabolized B[a]P coated on diesel particles to
proximate carcinogenic forms.  Unless present on the particle surface, B[a]P is more likely to
pass directly into the bloodstream and escape activation by phagocytic cells.
       The importance of DE-associated PAHs in the induction of lung cancer in humans may
be enhanced because of the possibility that the human lung is  more sensitive to these compounds
than are rat lungs.  Rosenkranz (1996) summarized information indicating that in humans and
mice, large proportions of lung cancers contain both mutated p53 suppressor genes and K-ras
genes.  Induction of mutations in these genes by genotoxins, however, is much lower in rats than
in humans or mice.
       B[a]P, although only one of many PAHs present in DE, is the one  most extensively
studied. Bond et al. (1983, 1984) demonstrated metabolism of particle-associated B[a]P and
free B[a]P by alveolar macrophages (AM)  and by type II alveolar cells. The respiratory tract
cytochrome P-450  systems have an even greater concentration in the nonciliated bronchiolar
cells (Boyd,  1984). It is worth noting that bronchiolar  adenomas that develop following diesel
exposure have been found to resemble both Type II and nonciliated bronchiolar cells.  It should

                                         7-128

-------
also be noted that any metabolism of procarcinogens by these latter two cell types probably
involves the preextraction of carcinogens in the extracellular lining fluid and/or other
endocytotic cells, as they are not especially important in phagocytosis of particles.  Thus,
bioavailability is an important issue in assessing the relative importance of PAHs.
       Additionally,  a report by Borm et al. (1997) indicates that incubating rat lung epithelial-
derived cells with human polymorphonucleocytes (PMNs) (either unactivated or  activated by
preexposure to phorbol myristate  acetate) increases DNA adduct formation caused by exposure
to B[a]P; at 0.05 to 0.5 micromolar concentration, addition of more activated PMN in relation to
the number of lung cells further increased adduct formation in a dose-dependent manner. The
authors suggest that "an inflammatory response in the lung may increase the biologically
effective dose of PAHs, and may  be relevant to data interpretation and risk assessment of PAH-
containing particles."  These data raise the possibility that DE exposure at low concentrations
may result in levels of neutrophil  influx that would not necessarily be detectable via
histopathological examination as  acute inflammation, but that might be effective  at amplifying
any potential DE genotoxic effect.
       Nitro-PAHs have also been implicated as potentially involved in diesel-exhaust-induced
lung cancer. Although the nitro-PAH fraction of diesel was less effective than PAHs in the
induction of lung cancer when implanted into the lungs of rats (Grimmer et al., 1987), in a study
of various extracts of DE particles, 30%-40% of the total  mutagenicity could be attributed to a
group of six nitroarenes (Salmeen et al., 1984).  Moreover, Gallagher et al. (1994) reported
results suggesting that DNA adducts are formed from nitro-PAHs present in DNA and may play
a role in the carcinogenic process. Nitroarenes, however, quantitatively represent a very small
percentage of diesel particle extract (Grimmer et al., 1987), making their role in the tumorigenic
response uncertain.
       The induction of DNA adducts in humans occupationally exposed to DE indicates the
likelihood that PAHs are participating in the tumorigenic  response, and that these effects can
occur at exposure levels less than those required to induce lung particle overload.  Distinct
adduct patterns were  found among garage workers occupationally exposed to DE when
compared with nonexposed controls (Nielsen and Autrup, 1994).  Furthermore, the findings
were concordant with the adduct patterns observed in groups exposed to low concentrations of
PAHs from combustion processes.  Hemminki et al. (1994) also reported significantly elevated
levels of DNA adducts in lymphocytes from garage workers with known DE exposure compared
with unexposed mechanics. Hou  et al. (1995) found elevated adduct levels in bus maintenance
workers exposed to DE. Although no difference in mutant frequency was observed between the
groups, the adduct levels were significantly different (3.2 vs. 2.3 x 10"8).  Nielsen et al. (1996b)
measured three biomarkers in DE-exposed bus garage workers: lymphocyte DNA adducts,
hydroxyethylvaline adducts in hemoglobin, and 1-hydroxypyrene in urine.  Significantly
increased levels were reported for all three. Qu et al. (1996)  detected increased adduct levels,  as

                                         7-129

-------
well as increases in some individual adducts, in the blood of underground coal miners exposed to
DE.

7.4.2.  Role of Inflammatory Cytokines and Proteolytic Enzymes in the Induction of Lung
       Cancer in Rats by Diesel Exhaust
       It is well recognized that the deposition of particles in the lung can result in the efflux of
PMNs from the vascular compartment into the alveolar space compartment in addition to
expanding the AM population size.  Following acute exposures, the influx of the PMNs is
transient, lasting only a few days (Adamson and Bowden, 1978; Bowden and Adamson, 1978;
Lehnert et al., 1988). During chronic exposure the numbers of PMNs lavaged from the lungs of
diesel-exposed rats generally increased with increasing exposure duration and inhaled DPM
concentration (Strom, 1984). Strom (1984) also found that PMNs in diesel-exposed lungs
remained persistently elevated for at least 4 months after cessation of exposure, a potential
mechanism that may be related to an ongoing release of  phagocytized particles. Evidence in
support of this possibility was reported by Lehnert  et al. (1989) in a study in which rats were
intratracheally instilled with 0.85, 1.06, or 3.6 mg of polystyrene particles. The PMNs were not
found to be abnormally abundant during the clearance  of the two lower lung burdens, but they
became progressively elevated in the lungs of the animals in which alveolar-phase clearance was
inhibited.  Moreover, the particle burdens in the PMNs became progressively greater over time.
Such findings are consistent with an ongoing particle relapse process, in which particles released
by dying phagocytes are ingested by new ones.
       The inflammatory response, characterized by efflux of PMNs from the vascular
compartment, is mediated by inflammatory chemokines.  Driscoll et al. (1996) reported that
inhalation of high concentrations of carbon black stimulated the release of macrophage
inflammatory protein 2 (MIP-2) and monocyte chemotactic protein 1 (MCP-1).  They also
reported a concomitant increase  in hprt mutants. In a following study it was shown that particle
exposure stimulates production of tumor necrosis factor TNF-a, an agent capable of activating
expression of several proteins that promote both adhesion of leucocytes and chemotaxis (Driscoll
et al., 1997a). In addition,  alveolar macrophages also have the ability to release several other
effector molecules or cytokines that can regulate numerous functions of other lung cells,
including their rates of proliferation (Bitterman et al., 1983; Jordana et al., 1988; Driscoll et al.,
1996).
       Another characteristic of AMs and PMNs under particle overload conditions is the
release of a variety of potentially destructive hydrolytic enzymes, a process known to occur
simultaneously with the phagocytosis of particles (Sandusky et al., 1977). The essentially
continual release of such enzymes during chronic particle deposition and phagocytosis in the
lung may be detrimental to the alveolar epithelium, especially to Type I cells. Evans et al.
(1986) showed that injury to Type I cells is followed shortly thereafter by a proliferation of Type

                                          7-130

-------
II cells. Type II cell hyperplasia is a common feature observed in animals that have received
high lung burdens of various types of particles, including unreactive polystyrene microspheres.
Exaggerated proliferation as a repair or defensive response to DPM deposition may have the
effect of amplifying the likelihood of neoplastic transformation in the presence of carcinogens
beyond that which would normally occur with lower rates of proliferation, assuming an increase
in the cycling of target cells and the probability of a neoplastic-associated genomic disturbance.

7.4.3. Role of Reactive Oxygen Species in Lung Cancer Induction by Diesel Exhaust
       Phagocytes from a variety of rodent species produce elevated levels of oxidant reactants
in response to challenges, with the physiochemical characteristics of a phagocytized particle
being a major factor in determining the magnitude of the oxidant-producing response. Active
oxygen species released by the macrophages and lymphatic cells can cause lipid peroxidation in
the membrane of lung epithelial cells. These lipid peroxidation products can initiate a cascade of
oxygen free radicals that progress through the cell to the nucleus, where they damage DNA. If
this damage occurs during the epithelial cell's period of DNA synthesis, there is some
probability that the DNA will be replicated unrepaired (Lechner and Mauderly, 1994).  The
generation of reactive oxygen species by both AMs and PMNs should therefore be considered as
one potential factor of what probably is a multistep process that culminates in the development
of lung tumors in response to chronic deposition of DPM.
       Even though products of phagocytic oxidative metabolism, including superoxide anions,
hydrogen peroxide, and hydroxyl radicals, can kill tumor cells (Klebanoff and Clark,  1978), and
the reactive oxygen species can peroxidize lipids to produce cytotoxic metabolites such as
malonyldialdehyde, some products of oxidative metabolism apparently can also interact with
DNA to produce mutations.  Cellular DNA is damaged by oxygen free radicals generated from a
variety of sources (Ames, 1983; Trotter, 1980).  Along this line, Weitzman and Stossel  (1981)
found that human peripheral leukocytes are mutagenic in the Ames assay.  This mutagenic
activity was related to PMNs and blood monocytes;  blood lymphocytes alone were not
mutagenic. These investigators speculated that the mutagenic activity of the phagocytes was a
result of their ability to produce reactive oxygen metabolites, inasmuch as blood leukocytes from
a patient with chronic granulomatous diseases, in which neutrophils have  a defect in the
NADPH oxidase generating system (Klebanoff and Clark,  1978), were less effective in
producing mutations than were normal leukocytes.  Of related significance, Phillips et al. (1984)
demonstrated that the incubation of Chinese hamster ovary cells with xanthine plus xanthine
oxidase (a system for enzymatically generating active oxygen species) resulted in genetic
damage hallmarked by extensive chromosomal breakage and sister chromatid exchange and
produced an increase in the frequency of thioguanidine-resistant cells (HGPRT test).  Aside from
interactions of oxygen species with DNA, increasing evidence also points to an important role of
                                         7-131

-------
phagocyte-derived oxidants and/or oxidant products in the metabolic activation of
procarcinogens to their ultimate carcinogenic form (Kensler et al., 1987).
       Driscoll et al. (1997b) have demonstrated that exposure to doses of particles producing
significant neutrophilic inflammation are associated with increased mutation in rat alveolar type
II cells. The ability  of particle-elicited macrophages and neutrophils to exert a mutagenic effect
on epithelial cells in vitro supports a role for these inflammatory cells for the in vivo mutagenic
effects of particle exposure.  The inhibition of bronchoalveolar lavage cell-induced mutations by
catalase implies a role for cell-derived oxidants in this response.
       Hatch and co-workers (1980) have demonstrated that interactions of guinea pig AMs
with a wide variety of particles, such as silica, metal oxide-coated fly ash,
polymethylmethacrylate beads, chrysotile asbestos, fugitive dusts, polybead carboxylate
microspheres, glass  and latex beads, uncoated fly ash, and fiberglass increase the production of
reactive oxygen species. Similar findings have been reported by numerous investigators for
human, rabbit, mouse, and guinea pig AMs (Drath and Karnovsky, 1975; Allen and Loose,
1976; Beall et al., 1977; Lowrie and Aber, 1977; Miles et al., 1977; Rister and Baehner, 1977;
Hoidal et al., 1978). PMNs are also known to increase production of superoxide radicals,
hydrogen peroxide,  and hydroxyl radicals in response to membrane-reactive agents and particles
(Goldstein et al., 1975; Weiss et al., 1978; Root and Metcalf, 1977).  Although these responses
may occur at any concentration, they are likely to be greatly enhanced at high exposure
concentrations with  slowed clearance and lung particle overload.
       Reactive oxygen species can also be generated from particle-associated organics. Sagai
et al. (1993) reported that DPM can nonenzymatically generate active oxygen species (e.g.,
superoxide [O2"] and hydroxyl radical [OH] in vitro without any biologically activating systems)
such as microsomes, macrophages, hydrogen peroxide,  or cysteine. Because DPM washed with
methanol could no longer produce these radicals, it was concluded that the active components
were compounds extractable with organic solvents.  However, the nonenzymatic contribution to
the DPM-promoted  active oxygen production was negligible compared with that generated via
an enzymatic route (Ichinose et al., 1997a). They reported that O2" and OH can be
enzymatically generated from DPM by the following process.  Soot-associated quinone-like
compounds are reduced to the semiquinone radical by cytochrome P-450 reductase. These
semiquinone radicals then reduce O2to O2", and the produced superoxide reduces ferric ions to
ferrous ions, which catalyzes the homobiotic cleavage of H2O2 dismutated from O2 by
superoxide dismutase or spontaneous reactions to produce OH.  According to Kumagai et al.
(1997), while quinones are likely to be the favored substrates for this reaction,  the participation
of nitroaromatics cannot be ruled out.
       One of the critical lesions to  DNA bases generated by oxygen free radicals is 8-
hydroxydeoxyguanosine (8-OHdG).  The accumulation of 8-OhdG as a marker of oxidative
DNA damage could be an important  factor in enhancing the mutation rate leading to lung cancer

                                          7-132

-------
(Ichinose et al., 1997a).  For example, formation of 8-OHdG adducts leads to G:C to T:A
transversions unless repaired prior to replication. Nagashima et al. (1995) demonstrated that the
production of (8-OHdG) is induced in mouse lungs by intratracheal instillation of DPM.
Ichinose et al. (1997b) reported further that although intratracheal instillation of DPM in mice
induced a significant increase in lung tumor incidence, comparable increases were not reported
when mice were instilled with extracted DPM (to remove organics).  Lung injury was  also less
in the mice instilled with extracted DPM.  Moreover, increases in 8-OHdG in the mice instilled
with unextracted DPM correlated very well with increases in tumor rates.  In a related study,
Ichinose et al. (1997a) intratracheally instilled small doses of DPM, 0.05, 0.1, or 0.2 mg weekly
for 3 weeks, in mice fed standard or high-fat diets either with or without P-carotene.  High
dietary fat enhanced DPM-induced lung tumor incidence, whereas P-carotene, which may act as
a free radical scavenger, partially reduced the tumorigenic response.  Formation of 8-OHdG was
again significantly correlated with lung tumor incidence in these studies, except at the highest
dose. Dasenbrock et al. (1996) reported that extracted DPM, intratracheally instilled into rats
(15 mg total dose) induced only marginal increases in lung tumor induction, while unextracted
DPM was considerably more effective.  Although adducts were not measured in this study, it
nevertheless provides support for the likelihood that activation of organic metabolites and/or
generation of oxygen free radicals from organics are involved in the carcinogenic process.
       Additional support for the involvement of particle-associated radicals in tissue damage
was provided by the finding that pretreatment with superoxide  dismutase (SOD), an antioxidant,
markedly reduced lung injury and death due to instillation of DPM.  Similarly, Hirafuji et al.
(1995) found that the antioxidants catalase, deferoxamine, and  MK-447 inhibited the toxic
effects of DPM on guinea pig tracheal cells and tissues in vitro.
       Although the data presented supported the hypothesis that generation of reactive oxygen
species resulting from exposure to DPM is involved in the carcinogenic process, it should be
noted that 8-OHdG is efficiently repaired and that definitive proof of a causal relationship in
humans is still lacking.  It is also uncertain whether superoxide or hydroxyl radicals chemically
generated by DPM alone promote 8-OHdG production in vivo  and induce lung toxicity, because
SOD is extensively located in mammalian tissues. Nevertheless, demonstration that oxygen free
radicals can be generated from  particle-associated organics,  that their presence will induce
adduct formation  and DNA damage unless repaired, that tumor induction in  experimental
animals correlates with OhdG adducts, and that treatment with antioxidant limits lung damage,
provides strong support for the involvement of oxygen free  radicals in the toxicologic  and
carcinogenic response to DE.

7.4.4.  Relationship of Physical Characteristics of Particles to Cancer Induction
       The biological potential of inhaled particles is strongly influenced by surface chemistry
and character. For example, the presence  of trace metal  compounds such as aluminum and iron,
as well as ionized or protonated sites, is important in this regard (Langer and Nolan, 1994). A
                                          7-133

-------
major factor is specific surface area (surface area/mg).  PMNs characteristically are increased
abnormally in the lung by DE exposure, but their presence in the lungs does not appear to be
excessive following the pulmonary deposition of even high lung burdens of spherical TiO2
particles in the 1-2 |lm diameter range (Strom, 1984). In these studies lung tumors were
detected only at an inhaled concentration of 250 |lg/m3. In a more recent study in which rats
were exposed to TiO2 in the 15-40 nm size range, inhibition of particle clearance  and
tumorigenesis were induced at concentrations of 10 mg/m3 (Heinrich et al., 1995). Comparison
of several chronic inhalation studies correlating particle mass and particle surface area  retained
in the lung with tumor incidence indicated that particle surface area is a much better dosimeter
than particle mass (Oberdorster and Yu, 1990; Driscoll et al., 1996).  Heinrich et  al.  (1995) also
found that lung tumor rates increased with specific particle surface area following exposure to
DE, carbon black, or titanium dioxide, irrespective of particle type.  Langer and Nolan (1994)
reported that the hemolytic potential of Min-U-Sill5, a silica flour, increased in direct
relationship to specific surface area at nominal particle diameters ranging from 0.5 to 20 |lm.
       Ultrafine particles appear to be more likely to be taken up by lung epithelial cells. Riebe-
Imre et al. (1994) reported that CB is taken up by lung epithelial cells in vitro, inducing
chromosomal damage and disruption of the cytoskeleton, lesions that closely resemble those
present in tumor cells.  Johnson et al. (1993) reported that 20-nm polytetrafluoroethylene
particles are taken up by pulmonary epithelial cells as well as polymorphonuclear leucocytes,
inducing an approximate 4-, 8-, and 40-fold increase in the release of interleukin-1 alpha and
beta, inducible nitric oxide synthetase, and macrophage inflammatory protein, respectively.
       The carcinogenic potency of diesel particles, therefore, appears to be related, at least to
some extent, to their small size and convoluted shape, which results in a large specific  particle
surface area.  Toxicity and carcinogenicity increased with decreasing particle size into  the
submicron range.  For example, Heinrich et al. (1995) have shown that ultrafine titanium dioxide
(approximately 0.2 |lm diameter) is much more toxic than particles with a 10-fold greater
diameter of the same composition used in an earlier study by Lee et al. (1986). This increase in
toxicity has been noted with even smaller particles.  For example, carbon  black particles 20 nm
in diameter were shown to be significantly more toxic than 50 nm particles (Murphy et al.,
1999). The relationship between particle size and toxicity is of concern because,  as noted in
Chapter 2, approximately 50%-90% of the number of particles in DE are in the size range from
5 to 50 nm.  Other than disruption of the cytoskeleton of epithelial cells, there is little
information regarding the means by which particle size influences carcinogenicity as well as
noncancer toxicity.
                                          7-134

-------
7.4.5.  Integrative Hypothesis for Diesel-Induced Lung Cancer
       The induction of lung cancer in rats by large doses of carbon black via inhalation
(Heinrich et al., 1995; Mauderly et al., 1991; Nikula et al., 1995) or intratracheal instillation
(Kawabata et al., 1994; Pott et al., 1994; Dasenbrock et al., 1996) led to the development of the
lung particle overload hypothesis. According to this hypothesis the induction of neoplasia by
insoluble low- toxicity particles is associated with an inhibition of lung particle clearance and the
involvement of persistent alveolar epithelial hyperplasia.  Driscoll (1995), Driscoll et al. (1996),
and Oberdorster and Yu (1990) outlined a proposed mechanism for the carcinogenicity of DE at
high doses that emphasizes the role of phagocytic cells. Following exposure, phagocytosis of
particles acts as a stimulant for oxidant production and inflammatory cytokine release by lung
phagocytes.  It was hypothesized that at high particle exposure concentrations the quantity of
mediators released by particle-stimulated phagocytes exceeds the inflammatory defenses of the
lung (e.g., antioxidants, oxidant-metabolizing enzymes, protease inhibitors, cytokine inhibitors),
resulting in tissue injury and inflammation. With continued particle exposure and/or the
persistence of excessive particle burdens, there then develops an environment of phagocytic
activation, excessive mediator release-tissue injury and, consequently, more tissue injury,
inflammation, and  tissue release. This is accompanied by cell proliferation.  As discussed in a
review by Cohen and Ellwein (1991), conceptually, cell proliferation can increase the likelihood
that any oxidant-induced or spontaneously occurring genetic damage becomes fixed in a
dividing cell and is clonally expanded. The net result of chronic particle exposures sufficient to
elicit inflammation and cell proliferation in the rat lung is an increased probability that the
genetic changes necessary for neoplastic transformation will occur.  A schematic of this
hypothesis has been outlined by McClellan (1997) (see Figure 7-3). In support of this
hypothesis, it was reported that concentrations of inhaled CB resulted in increased cytokine
expression and inflammatory influx of neutrophils (Oberdorster et al., 1995), increased
formation of 8-OhdG (Ichinose et al., 1997b), and increase in the yield of hprt mutants, an effect
ameliorated  by treatment with antioxidants (Driscoll, 1995; Driscoll et al., 1996).  Metabolism
of carcinogenic organics to active forms  as well as the generation of reactive oxygen species
from certain organic species are likely to contribute to the toxic and carcinogenic  process.
       At low exposure concentrations, the lung  particle overload condition is not present and
the overload-induced inflammatory effects are not present. Note, however, as discussed in
Chapters 5 and 6, that other types of inflammation are present in the rat lung at exposures below
those inducing lung particle overload.  However, at low exposures, activation of organic
carcinogens  and generation of oxidants from the organic fraction can still be expected.  Actual
contribution depends upon elution/bioavailability and the  effectiveness of antioxidants.  Direct
effects of ultrafine diesel particles taken  up by epithelial cells are also likely to play a role.
       Although high-dose induction of cancer is logically explained by this hypothesis, particle
overload has not been clearly shown to induce lung cancer in other species. As noted in the
quantitative  chapter, the relevance of the rat pulmonary response is therefore problematic. The

                                          7-135

-------
rat pulmonary noncancer responses to DPM, however, have fairly clear interspecies and human
parallels.  In response to poorly soluble particles such as DPM, humans and rats both develop an
alveolar macrophage response, accumulate particles in the interstitium, and show mild interstitial
fibrosis (ILSI, 2000).  Other species (mice, hamsters) also have shown similar noncancer
pulmonary responses to DPM, but without accompanying cancer response. The rat response for
noncancer pulmonary histopathology, however, seems to be more pronounced compared with
humans or other species,  i.e., rats appear to be more sensitive. Although many critical elements
of interspecies comparison, such as the role of airway geometry and patterns of particle
deposition, need further elucidation, this basic interspecies  similarity and the possible greater
sensitivity of pulmonary response seen after longer exposures at high doses make pulmonary
histopathology in rats a valid basis for noncancer dose-response assessment.
Figure 7-3. Pathegenesis of lung disease in rats with chronic, high-level exposures to
particles.

                                          Diesel exhaust
                                          Particulate matter
          Carbon black                               Exposure

               Exposure
 Clearance
                                              Organic chemicals
                                                     I
               I
             Cytokines
             Growth factors
             Proteases
   Macrophage
              Inflammation
              Cell injury
              Cell proliferation
              Hyperplasia
                            Fibrosis
                                                                            Initiated cell
Preneoplastic
lesion
Malignant
tumor
Source: Modified from McClellan, 1997.
                                          7-136

-------
7.4.6. Summary
       Recent studies have shown rat lung tumor rates resulting from exposures to nearly
organic-free carbon black (CB) particles at high concentrations to be similar to those observed
for DE exposures, thus providing strong evidence for a particle overload mechanism for DE-
induced pulmonary carcinogenesis in rats. Such a mechanism is also supported by the fact that
carbon particles per se cause inflammatory responses and increased epithelial cell proliferation
and that AM function may be compromised under conditions of particle overload.
       The particle overload hypothesis appears sufficient to account for DE-induced lung
cancer in rats.  However, there is also biological plausibility for lung cancer induction in humans
at concentrations insufficient to induce lung particle overload as seen in rats (Chapter 3, Section
3.4 and ILSI, 2000).  The uptake of particles by epithelial cells at  ambient or occupational
exposure levels, DNA damage resulting from oxygen-free radicals generated from organic
molecules, and the gradual in situ extraction and activation of procarcinogens associated with the
diesel particles may play a role in this response and provide a basis for the plausibility. The
slower particle clearance rates in humans (up to a year or more) may result in greater extraction
of organics.  This is supported by reports of increased DNA adducts in humans occupationally
exposed to DE at concentrations unlikely to induce lung particle overload.  Although these
modes of action can be expected to function at lung overload conditions also, they are likely to
be overwhelmed by inflammatory associated effects.
       The evidence  to date indicates that caution must be exercised in extrapolating
observations made in animal models  to humans when assessing the potential for DE-induced
pulmonary carcinogenesis. The  carcinogenic response and the formation of DNA adducts in rats
exposed to DE and other particles at high exposure concentrations may be species-specific and
not DPM specific.  The likelihood that different modes of action predominate at high and low
doses, such as  lung particle overload, also renders high-dose extrapolation to lower ambient
concentrations uncertain.

7.5.  WEIGHT-OF-EVIDENCE EVALUATION FOR POTENTIAL HUMAN
     CARCINOGENICITY
       A carcinogenicity weight-of-evidence evaluation  is a synthesis of all pertinent
information  addressing the question of how likely an agent is to be a human carcinogen.  EPA's
1986 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 1986) provide a classification
system for the  characterization of the overall weight of evidence for potential human
carcinogenicity based on human evidence, animal evidence, and other supportive data. This
system includes Group A: Human Carcinogen; Group B: Probable Human Carcinogen; Group
C: Possible Human Carcinogen; Group D: Not Classifiable as to Human Carcinogenicity; and
Group E: Evidence for Noncarcinogenicity to Humans.
                                         7-137

-------
       As part of the guidelines development and updating process, the Agency has developed
revisions to the 1986 guidelines to take into account knowledge gained in recent years about the
carcinogenic processes. With regard to the weight-of-evidence evaluation for potential human
carcinogenicity, EPA's 1996 Proposed Guidelines for Carcinogen Risk Assessment (U.S. EPA,
1996b) and the subsequent revised external review draft (U.S. EPA, 1999) emphasize the need
for characterizing cancer hazard, in addition to hazard identification. To express the weight of
evidence for potential human carcinogenicity, EPA's proposed 1996 and 1999 guidelines utilize
a hazard narrative in place of the 1986 A-E classification system. In order to provide some
measure of consistency in using the 1996 and 1999 draft guidelines, standard hazard descriptors
are used as part of the hazard narrative.  The revised guidelines also stress the importance of
considering the mode(s) of action information for making an inference about potential cancer
hazard beyond the range of observation, typically encountered at levels of exposure in the
general  environment. "Mode of action" refers to a series of key biological events and processes
that are  critical to the development of cancer. This is contrasted with "mechanisms of action,"
which is defined as a more detailed description of the complete sequence of biological events at
the molecular level that must occur to produce a carcinogenic response.
       The sections to follow evaluate and weigh the individual lines of evidence and combine
all evidence to make an informed judgment about the carcinogenicity hazard of DE. A
conclusion in accordance with EPA's 1986 classification system (U.S. EPA, 1986) is provided,
as well as a hazard narrative  along with appropriate hazard descriptors according to EPA's
Proposed Guidelines (U.S. EPA, 1996b, 1999).  These sections draw on information reviewed in
Chapters 2, 3, 4, and 7.

7.5.1. Human Evidence
       Twenty-two epidemiologic studies about the carcinogenicity of workers exposed to DE
in various occupations are reviewed in Section 7.2. Exposure to DE has typically been inferred
based on job classification within an industry. Increased lung cancer risk, although not always
statistically significant, has been observed in 8 out of 10 cohort and 10 of 12 case-control studies
within several industries, including railroad workers, truck drivers, heavy equipment operators,
and professional drivers. The increased lung cancer relative risks generally range from 1.2 to
1.5, though a few studies show relative risks as high as 2.6. Statistically significant increases in
pooled relative risk estimates (1.33 to 1.47) from two independent meta-analyses further support
a positive relationship between DE exposure and lung cancer in a variety of DE-exposed
occupations.
       The generally small increased lung cancer relative risk (less than 2) observed in the
epidemiologic studies and meta-analyses potentially weakens  the evidence of causality. When a
relative  risk is less than 2, if  confounders (e.g., smoking, asbestos exposure) are having an effect
on the observed risk increases, it could be enough to account for the increased  risk.  With the
                                         7-138

-------
strongest risk factor for lung cancer being smoking, there is a concern that smoking effects may
be influencing the magnitude of the observed increased relative risks. However, in studies for
which the effects of smoking were accounted for, increased relative risks for lung cancer
prevailed. Though some studies did not have information on smoking, significant confounding
by smoking is unlikely because the comparison populations were from the same socioeconomic
class. Moreover, when the meta-analysis focused only on the smoking-controlled studies, the
relative risks tended to increase.
       As evaluated in Section 7.2.4.5, application of the criteria for causality (including the
biological plausibility) leads to the conclusion that the increased risks observed in available
epidemiologic studies are consistent with a causal association between exposure to DE and
occurrence of lung cancer. Overall,  the human evidence for potential carcinogenicity for DE is
judged to be strong, but less than sufficient for DE to be considered as a human carcinogen
because of exposure uncertainties (lack of historical exposure data for workers exposed to DE)
and an inability to reach a fully and direct accounting for  all possible confounders.

7.5.2. Animal Evidence
       DE and its organic constituents, both in the gaseous and particle phase, have been
extensively tested for carcinogenicity in many experimental studies using several animal species
and with different modes of administration.
       Several well-conducted lifetime rat inhalation studies have consistently demonstrated that
chronic inhalation exposure to sufficiently high concentrations  of DE produced dose-related
increases in lung tumors (benign and malignant).  However, the lung cancer responses in rats
from high-concentration exposures appear to be mediated by impairment of lung clearance
mechanisms through particle overload, resulting in persistent chronic inflammation and
subsequent pathologic and neoplastic changes in the lung.  Overload conditions are not expected
to occur in humans as a result of environmental or most occupational exposures to DE. Thus,
the rat lung tumor response is not considered relevant to an evaluation of the potential for a
human environmental exposure-related hazard (Section 7.4).
       The chronic inhalation studies of DE in mice showed equivocal results, whereas negative
findings were consistently seen in hamsters. The gaseous phase of DE (filtered exhaust without
particulate fraction) was found not to be carcinogenic in rats, mice, or hamsters.
       In  several intratracheal instillation studies, diesel particulate matter (DPM), carbon black,
and the organic DPM extracts which were virtually devoid of PAHs, have been found to produce
increased  lung tumors in rats. When directly implanted into the rat lung, DPM condensate
containing mainly four- to seven-ring PAHs induced increases  in lung tumors. In several dermal
studies in  mice, DPM extracts have also been shown to cause skin tumors and sarcomas in mice
following subcutaneous injection.
                                          7-139

-------
       Available data and hypotheses suggest that both the carbon core and the adsorbed
organics have potential roles in inducing lung tumors in the rat, although their relative
contribution to the carcinogenic response remains to be determined.
       The consistent findings of carcinogenic activity by DPM and the organic extracts of
DPM in noninhalation studies (intratracheal instillation, lung implantation, skin painting)
contribute to the overall evidence for a human hazard potential for DE. The lack of a tumor
response from traditional animal inhalation studies in other rodent species is noted. Without
understanding the mode(s) of action of DE's carcinogenicity in humans it is difficult to assess
the meaning of nonpositive results from the mouse and hamster inhalation bioassays, and the
unusable results from the rat, while having other evidence of carcinogenic potential and
plausibility.
       It should be noted that the animal  studies used DE from engines available in the  1980s,
and that present-day engine emissions have different characteristics (e.g., higher elemental
carbon content and lesser amounts of adsorbed organics on the carbon particles), with uncertain
impact on  the outcome of the experimental studies. The same point can be made for the
occupational epidemiologic  studies.

7.5.3. Other Key Data
       Other key data are judged to be supportive of potential carcinogenicity of DE. As
discussed in Chapter 2, DE is a complex mixture of hundreds of constituents in either gaseous
phase or particle phase. Although present in small amounts, several organic compounds in the
gaseous phase (e.g. PAHs, formaldehyde, acetaldehyde, benzene, 1,3-butadiene) are known to
exhibit mutagenic and/or carcinogenic activities. PAHs and PAH derivatives, including nitro-
PAHs, present on the diesel particle are also known to  be mutagenic and carcinogenic.  As
reviewed in  Chapter 4, DPM and DPM organic extracts have been shown to induce gene
mutations  in a variety of bacteria and mammalian cell test systems. In addition, DE,  DPM and
DPM extracts have been found to cause chromosomal aberrations, aneuploidy, and sister
chromatid exchange in both in vivo and in vitro tests.
       There is also suggestive evidence  for the bioavailability of the organics from DE
(Chapter 3, Sections.5). Elevated levels of DNA adducts in lymphocytes have been reported in
workers exposed to DE.  In addition, animal  studies showed that some of the radiolabeled
organic compounds are eluted from DE particles following deposition in the lungs.

7.5.4. Mode of Action
       As discussed in Section 7.4, the modes of action of DE-induced carcinogenicity in
humans is not understood. It can be suggested that one or multiple modes of action may be
involved.  These may include: (a) mutagenic and genotoxic events (e.g., direct and indirect
effects on DNA and effects on chromosomes) by organic compounds in the gaseous and particle
                                         7-140

-------
phases; (b) indirect DNA damage via the production of reactive oxygen species (ROS) induced
by particle-associated organics; and (c) particle-induced chronic inflammatory response leading
to oxidative DNA damage through the release of cytokines, ROS, etc., and an increase in cell
proliferation.
       The particulate phase or whole DE exposure, as measured by DPM, appears to have the
greatest observable contribution to the carcinogenic effects, and both the particle core and the
associated organic compounds have demonstrated carcinogenic properties, although a role for
the gas-phase components cannot be ruled out. The carcinogenic activity of DE may also be
related to the small size of the particles. Moreover, the relative contribution of the possible
mode(s) of action may be different at different exposure levels. For example, available evidence
from rat studies indicates the importance of the role of the DPM in mediating lung tumor
response at high exposure levels.  Thus, the role of the adsorbed organic compounds may take on
increasing importance at lower exposure levels.

7.5.5.  Characterization of Overall Weight of Evidence: EPA's 1986 Guidelines for
       Carcinogen Risk Assessment
       The totality of evidence supports the conclusion that DE is a.probable human carcinogen
(Group Bl). This conclusion is based on:

       •       "Limited"evidence (i.e., strong but less than sufficient evidence for "known
             human carcinogen"), for a causal association between DE exposure and increased
             risk of lung cancer among workers in different occupations;
       •       Evidence of carcinogenicity of DPM in rats and mice by noninhalation routes of
             exposure (intratracheal instillation, lung implantation, skin painting, and
             subcutaneous injection); and

             Extensive supporting data including the demonstrated mutagenic and/or
             chromosomal effects of DE and its organic constituents,  suggestive evidence for
             the bioavailability of the organics from DE, and knowledge of the known
             mutagenic and/or carcinogenic activity of a number of individual organic
             compounds present on the particles (e.g,. PAH and derivatives) and in the DE
             gases (e.g., benzene, 1,3-butadiene, and aldehydes).
                                         7-141

-------
7.5.6.  Weight-of-Evidence Hazard Narrative:  EPA's Proposed Guidelines for Carcinogen
       Risk Assessment (1996b, 1999)
       The combined evidence supports the conclusion that DE is likely to be carcinogenic to
humans by inhalation and that this hazard applies to environmental exposure conditions.  The
spectrum of evidence and the inferences drawn provide a substantial case for this hazard
potential.  The weight of evidence of human carcinogenicity is based on:

              Strong but less than sufficient epidemiologic evidence for a causal association
              between DE exposure and increased risk of lung cancer among workers in
              different occupations;
       •      Evidence of carcinogenicity of DPM in rats and mice by noninhalation routes of
              exposure (intratracheal instillation, lung implantation, skin painting, and
              subcutaneous injection); and
              Extensive supporting data including the demonstrated mutagenic and/or
              chromosomal effects of DE and its organic constituents, suggestive evidence for
              the local and systemic bioavailability of the organics from DE, and knowledge of
              the known mutagenic and/or carcinogenic activity of a number of individual
              organic compounds present on the particles (e.g., PAH and derivatives) and in the
              DE gases (e.g., benzene, 1,3 butadiene, and aldehydes).

       The weight-of-evidence for the lung cancer hazard is considered strong, even though
inferences and uncertainties are involved.  Major uncertainties include:

              There is scientific debate about the significance of the occupational epidemiologic
              evidence for a causal association between occupational exposure and increased
              lung cancer risk.  Some experts view the evidence as weak given that most of the
              relative risk increases are <2.0, whereas others consider the evidence as more than
              adequate and compelling. With relatively low relative risks (<2.0), the effects of
              possible confounding exposures or other factors could play a significant role in
              the risk increases. For example, there is specific concern about whether the
              effects of smoking, a known cause of lung cancer, has been adequately or fully
              accounted for in the key studies.  In more general terms, the lack of historical
              exposure data to retrospectively validate estimated DE exposure levels is also a
              limitation.
              A lack of knowledge about the mode(s) of action of DE lung cancer in humans
              results in the use of a number of default risk assessment assumptions which, while
              justifiable by evidence or policy choice, introduce uncertainty. To date, available
              evidence for the role of DPM, both the adsorbed organics and the carbon core
                                         7-142

-------
             particle, has been shown only for high exposure conditions in the rat lung. The
             tumor inducing mode-of-action in the rat lung appears to depend on particle
             overloading of the lung and subsequent pathology.  This sequence is judged not to
             be relevant for assessing the hazard to humans exposed in the ambient
             environment.  There is virtually no information about the relative role of DE
             constituents in mediating the carcinogenic effects at lower experimental exposure
             levels, though hypotheses exist.

       While a major uncertainty relates to the incomplete understanding of DE's mode(s) of
action for the induction of lung cancer in humans, available data and hypotheses suggest that
DE-induced lung carcinogenicity may be mediated by mutagenic and nonmutagenic events from
both the particles and the associated organic compounds, although a role for the organics in the
gaseous phase cannot be ruled out. Given that there is some evidence for a mutagenic mode of
action, a cancer hazard is presumed at environmental exposure levels.  This is consistent with
EPA's science policy position, which assumes a nonthreshold effect for carcinogens in the
absence of definitive data demonstrating a  nonlinear or threshold mechanism. It should also be
noted that there are not orders of magnitude differences between lower level occupational and
higher end environmental exposure levels,  in fact, there appears to be exposure overlap.  This
observation means that an extrapolation of the occupational hazard to lower environmental
exposure levels is minimal, and thus, the conclusion of an environmental hazard is supported.
Given these circumstances, linear low-dose extrapolation also would be an appropriate default
choice in dose-response assessment that is focused on environmental levels of exposure (Chapter
8, Section 8.2). Because of insufficient information, the human carcinogenic potential of DE by
oral and dermal exposures cannot be  determined.

7.6.  EVALUATIONS BY OTHER ORGANIZATIONS
       Several organizations have reviewed the  relevant data and evaluated the potential human
carcinogenicity of DE or its particulate component. The conclusions reached by these
organizations are generally comparable to the evaluation made in this assessment using EPA's
Carcinogen Risk Assessment Guidelines.  A summary of available evaluations conducted by
other organizations is provided in Table 7-9.

7.7.  CONCLUSION
       It is concluded that environmental exposure to DE may present a lung cancer hazard to
humans.  The particulate phase appears to have the greatest contribution to the carcinogenic
effect, both the particle core and the associated organic compounds have demonstrated
                                         7-143

-------
       Table 7-9. Evaluations of DE as to human carcinogenic potential
Organization
NIOSH (1988)
IARC (1989)
IPCS (1996)
California EPA (1998)
NTP (2000)
Human data
Limited
Limited
N/Aa
"Consistent evidence for
a causal association"
"Elevated lung cancer in
occupationally exposed
groups"
Animal data
Confirmatory
Sufficient
N/A
"Demonstrated
carcinogenicity"
"Supporting animal and
mechanistic data"
Overall
evaluation
Potential occupational
carcinogen
Probably carcinogenic to
humans
Probably carcinogenic to
humans
DPM as a "toxic air
contaminant" (California
Air Resources Board)
DPM-Reasonably
anticipated to be a
carcinogen
aNot applicable.

carcinogenic properties, although a role for the DE gas-phase components cannot be ruled out.
Using either EPA's 1986 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 1986) or the
proposed revisions (U.S. EPA, 1996b, 1999), DE is judged to be a probable human carcinogen,
or likely to be carcinogenic to humans by inhalation, respectively.  The weight of evidence for
potential human carcinogenicity for DE is considered strong, even though inferences are
involved in the overall assessment. Major uncertainties of the hazard assessment include the
following unresolved issues:

       •      There has been a considerable scientific debate about the significance of the
             available human evidence for a causal association between occupational exposure
             and increased lung cancer risk. Some experts view the evidence as weak given
             that most of the relative risk increases are <2.0 whereas others consider the
             evidence as more than adequate and compelling. Additionally, there is debate
             about whether the effects of smoking have been adequately accounted for in key
             studies, as well as the lack of historical DE exposure data to retrospectively
             validate estimated DE exposure levels for the available studies.
       •      A lack of knowledge about the mode(s) of action for DE lung cancer in humans
             results in the use of a number of default risk assessment  assumptions which, while
             justifiable by evidence or policy choice, introduce uncertainty.  To date, available
             evidence for the role of DPM, both the adsorbed organics and the carbon core
             particle, has been  shown only for high-exposure conditions in the rat lung. The
                                        7-144

-------
               tumor-inducing mode of action in the rat lung appears to depend on particle

               overloading of the lung but this is judged to be not relevant for assessing the

               human hazard of ambient exposures. There is virtually no information about the

               relative role of DE constituents in mediating carcinogenic effects at lower

               experimental  or environmental exposure levels. Furthermore, there is only a

               limited understanding regarding the relationship between DPM particle size and

               carcinogenicity.

               DE is present in ambient PM (e.g., PM2 5 or PM10); however,  a cancer hazard for

               ambient PM has not been identified, as of 1996 (EPA 1996b). An updated

               evaluation is expected in 2002.


Additional research is needed to address these issues to reduce the uncertainty associated with

the potential cancer hazard of exposure to DE.


                             REFERENCES FOR CHAPTER 7

Abbey, DE; Mills, PK; Petersen, FF; et al. (1991) Long-term ambient concentrations of total suspended particulates
and oxidants as related to incidence of chronic disease in California Seventh-Day Adventists. Environ Health
Perspect 94:43-50.

Abbey DE; Lebowitz MD; Mills PK; et al. (1995) Long-term ambient concentrations of particulates and
development of chronic disease in a cohort of nonsmoking California residents.  Inhal Toxicol 7:19-34.

Adamson, IYR; Bowden, DH. (1978) Adaptive response of the pulmonary macrophagic system to carbon. II.
Morphologic studies. Lab Invest 38:430-438.

Ahlberg, J; Ahlbom, A; Lipping, H; et al. (1981) [Cancer among professional drivers—a problem-oriented
register-based study]. Lakartidningen 78:1545-1546.

Allen, RC; Loose, LD. (1976) Phagocytic activation of a luminol-dependent chemiluminescence in rabbit alveolar
and peritoneal macrophages. Biochem Biophys Commun 69:245-252.

Ames, BN. (1983) Dietary carcinogens and anticarcinogens. Science 221:1256-1264.

Aronsen, KJ; Siemiatycki,  J; Dewar, R; et al. (1996) Occupational risk factors for prostate cancer:  results from a
case-control study in Montreal, Quebec, Canada. Am J Epidemiol 143(4):363-373.

Balarajan, R; McDowell, ME. (1983) Malignant lymphomas and road transport workers. J Epidemiol Community
Health 37:316-317.

Balarajan, R; McDowall, ME. (1988) Professional drivers in London: a mortality study. Br J Ind Med 45:483-
486.

Ball, JC; Green, B; Young, WC; et al., (1990) S9-activated Ames assays of diesel particle extracts:  detecting
indirect-acting mutagens in samples that are direct-acting.  Environ Sci Technol 24:890-894.

Beall, GD; Repine, JE; Hoidal, JR; et al. (1977) Chemiluminscence by human alveolar macrophages: stimulation
with heat killed bacteria or phorbol myristate acetate. Infect Immunol 17:117-120.
                                             7-145

-------
Benhamou, S; Benhamou, E; Flamant, R. (1988) Occupational risk factors of lung cancer in a French case-control
study. Br J Ind Med 45:231-233.

Bender, AP; Parker, DL; Johnson, RA; et al. (1989) Minnesota Highway Maintenance Workers Study: cancer
mortality. Am J Ind Med 15:545-556.

Bhatia, R; Lopipero, P; Smith, A. (1998) Diesel exhaust exposure and lung cancer. Epidemiology 9(1):84-91.

Bitterman, PB; Aselberg, S; Crystal, RG. (1983) Mechanism of pulmonary fibrosis: spontaneous release of the
alveolar macrophage-derived growth factor in interstitial lung disorders. J Clin Invest 72:1801-1813.

Boffetta, P; Stellman, SD. (1988) Association between diesel exhaust exposure and multiple myeloma: an
example of confounding. Prev Med 17:236-237.

Boffetta, P; Harris, RE; Wynder, EL. (1990) Case-control study on occupational exposure to diesel exhaust and
lung cancer risk. Am J Ind Med 17:577-591.

Bohning, DE; Atkins, HL; Cohn, SH. (1982) Long-term particle clearance in man: normal and impaired. Ann
OccupHyg 26:259-271.

Bond, JA; Mitchell, CE; Li, AP. (1983) Metabolism and macromolecular covalent binding of benzo[a]pyrene in
cultured Fischer-344 rat lung type II epithelial cells.  Biochem Pharmacol 32:3771-3776.

Bond, JA; Butler, MM; Medinsky, MA; et al. (1984) Dog pulmonary macrophage metabolism of free and particle-
associated [14C]benzo[a]pyrene. J Toxicol Environ Health 14:181-189.

Bond, JA; Johnson, NF; Snipes, MB; et al. (1990) DNA adduct formation in rat alveolar type II cells: cells
potentially at risk for inhaled DE. Environ Mol Mutagen 16:64-69.

Boorman, GA; Brockman, M; Carlton, WW; et al. (1996)  Classification of cystic keratinizing squamous lesions of
the rat lung:report of a workshop.  Toxicol Pathol 24:564-572.

Borm, PJA; Knaapen, AM; Schins, RFP; et al. (1997) Neutrophils amplify the formation of DNA adducts by
benzo[a]pyrene in lung target cells. Environ Health  Perspect 105 (Suppl 5): 1089-1093.

Bowden, DH; Adamson, IYR. (1978) Adaptive responses  of the pulmonary macrophagic system to carbon. I.
Kinetic studies. Lab Invest 38:422-438.

Boyd, MR.  (1984) Metabolic activation and lung toxicity: a basis for cell-selective pulmonary damage by foreign
chemicals. Environ Health Perspect 55:47-51.

Brightwell, J; Fouillet, X; Cassano-Zoppi, AL; et al. (1986) Neoplastic and functional changes in rodents after
chronic inhalation of engine exhaust emissions.  In: Carcinogenic and mutagenic effects of diesel engine exhaust.
Ishinishi, N; Koizumi, A; McClellan, R; et al., eds. Amsterdam: Elsevier; pp. 471-485.

Brightwell, J; Fouillet, X; Cassano-Zoppi, AL; et al. (1989) Tumors of the respiratory tract in rats and hamsters
following chronic inhalation of engine exhaust emissions.  J Appl Toxicol 9:23-31.

Briiske-Hohlfeld, I; Mohner, M; Ahrens, W; et al. (1999) Lung cancer risk in male workers occupationally
exposed to diesel motor emissions in Germany.  Am J Ind Med 36:405-414.

Buiatti, E; Kriebel, D; Geddes, M; et al. (1985) A case control study of lung cancer in Florence, Italy.  I.
Occupational risk factors. J Epidemiol Commun Health 39:244-250.

California Environmental Protection Agency. (Cal EPA, OEHHA) (1998) Health risk assessment for diesel
exhaust.  Public and Scientific Review Draft. February 1998.
                                                7-146

-------
Claxton, LD. (1983) Characterization of automotive emissions by bacterial mutagenesis bioassay: a review.
Environ Mutagen 5:609-631.

Coggon, D; Pannett, B; Acheson, ED. (1984) Use of job-exposure matrix in an occupational analysis of lung and
bladder cancers based on death certificates. J Natl Cancer Inst 72:61-65.

Cohen, SM; Ellwein, LB. (1991) Genetic error, cell proliferation and carcinogenesis. Cancer Res 51:6493-6505.

Cohen, AJ; Higgins, MWP. (1995) Health effects of diesel exhaust: epidemiology. In: Diesel exhaust: a critical
analysis of emissions, exposure, and health effects. Cambridge, MA:  Related HEI Publications; pp. 251-292.

Commins, BT; Waller, RE; Lawther, PJ. (1957) Air pollution in diesel bus garages. Br J Ind Med 14:232-239.

Cordier, S; Clavel, J; Limasset, JC; et al. (1993) Occupational risks of bladder cancer in France: a multicenter
case-control study. Int J Epidemiol 22:403-411.

Cox, LA.  (1997) Does diesel exhaust cause human lung cancer? J Risk Anal 17(6):807-829.

Crump, KS. (1999) Lung cancer mortality and diesel exhaust: reanalysis of a retrospective cohort study of U.S.
railroad workers. Inhal Toxicol 11:101-107.

Crump, KS; Lambert, T; Chen, C; et al. (1991) Assessment of risk from exposure to diesel engine emissions.
Clement International Corporation. Prepared for U.S. EPA under contract no. 68-02-4601; 56.

Damber, LA; Larsson, LG. (1987) Occupation and male lung cancer: a case-control study in northern Sweden. Br
J Ind Med 44:446-453.

Dasenbrock, C; Peters, L; Creutzenberg, O; et al. (1996) The carcinogenic potency of carbon particles with and
without PAH after repeated intratracheal administration in the rat.  Toxicol Lett 88:15-21.

Department of Labor, Mine Safety and Health Administration (2001, Jan. 19) Federal Register 66(13):5526-5706.

Depass, LR; Chen, KC; Peterson, LG. (1982) Dermal carcinogenesis bioassays of diesel particulates and
dichloromethane extract of diesel particulates in C3H mice. In:  Toxicological effects of emissions from diesel
engines. Lewtas, J, ed.  Amsterdam: Elsevier; pp. 321-326.

Dockery, DW;  Pope, CA, III; Xu, X; et al. (1993) An association between air pollution and mortality in six U.S.
cities. N Engl J Med 329:1753-1759.

Dontenwill, W; Chevalier, H-J; Harke, H-P; etal. (1973) Investigations on the effects of chronic cigarette smoke
inhalationin Syrian golden hamsters. JNatl Cancer Inst 51:1681-1832.

Drath, DB; Karnovsky, ML. (1975) Superoxide production by phagocytic leukocytes. J Exp Med 141:257-262.

Driscoll, K. (1995) Role of inflammation in the development of rat lung tumors in response to chronic particle
exposure. In: Proceedings of the MIT toxicology symposium: particle overload in the rat lung and lung cancer:
relevance for human risk assessment. Cambridge, MA: Massachusetts Institute of Technology; 29 pp.


Driscoll, KE; Carter, JM; Howard, BW; et al. (1996) Pulmonary inflammatory, chemokine, and mutagenic
responses in rats after subchronic inhalation of carbon black. Toxicol Appl Pharmacol 136:372-380.

Driscoll, KE; Carter, JM; Hassenbein, DG; et al. (1997a) Cytokines and particle-induced inflammatory cell
recruitment. Environ Health Perspect 105 (Suppl 5): 1159-1164.

Driscoll, KE; Deyo, LC; Carter, JM; et al. (1997b) Effects of particle exposure and particle-elicited inflammatory
cells on mutation in rat alveolar epithelial cells. Carcinogenesis 18:423-430.


                                                7-147

-------
Edling, C; Anjou, C-G; Axelson, O; et al. (1987) Mortality among personnel exposed to diesel exhaust. Int Arch
Occup Environ Health 59:559-565.

El-Bayoumy, K; Rivenson, A; Johnson, B; et al. (1988) Comparative tumorigenicity of 1-nitropyrene, 1-
nitrosopyrene, and 1-aminopyrene administered by gavage to Sprague-Dawley rats. Cancer Res 48:4256-4260.

Emmelin, A; Nystrom, L; Wall, S. (1993) Diesel exhaust exposure and smoking: a case-referent study of lung
cancer among Swedish dock workers. Epidemiology 4:237-244.

Enya, T;  Suzuki, H; Watanabe, T; et al. (1997)  3-Nitrobenzanthrone, a powerful bacterial mutagen and suspected
human carcinogen found in diesel exhaust and airborne particulates. Environ Sci Technol 31:2772-2776.

Evans, M; Shami, SG; Martinez, LA. (1986) Enhanced proliferation of pulmonary alveolar macrophages after
carbon instillation in mice depleted of blood monocytes by strontium-89. Lab Invest 54:154-159.

Farrell, RL; Davis, GW. (1974) Effect of paniculate benzo(a)pyrene carrier on carcinogenesis in the respiratory
tract of hamsters. In: Experimental lung cancer. Karbe, E; Davis, GW, eds. Berlin: Springer; pp. 186-198.

Flodin, U; Fredriksson, M; Persson,  B. (1987) Multiple myeloma and engine exhausts, fresh wood, and creosote: a
case-referent study. Am J Ind Med 12:519-529.

Gallagher, J; George, M; Kohan, M; et al. (1993) Detection and comparison of DNA adducts after in vitro and in
vivo diesel emission exposures. Environ Health Perspect 99:225-228.

Gallagher, J; Heinrich, U; George, M; et al. (1994) Formation of DNA adducts in rat lung following chronic
inhalation of diesel emissions, carbon black, and titanium dioxide particles.  Carcinogenesis 15:1291-1299.

Garland,  FC; Gorham, ED; Garland, CF; et al. (1988) Testicular cancer in US Navy personnel. Am J Epidemiol
127:411-414.

Garshick, E; Schenker, MB; Munoz, A; et al. (1987) A case-control study of lung cancer and diesel exhaust
exposure in railroad workers. Am Rev Respir Dis  135:1242-1248.

Garshick, E; Schenker, MB; Munoz, A; et al. (1988) A retrospective cohort study of lung cancer and diesel
exhaust exposure in railroad workers. Am Rev Respir Dis 137:820-825.

Gerde, P; Medinsky, MA; Bond, JA. (1991) Particle-associated polycyclic aromatic hydrocarbons~a reappraisal
of their possible role in pulmonary carcinogenesis. Toxicol Appl Pharmacol 188:1-13.

Goldstein, IM; Roos, D; Kaplan, HB; et al. (1975) Complement and immunoglobulins stimulate superoxide
production by human leukocytes independently  of phagocytosis. J Clin Invest 56:1155-1163.

Grimmer, G; Brune, H; Deutsch-Wenzel, R; et al. (1987) Contribution of polycyclic aromatic hydrocarbons and
nitre-derivatives to the carcinogenic impact of diesel engine exhaust condensate evaluated by implantation into
the lungs of rats.  Cancer Lett 37:173-180.

Guberan, E; Usel, M; Raymond, L; et al. (1992) Increased risk for lung cancer and for cancer of the
gastrointestinal tract among Geneva professional drivers. Br J Ind Med 49:337-344.

Gustafsson, L; Wall, S; Larsson, LG; et al. (1986) Mortality and cancer incidence among Swedish dock
workers—a retrospective cohort study. Scand J  Work Environ Health 12:22-26.

Gustavsson, P; Plato, N; Lidstrom, EB; et al. (1990) Lung cancer and exposure to diesel exhaust among bus
garage workers. Scand J Work Environ Health 16:348-354.

Hall, NEL; Wynder, EL. (1984) Diesel exhaust  exposure and lung cancer: a case-control study. Environ Res
34:77-86.
                                                7-148

-------
Hammond, SK. (1998) Occupational exposure to diesel exhaust. Presented at the Scientific Review Panel
Meeting, California Air Resources Board, March 11, 1998.

Hansen, ES. (1993) A follow-up study on the mortality of truck drivers. Am J Ind Med 23:811-821.

Hansen, J; Raaschou-Nielsen, O; Olsen, JH; et al. (1998) Increased risk of lung cancer among different types of
professional drivers in Denmark.  Occup Environ Med 55:115-118.

Hatch, GE; Gardner, DE; Menzel, DB. (1980) Stimulation of oxidant production in alveolar macrophages by
pollutant and latex particles. Environ Res 32:121-136.

Hayes, RB; Thomas, T; Silverman, DT; et al. (1989) Lung cancer in motor exhaust-related occupations. Am J Ind
Med 16:685-695.

Health Effects Institute (HEI). (1999) Diesel emissions and lung cancer and epidemiology and quantitative risk
assessment-a special report of the institutes diesel epidemiology expert panel. Cambridge, MA: June, 1995.


Heinrich, U. (1990a) Presentation at workshop on health effects of diesel emissions. U.S. EPA, Research Triangle
Park, July 1990.

Heinrich, U. (1990b) Carcinogenicity of carbon black "Printex 90" in female Wistar rats. Presented at: U.S.
Environmental Protection Agency Diesel Emissions Peer Review Workshop; July 18-19; Chapel Hill, NC.

Heinrich, U; Peters, L; Funcke, W; et al. (1982) Investigation of toxic and carcinogenic effects of diesel exhaust
in long-term inhalation exposure of rodents. In: Lewtas, J, ed. Toxicological effects of emissions from diesel
emissions. Amsterdam: Elsevier; pp. 225-242.

Heinrich, U; Pott, F; Rittinghausen, S. (1986a) Comparison of chronic inhalation effects in rodents after long-term
exposure to either coal oven flue gas mixed with pyrolized pitch or diesel engine exhaust. In: Carcinogenic and
mutagenic effects of diesel engine exhaust. Ishinishi, N; Koizumi, A; McClellan, R; Stober, W, eds. Amsterdam:
Elsevier; pp. 441-457.

Heinrich, U; Muhle, H; Takenaka, S; et al. (1986b) Chronic effects on the respiratory tract of hamsters, mice and
rats after long-term inhalation of high concentrations of filtered and unfiltered diesel engine emissions. J Appl
Toxicol 6:383-395.

Heinrich, U; Peters, L; Fuhst, R; et al. (1989a) The effect of automotive exhaust exposure on the carcinogenicity
of diphenylnitrosamine (DPN) in the respiratory tract of rats.  Exp Pathol 37:51-55.

Heinrich, U; Mohr, U; Fuhst, R; et al. (1989b) Investigation of a potential cotumorigenic effect of the dioxides of
nitrogen and sulfur, and of diesel-engine exhaust, on the respiratory tract of Syrian golden hamsters. Research
Report No. 26. Cambridge, MA: Health Effects Institute; pp.  27.

Heinrich, U; Dungworth, DL; Pott, F; et al. (1994) The carcinogenic effects of carbon black particles and tar-pitch
condensation aerosol after inhalation exposure of rats. Ann Occup Hyg 48 (suppl. 4):351-356.

Heinrich, U; Fuhst, R; Rittinghausen, S; et al. (1995) Chronic inhalation exposure of Wistar rats and two different
strains of mice to diesel engine exhaust, carbon black, and titanium dioxide. Inhal Toxicol 7:533-556.

Hemminki, K; Soderling, J; Ericson, P; et al. (1994) DNA adducts among personnel servicing and loading diesel
vehicles. Carcinogenesis 15:767-769.

Hill, AG. (1965) The environment and disease: association or causation? Proc R Soc Med 58:295-300.

Hirafuji, M; Sakakibara, M; Endo, T; et al.  (1995) Biological effects of diesel exhaust particles (DEP)  on tissues
and cells isolated from respiratory tracts of guinea pigs. Res Commun Mol Pathol Pharmacol 90:221-233.


                                                7-149

-------
Hirose, M; Lee, M-S; Wang, CY; et al. (1984) Induction of rat mammary gland tumors by 1-nitropyrene, a
recently recognized environmental mutagen. Cancer Res 44:1158-1162.

Hoar, SK; Hoover, R. (1985) Truck driving and bladder cancer mortality in rural New England. J Natl Cancer Inst
74:771-774.

Hoidal, JR; Repine, JE; Beall, GD; et al. (1978) The effect of phorbol myristate acetate on metabolism and
ultrastructure of human alveolar macrophages. Am J Pathol 91:469-482.

Hou, S-M; Lambert, B; Hemminki, K. (1995) Relationship between hprt mutant frequency, aromatic DNA
adducts and genotypes for GSTM1 andNAT2 in bus maintenance workers. Carcinogenesis 16:1913-1917.

Howard, PC; Gerrard, JA; Milo, GE; et al. (1983) Transformation of normal human skin fibroblasts by 1-
nitropyrene and 6-nitrobenzo[a]pyrene. Carcinogenesis 4:353-355.

Howe, GR; Lindsay, JP. (1983) A follow-up study of a ten percent sample of the Canadian labour force: I.  Cancer
mortality in males 1965-1973.

Howe, GR; Burch, JD; Miller, AB; et al. (1980) Tobacco use, occupation, coffee, various nutrients, and bladder
cancer. J Natl Cancer Inst 64:701-713.

Howe, GR; Fraser, D; Lindsay, J; et al. (1983) Cancer mortality (1965-77)  in relation to diesel fume and coal
exposure in a cohort of retired railway workers. J Natl Cancer Inst 70:1015-1019.

Huisingh, J; Bradow, R; lungers, R; et al. (1978) Application of bioassay to the characterization of diesel particle
emissions. In: Application of short-term bioassay in the fractionation and analysis of complex mixtures. Waters,
MD; Nesnow, S; Huisingh, JL; et al., eds. New York: Plenum; pp. 381-418.

Humble, CG; Samet, JM; Skipper, BE. (1984) Comparison of serf- and surrogate-reported dietary information.
Am J Epidemiol 119:86-98.

IARC. (International Agency for Research on Cancer) (1989) Monographs  on the evaluation of carcinogenic risks
to humans. Vol. 46. Diesel and gasoline engine exhaust and some nitroarenes. Lyon, France: IARC.

IARC. (1997) Monographs on the evaluation of carcinogenic risks to humans.  Vol. 68. Silica, some silicates, coal
dust and para-aramid fibrils. Lyon, France: IARC, pp. 362-375.

Ichinose, T; Yajima, Y; Nagashima, M; et al. (1997a) Lung Carcinogenesis and formation of 8-hydroxy-
deoxyguanosine in mice by diesel exhaust particles. Carcinogenesis 18:185-192.

Ichinose, T; Yamanushi, T; Seto, H; et al. (1997b) Oxygen radicals in lung Carcinogenesis accompanying
phagocytosis of diesel exhaust particles. Int JOncol 11:571-575.

International Life  Sciences Institute (ILSI). (2000) ILSI Risk Science Institute workshop: the relevance of the rat
lung response to particle overload for human risk assessment. Gardner, DE, ed. Inhal Toxicol 12(1-2).

IPCS. (1996) International Program on Chemical Safety Environmental Health Criteria 171. Diesel fuel and
exhaust emissions. Geneva: World Health Organization, pp. 172-176.

Iscovich, J; Castelleto, R; Esteve, J; et al. (1987) Tobacco smoking, occupational exposure and bladder cancer in
Argentina. Int J Cancer 40:734-740.

Ishinishi, N; Kuwabara, N; Takaki, Y; et al. (1988a) Long-term inhalation experiments on diesel exhaust. In:
Diesel exhaust and health risks. Results of the HERP studies. Ibaraki, Japan: Research Committee for HERP
Studies; pp. 11-84.

Ishinishi, N; Inamasu, T; Hisanaga, A; et al. (1988b) Intratracheal instillation study of diesel paniculate extracts in
hamsters. In: Diesel exhaust and health risk. Ibaraki, Japan: Research Committee for HERP Studies; pp. 209-216.

                                                7-150

-------
Iwai, K; Udagawa, T; Yamagishi, M; et al. (1986a) Long-term inhalation studies of diesel exhaust on F344 SPF
rats. Incidence of lung cancer and lymphoma. In: Ishinishi, N; Koizumi, A; McClellan, R; Stober, W, eds.
Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam: Elsevier; pp. 349-360.

Iwai, K; Udwaga, T; Yamagishi, M; et al. (1986b) Long-term inhalation studies of diesel exhaust on F344 SPF
rats. Dev Toxicol Environ Sci 13:349-360.

Iwai, K; Higuchi, K; Udagawa, T; et al. (1997) Lung tumor induced by long-term inhalation or intratracheal
instillation of diesel exhaust particles. Exp Pathol Toxicol 49:393-401.

Iyer, V; Harris, RE; Wynder, EL. (1990) Diesel exhaust  exposure and bladder cancer risk. Eur J Epidemiol 6:49-
54.

Jensen, OM; Wahrendorf, J; Knudsen, JB; et al. (1987) The Copenhagen case-referent study on bladder cancer:
risks among drivers, painters, and certain other occupations. Scand J Work Environ Health 13:129-134.

Jockel, KH; Ahrens, W; Jahn, I; et al. (1995) Untersuchungen zu Lungenkrebs und Risiken am Arbeitsplatz
(Schlussbericht). Berlin:  Wirtschaftsverlag NW. Fb01HK546.

Jockel, KH; Ahrens, W; Jahn, I; et al. (1998) Occupational risk factors for lung cancer—a case-control study in
West Germany. Int J Epidemiol 27:549-560.

Johnson, C; Finkelstein, J; Gelein, R; et al. (1993) Early alterations in the mRNA abundance of IL-loc, IL-lp,
iNOS, MIP-2, TGFP, and VEGF associated with ultrafine particle  exposure. Fundam Appl Toxicol 50:138 (abstr).

Jordana, M; Richards, C; Irving, LB; et al. (1988) Spontaneous in vitro release of alveolar macrophage cytokines
after the intratracheal instillation of bleomycin in rats: characterization and kinetic studies. Am Rev Respir Dis
137:1135-1140.
Kaplan, I. (1959) Relationship of noxious gases to carcinoma of the lung in railroad workers. J Am Med Assoc
171:2039-2043.

Kaplan, HL; Mackenzie, WF; Springer, KJ; et al. (1982) A subchronic study of the effects of exposure of three
species of rodents to diesel exhaust. In: Toxicological effects of emissions from diesel engines.  Lewtas, J, ed.
New York: Elsevier Biomedical; pp. 161-220.

Kaplan, HL; Springer, KJ; MacKenzie, WF. (1983) Studies of potential health effects of long-term exposure to
diesel exhaust emissions. Final Report No. 01-0750-103 (SWRI) and No. 1239 (SFRE). San Antonio, TX:
Southwest Research Institute.

Karagianes, RF; Palmer, RF; Busch, RH. (1981) Effects of inhaled diesel emissions and coal dust in rats. Am Ind
Hyg Assoc J. 42:382-391.

Kawabata, Y; Iwai, K; Udagawa, T; et al. (1986) Effects of diesel soot on unscheduled DNA synthesis of tracheal
epithelium and lung tumor formation. In: Carcinogenic and mutagenic effects of diesel engine exhaust. Ishinishi,
N; Koizumi, A; McClellan, R; et al., eds. Amsterdam: Elsevier; pp. 213-232.

Kawabata, Y; Udagawa, T; Higuchi, K; et al. (1994) Early one year exposure to diesel engine exhaust causes lung
tumors.  Toxic and carcinogenic effects of solid particles on the respiratory tract.  In: Toxic and carcinogenic
effects of solid particles in the respiratory tract. Mohr, U; Dungworth, DL; Mauderly, JL; et al., eds. Washington,
DC: International Life Sciences Institute Press, pp. 429-431.

Kensler, TW; Egner, PA; Moore, KG; et al. (1987) Role of inflammatory cells in the metabolic activity of
poly cyclic aromatic hydrocarbons in mouse skin. Toxicol Appl Pharmacol 90:337-346.

Klebanoff, SJ; Clark, RA. (1978) The metabolic burst. In: The neutrophil: function and  clinical disorders.
Amsterdam: North-Holland; pp. 283-368.

                                                7-151

-------
Kolonel, LN; Hirohata, T; Nomura, AMY. (1977) Adequacy of survey data collected from substitute respondents.
Am JEpidemiol 106:476-484.

Kotin, P; Falk, H; Thomas, M. (1955) Aromatic hydrocarbons. III. Presence in the paniculate phase of diesel
engine exhausts and the carcinogenicity of exhaust extracts. Arch Ind Health 11:113-120.

Kumagai, Y; Arimoto, T; Shinyashiki, M; et al. (1997) Generation of reactive oxygen species during interaction
of diesel exhaust particle components with NADPH cytochrome P450 reductase and involvement of the
bioactivation in the DNA damage. Free Radical Biol Med 22:479-487.

Kunitake, E; Shimamura, K; Katayama, H; et al. (1986) Studies concerning carcinogenesis of diesel paniculate
extracts following intratracheal instillation, subcutaneous injection, or skin application. In: Carcinogenic and
mutagenic effects of diesel engine exhaust. Ishinishi, N; Koizumi, A; McClellan, R; et al., eds.  Amsterdam:
Elsevier; pp. 235-252.

Kunitake, E; Imase, A; Shimamura, K; et al. (1988) Skin application and subcutaneous injection experiments of
diesel paniculate extracts using ICR mice and nude mice. In: Diesel exhaust and health risks. Results of the HERP
studies.  Ibaraki, Japan: Research Committee for HERP Studies; pp. 217-225.

Langer,  AM; Nolan, RP. (1994) Factors controlling the biological potential of inorganic dusts:  surface area,
chemistry and character. In: Toxic and carcinogenic effects of solid particles in the respiratory tract. Mohr, U;
Dungworth, DL; Mauderly, JL; et al., eds. Washington, DC: International Life Sciences Institute Press, pp. 147-
156.

Lechner, JF; Mauderly, JL. (1994) Sequence of events in lung carcinogenesis: initiation and promotion,
protooncogenes and tumor suppressor genes and particulates. In:  Toxic and carcinogenic effects of solid particles
in the respiratory tract. Mohr, U; Dungworth, DL; Mauderly, JL; et al., eds. Washington, DC: International Life
Sciences Institute Press, pp. 235-251.

Lee, KP; Henry, NW, III; Trochimowicz, HJ; et al. (1986) Pulmonary response to impaired lung clearance in rats
following excessive TiO2 dust deposition. Environ Res 41:144-167.

Lehnert, BE; Valdez, YE; Bomalaski, SH. (1988) Analyses of particles in the lung free cell, tracheobronchial
lymph nodal, and pleural space compartments following their deposition in the lung as related to lung clearance
mechanisms. Ann Occup Hyg 32:125-140.

Lehnert, BE; Ortiz, JB; London, JE; et al. (1989) Migratory behavior of alveolar macrophages during clearance of
light to heavy lung burdens of particles. 1989 Annual Meeting of  the American Thoracic Society, Cincinnati, OH,
May 14-17.

Lerchen, ML; Samet, JM. (1986) An assessment of the validity of questionnaire responses provided by a surviving
spouse.  Am JEpidemiol 123:481-489.

Lerchen, ML; Wiggins, CL; Samet, JM. (1987) Lung cancer and occupation in New Mexico. J  Natl Cancer Inst
79:639-645.

Lewis, TR; Green, FHY; Moorman, WJ; et al. (1989) A chronic inhalation toxicity study of diesel engine
emissions and coal  dust, alone and combined. J Am Coll Toxicol  8:345-375.

Lipsett,  M; Campleman, S. (1999) Occupational exposure to diesel exhaust and lung cancer: a meta-analysis.
Am J Public Health 80(7):1009-1017.

Lloyd, JW. (1971) Long-term mortality of steelworkers. V. Respiratory cancer in coke plant workers. J Occup
Med 13:53-68.

Lowrie, DB; Aber,  VR. (1977) Superoxide production by rabbit alveolar macrophages. Life Sci 21:1575-1584.
                                                7-152

-------
Mauderly, JL. (1994) Toxicological and epidemiological evidence for health risks from inhaled engine emissions.
Environ Health Perspect 102(4): 165-171.

Mauderly, JL; Jones, RK; Griffith, WC; et al. (1987) Diesel exhaust is a pulmonary carcinogen in rats exposed
chronically by inhalation.  Fundam Appl Toxicol 9:208-221.

Mauderly, JL; Snipes, MB; Barr, EB; et al. (1991) Influence of particle-associated organic compounds on
carcinogenicity of diesel exhaust. Eighth Health Effects Institute Annual Conference, April 21-24, Colorado
Springs, CO.

Mauderly, JL; Banas, DA; Griffith, WC; et al. (1996)  Diesel exhaust is not a pulmonary carcinogen in CD-I
mice exposed under conditions carcinogenic to F344 rats. Fundam Appl Toxicol 30:233-242.

McClellan, R. (1997) Use of mechanistic data in assessing human risks from exposure to particles.  Environ
Health Perspect 105(Suppl 5): 1-17.

McLaughlin, JK; Dietz, MS; Mehl, ES; et al. (1987) Reliability of surrogate information on cigarette smoking by
type of informant. Am J Epidemiol 126:144-146.

Miles, PR; Lee, P; Thrush, MA; et al. (1977) Chemiluminescence associated with phagocytosis of foreign
particles in rabbit alveolar macrophages. Life Sci 20:165-170.

Mohr, U; Takenaka, S; Dungworth, DL. (1986) Morphologic effects of inhaled diesel engine exhaust on lungs of
rats: comparison with effects of coal oven flue gas mixed with pyrolized pitch. In: Carcinogenic and mutagenic
effects of diesel engine exhaust. Ishinishi, N; Koizumi, A; McClellan, R;  et al., eds. Amsterdam: Elsevier; pp.
459-470.

Mokler, BV; Archibeque,  FA; Beethe, RL; et al. (1984)  Diesel exhaust exposure system for animal studies.
Fundam Appl Toxicol 4:270-277.

Mumford, JL; Chapman, RS; Harris, DB. (1989) Indoor air exposure to coal and wood combustion emissions
associated with a high lung cancer rate in Xuan Wei, China. Environ Int  15:315-320.

Murphy, SAM; Berube, KA; Richards, RJ. (1999) Bioreactivity of carbon black and diesel-exhaust particles to
primary Clara and type II  epithelial cell cultures.  Occup Environ Med 56:813-819.

Muscat, JE; Wynder, EL.  (1995) Diesel engine exhaust and lung cancer: an unproven association. Environ Health
Perspect 103:812-818.

Nagashima, M; Kasai, H;  Yokota, J; et al. (1995) Formation of an oxidative DNA damage,  8-
hydroxydeoxyguanosine, in mouse lung DNA after intratracheal instillation of diesel exhaust particles and effects
of high-dietary fat and beta-carotene on this process. Carcinogenesis 16:1441-1445.

National Institute for Occupational Safety and Health (NIOSH). (1988) Carcinogenic effects of exposure to DE.
NIOSH  Current Intelligence Bulletin 50. DHHS (NIOSH) Publication No. 88-116. Atlanta,  GA: Centers for
Disease Control.

National Toxicology Program (NTP). (2000) 9th report on carcinogens. Public Health Service, U.S. Department of
Health and Human Services, Research Triangle Park, NC.  Available from: http://ntp-server.niehs.nih.gov.

Nesnow, S; Evans, C; Stead, A; et al. (1982) Skin carcinogenesis studies of emission extracts. In: Toxicological
effects of emissions from  diesel engines. Lewtas,  J, ed. Amsterdam: Elsevier; pp. 295-320.

Nesnow, S; Triplett, LL; Slaga, TJ. (1984) Tumor initiating activities of 1-nitropyrene and its nitrated products in
Sencarmice. Cancer Lett  23:1-8.

Nielsen, PS; Autrup, H. (1994) Diesel exhaust-related DNA adducts in garage workers. Clin Chem 40:1456-1458.


                                                7-153

-------
Nielsen, PS; Andreassen, A; Farmer, PB; et al. (1996a) Biomonitoring of diesel exhaust-exposed workers. DNA
and hemoglobin adducts and urinary 1-hydroxypyrene as markers of exposure.  Toxicol Lett 86:27-37.

Nielsen, PS; de Pater, N; Okkels, H; et al. (1996b) Environmental air pollution and DNA adducts in Copenhagen
bus drivers—effect of (incomplete entry)

Nikula, KJ; Snipes, MB; Barr, EB; et al. (1995) Comparative pulmonary toxicities and carcinogenicities of
chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl Toxicol 25:80-94.

Notani, PN; Shah, P; Jayant, K; et al. (1993) Occupation and cancers of the lung and bladder: a case-control study
in Bombay. Int J Epidemiol 22:185-191.

Oberdorster, G; Yu, CP. (1990) The carcinogenic potential of inhaled diesel exhaust: a particle effect? J Aerosol
Sci21(suppl. 1):S397-S401.

Oberdorster, G; Baggs,  R; Gelein, R; et al. (1995) Pulmonary effects of inhaled carbon black in rats. Toxicologist
15:46.

Orthoefer, JG; Moore, W; Kraemer, D. (1981) Carcinogenicity of diesel exhaust as tested in strain A mice.
Environ Int 5:461-471.

Pepelko, WE; Peirano, WB. (1983) Health effects of exposure to diesel  engine emissions. J Am Coll Toxicol
2:253-306.

Percy,  C; Stanek, E, III; Gloeckler, L. (1981) Accuracy of cancer death certificates and its effect on cancer
mortality statistics. Am J Public Health 71:242-250.

Phillips, BJ; James, TEB; Anderson, D. (1984) Genetic damage in CHO cells exposed to enzymatically generated
active oxygen species. Mutat Res  126:265-271.

Pope, CA, III; Thun, MJ; Namboordiri, MM; et al. (1995) Paniculate air pollution as a predictor  of mortality in a
prospective study of U.S. adults. Am J Respir Crit Care Med 151:669-674.

Pott, F; Stober, W. (1983) Carcinogenicity of airborne combustion products observed in subcutaneous tissue and
lungs of laboratory rodents. Environ Health Perspect 47:293-303.

Pott, F; Heinrich, U. (1990) Relative significance of different hydrocarbons for the carcinogenic potency of
emissions from various  incomplete combustion processes. In: complex mixtures and scientific risk. Scientific
publication 104. Vaino, A; Sorsa, M; McMichael, AJ; eds. Lyon, France: IARC; pp. 288-297.

Pott, F; Heinrich, U; Vainio, H; et al. (1990) Complex mixtures and cancer risk.  In: Relative significance of
hydrocarbons for the carcinogenic potency of emissions from various incomplete combustion processes. IARC,
288-297.

Pott, F; Dungworth, DL; Heinrich, U; et al. (1994) Lung tumours in rats after intratracheal instillation of dusts.
Ann OccupHyg 38:357-363, Suppl. 1.

Qu, SX; Leigh, J; Koelmeyer, H; et al. (1996) Detection of DNA adducts in coal miners exposed to diesel
emissions in Australia. Fundam Appl Toxicol 50:237 (abstr).

Raffle, PAB. (1957) The health of the worker. Br J Ind Med 14:73-80.

Riebe-Imre, M; Aufderheide, M; Gartner-Hubsch, S; et al. (1994) Cytotoxic and genotoxic effects of insoluble
poarticles in vitro. In: Toxic and carcinogenic effects of solid particles in the respiratory  tract. Mohr, U;
Dungworth, DL; Mauderly, DL; et al., eds. Washington, DC: International Life Sciences Institute Press, pp. 519-
523.
                                                7-154

-------
Risch, HA; Burch, JD; Miller, AB; et al. (1988) Occupational factors and the incidence of cancer of the bladder in
Canada. Br J Ind Med 45:361-367.

Rister, M; Baehner, RL. (1977) Effect of hyperoxia on superoxide anion and hydrogen peroxide production of
polymorphonuclear leukocytes and alveolar macrophages. Br J Haematol 36:241-248.

Rittinghausen, S; Mohr, U; Dungworth, DL. (1997) Pulmonary cystic keratinizing squamous cell lesions after
inhalation/instillation of different particles. Exper Toxicol Pathol 49:433-446.

Rogot, E; Reid, DD. (1975) The validity of data from next-of-kin in studies of mortality among migrants. Int J
Epidemiol 4:51-54.

Root, RK; Metcalf, JA. (1977) H2O2 release from human granulocytes during phagocytosis. J Clin Invest
60:1266-1279.

Rosenblatt, MB; Teng, PK; Kerpe, S; etal. (1971) Causes of death in 1,000 consecutive autopsies. NY State J
Med 71:2189-2193.

Rosenkranz, HS. (1996) Mutagenic nitroarenes, diesel emissions, particle-induced mutations and cancer: an essay
on cancer causation by a moving target. Mutat Res 367:65-72.

Rosenkranz, HS; Mermelstein, R. (1983) Mutagenicity and genotoxicity of nitroarenes: all nitro-containing
chemicals were not created equal. Mutat Res  114:217-267.

Rothman, KJ. (1986) Modern epidemiology. Boston/Toronto: Little, Brown and Co.; pp. 16-21.

Rushton, L; Alderson, MR. (1983) Epidemiological survey of oil distribution centres in Britain. Br J Ind Med
40:330-339.

Rushton, L; Alderson, MR; Nagarajah, CR. (1983) Epidemiological survey of maintenance workers in London
Transport Executive bus garages at Chiswick Works. Br J Ind Med 40:340-345.

Sagai, M; Saito, H; Ichinose, T; et al. (1993) Biological effects of diesel exhaust particles. I. In vitro production of
superoxide and in vivo toxicity in mouse. Free Radical Biol Med 14:37-47.

Salmeen, IT; Pero, AM; Zator, R; et al. (1984) Ames assay chromatograms and the identification of mutagens in
diesel exhaust particle extracts.  Environ Sci Technol 18:375-382.

Sandusky, CB; Cowden, MW; Schwartz, SL. (1977) Effect of particle size on regurgitive exocytosis by rabbit
alveolar macrophages. In: Pulmonary macrophages and epithelial cells. CONF-760927, National Technical
Information Service, U.S. Dept. of Commerce, pp. 85-105.

Saverin. R; Braunlich, A; Dahman, D; et al. (1999) Diesel exhaust and lung cancer mortality in potash mining.
Am J Ind Med 36:415-422.

Scheepers, PJT; Bos, RP. (1992) Combustion of diesel fuel from a toxicological perspective. II. Toxicity. Int
Arch Occup Environ Health 64:163-177.

Schenker, MB; Smith, T; Munoz, A; et al. (1984) Diesel exposure and mortality among railway workers: results of
a pilot study. Br J Ind Med 41:320-327.

Schottenfeld,  D; Eaton, M; Sommers,  SC; et al. (1982) The autopsy as a measure of accuracy of the death
certificate. Bull NY Acad Med 58:778-794.

Sera, N; Fukuhara, K; Miyata, N; et al. (1994) Detection of nitro-azabenzol[a]pyrene derivatives in the
semivolatile phase originating from aiborne paniculate matter, diesel and gasoline vehicles. Mutagenesis 9:47-52.
                                                7-155

-------
Shefner, AM; Collins, SC; Dooley, L; et al. (1982) Respiratory carcinogenicity of diesel fuel emissions interim
results. In: Toxicological effects of emissions from diesel engines. Lewtas, J, ed. Amsterdam: Elsevier; pp. 329-
350.

Siemiatycki, J; Germ, M;  Stewart, P; et al. (1988) Associations between several sites of cancer and ten types of
exhaust and combustion products: results from a case-referent study in Montreal. Scand J Work Environ Health
14:79-90.

Silverman, DT; Hoover, RN; Albert, S; et al. (1983) Occupation and cancer of the lower urinary tract in Detroit. J
Natl Cancer Inst 70:237-245.

Silverman, DT; Hoover, RN; Mason, TJ; et al. (1986) Motor exhaust-related occupations and bladder cancer.
Cancer Res 46:2113-2116.

Steenland, K; Burnett, C;  Osoria, AM. (1987) A case-control study of bladder cancer using city directories as a
source of occupational data. Am J Epidemiol 126:247-257.

Steenland, NK; Silverman, DT; Hornung, RW. (1990) Case-control study of lung cancer and truck driving in the
Teamsters Union. Am J Public Health 80:670-674.

Steenland, K; Deddens, J; Stayner, L. (1998) Diesel exhaust and lung cancer in the trucking industry:  exposure-
response analyses and risk assessment.  AmJIndMed  34:220-228.

Steineck, G; Plato, N; Gerhardsson,  M; et al. (1990) Increased risk of urothelial cancer in Stockholm during 1985-
87 after exposure to benzene and exhausts. Int J Cancer 45:1012-1017.

Stenback, F; Rowland, J; Sellakumar, A. (1976)  Carcinogenicity of benzo[a]pyrene and dusts in the hamster lung
(instilled intratracheally with titanium oxide, aluminum oxide, carbon and ferric oxide). Oncology 33:29-34.

Stern, FB; Lemen, RA; Curtis, RA. (1981) Exposure of motor vehicle examiners to carbon monoxide:  a historical
prospective mortality study. Arch Environ Health 36:59-66.

Stober, W. (1986) Experimental induction of tumors in hamsters, mice and rats after long-term inhalation of
filtered and unfiltered diesel engine  exhaust. In:  Carcinogenic and mutagenic effects of diesel engine exhaust.
Ishinishi, N; Koizumi, A;  McClellan, R; et al., eds. Amsterdam: Elsevier; pp. 421-429.

Stober, W; Abel, UR. (1996) Lung cancer due to diesel soot particles in ambient air? A critical appraisal of
epidemic logical studies addressing this question. Int Arch Occup Environ Health 68(Suppl):S3-S61.

Strom, KA. (1984) Response of pulmonary cellular defenses to the inhalation of high concentrations of diesel
exhaust. J Toxicol Environ Health 13:919-944.

Swanson, GM; Lin, CS; Burns, PB.  (1993) Diversity in the association between occupation and lung cancer
among black and white men. Cancer Epidemiol Biomarkers Prev 2:313-320.

Takaki, Y; Kitamura, S; Kuwabara,  N; et al. (1989) Long-term inhalation studies of exhaust from diesel engine in
F-344 rats: the quantitative relationship between pulmonary hyperplasia and anthracosis. Exp Pathol 37:56-61.

Takemoto, K; Yoshimura, H; Katayama, H. (1986) Effects of chronic inhalation exposure to diesel exhaust on the
development of lung tumors in di-isopropanol-nitrosamine-treated F344 rats and newborn C57BL and  ICR mice.
In: Carcinogenic and mutagenic effects of diesel engine exhaust. Ishinishi, N; Koizumi, A; McClellan, R; et al.,
eds. Amsterdam: Elsevier; pp. 311-327.

Takemoto, K; Katayama, H; Kuwabara, T; et al. (1988) Carcinogenicity by subcutaneous administration of diesel
paniculate extracts in mice. In: Diesel exhaust and health risks. Ibaraki, Japan:  Research Committee for HERP
Studies; pp. 227-234.
                                                7-156

-------
Thyssen, J; Althoff, J; Kimmerle, G; et al. (1981) Inhalation studies withbenzo[a]pyrene in Syrian golden
hamsters. J Natl Cancer Inst 66:575-577.

Trotter, JR. (1980) Spontaneous cancer and its possible relationship to oxygen metabolism.  Proc Natl Acad Sci
77:1763-1767.

U.S. Environmental Protection Agency (U.S. EPA). (1986) Guidelines for carcinogen risk assessment. Federal
Register 51(185):33992-34003.

U.S. EPA. (1993) Motor vehicle-related air toxics study.  Ann Arbor, MI: Office of Mobile Sources. EPA/420/R-
93/005. Available from National Technical Information Service, Springfield, VA, PB93-182590/XAB.

U.S. EPA. (1996a) Air quality criteria for paniculate matter.  Prepared by the National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC. EPA/600/P-95/001bF.

U.S. EPA. (1996b) Proposed guidelines for carcinogen risk assessment. Office of Research and Development.
Federal Register 61(79): 17960-18011. EPA/600/P-92/003C.

U.S. EPA. (1999) Guidelines for carcinogen risk assessment. Review Draft. NCEA-F-0644, July. Risk Assessment
Forum, Washington, DC.  http://www.epa.gov/ncea/raf/cancer.htm.

Vineis, P; Magnani, C. (1985) Occupation and bladder cancer in males: a case-control study. Int J Cancer 35:599-
606.

Vostal, JJ. (1986) Factors limiting the evidence for chemical carcinogenicity of diesel emissions in long-term
inhalation experiments. In: Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the
international satellite symposium on toxicological effects of emissions from diesel engines; July 1986; Tsukuba
Science City,  Japan. Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. (Developments in toxicology and
environmental science: v.  13.) Amsterdam: Elsevier Science Publishers BV; pp. 381-396.

Waller, RE. (1981) Trends in lung cancer in London in relation to exposure to diesel fumes. Environ Int 5:479-
483.

Ward, JM. (1983) Background data and variations in tumor rates of control rats and mice. Prog Exper Tumor Res
26:241-258.

Waxweiler, RJ; Wagoner,  JK; Archer, VE. (1973) Mortality of potash workers. J Occup Med 15:486-489.

Weinstein, IB; Troll, W. (1977) National Cancer Institute workshop on tumor promotion and cofactors in
carcinogenesis. Cancer Res 37:3461-3463.

Weiss, SJ; Rustagei, PK; LoBuglio, AF. (1978) Human granulocyte generation of hydroxyl radical.  J Exp Med
147:316-323.

Weitzman,  SA; Stossel, TP. (1981) Mutation caused by human phagocytosis.  Science 212:546-547.

White, HJ; Vostal, JJ; Kaplan, HL; et al. (1983) A long-term inhalation study evaluates the pulmonary effects of
diesel emissions. J Appl Toxicol 1:332.

Wichmann, HE; Kreienbrock, L; Kreuzer, M; et al. (1998) Lungenkrebrisiko durch Radon in der Bundesrepublik
Deutschland (West).  In: Fortschritte in der Umwelmedizin, ecomed Verlagsgesellschaft
Landsberg/Lech.Wichmann, HE; Schliepkoter, HW; Fulgraff, G; et al., eds.

Williams, RR; Stegens, NL; Goldsmith, JR. (1977) Associations of cancer site and type with occupation and
industry from the Third National Cancer Survey interview. J Natl Cancer Inst 59:1147-1185.

Wong, O; Morgan, RW; Kheifets, L; et al. (1985) Mortality among members  of a heavy construction equipment
operators union with potential exposure to diesel exhaust emissions. Br J Ind Med 42:435-448.

                                                7-157

-------
Woskie, SR; Smith, TJ; Hammond, SK; et al. (1988a) Estimation of the diesel exhaust exposures of railroad
workers: I. Current exposures. Am J Ind Med 13:381-394.

Woskie, SR; Smith, TJ; Hammond, SK; et al. (1988b) Estimation of the diesel exhaust exposures of railroad
workers: II. National and historical exposures. Am J Ind Med 13:395-404.

Wynder, LE; Dieck, G; Hall, NE.  (1985) A case control study of diesel exhaust exposure and bladder cancer.
Environ Res 34:475-489.

Zaebst, DD; Clapp, DE; Blade, LM; et al. (1991) Quantitative determination of trucking industry workers'
exposures to diesel exhaust particles. Am Ind Hyg Assoc J 52:529-541.

Zamora, PO; Gregory, RE; Brooks, AL. (1983) In vitro evaluation of the tumor-promoting potential of diesel-
exhaust-particle extracts. J Toxicol Environ Health 11:188-197.
                                                7-158

-------
       8.  DOSE-RESPONSE ASSESSMENT:  CARCINOGENIC EFFECTS

8.1.  INTRODUCTION
       Dose-response assessment for carcinogenicity defines the relationship between the
exposure/dose of an agent and the degree of carcinogenic response, and evaluates potential
cancer risks to humans at exposure/dose levels of interest.  Most often, the exposure/dose
response of interest is well below the range of observation. As a result, dose-response
assessment usually entails an extrapolation from the generally high exposures in studies on
humans or laboratory animals to the exposure levels expected from human contact with the agent
in the environment. It also includes considerations of the scientific validity of these
extrapolations based on available knowledge about the underlying mechanisms or modes of
carcinogenic action. The complete sequence of biological events that must occur to produce an
adverse effect is defined as "mechanism of action."  In cases where only partial  information is
available,  the term "mode of action" is used to refer to the mechanisms for key events that are
judged to be sufficient to inform about the shape of the dose-response curve beyond the range of
observation.
       This chapter evaluates the available exposure/dose-response data and discusses
extrapolation issues in estimating the cancer risk of environmental exposure to diesel engine
exhaust (DE).  It concludes that available data are inadequate to confidently derive a cancer unit
risk estimate for DE or its component, diesel particulate matter (DPM). Unit risk is one possible
output from a dose-response assessment and is defined as the estimated upper-bound  cancer risk
at a specific exposure or dose from a continuous average lifetime exposure to a carcinogen (in
this case, cancer risk per |lg/m3 of DPM).  In lieu of unit-risk-based quantitative risk estimates,
this chapter provides a perspective about potential risk at environmental levels.  Subsequent
sections of this chapter discuss issues related to dose-response evaluation of human cancer risk
for DE exposure, including the target tumor site and underlying mode of action, suitable
measures of dose, approaches to low-dose extrapolation, and appropriate data to be used in the
dose-response analysis.  This is followed by a simple analysis  of the possible degree and extent
of risk from environmental exposure to DE.
       Appendix C provides a summary review of dose-response assessments conducted to date
by other organizations and investigators.  These risk estimations were performed on the basis of
either epidemiologic and/or experimental data. As concluded in Section 8.5, EPA finds that
available epidemiologic data are too uncertain to confidently derive a unit risk estimate for DE-
induced lung cancer, and that rat data are not suitable for estimating human risk. Nevertheless, a
review of dose-response evaluations is provided in the appendix for historical context.
                                           8-1

-------
8.2.  MODE OF ACTION AND DOSE-RESPONSE APPROACH
       According to EPA's 1996 Proposed Guidelines for Carcinogen Risk Assessment (U.S.
EPA, 1996), dose-response assessment is performed in two steps: assessment of observed data
to derive a point of departure, followed by extrapolation to lower exposures to the extent
necessary. Human data are always preferred over animal data, if available, as their use obviates
the need for extrapolation across species. Mode-of-action information is important to dose-
response evaluation, as it informs about the relevance of animal data to assessment of human
hazard and risk, the shape of the dose-response curve at low doses, and the most appropriate
measure(s) of exposure/dose and response.
       If there are sufficient quantitative data (humans and/or animals) and adequate
understanding of the carcinogenic process, the preferred approach is to use a biologically based
model for both the range of observation and extrapolation below that range. Otherwise, as a
default procedure, a standard mathematical model is used to curve-fit the observed dose-
response data to obtain a point of departure, which is the lower 95% confidence limit of the
lowest exposure/dose that is associated with a selected magnitude of excesses of cancer risk in
human or animal studies.  Default approaches for low-dose extrapolation should be consistent
with the current understanding of the mode(s) of action.  These include approaches that assume
linearity or nonlinearity, or both.  Linear extrapolation is used when there is insufficient
understanding of the modes of action, or the mode-of-action information indicates that the dose-
response curve at low dose is, or is expected to be, linear. Linear extrapolation involves the
calculation of the slope of the line drawn from the point of departure to zero exposure or dose
(i.e., above background).  When there is sufficient evidence for a nonlinear mode of action but
not enough data to construct a biologically based model for the relationship, a margin of
exposure is used as a default approach. A margin-of-exposure analysis compares the point of
departure (i.e., the lowest exposure associated with some cancer  risk) with the dose associated
with the environmental exposure(s) of interest and determines whether or not the exposure
margins are adequate.  Both default approaches may be used for  a tumor response if it is
mediated by linear and nonlinear modes of action. The dose-response approaches considered in
this chapter follow the principles of EPA's guidelines for carcinogen risk assessment (U.S. EPA,
1986, 1996, 1999).
       As reviewed in Chapter 7, there is substantial evidence from combined human and
experimental evidence that DE is likely to pose a cancer hazard to humans at anticipated levels
of environmental exposure. The critical target organ is the lung.  Evidence exists for a causal
relationship between risk for lung cancer and occupational exposure to DE in certain
occupational workers such as railroad workers, truck drivers, heavy equipment operators, transit
workers, etc.
                                          8-2

-------
       The mechanism(s) by which DE induces lung cancer in humans has not been established.
As discussed in Chapter 7, Section 7.4, several modes of action have been postulated on the
basis of available mechanistic studies, including direct DNA effects (gene mutations) by the
adsorbed organic compounds and the gaseous fractions, indirect DNA effects (e.g., chromosomal
aberrations, sister chromatid exchange [SCE], micronuclei) by DE and DPM, oxidative DNA
damage by DPM via release of reactive oxygen species (ROS), and particle-induced chronic
inflammatory response leading to epithelial cell cytotoxicity and regenerative cell proliferation
via release of cytokines, growth factors, and ROS. It is likely that a combination of modes of
action contributes to the overall carcinogenic activity of DE, and that the relative contribution of
the various modes of action may vary with different exposure levels.
       In the absence of a full understanding of the relative roles of DE constituents in inducing
lung cancer in humans,  and because there is some evidence for a mutagenic mode of action,
linear low-dose extrapolation is an appropriate and prudent default choice for modeling dose-
response, and if needed, risk extrapolation from high to lower exposures (U.S. EPA, 1986, 1996,
1999). It also should be noted that there are not order of magnitude differences between lower
levels of occupational and higher end environmental exposure estimates. In fact, there appears
to be exposure overlap.  This means that an extrapolation of the occupational hazard to lower
environmental exposure levels is minimal. Other individuals and organizations have used either
linear risk extrapolation models and/or mechanistically based models to estimate cancer risk
from environmental exposure to DE (e.g., IPCS, 1996; Cal EPA, 1998; also see Appendix C).
These were examined but not found to provide a compelling basis for unit risk derivation
because of database uncertainty and/or recent understandings about the suitability of the rat data
for modeling a dose-response at environmental levels of exposure.
       For example, there are an observable series of events showing how DE causes lung
tumors in the rat under high exposure experimental conditions. Prolonged exposure to high
concentrations of a variety of poorly soluble particles including DPM (and its carbon core,
devoid of organics) causes lung tumors in rats through a mode of action that involves impairment
of lung clearance mechanisms (referred to as "lung overload response"), leading to persistent
chronic inflammation, cell proliferation, metaplasia, and ultimately the development of lung
tumors (ILSI, 2000). Because this  mode of action is not expected to be operative at
environmental exposure conditions, the rat lung tumor dose-response data are not considered
suitable for predicting human risk at low environmental exposure concentrations.

8.3. USE OF EPIDEMIOLOGIC STUDIES FOR QUANTITATIVE RISK ASSESSMENT
       As discussed above, human data are considered more appropriate than animal data in
estimating environmental  cancer risk for DE.  Still, there are many uncertainties in using the

                                          8-3

-------
available epidemiologic studies that have quantitative exposure data to extrapolate the risk to the
general population for ambient-level DE exposure.

8.3.1.  Sources of Uncertainty
       The greatest uncertainty in estimating DE-induced cancer risk from epidemiologic
studies is the lack of knowledge of actual historical exposures for individual workers,
particularly for the early years. Reconstruction of historic exposures is based on job exposure
categories, industrial hygiene measurements, and assumptions made about exposure patterns.
       Another related uncertainty is the choice of markers of exposure to DE. As discussed
above, the modes of action for DE-induced lung cancer in humans are not fully understood, and
thus the best measure of DE exposure is unknown. Various markers of DPM (e.g., respirable-
sized particles, elemental carbon [EC]) have been used as dosimeters for DE.  Though EC is
more sensitive and more specific than respirable-sized particles, both are considered appropriate
dosimeters.  Related to the choice of dosimeter, having a relatively constant relationship between
the organics (on the particle) and the particle mass would be consistent with a possible mode-of-
action role for both the particle and organic components. However, evidence of such a constant
historic relationship remains unclear.  As discussed in Chapter 2 (Section 2.5.2), it appears that
newer model on-road engine exhaust  has a lesser quantity of organics adsorbed onto the particle
compared to older model engines.  On the other hand, with regard to DE in the ambient air,
there is significant variation in the amounts of DPM organic components emitted because of
aged vehicles in the on-road fleet, driving patterns, and the additional presence of nonroad DE
(e.g., marine vessels and locomotives, which generally use older technology than on-road
engines).
       Another major uncertainty associated with many of the DE epidemiologic studies was the
inability to fully control for smoking effects, resulting in possible errors in estimating relative
risk increases.  Changes in adjustments for smoking could result in considerable changes in
relative risk, because smoking has a much larger effect on relative lung cancer risk than is likely
for DE. It is difficult to effectively control for a smoking effect in a statistical analysis because
cigarette smoke contains an array of biologically active compounds and affects multiple steps of
carcinogenesis, thus probably making smokers more susceptible to DE-induced lung cancer than
are nonsmokers. A statistical analysis would not be able to  adjust for such an  effect without
having a detailed record of the smoking history of individuals.
       A potential uncertainty involves the use of occupational worker data to extrapolate cancer
hazard risk to the general population and sensitive subgroups. By sex, age, and general health
status, workers are not fully representative of the general population. For example, there is
virtually no information to determine whether infants and children or people in poor health

                                           8-4

-------
respond differently to DE exposure than do workers.  Finally, the use of linear low-dose
extrapolation may contribute to uncertainty in estimating environmental risks.

8.3.2.  Evaluation of Key Epidemiologic Studies for Potential Use in Quantitative Risk
       Estimates
       Among the available epidemiologic studies, only the railroad worker studies and the
Teamster truck driver studies have reconstructed quantitative historical exposure data for
possible use in deriving a unit risk estimate for DE-induced lung cancer.  This section evaluates
the strengths and limitations of these data and their suitability for dose-response analysis.

8.3.2.1. Railroad Worker Studies
       Garshick and colleagues conducted both cohort and case-control studies of lung cancer
mortalities among U.S. railroad workers registered with the U.S. Railroad Retirement Board
(RRB).
       In the cohort study (Garshick et al., 1988), lung cancer mortality was ascertained through
1980 in 55,407 railroad workers, age 40 through 64 in 1959, with at least 10 years of work in
selected railroad jobs (39 job titles).  The cohort was selected on the basis of job titles in 1959.
Industrial hygiene evaluations and descriptions of job activities were used to classify jobs as
exposed or unexposed to diesel emissions. Workers with recognized asbestos exposure were
excluded from the job categories selected for study. However, a few jobs with some potential
for asbestos  exposure were included in the cohort. Each subject's work history was determined
from a yearly job report filed by his employer with the RRB from 1959 until death or retirement.
The year 1959 was chosen as the effective start of DE exposure for this study because by this
time 95% of the locomotives in the United States were diesel powered. The author reported
statistically significant relative risk increases of 1.57 for the 40-44 year age group and  1.34 for
the 45-49 year age group, after exclusion of workers exposed to asbestos and controls for
smoking.  Age groups were determined by their ages in 1959.
       A main strength of the cohort study is the large sample size (55,407), which allowed
sufficient power to detect small risks. This study also permitted the exclusion of workers with
potential past exposure to asbestos. The stability of job career paths in the cohort ensured that of
the workers  40 to 64 years of age in 1959 classified as DE-exposed, 94% of the cases were still
in DE-exposed jobs 20 years later.
       The main limitation of the cohort study is the lack of quantitative data on exposure to
DE. The number of years exposed to DE was used as a surrogate for dose.  The dose, based on
duration of employment, has inaccuracies because individuals were working on both steam and
diesel locomotives during the transition period.  It should be noted that the investigators included

                                          8-5

-------
only exposures after 1959; the duration of exposure prior to 1959 was not known.  Other
limitations of this study include its inability to examine the effect of years of exposure prior to
1959 and the less-than-optimal latency period for lung cancer expression. No adjustment for
smoking was made in this study.  For a detailed description of this study please refer to Chapter
7, Section 7.2.1.7.
       Garshick and colleagues also conducted a case-control study of railroad workers who
died of lung cancer between 1981 and 1982 (Garshick et al., 1987). The author reported
statistically significant increased  odds ratios (with asbestos exposure accounted for) of 1.41
(95% confidence interval [CI] = 1.06, 1.88) for the <64 year age group and 1.64 (95% CI = 1.18,
2.29) for the <64 year age group  with >20 years of exposure when compared with the 0-4 year
exposure group.  The population base for this case-control study was approximately 650,000
active and retired male U.S. railroad workers with 10 years or more of railroad service who were
born in 1900 or later. The cases were selected from deaths with primary lung cancer, which was
the underlying cause of death in most cases.  Each case was matched to two deceased controls
whose dates of birth were within  2.5 years of the date of birth of the case and whose dates of
death were within 31 days of the  date of death noted in the case. Controls were selected
randomly from workers who did not have cancer noted anywhere on their death certificates and
who did not die of suicide or of accidental or unknown causes.  A total of 1,256 cases and 2,385
controls were selected for the study. Among younger workers,  approximately 60% had exposure
to DE, whereas among older workers, only 47% were exposed to DE. DE exposure surrogates
for workers were similar to those in the cohort study. Asbestos exposure was categorized on the
basis of jobs held in 1959, or on the last job held if the subject retired before 1959.  Smoking
history information was obtained from the next  of kin.
       The strengths of the case-control study are consideration of confounding factors such as
asbestos exposure and smoking; classification of DE exposures by job titles and industrial
hygiene sampling; and exploration of interactions between smoking, asbestos exposure, and DE
exposure.  Major limitations of this study include (a) possible overestimation of cigarette
consumption by  surrogate respondents; (b) use of the Interstate Commerce Commission (ICC)
job classification as a surrogate for exposure, which may have led to misclassification of DE
exposure jobs  with low intensity and intermittent exposure, such as railroad police and bus
drivers, as unexposed; (c) lack of data on the contribution of unknown occupational or
environmental exposures and passive smoking;  and (d) a suboptimal latency period of 22 years,
which may not be long enough to observe a full expression of lung cancer. For a detailed
description of this study, please see Chapter 7, Section 7.2.2.4.
       As a part of these epidemiologic studies, Woskie et al. (1988a) conducted an industrial
hygiene survey in the early 1990s for selected jobs in four small northern railroads. DE

                                          8-6

-------
exposure was considered as a yes/no variable based on job in 1959 and estimated years of work
in a diesel- exposed job as an index of exposure. Thirty-nine job titles were originally identified
and were then collapsed into 13 job categories and, for some statistical  analyses, into 5
categories (clerks, signal maintainers, engineers/firers, brakers/conductors/hostlers, and shop
workers) (Woskie et al., 1988b; Hammond et al., 1988).  As discussed below, these exposure
estimations were used by Crump et al. (1991) and by Cal EPA (1998) for their dose-response
analyses.

8.3.2.1.1.  Potential for the data to be used for dose-response modeling. Both case-control and
cohort studies can be used for dose-response analysis if exposure for each worker is available.
Control of a smoking effect is important when lung cancer is the disease of interest. However, as
discussed previously (see Section 8.3.1), one may not be able to control smoking completely in a
dose-response analysis because of the lack of detailed records of the smoking history of
individuals.
       Garshick et al. (1988) reported a positive relationship of relative risk and duration of
exposure by modeling age in 1959 as a covariate in an exposure-response analysis.  The positive
relationship disappeared when attained age was used instead of age in 1959, and a negative dose-
response was observed (Crump et al., 1991). This negative dose-response continued to be
upheld in a subsequent reanalysis (Crump, 1999). Garshick (letter from Garshick, Harvard
Medical School, to Chao Chen, U.S. EPA, dated August  15, 1991) performed further analysis
and reported that the relationship between years of exposure and risk of lung cancer, when
adjusted for attained age and calendar year, was flat to negative depending upon which model
was used. In contrast, California EPA (Cal EPA, 1998) found a positive dose-response by using
age in 1959 but allowing for an interaction term of age and calendar year in the model.
       Crump et al. (1991)  also found, and Garshick (letter from Garshick, Harvard Medical
School, to Chao Chen, U.S. EPA, dated August 15, 1991) confirmed, that in the years 1977-1980
the death ascertainment was not complete. About 20% to 70% of deaths were unaccounted for,
depending upon the calendar year. Further analysis, based on job titles in 1959 and limited to
deaths occurring through 1976, showed that the youngest workers still had the highest risk of
dying of lung cancer.
       Extensive statistical  analyses were conducted by a panel convened by HEI (1999) to
investigate the utility of the railroad worker cohort for use in dose-response based quantitative
risk assessment. Seven models were used to test the data, and the models were formed by
varying a number of covariates in different combinations. The covariates included employment
duration, cumulative exposure with and without correction for background exposure, and three
job categories: clerks and signalmen, train workers (which include engineers/firers/brakers/

                                          8-7

-------
conductors), and shop workers.  The coefficient for each covariate in a model is used to calculate
relative risk for the associated covariate.  In summary, the panel found that effects of exposure as
defined by an exposure-response curve were either flat or negative in all of the models. In these
analyses, relative risk for each job category was assumed to be constant with respect to age.
Further exploration of the data showed that the relative risk for train workers was not constant.
The panel's statistical analyses also revealed the complexity of the data and difficulties of
providing an adequate summary measure of effect, probably because calendar year and
cumulative exposure are highly correlated, which makes it especially difficult to sort out their
separate effects. The difficulty of providing an adequate measure of DE effect was further
demonstrated in Table C.3 of the HEI report, in which negative or positive effects for cumulative
exposure (with background exposure adjustment) were obtained depending on whether or not job
category was included in the model.  A similar review of the divergent views about the railroad
worker dose response also can be found in Chapter 7, Section 7.2.1.7.
       The diverging results about the presence or absence of exposure response for the railroad
worker data have become a source of continuing debate about the suitability of these data for
estimating DE cancer risk. Although it is difficult to identify the exact reason for the diverging
findings, the "age effect" appears to be a main source of uncertainty because age, calendar year,
and cumulative exposure are not mutually independent. Therefore, an ideal dose-response
analysis would account for the ages when exposure to DE began and terminated, along with the
attained age and other covariates for each person, using age-dependent exposure concentration
rather than cumulative exposure over lifetime as a dosimeter.  This analysis would be possible
for the railroad workers if information were available on the ages when exposure began and
terminated.
       Given the equivocal evidence for positive  exposure response, EPA has not derived a unit
risk on the basis of the available railroad worker data.  This determination should not be
construed, however, to imply that the railroad worker studies contain no useful information on
lung cancer risk from exposure to DE.

8.3.2.2. Teamsters Union Trucking Industry  Studies
       Steenland et al. (1990) conducted a case-control study of lung cancer deaths in the
Central States Teamsters Union to determine the risk of lung cancer among different trucking
industry occupations.  The study found statistically significant increased odds ratios for lung
cancer of  1.89 and 1.64, depending on years of employment. Cases comprised all deaths from
lung cancer (1,288).  The 1,452 controls comprised every sixth death from the entire file,
excluding deaths from lung cancer, bladder cancer, and motor vehicle accidents. Individuals
were required to have 20 years tenure in the union to be eligible to claim benefits.

-------
       Detailed information on work history and potential confounders such as smoking, diet,
and asbestos exposure was obtained by questionnaire.  On the basis of interview data and the
1980 census occupation and industry codes, subjects were classified either as nonexposed or as
having held other jobs with potential DE exposure. The Teamsters Union work history file did
not have information on whether men drove diesel or gasoline trucks,  and the four principal
occupations were long-haul drivers, short-haul or city drivers, truck mechanics, and
dockworkers.  Subjects were assigned the job category in which they had worked the longest.
       The main strengths of the study are the availability of detailed records from the
Teamsters Union, a relatively large sample size, availability of smoking data, and measurement
of possible asbestos exposures.  Some limitations of this study include possible
misclassifications of exposure and smoking habits, as information was provided by next-of-kin
and lack of sufficient latency to observe lung cancer excess.
       Steenland et al. (1998) conducted an exposure-response analysis by supplementing the
data from their earlier case-control study of lung cancer and truck drivers in the Teamsters Union
with exposure estimates based on a 1990 industrial hygiene survey of EC exposure (Zaebst et al.,
1991), a surrogate for DE in the trucking industry. Available data indicate that exposure to
workers in the trucking industry in 1990 averaged 2-27 |lg/m3 of EC.  The 1990 exposure
information was used by Steenland as a baseline exposure measurement to reconstruct past
exposure (in the period of 1949 to  1983) by assuming that the exposure for workers in different
job categories is a function of highway mileages traveled by heavy-duty vehicles and efficiency
of the engine over the years.
       The industrial hygiene survey by Zaebst et al. (1991) of EC exposures in the trucking
industry provided exposure estimates for each job category in 1990. The EC measurements were
generally consistent with the epidemiologic results, in that mechanics were found to have the
highest exposures and relative risk, followed by long-haul and short-haul drivers. Dockworkers
who had the lowest exposures also had the lowest relative risks.
       Past exposures were estimated assuming that they were a function of (a) the number of
heavy-duty trucks on the road, (b) the particulate emissions (grams/mile) of diesel engines over
time, and (c) leaks from truck exhaust systems for long-haul drivers. Estimates of past exposure
to EC (as a marker for DE exposure) were made based on the assumption that average 1990
levels for a particular job  category could be assigned to all subjects in that category, and that
levels prior to 1990 were directly proportional to vehicle miles traveled by heavy-duty trucks
and the estimated emission levels of diesel engines. For example, a 1975 exposure level was
estimated by the following equation: 1975 level = 1990 level x (vehicle miles 1975/vehicle miles
1990) x (emissions 1975/emissions 1990).  Once estimates of exposure for each year of work
history were derived for each subject, analyses were conducted by  cumulative level of estimated

                                           8-9

-------
carbon exposure.  As with most epidemiologic studies, the endeavors to reconstruct exposures
for epidemiologic studies are subject to uncertainties.

8.3.2.2.1.  Potential for the data to be used for dose-response modeling.  Steenland et al. (1998)
analyzed their case-control data and showed a significant positive trend in lung cancer risk with
increasing cumulative exposure to DE.  The study by Steenland et al. (1998) could provide a
valuable database for calculating unit risk for DE emissions. The strength of this data set is that
the smoking histories of workers were obtained to the extent possible. Smoking is especially
important  in assessing the lung cancer risk due to DE exposure because smoking has much
higher relative risk (or odds ratio) of lung cancer than does DE. In the Steenland et al. (1998)
study, the  overall (ever-smokers vs. nonsmokers) odds ratio for developing lung cancer from
smoking is about 7.2, which is about fivefold larger than the 1.4 relative risk increase from a
large synthesis of many DE epidemiologic studies.  It is possible that a modest change of
information on smoking and diesel exposure might alter the conclusion and risk estimate.
       Another strength of the Teamster data for use in environmental risk assessment for the
general  population is that exposures of Teamsters are closer to ambient exposures than are those
of railroad workers. The  Teamsters Union truck driver case-control workers had cumulative
exposure ranging from 19 to 2,440 |lg/m3-years of EC, with the median and 95th percentile,
respectively, of 358 and 754 (Ig/m3-years of EC.  The median and 95th percentile of an
environmentally equivalent exposure would be 3 and 6 |lg/m3, respectively.1 These
environmental equivalent exposures for the Teamsters Union truck drivers are close to the
estimated  ambient exposures of <1.0 |ig/m3 to 4.0 |ig/m3 (see Table 2-31).
       Steenland et al. (1998) stated that their risk assessment is exploratory because it depends
on estimates about unknown past exposures.  An EPA reanalysis of DE exposure for this study is
underway. With a revised exposure assessment, a reevaluation of the dose-response would be
appropriate. In a recent review, HEI (1999) concluded that the Teamsters studies may be useful
for quantitative risk assessment, but significant further evaluation and development are needed.
Given the  ongoing reanalysis of exposure, EPA will not, at this time, use the Steenland et al.
(1998) occupational risk assessment findings to derive equivalent environmental parameters and
cancer unit risk estimates.
        The conversion assumes (a) DPM = 40% EC as reported by Steenland et al. (1998), (b) environmental
equivalent exposure is approximately = 0.21 x occupational exposure, and (c) 70 |Ig/m3-years is equivalent to a lifetime
of exposure at 1 |Ig/m3.

                                           8-10

-------
8.3.3. Conclusion
       Even though available evidence supports a conclusion that DE is likely to be a human
lung carcinogen, the conclusion of the dose-response evaluation is that the available data are not
sufficient to confidently estimate a cancer unit risk or unit risk range.  The absence of such a
cancer unit risk for DE limits the ability to quantify, with confidence, the potential impact of the
hazard on exposed populations. Two significant short-term activities are underway to improve
the epidemiologic database for dose-response assessment: (1) a follow-up study to correct the
undercounting of mortality in the Garshick et al. (1988) railroad worker study, and (2) an EPA-
sponsored effort to improve the exposure estimates for Teamsters Union truck drivers (Steenland
et al., 1998). EPA will monitor this ongoing research as well as the ongoing NCI-NIOSH study
of nonmetal miners and the recently NCI-funded epidemiologic study  of truck drivers.  As these
studies or other new data become publicly available, EPA will reconsider the merit of
conducting additional dose-response analysis and unit risk derivation.

8.4.  PERSPECTIVES ON CANCER RISK
       Although the available data are considered inadequate to confidently estimate a cancer
unit risk, this does not mean that there is no information about the possible cancer risk of DE.
To examine the significance of the potential cancer hazard from environmental exposure to DE,
all relevant epidemiologic and exposure data as well as simple risk assessment tools can be used.
Such an approach does not produce confident estimates of cancer unit risk.  Rather,  these
exploratory approaches provide a perspective on the possible magnitude of cancer risk and thus
insight about the potential significance of the hazard. This section describes approaches and
methods that are used to gauge the magnitude of possible cancer risk from ambient exposure to
DE.
       The first approach involves examining the differences between the levels of occupational
and ambient environmental exposures, while assuming that cancer risk to DE is linearly
proportional with cumulative lifetime exposure.  Risks to the general public would be low in
comparison to occupational risk if the differences in exposure are large (e.g., about three orders
of magnitude or more). On the other hand, if the exposure differences are smaller (i.e., within
one to two orders of magnitude), environmental risks become more of a concern as they
approach the range of workers' risk observed in epidemiologic studies of past occupational
exposures.  This assumes that the carcinogenic potency of historical and current-day DE is not
significantly different, a reasonable assumption, though not without uncertainty.
       Table 8-1 shows occupational exposure estimates for some of the occupational groups
where increased relative risks of lung cancer (e.g., meta-analyses) have been analyzed. Given
that no statistical properties associated with these exposure estimates are known, their use here is

                                          8-11

-------
Table 8-1.  DPM exposure margins (ratio of occupational -=- environmental exposures)
Occupational group

Public transit

U.S. railroad workers

Fork Lift Operators

High end boundary

Estimated occupational
exposure/concentration
Environmental equivalent a
15-98 |J,g/m3
3-21 |ig/m3
39-191 |lg/m3
8-40 |lg/m3
7-403 |lg/m3
1-85 |lg/m3
1200 |ig/m3
252 |J,g/m3
Exposure margin
ratio - average
exposure
for 0.8 |J,g/m3 of
environmental
exposure1"
4-96


10-50
2-106

315

Exposure margin
ratio - high-end
exposure
for 4.0 |Ig/m3 of
environmental
exposure1"
0 8-5


2-10
0.37-21

63

Reference0

Birch and Gary,
1996

Woskieetal., 1988b

Groves and Cain,
9000d

see text discussion

a Equivalent environmental exposure = occupational exposure x 0.21, see Chapter 2, Section 2.4.3.1, some values are rounded.
b 0.8 |-lg/m3 = average 1990 nationwide on-road exposure estimate from HAPEM model; the companion rural estimate is 0.5 |_lg/m3, and 4 |_lg/m3
is
a high-end estimate. The 1996 nationwide average is 0.7 |_lg/m3, the companion rural estimate is 0.2 |_lg/m3; however, a high-end estimate is not
available for 1996. See Chapter 2, Sections 2.4.3.2.1 and 2.4.3.2.2.
0 See Table 2-27 for more details about Birch and Gary, Woskie.
d 403 |lg/m3 is a 99 percentile estimate of EC/|lg/m3, the DPM equivalent of the EC measurement can be estimated as DPM = EC x 0.62 to 1.31.

not intended to be precise or to match with specific epidemiologic data, but rather to provide a
broad range of possible exposures in the workplace. The purpose is to identify a high- and low-
end occupational exposure consistent with the occupational groups of interest and then to
compare these to estimates of environmental exposure.  Given the special interest in discerning
the lower risk magnitude, especially to see if the lower risk might be above or below one in 1
million, a high-end exposure estimate would be used, and as discussed later, the occupational
exposure can be arbitrarily increased (e.g., toward an extreme value) to ensure that a low end of
risk is identified, consistent with the reported occupational risk increases. Environmental
exposure data from on-road vehicle emissions are based on the 1990 nationwide exposure
estimates from the HAPEM model (see Chapter 2, Section 2.4.3.2.1). Both average (0.8 |ig/m3)
and high-end exposures (4 |ig/m3) are used.
       In order to compare differences between occupational  and environmental exposures, it is
necessary to convert occupational exposure to continuous exposure (i.e.,  environmental
equivalent exposure = 0.21 x occupational exposure, see Section 2.4.3.1). Accordingly, Table 8-
1 shows equivalent environmental levels and the ratios of occupational to environmental
                                             8-12

-------
exposures, referred to as exposure margins (EMs). An EM of 1 or less indicates that
environmental exposure is comparable to occupational exposure. An EM >1 means that the
occupational equivalent exposure is greater than the benchmarked environmental exposure.
       Table 8-1 shows that the EMs based on the average nationwide environmental exposure
(0.8 |ig/m3) approach three orders of magnitude.  EM's that range from 1 to 10 also can be
viewed as showing that adjusted occupational exposures are relatively close to the ambient
environmental levels that were chosen as benchmarks. This closeness sets the stage for less
uncertainty in hazard and risk extrapolation from the occupational to environmental setting.  It
also raises a concern that risks to the general public could approach worker risks that were
observed in the occupational epidemiologic studies.  Table 8-1 is based upon DE exposure
estimates from on-road sources only. With the addition of exposure from nonroad sources, the
average nationwide-based EM ratios would be lower. For example, using 1996 exposure data
for urban populations (Table 2-30), the exposure from on-road sources is 0.5 |lg/m3, whereas
nonroad sources contribute 0.9 |lg/m3, for a total of 1.4 |lg/m3. Using this exposure value in
place of the EM calculation of Table 8-1 (1990 estimate of 0.8) produces a nearly 43% reduction
in the EM ratio. A comparison  of EM changes for the high-end on-road plus nonroad exposure
is not possible at the present time because the 1996 data have not yet been modeled to obtain a
high-end value similar to the  1990 value of 4.0 |lg/m3.
       A second approach to explore the possible cancer risk to the general population from
environmental exposure to DE is more quantitative.  One begins by examining the risk observed
in DE-exposed workers and then making reasoned assumptions as to how these risks can be
translated to environmental exposure conditions. Such an  approach involves three major
assumptions:  (1) the excess lung cancer risk as shown in numerous epidemiologic studies and in
two meta-analyses is indeed due to DE exposure, (2) the increased lung cancer risk over
background is linearly proportional to the lifetime exposure to DE, and (3) the past DE exposure
for workers has the same cancer-inducing potential as the current DE in ambient air. Any of
these assumptions could have an impact on the possible environmental risk by either increasing
or decreasing the risk estimates, including the possibility of a lower or zero risk at environmental
levels.
       As reviewed in Chapter  7, Section 7.2, numerous epidemiologic studies have shown
increased lung cancer risks (i.e., some are deaths, some are cases) among workers in certain
occupations.  The  relative risks  or odds ratios range from 1.2 to 2.6.  Two independent meta-
analyses show smoking-adjusted relative risk increases of  1.35 (Bhatia et al., 1997) and 1.47
(Lipsett and Campleman, 1999). For this analysis, a  relative risk of 1.4 is selected as a
reasonable estimate. This risk means that the workers faced an extra risk 40% higher than the
                                          8-13

-------
5% background lifetime lung cancer risk in the U.S. population.2  Thus, using the relationship
[excess risk = (relative risk-1) x background risk], these DE-exposed workers would have an
excess risk of 2% (10"2) (i.e., to develop lung cancer) due to occupational exposure to DE  [(1.4 -
1) x (0.05) = 0.02]. The validity of this interpretation depends on an important assumption:  that
the observed incremental risk of 40% was due to DE exposure alone and not to other unknown
extraneous factors. It should be noted, however, that the conclusion about the risk perspective
would not be changed even if the incremental risk of 40% were greatly reduced (e.g., to 20%);
the conclusion would be changed only if almost all of the incremental risk were due to nondiesel
factors.
       Next, one would consider the EM (i.e., the EM ratio) between the occupational exposures
and general-population environmental  exposures. DPM concentrations in the workplace, used as
a surrogate for worker exposure, are shown for three occupational worker groups in Table 8-1.
These range from 7-403 |lg/m3 (with an equivalent continuous exposure of 1-85 |lg/m3). These
worker groups are consistent with many of those cited in the meta-analyses. For this
exploratory risk estimation approach, we want to intentionally adopt a high-end boundary
exposure estimate  that is unlikely to  be exceeded, so that a lower bounding of the risk would be
identified.  An occupational boundary exposure of 1,200 |lg/m3 with its environmental
equivalent estimated value of 252 |ig/m3 has been purposefully adopted to represent a high-end
boundary estimate. This happens to  be about three times the forklift operator value shown in
Table 8-1, and clearly a high-end estimate when Table 2-27 is examined, exclusive of the
estimates for miners which are not included  in the meta-analyses. It should be noted once again
that none of these estimates are intended to be precise.
       Table 8-1 shows that the DPM  exposure margin ratio between occupational and
environmental exposure, using the nationwide average exposure value of 0.8 |lg/m3, may range
from 2 to 315 when the boundary estimate is used, and range from 0.37-63 when 4.0 |lg/m3 is
used as a high-end environmental benchmark exposure.   Some of these extreme values will be
used, as discussed in the next paragraph.
       Risks from environmental exposure depend on the shape of the dose-response curve in
the range between occupational and  environmental exposures. If lifetime risks in this range were
       2The background rate of 0.05 is an approximated lifetime risk calculated by the method of lifetable analysis
using age-specific lung cancer mortality data and probability of death in the age group taken from the National
Health Statistics (HRS) monographs of Vital Statistics of the U.S. (Vol. 2, Part A, 1992). Similar values based on
two rather crude approaches also can be obtained: (1) 59.8 x l(r5 / 8.8 x lO'3 = 6.8 x 10'2 where 59.8 x 10'5 and 8.8 x
10"3 are, respectively, the crude estimates of lung cancer deaths (including intrathoracic organs, estimated to be less
than 105 of the total cases) and total deaths for 1996 reported in Statistical Abstract of the U.S. (Bureau of the
Census, 1998, 118th Edition), and (2) 156,900/270,000,000 x 76 = 0.045, where 156,900 is the projected lung cancer
deaths for the year 2000 as reported in Cancer Statistics J of the American Cancer Society, Jan/Feb 2000),
270,000,000 is the current U.S. population, and 76 is the expected lifespan.

                                            8-14

-------
to fall proportionately with reduced exposure, and if one assumes that past occupational
exposures were at the high-boundary end, then the risk from average environmental exposure
could be between 10"5 and 10"4 (0.02 ^ 315 = 6 x 10"5). On the other hand, if occupational
exposures for different groups were lower, risks from environmental exposure would be higher.
For example, if occupational concentrations or exposures were closer to 100 |ig/m3, a value that
is represented in several data sets shown in Table 8-1 (with an equivalent environmental
exposure of 21 |lg/m3 and a corresponding EM of 26), then risks from environmental exposure
would approach 10"3 (0.02 + 26 = 8 x  10"4). If lifetime risks were to fall more than
proportionately, then risks would be lower. The latter two sources of dose-response uncertainty
(i.e., the actual occupational exposures and the shape of the dose-response curve at low
exposures) cannot be further defined with currently available information. Use of higher
environmental exposures  (>0.8 up to 4.0 |lg/m3) lowers the EM value and increases the
estimated risk.
       The magnitude of the estimated lifetime cancer risk derived from using a very high-end
occupational-to-environmental exposure difference, establishes a reasonable basis to  believe that
the general population could face possible risks high enough to be of concern.  This does not
directly address the segments of the population that may be at highest risk: those who are
additionally exposed to nonroad sources of DE and children who may be more sensitive to early-
life DE  exposure, if not in fact, a longer period of potential lifetime exposure.
       The analyses presented above are not intended to be precise but are useful  in gauging the
possible risk range using simple risk exploration methods.  Best scientific judgment guided the
selection of assumptions and other elements of this analysis which are deemed reasonable and
appropriate for identifying possible risks based on the information currently available. These
analyses provide a sense of where an upper limit (or "upper bound") of the cancer risk may be.
The simple methodologies used are generic and are valid for any increased relative risk data,
however, they are not unique to the DE  data.  These analyses are subject to numerous
uncertainties, particularly the lack of actual exposure information in the epidemiologic studies
and uncertainties related to the three major underlying assumptions mentioned earlier. Any of
these uncertainties could have an  impact on the possible  risk levels discussed above.  Lower
risks are possible and one cannot  rule out zero risk. The risks could be zero because (a) some
individuals within the population may have a high tolerance to exposure from DE and therefore
not be susceptible to the cancer risk from environmental  exposure, and (b) although evidence of
this has not been seen, there could be a threshold of exposure below which there is no cancer
risk.
       The estimated possible risk ranges (10"5 to 10"3 as well as lower and zero risk) provide a
perspective of the potential significance of the lung cancer hazard. This perspective should not

                                          8-15

-------
be viewed as a definitive quantitative characterization of risk. The development of risk estimates
does not constitute endorsement of their validity as surrogates for cancer unit risk or their
suitability for estimating numbers of cancer cases.  Further research is needed to more accurately
assess and characterize environmental cancer risks of DE.

8.5.  SUMMARY AND DISCUSSION
       There are numerous published quantitative dose-response assessments to estimate human
cancer risk from exposure to DE that use epidemiologic and/or experimental animal data (see
Appendix C). These dose-response assessments were considered but failed to present a fully
sufficient basis for a confident derivation of a cancer unit risk.  This is because of epidemiologic
exposure-related database uncertainties and the recent understanding about the lack of relevance
of the rat lung cancer response to environmentally exposed humans.  The dose-response analysis
in this chapter has focused on the feasibility of using the occupational epidemiologic data to
develop dose-response relationships and extrapolating them to the presumably lower levels of
environmental exposure. Typically, this would result in the derivation of an exposure/dose-
specific cancer unit risk. In the absence of an understanding about the mode(s) of action for DE-
induced lung cancer in humans, coupled with the consideration that DE contains many
mutagenic and carcinogenic constituents, linear low-dose extrapolation  is judged to be an
appropriate default choice for dose-response analysis, should there be satisfactory data to
perform such an analysis.
       This chapter specifically evaluated the suitability of using the railroad worker studies
(Garshick et al., 1987, 1988) and the Teamster Union truck driver studies (Steenland et al.,  1990,
1998) for dose-response analysis. However, because of uncertainties about the exposure
response for the railroad workers and exposure uncertainties for the truck drivers, a dose-
response-based cancer unit risk estimate for DE is not developed from these data sets at this
time.
       In the absence of a unit risk to assess environmental cancer risk, some insight about the
possible significance of the hazard can be drawn from the available epidemiologic data and
exploratory risk evaluation techniques. A discussion of possible risk is  presented in the form of
a perspective on the possible magnitude of risk from environmental exposure.  The perspective
discussion notes the small exposure margins and possible overlap between some occupational
and environmental exposure levels. This lessens the uncertainty of extrapolating the
occupational hazard and observed risk into the environmental setting. Furthermore, based on a
more quantitative approach involving the observed lung cancer from occupational exposures and
the magnitude of occupational and environmental exposure differences, an exploratory risk
analysis shows that environmental cancer risks possibly range from 10"5 to nearly 10"3, while a

                                          8-16

-------
consideration of the numerous uncertainties and assumptions also indicates that lower risk is
possible and zero risk cannot be ruled out. These risk findings are only general indicators of the
potential significance of the lung cancer hazard and should not be viewed as a definitive
quantitative characterization of risk or be used to estimate an exposure-specific population
impact, i.e., estimating numbers of cancer deaths. Best scientific judgment guided the selection
of assumptions and other elements of the analysis which are deemed reasonable and appropriate
for identifying possible risks based on the information currently available.  Further research is
needed to more accurately assess and characterize environmental cancer risks of DE.
        This exploratory risk analysis uses data collected from engines built prior to the mid-
1990s.  While engine exhaust emissions have been decreasing and exhaust composition has been
changing in recent years, particularly for on-road  engines, EPA believes that the insight gained
from the risk perspective is pertinent to current on-road and nonroad engine use. New and
cleaner engines will become available in response to environmental concerns and strict  new
regulations.  As cleaner engines replace a substantial number of existing engines, the risk
perspective will need to be reevaluated.
                             REFERENCES FOR CHAPTER 8

Bhatia, R; Lopipero, P; Smith, A. (1997) Diesel exhaust exposure and lung cancer. Epidemiology 9(1):84-91.

Birch, ME; Gary, RA. (1996) Elemental carbon-based method for monitoring occupational exposures to paniculate
diesel exhaust. Aerosol Sci Technol 25:221-241.

California Environmental Protection Agency (Cal EPA). (1998) Health risk assessment for DE. Public and
Scientific Review Draft.
Crump, KS. (1999) Lung cancer mortality and DE: reanalysis of a retrospective cohort study of U.S. railroad
workers. Inhal Toxicol 11:1-17.

Crump, KS; Lambert, T;  Chen, C. (1991) Assessment of risk from exposure to diesel engine emissions. Clement
International Corporation. Prepared for U.S. EPA under contract no. 68-02-4601.
Garshick, E; Schenker, MB; Munoz, A; et al. (1987) A case-control study of lung cancer and DE exposure in
railroad workers. Am Rev Respir Dis 135:1242-1248.
Garshick, E; Schenker, MB; Munoz, A; et al. (1988) A retrospective cohort study of lung cancer and DE exposure in
railroad workers. Am Rev Respir Dis 137:820-825.
Hammond, SK; Smith, TJ; Woskie, SR; et al. (1988) Markers of exposure to diesel exhaust and cigarette smoke in
railroad workers. Am Ind Hyg Assoc J 49:516-522.
Health Effects Institute (HEI). (1999) Diesel emissions and lung cancer: epidemiology and quantitative risk
assessment. A special report of the Institute's Diesel Epidemiology Expert Panel. Cambridge, MA: Health Effects
Institute.
                                             8-17

-------
International Life Sciences Institute (ILSI). (2000) ILSI Risk Science Institute workshop: The relevance of the rat
lung response to particle overload for human risk assessment. Gardner, DE, ed. Inhal Toxicol: 12(1-2).

International Programme on Chemical Safety: World Health Organization (IPCS). (1996) Diesel fuel and exhaust
emissions. Environmental Health Criteria 171. Geneva: World Health Organization.

Lipsett, M; Campleman, S. (1999) Occupational exposure to DE and lung cancer: a meta-analysis. Am J Publ Health
89(7):1009-1017.

Steenland, NK; Silverman, DT; Hornung, RW. (1990) Case-control study of lung cancer and truck driving in the
Teamsters Union. Am J Publ Health 80:670-674.

Steenland, K; Deddens, J; Stayner,  L. (1998) DE and lung cancer in the trucking industry: exposure-response
analysis and risk assessment. Am J Ind Med 34:220-228.

U.S. Environmental Protection Agency (EPA). (1986) Guidelines for carcinogen risk assessment. Federal Register
51(185):33992-34003.

U.S. EPA. (1996) Proposed guidelines for carcinogen risk assessment. Federal Register61(79):17960-18011.

U.S. EPA. (1999) Guidelines for Carcinogen Risk Assessment. Review Draft. NCEA-F-0644, July. Risk Assessment
Forum, Washington, DC.  http://www.epa.gov/ncea/raf/cancer.htm.

Woskie, SR; Smith, TJ; Hammond, SK; et al. (1988a) Estimation of the DE exposures of railroad workers: II.
National and historical exposures. Am J Ind Med 13:395-404.

Woskie, SR; Smith, TJ; Hammond, SK; et al. (1988b) Estimation of the DE exposures of railroad workers: I. Current
exposures. Am J Ind Med 13:381-394.

Zaebst, D; Clapp, D; Blade, L; et al. (1991) Quantitative determination of trucking industry workers' exposures to
diesel particles.  Am Ind Hyg Assoc J 52:529-541.
                                                  8-18

-------
    9. CHARACTERIZATION OF POTENTIAL HUMAN HEALTH EFFECTS OF
       DIESEL EXHAUST: HAZARD AND DOSE-RESPONSE ASSESSMENTS
9.1. INTRODUCTION
       Human health risk assessment entails the evaluation of all pertinent information on the
hazardous nature of environmental agents, on the extent of human exposure to them, and on the
characterization of the potential risk to an exposed population. The information is typically
organized into four components: hazard assessment, dose-response assessment, exposure
assessment, and risk characterization. This health assessment document focuses only on the
hazard and dose-response assessment components.  The overall objectives of this diesel engine
exhaust (DE) assessment are:

       •   to identify and characterize the human health effects, i.e., hazards that may result
          from environmental exposure to DE;
       •   to determine whether there is a quantitative exposure/dose-response relationship for
          DE exposure and the health effect in the range of observation, and if sufficient data
          are available (1) for noncancer effects to derive estimates of exposure that are
          believed to be without appreciable risk, and (2) for carcinogenicity to derive an
          exposure/dose-specific cancer unit risk; and
       •   to summarize and integrate the findings of the assessment into a characterization and
          discuss the uncertainties.

       This chapter summarizes and integrates the key findings about the nature and
characteristics of environmental exposure to DE (Chapter 2), health hazard information
(Chapters 3, 4, 5, and 7), and dose-response analyses (Chapters 6 and 8) that are relevant to the
potential  human health effects associated with current-day environmental exposure to DE.  It
also discusses the uncertainties associated with the key findings, including critical data and
knowledge gaps, key assumptions, and EPA's science policy choices that are used to bridge the
data and knowledge gaps.
       This assessment is the Agency's first health assessment for DE emissions and was
developed to provide information about the potential for DE-related environmental health
hazards that could be used in evaluating regulatory initiatives under provisions of the Clean Air
Act.
                                          9-1

-------
9.2. PHYSICAL AND CHEMICAL COMPOSITION OF DIESEL EXHAUST
       As reviewed in Chapter 2, DE is a complex mixture of hundreds of constituents in gas or
particle phases.  Gaseous components of DE include carbon dioxide, oxygen, nitrogen, water
vapor, carbon monoxide, nitrogen compounds, sulfur compounds, and low molecular-weight
hydrocarbons and their derivatives. The particulate matter of DE, diesel particulate matter
(DPM), is composed of elemental carbon (EC), adsorbed organic compounds, and small amounts
of sulfate, nitrate, metals, trace elements, water, and unidentified compounds.  Incomplete
combustion of fuel hydrocarbons as well as engine oil and other fuel components such as sulfur
leads to the formation of DPM. DPM is either directly emitted from diesel-powered engines
(primary particulate matter) or is formed from the gaseous compounds emitted by a diesel engine
(secondary particulate  matter).
       DE emissions vary in chemical composition and particle sizes among different engine
types, fuel formulations, and within engine types according to operating conditions. As the
emissions age in the environment they also change. There also  have been changes in DE
emissions over time as a result of changes in engine technology and fuel reformulation.  The
following sections identify and characterize the key components of DE that are of concern in
possible health outcomes, and discuss the changes in the composition of DE over time. The
latter information is critical for making a scientific judgment about the appropriateness of using
epidemiologic and toxicologic findings from past DE exposures to assess hazard and risk from
current-day environmental exposures.  It should be noted that available animal studies are based
on exhaust exposures from various  model year on-road diesel engines since  1980, whereas many
of the epidemiologic studies refer to exposures from on-road and nonroad diesel engines in use
from the 1950s through the mid-1990s.
       After emission  from the tailpipe, DE undergoes dilution, chemical and physical
transformations, and dispersion and transport into the atmosphere.  After a day or so in the
ambient environment, the exhaust mixture is said to be aging, a recognition of the atmospheric
transformation processes, mostly focused on the organics present, though some particle size
changes also may occur.  The public health impact of DE mixture transformations is not clear, as
some atmospheric processes alter chemical forms to a less toxic form whereas others seem to
produce a chemical form with increased toxicity (Chapter 2, Section 2.3).

9.2.1. Diesel Exhaust Components of Possible Health Concern
       The components of DE that are of health concern for this assessment are the particles
(elemental carbon core), the organic compounds adsorbed to the particles, and the organic
compounds present in the gas phase. The gaseous oxides of carbon, nitrogen,  and sulfur are also
                                          9-2

-------
of public health interest and the relevant health considerations for these are reviewed separately
in EPA's Ambient Air Quality Criteria Documents.

9.2.1.1. Diesel Particles
       Approximately 80%-95% of DPM mass is in the fine particle size range (<2.5
micrometers, ambient paniculate matter [PM]), with a mean particle aerodynamic diameter of
about 0.2 micrometers.  Ultrafine particles (<0.1  micrometers), a smaller size component of the
fine particles, average about 0.02 micrometers in aerodynamic diameter and account for about
1%-20% of the DPM mass and 50%-90% of the  total number of particles in DPM (Chapter 2,
Section 2.2.8.3).
       Particle size is important for a number of reasons. Particles with aerodynamic diameters
>2.5 micrometers (i.e., >PM25) tend to be retained in the upper portions of the respiratory tract,
whereas particles with diameters <2.5  micrometers (i.e., PM2 5) are deposited in all areas,
especially into the lower portions of the respiratory tract, including the deep lung. These fine
and ultrafme particles have a very large surface area per gram mass (Chapter 2, Section 2.2.2),
which enables them to adsorb and transport inorganic and organic compounds into the lung
(Chapters, Section 3.3).
       DPM is part of ambient particulate matter (PM). The EPA Emissions Trends Report
(U.S. EPA, 2000) indicates that annual nationwide emissions of diesel PM25 (on-road and
nonroad) in 1998 were 77% of all mobile-source emissions in 1998, 23% of the total PM25
inventory excluding natural and miscellaneous sources, and 6% if the natural and miscellaneous
sources are included.  Some geographic areas have a higher percentage of DPM in ambient PM25
because of differences in the number and types of diesel engines present in the area (e.g., on-
road engines as well as nonroad engines). For instance, in Manhattan, New York, on-road diesel
PM was reported to contribute about 53% of ambient PM10 during 3 days in 1993, whereas 1996-
1997 studies in the Phoenix and Denver areas showed diesel PM to be 10%-15%  of total PM25
mass (Chapter 2, Section 2.4.2.1).
       DPM generally contains a high percentage of EC per unit mass, which can be used as a
distinguishing feature from noncombustion sources of PM25 and, to an extent, other combustion
sources.  The DPM EC content can range from more than 50% to approximately 75% of the
DPM mass depending on age of engine, type of engine (heavy-duty versus light-duty), fuel
characteristics, and driving conditions. The organic carbon portion of DPM can range
approximately from 19% to 43%, though higher  and lower values also have been reported.  In
comparison, gasoline engine exhaust generally has a reverse pattern of low EC content and a
high percentage of organics on the particle mass  (see Chapter 2, Table 2-13).
                                          9-3

-------
9.2.1.2. Organic Compounds
       The organic compounds present in the gases and adsorbed onto the particles include a
wide spectrum of compounds related to unburned diesel fuel, lube oil, low levels of partial
combustion, and pyrolysis products (see Chapter 2, Table 2-19). The organic compounds
present in the gaseous phase include alkanes, alkenes, aldehydes, monocyclic aromatic
compounds, and polycyclic aromatic hydrocarbons (PAHs). Among the gaseous components of
DE, the aldehydes are particularly important because of their potential carcinogenic effects and
because they make up an important fraction of the gaseous emissions.  Formaldehyde accounts
for a majority of the aldehyde emissions (65%-80%) from diesel engines. Acetaldehyde and
acrolein are the next most abundant aldehydes.  Other gaseous components of DE that are
notable for their carcinogenic effects include benzene,  1,3-butadiene, PAHs, and nitro-PAHs
(including those with <4 rings and nitro-PAHs with 2 and 3 rings).  A number of the gaseous
compounds (e.g., aldehydes, alkanes, alkenes, NOX, SOX) also are known to induce respiratory
tract irritation given sufficient exposure (see Chapter 2, Table 2-21). Very small amounts of
dioxins have been measured in heavy-duty diesel truck exhaust. These emissions are estimated
to represent about 1.2% of the 1995 national dioxin inventory; dioxin emissions from nonroad
exhausts have not been estimated (Chapter 2, Section 2.2.7.2).
       Organic substances adsorbed onto DPM include C14.35 hydrocarbon compounds, PAHs
with >4 rings, and nitro-PAHs. PAHs and their derivatives comprise <1% of the DPM mass
(Chapter 2, Section 2.2.8). Many of these hydrocarbons are known to have mutagenic and
carcinogenic properties.  California EPA (Cal EPA, 1998) identified at least 19 hydrocarbons
present in DE that are known or suspected carcinogens, according to evaluations by the
International Agency for Research on Cancer (IARC).

9.2.2. "Fresh" Versus "Aged" Diesel Exhaust
       Newly emitted exhaust is termed "fresh," whereas exhaust that is more than 1 or 2 days
old is referred to as "aged" because of alterations caused by sunlight and other chemical physical
reactions that occur in the atmosphere. The overall toxicological consequence of DE aging is
unclear because during aging some compounds  in the DE mixture are altered to more toxic forms
while others are made less toxic. For example, PAHs present in fresh emissions may be nitrated
by atmospheric NO3 to form nitro-PAHs, thus adding to the existing burden of toxic nitro-PAHs
present in fresh exhaust.  On the other  hand, PAHs present in the gas phase can react with
hydroxyl radicals present in the ambient air, leading to a reduced atmospheric lifetime of the
original PAHs.  Alkanes and alkenes may be converted to aldehydes, and oxides of nitrogen to
nitric acid (Chapter 2, Section 2.3).
                                          9-4

-------
9.2.3. Changes in Diesel Exhaust Emissions and Composition Over Time
       Chapter 2, with its Summary in Section 2.5, provides a full review of emissions trends
and a complete characterization of the physical and chemical changes in DE over the years,
taking into consideration the lack of consistent analytical and measurement techniques and the
variability in emissions based on vehicle mix, driving cycles, engine deterioration, and other
factors.  Key findings and inferences relevant to the potential health effects of DE are discussed
below.
       As discussed in Chapter 2, Section 2.2.3, the EPA Emissions Trend Report estimates that
DPM10 on-road emissions decreased 27% between 1980 and 1998.  DPM emission factors
(g/mile by model year) from new on-road diesel vehicles decreased on average by a factor of six
from the mid-1970s to the mid-1990s. These significant reductions are largely attributable to
reductions in three PM components: EC, organic carbon, and sulfate.  Limited data are available
to assess the changes in emission rates from locomotive, marine, or other nonroad diesel sources
over time, although it is estimated that DPM10 (< 10 |lm) emissions from nonroad diesel engines
increased 17% between 1980 and 1998 (Chapter 2, Section 2.2.5).
       Because of changes in engine technology and fuel composition, the chemical
composition of DPM from on-road vehicles has also changed over time. The percentage of
soluble organic material associated with DPM decreased by model year from the 1980s to the
1990s, and the proportion of EC is correspondingly higher. PAHs and nitro-PAHs are present in
DPM from both new and older diesel engines.  There are insufficient data to provide clear
insight into the potential for changes in total PAH  emissions over time or specific PAHs such as
benzo[a]pyrene and 1-nitropyrene.  It should be noted that the chemical composition of ambient
DPM to which people are currently exposed is determined by a combination of exhaust from
older and newer engines as well as on-road and nonroad applications of those engines.
Consequently, the decrease in the soluble organic fraction of DPM by model year for on-road
engines does not directly translate into a proportional decrease in DPM-associated organic
material to which people are exposed. In addition, the contributions from high-emitting and/or
smoking diesel engines have not been quantified (Chapter 2, Section 2.5.2).
       Because of these uncertainties, the exposure impact of changes in DPM composition over
time cannot be confidently characterized.  Available data clearly indicate that lexicologically
significant organic components of DE (e.g., PAHs, PAH derivatives, nitro-PAHs) were present
in DPM and DE in the 1970s and are still  present.  Even though a significant fraction of ambient
DPM (possibly more than 50%) is emitted by nonroad equipment, data are currently inadequate
to characterize changes in the chemical composition of DPM from nonroad equipment over time.
Given the variation in fuel, engine technology,  and in-use operational factors over the years,
caution should be exercised in presuming that a decrease in the amount of emissions or emission

                                          9-5

-------
constituents from older engines to present day in-use engines will result in a decrease in
hazard/risk. In meeting the 2007 federal regulations for heavy-duty DE, the exhaust composition
will be markedly changed with a consequence that health hazards are expected to be significantly
reduced.

9.3. AMBIENT CONCENTRATIONS AND EXPOSURE TO DIESEL EXHAUST
       Chapter 2, Section 2.4 provides information on occupational and environmental
exposures to DE in order to provide a context for the hazard assessment and dose-response
analysis.  Highlights of the available information are discussed below.
       DE is emitted from a variety of sources, both on-road (e.g.,  motor vehicles, construction
equipment) and nonroad (e.g., farm equipment, railway locomotives,  or marine uses).
Environmental exposure to DE is generally higher in urban areas than in rural areas. The
concentration of DE in the air will vary within any geographic area depending on the number
and types of diesel engines in the area and the atmospheric patterns of dispersal.  Some
important factors that determine the difference between the ambient concentration of DE and the
resultant exposure to an individual include the proximity of a person to the DE source and
his/her pattern of activity which, for example, includes outdoor versus indoor activities as well as
related breathing rates. Certain occupational populations (e.g., transportation and garage
workers, heavy-equipment operators, and others who spend considerable time outdoors) can be
exposed to much higher levels of DE than the general population. The amount or number of
particles delivered and retained in the lung is one factor that could contribute to differential
human susceptibility to DPM. For example, children have smaller  lungs than adults and thus
could have a higher lung burden of inhaled DPM per lung surface area if their activity pattern
results in a high breathing frequency.
       As DE is a complex mixture of many constituents, environmental concentration
measurements and related human exposure is difficult to precisely measure. Even though levels
of a number of DE constituents are generally known, it is difficult to quantify the portion that
comes from DE since other types of emission sources also may emit the same constituent.
Moreover, there is still incomplete knowledge about the relative roles of the relevant DE
constituents in mediating the potential health effects of DE.  Historically, exposure levels to
DPM have been used as a surrogate marker/dosimeter for whole DE.  Although uncertainty
exists as to whether DPM mass (expressed as |ig/m3 of DPM) is the most appropriate dosimeter
for health effect purposes, it is considered to be a reasonable choice until more definitive
information is available about the mechanisms or mode(s) of toxicity  action of DE.
       Several techniques exist for estimating ambient concentrations of DPM, including
chemical mass balance (CMB) source apportionment, dispersion modeling, and using EC as a

                                          9-6

-------
surrogate for DPM.  DPM concentrations reported from CMB and dispersion modeling studies
in the 1980s suggest that in urban and suburban areas (Phoenix and Southern California), the
annual average DPM concentration ranged from 2 to 13 |lg/m3.  In the 1990s, annual or seasonal
average DPM concentrations in suburban or urban locations have ranged from 1.2 to 4.5 |lg/m3.
DPM concentrations at a major bus stop in downtown Manhattan ranged from 13.2 to 46.7
|lg/m3 over a 3-day period in 1993. In nonurban and rural areas in the 1980s, DPM
concentrations were reported to range from 1.4 to 5 |lg/m3. In the 1990s, nonurban air basins in
California were reported to have DPM concentrations ranging from 0.2 to 2.6 |lg/m3 (Chapter 2,
Section 2.4.2).
       A comprehensive exposure assessment is not presented in this assessment, though EPA is
developing this  in an analysis called the National Air Toxics Assessment.  Interim exposure
estimation based on EPA's Hazardous Air Pollutant Exposure Model (HAPEM-MS3 model), for
on-road sources only, suggests that in 1996 annual average DPM exposure in urban areas from
only on-road engines was 0.7 |lg/m3, while in rural areas exposure was 0.3 |lg/m3. Among 10
urban areas, the 1996 annual average estimated exposure ranged from 0.5 to 1.2 |ig/m3. A high-
end exposure estimate for 1996 is not yet available. Comparable 1990 exposure estimates for
on-road sources ranged from 0.9 |lg/m3 for urban areas to 0.5 |lg/m3 for rural areas.  In 1990
exposure estimates for the most highly exposed individuals (e.g., outdoor workers and children
who spend large amounts of time outdoors) were estimated to be up to 4.0 |ig/m3 (Chapter 2,
Section 2.4.3.2,  Table 2-29). Nationwide level nonroad emission exposures are estimated to be
nearly double those from on-road sources.
       Estimates for occupational exposures to DE as DPM mass are generally higher than
environmental exposures. Tables 2-27 and 2-28 provide historic exposure estimates for specific
worker categories.  For example, historic DPM exposure estimates range from 39-191 |lg/m3 for
railroad workers, 4-748 |lg/m3 for firefighters, 7-98 |lg/m3 for public transit workers and airport
crews, 5-61 |lg/m3 for mechanics and dock workers, and 2-7 |lg/m3 for long- and short-haul
truck drivers.  For a direct comparison of lifetime exposures between an occupational setting (8
hours per day, 5 days per week, for 45 years) and environmental exposure (continuous exposure
for 70 years), the occupational estimates are converted to an equivalent environmental lifetime
estimate,1 which is also shown in Table 2-28. A conversion of EC-based measurements to total
DPM also may be needed for some estimates. The estimated 70-year lifetime exposures
equivalent to those for the occupational groups discussed above range from about 0.4-157
|ig/m3. These data indicate that some lower-end occupational estimates of DPM, when
converted to environmental equivalents, overlap the range of estimated environmental exposures
       Environmental equivalent occupational exposure = 0.21 x occupational exposure.

                                          9-7

-------
to DPM from on-road emissions (national average in 1990 of 0.8 |lg/m3, with high-end
exposures up to about 4 |lg/m3). The addition of nonroad emission exposures, when appropriate,
makes the case for overlap of occupational and environmental exposure more prevalent.

9.4. HAZARD CHARACTERIZATION
        The primary health effects of concern from environmental exposure to DE include
effects associated with both acute and short-term exposures as well as chronic exposures. It is
recognized that acute exposures may produce transitory physiological symptoms of varied
severity as well as exacerbation of allergenic effects from acute and repeated exposures. On the
basis of combined human and experimental evidence from chronic exposure studies, noncancer
respiratory effects and lung cancer are observed.
      The health effects data are based on DE from a variety of engines existing before the
mid-1990s. There have been changes in the physical and chemical composition of some DE
emissions (on-road vehicle emissions) over time, though there is no definitive information to
show that the emission changes portend significant toxicological  changes. The mode(s) of action
for DE toxicity in humans is not understood, and hence knowledge is lacking about the role of
exhaust  mixture components in modulating the toxicity. Taken together, these considerations
have lead to a judgment that the hazards identified from older technology-based exposures are
applicable to current-day exposures. As new and cleaner diesel engines replace a substantial
number  of existing engines, the general applicability of the older data will need to be
reevaluated.
      As discussed in Chapter 6 (Section 6.4), it is also reasonable to expect that DPM, being a
constituent of ambient fine PM  (PM2 5), would contribute to the wider spectrum of effects that
have been associated with ambient PM2 5. Community epidemiologic studies have shown that
ambient PM2 5 exposure is statistically associated with increased mortality (especially among
people over 65 years of age with preexisting cardiopulmonary conditions) and morbidity as
measured by increases in hospital admissions, respiratory symptom rates, decrements in lung
function, and exacerbation of asthma, and possibly immunological effects in the respiratory
system.  There continues to be little epidemiologic evidence for an effect of ambient exposure to
PM on cancer rates (U.S. EPA,  1996a,b), though U.S. EPA's Criteria Document for Ambient PM
(expected to be released in 2002) will examine the question further.

9.4.1.  Acute and Short-Term  Exposures
      The combined human and animal evidence indicates that DE can induce irritation to the
eye, nose, and throat, as well as inflammatory responses in the  airways and the lung following
                                          9-8

-------
acute and/or short-term exposure to high concentrations. There also is suggestive evidence for
possible immunological and allergenic effects of DE.

9.4.1.1. Acute Irritation
       DE contains various respiratory irritants in the gas phase and in the particulate phase
(e.g., SOX, NOX, aldehydes). Acute exposure to DE has been associated with irritation of the eye,
nose, and throat, respiratory symptoms (cough and phlegm), and neurophysiological symptoms
such as headache, lightheadedness, nausea, vomiting, and numbness or tingling of the
extremities.  Such symptoms have been described mainly in reports of individuals exposed to DE
in the workplace, or in clinical studies in humans exposed acutely to high concentrations of DE.
Because of the general lack of validating exposure information in the reports, the role of DE in
causing these effects is unknown.  An exposure-response relationship for these acute irritation
and respiratory symptoms has not been demonstrated (Chapter 5, Section 5.1.1.1).

9.4.1.2. Respiratory Effects
       Available studies of occupational exposure to DE have not provided evidence for
significant decrements of lung function in workers over a work shift or after a short-term
exposure period. Short-term and subchronic inhalation studies of DE in animals (rats, mice,
hamsters, cats, guinea pigs) showed inflammation of the airways and minimal or no lung
function changes. These effects were associated with high DE exposures (up to 6 mg/m3).
Exposure-response relationships have not been established for these responses (Chapter 5,
Sections 5.1.1, 5.1.2,  and 5.1.3).

9.4.1.3. Immunological Effects
       Recent human and animal studies show that acute DE exposure episodes can exacerbate
immunological reactions to other allergens or initiate a DE-specific allergenic reaction.  The
effects seem to be associated with both the organic and carbon core fraction of DPM. In human
subjects, intranasal administration of DPM has resulted in measurable increases of IgE antibody
production and increased nasal mRNA for some proinflammatory cytokines. These types of
responses  also are markers typical of asthma, though for DE, evidence has not been produced in
humans that DE  exposure results in asthma.  The ability of DPM to act as an adjuvant to other
allergens also has been demonstrated in human subjects. For example, co-exposure to DPM and
ragweed pollen was reported to significantly enhance the IgE antibody response and cytokine
expression relative to ragweed pollen alone.  Available animal studies also demonstrate the
potential adjuvant effects of DPM with model allergens, e.g., in mouse studies the allergenic
reaction to ovalbumin and Japanese cedar pollen (Chapter 5, Sections 5.1.1.1.3 and 5.1.1.1.4).

                                           9-9

-------
Additional research is needed to further characterize immunological effects of DE and to
determine whether or not the immunological effects constitute a low-exposure hazard.  This
health endpoint is of considerable public health concern, given the increases in allergic
hypersensitivity in the U.S. population (Chapter 5, Section 5.6.2.6).

9.4.2. Chronic Exposure
9.4.2.1. Noncancer Effects
       Available long-term and cross-sectional human studies have provided evidence for an
association between respiratory symptoms (cough and phlegm) and DE exposure, but there was
no consistent effect on lung function. DE has been shown in many animal studies of several
species to induce lung injury (chronic inflammation and histopathologic changes) following
long-term inhalation exposure. DE also has been tested in laboratory animals for other health
effects, but no significant effects have been found. Overall, available data lead to the conclusion
of a potential chronic respiratory hazard to humans from long-term exposure to DE.

9.4.2.1.1.  Respiratory effects. A few human studies in various diesel occupational settings
suggest that DE exposure may impair pulmonary function, as evidenced by increases in
respiratory symptoms  and some reductions in baseline pulmonary function consistent with
restrictive airway disease.  Other studies found no particular effects. The methodologic
limitations in available human studies limit their usefulness in drawing any firm  conclusions
about DE  exposure and noncancer respiratory effects (Chapter 5, Section 5.1.1.2).
       Available studies in animals, however, provide  a large body of evidence demonstrating
that prolonged inhalation exposure to high concentrations of DE can result in pulmonary injury.
A number of long-term laboratory studies in rats, mice, hamsters, cats, and monkeys found
varying degrees of adverse lung pathology including focal thickening of the alveolar walls,
replacement of Type I alveolar cells by type II cells, and fibrosis. The rat is the most sensitive
animal species to DE-induced pulmonary toxicity (Chapter 5, Sections 5.1.3 and 5.4).
       Available mechanistic data, mainly in rats,  indicates that the DPM fraction of DE is a
controlling factor in the etiology  of pulmonary toxicity, although a role  for the adsorbed organic
compounds on the  particles and in the gaseous phase cannot be ruled out.  The lung injury
appears to be mediated by an invasion  of alveolar macrophages that release chemotactic factors
that attract neutrophils and additional alveolar macrophages, which in turn release mediators
(e.g., cytokines, growth factors) and oxygen radicals. These mediators result in persistent
inflammation, cytotoxicity, impaired phagocytosis, clearance of particles, and eventually
deposition of collagen by activated fibroblasts.  This mode of action seems to be operative for a
variety of poorly soluble particles in addition to DPM (ILSI, 2000).  Because long-term exposure

                                          9-10

-------
to DE has been shown to induce exposure-dependent chronic respiratory effects in a wide range
of animal species, and the mode of action is deemed relevant to humans, there is a sufficient
scientific basis to support a conclusion that humans also could be at hazard for these effects
under a chronic exposure condition.

9.4.2.1.2. Other noncancer effects.  The negative results from available studies in several
animal species (rats, mice, hamsters, rabbits, monkeys) indicate that DE is not likely to pose a
reproductive or developmental hazard to humans.  There has been some evidence from animal
studies indicating possible neurological and behavioral effects, as well as liver effects.  These
effects, however, are seen at exposures higher than the respiratory effects. Overall, there is
insufficient evidence to conclude that a low-exposure hazard exists for these endpoints (Chapter
5, Section 5.1.3.3).

9.4.2.2. Carcinogenic Effects
       Many epidemiologic and toxicologic studies have been conducted to examine the
potential for DE to cause or contribute to the development of cancer in humans and animals,
respectively.  In addition, there are some mode-of-action studies that seek to provide an
improved understanding about the underlying carcinogenic process and thus contribute to a
better understanding of the likelihood of hazard to humans. The available evidence indicates
that chronic inhalation of DE is likely to pose a lung cancer hazard to humans.  There is
insufficient information for an evaluation of the potential cancer hazard of DE by oral and
dermal routes of exposure.

9.4.2.2.1. Epidemiologic studies. Twenty-two epidemiologic studies about the carcinogenicity
of workers exposed to DE in various occupations are reviewed in Chapter 7, Section 7.2.
Exposure to DE has typically been inferred on the basis of job classification within an industry,
with cumulative exposure based on duration of employment  or age. Increased lung cancer risk,
although not always statistically significant, has been  observed in 8 out of 10 cohort studies and
10 of 12  case-control studies within several  industries, including railroad workers, truck drivers,
heavy-equipment operators, farm tractor operators, and professional diesel vehicle drivers.  The
increased lung cancer relative risks generally range from 1.2 to 1.5, although a few studies show
relative risks as high as 2.6. Statistically significant increases in relative risk,  1.33 to 1.47, are
also shown in two independent meta-analyses. The meta-analyses demonstrate the effect of
pooling many  studies and in this case show the positive relationship between DE exposure and
lung cancer across  a variety of DE-exposed  occupations.
                                          9-11

-------
       The generally small increases in lung cancer relative risk (1.2 to 1.5, i.e., less than 2)
observed in the epidemiologic studies potentially weakens the evidence of causality. This is
because with a relative risk of less than 2, if confounders (e.g., smoking, asbestos exposure) were
having an effect on the observed risk increases, then it could be enough to account for the
increased risk. With the strongest risk factor for lung cancer being smoking, there is a lingering
uncertainty as to whether smoking effects may be influencing the magnitude of the observed
increased relative risks, in spite of the fact that in key studies the investigating epidemiologists
assert that they have effectively controlled for smoking. In studies in which the effects of
smoking were controlled, increased relative risks for lung cancer prevailed.  While some studies
did not have information on smoking, confounding by smoking is judged unlikely to be
significant if the comparison populations were from the same socioeconomic class.
       As evaluated in Chapter 7 (Section 7.2.4.5), application of the criteria for causality
provides a rational basis to conclude that the increased risks observed in available epidemiologic
studies are consistent with a causal association between exposure to DE and occurrence of lung
cancer.  Overall, the human evidence for potential  carcinogenicity for DE is judged to be strong
but less than sufficient to satisfy the criteria for a "known" human carcinogen because of
exposure uncertainties (lack of historical exposure of workers to DE) and residual uncertainty as
to whether all confounders have been satisfactorily accounted for.  The epidemiologic evidence
is inconclusive for DE being associated with other forms of cancer.

9.4.2.2.2. Animal studies.  DE  and its organic constituents, both in the gaseous and particle
phase, have been extensively tested for carcinogenicity in many experimental studies using
several animal species and with different modes of administration.  Several well-conducted
studies have consistently demonstrated that chronic inhalation exposure to sufficiently high
concentrations of DE produced  dose-related increases in lung tumors (benign and malignant) in
rats.  In contrast, chronic inhalation studies of DE in mice showed equivocal results, whereas
negative findings were consistently seen in hamsters.  The gaseous phase of DE (filtered exhaust
without particulate fraction), was found not to be carcinogenic in rats, mice, or hamsters. The
available data indicate that among the traditional animal test species, the rat is the most sensitive
species to DE.  As reviewed in Chapter 7, Section  7.4, the lung cancer response in rats from
high-concentration exposures to DE appears to be mediated by impairment of lung clearance
mechanisms owing to particle overload, resulting in persistent chronic inflammation and
subsequent pathologic and neoplastic changes (i.e., cancer) in the rat lung. Particle overload
conditions in the human lung are not expected to occur as a result of environmental or most
occupational exposures to DE.   Thus, the increased lung tumors in  the rat are not an appropriate
                                          9-12

-------
basis from which to judge the potential for a human hazard or perform a dose-response analysis
to derive a cancer unit risk for humans.
       In several intratracheal instillation studies, DPM, DPM organic extracts, and carbon
black, which is virtually devoid of PAHs, have been found to produce increased lung tumors in
rats. When directly implanted into the rat lung, DPM condensate containing mainly four- to
seven-ring PAHs induced increases in lung tumors. DPM extracts also have been shown to
cause skin tumors in several dermal studies in mice and sarcomas in mice following
subcutaneous injection. Overall there are consistent findings of carcinogenic activity by the
organic extracts of DPM in noninhalation studies (i.e., intratracheal instillation, lung
implantation, skin painting). This contributes to the evidence for a potential human hazard.

9.4.2.2.3.  Other key data. While not as extensive as the human and  animal carcinogenicity data,
other types of data are judged to be supportive of DE's potential carcinogenicity in humans.  As
mentioned previously, DE is a complex mixture of hundreds of constituents in either the gaseous
phase or particle phase. Although present in small amounts, several organic compounds in the
gaseous phase (e.g., PAHs, formaldehyde, acetaldehyde, benzene, 1,3-butadiene) are known to
exhibit mutagenic and/or carcinogenic activities.  PAHs and PAH derivatives, including nitro-
PAHs present on the diesel particle, also are known to be mutagenic and carcinogenic. As
reviewed in Chapter 4, DPM and DPM organic extracts have been shown to induce gene
mutations in a variety of high-dose bacteria and mammalian cell test systems.  DPM and DPM
organic extracts also have been shown to induce chromosomal aberrations, aneuploidy, and sister
chromatid exchange in both rodent and human in vitro tests.
       There also is suggestive evidence for the bioavailability of organic compounds from the
DE mixture. Elevated levels of DNA adducts in lymphocytes have been reported in workers
exposed to DE.  In addition, inhalation studies of animals using radio-labeled materials indicate
some elution of organic compounds from DPM after deposition in the lung as measured by their
presence in biological tissue and fluids (Chapter 3, Section 3.5).

9.4.2.2.4. Modes of carcinogenic action.  The term "mode of action" refers to a series of key
biological events and processes that are critical to the  development of cancer.  As discussed in
Section 9.4.2.2.2, there is an understanding of the modes of action for the DE-induced lung
tumors in the rat. However, the modes of action by which DE increases lung cancer risks in
humans are unknown, and the  evidence in rats is not applicable to environmentally exposed
humans.
       As discussed in Chapter 7, Section 7.4, it is hypothesized that multiple modes of action
could be involved in mediating the carcinogenic effect of DE.  These modes of action may

                                          9-13

-------
include: (a) mutagenic events (e.g., direct effects on DNA and effects on chromosomes) by
organic compounds in the gas and particle phase, (b) indirect DNA damage via the production of
reactive oxygen species (ROS) induced by particle-associated organics, and (c) particle-induced
chronic inflammatory response leading to oxidative DNA damage through the release of
cytokines, ROS, etc., and an increase in cell proliferation.
       In rats, the particulate phase appears to have the greatest contribution to the carcinogenic
effects, and both the particle core and the associated organic compounds have  demonstrated
carcinogenic properties in one or more test systems.  While limited rat data and comparative
potency calculations suggest that gas-phase components are not the primary factors in the
development of lung cancer, a contributory role of the recognized toxic components cannot be
dismissed.  The relative importance of the various modes of action may be different at different
exposure levels. Evidence from rat studies indicates the importance of the EC component of the
DE particle in mediating lung tumor response at high exposure levels. As for the particle-
absorbed organics, their inherent toxicity potential gives rise to a hypothesis that they may play a
role in low or high exposures to DE.

9.4.2.2.5.  Weight-of-evidence evaluation. Chapter 7,  Section 7.5, provides an evaluation of the
overall weight of evidence for human carcinogenicity in accordance with EPA's Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 1986, 1996a,  1999). The totality of evidence supports
the conclusion that DE is ^probable human carcinogen (Group Bl) by inhalation exposure
using the criteria in the 1986 guidelines. A cancer hazard narrative for DE also is provided in
accordance with the revised draft 1996/1999 guidelines, which concludes that  DE is likely to be
carcinogenic to humans by inhalation from environmental exposures.  The common bases for
either conclusion include the following lines of evidence:

          strong but less than sufficient evidence for a causal association between DE exposure
          and increased lung cancer risk among workers in varied occupations where exposure
          to DE occurs;
       •   extensive supporting data including the demonstrated mutagenic and/or chromosomal
          effects of DE and its organic  constituents, and knowledge of the known mutagenic
          and/or carcinogenic activity of a number of individual organic compounds present
          with particles and in the DE gases;
       •   evidence of carcinogenicity of DPM and the associated organic compounds in rats
          and mice by noninhalation routes of exposure; and
       •   suggestive evidence for the bioavailability of DE organics from DE in humans and
          animals.
                                          9-14

-------
       A notable uncertainty in the characterization of the potential cancer hazard of DE at low
levels of environmental exposure is the incomplete understanding of about its mode(s) of action
for the induction of lung cancer in humans.  Available data suggest that DE-induced lung
carcinogenicity may be mediated by mutagenic and nonmutagenic events by both the particles
and the associated organic compounds, and that a role for the organics in the gaseous phase
cannot be ruled out.  Given that there is some evidence for a mutagenic mode of action, a cancer
hazard is presumed possible at environmental levels of exposure. This is consistent with EPA's
science policy position that assumes a nonthreshold effect for carcinogens with a mutagenic
component in the absence of definitive data demonstrating a threshold or nonlinear mechanism.
Additional support for an environmental hazard also comes from a comparison of the estimated
environmental levels to the estimated occupational exposure levels where risk is seen.  Given
that there is only a minimal margin between environmental and occupational exposure ranges, if
not an overlap, the extrapolation of observable hazard from the occupational setting to the
ambient environment is relatively confident. Because of insufficient information, the human
carcinogenic potential of DE by oral and dermal exposures cannot be determined.
       Several organizations previously have reviewed available relevant data and evaluated the
potential human carcinogenicity of DE or the particulate component (DPM) of DE. Similar
conclusions were reached by various organizations (see Table 7-9). For example, some
organizations have concluded that DE is probably carcinogenic to humans (IARC, 1989; IPCS,
1996),  or reasonably is anticipated to be a carcinogen (NTP, 2000).
       Overall, the weight of evidence for potential human carcinogenicity for DE is considered
strong, even though inferences are involved. Uncertainties are present, however, and include the
following unresolved issues.
       First, there has been a considerable scientific debate about the significance of the
available human evidence for a causal association between occupational exposure and  increased
lung cancer risk.  Some experts view the evidence as weak and/or inconsistent while others
consider the evidence compelling.  This is due to a lack of consensus  about whether the effects
of smoking and other potential confounders have been adequately accounted for in key studies,
and the lack of agreed-upon historical  DE exposure data for the available studies.
       Second, while the mode of action for DE-induced lung tumors in rats from high exposure
is sufficiently understood, the mode of action for the DE lung cancer risk in humans is not
known. To date, available evidence for the role of both the adsorbed  organics and the  carbon
core particle has only been shown under high-exposure experimental  animal test conditions.
There is virtually no information about the relative role of DE constituents in mediating
carcinogenic effects at the low-exposure levels.

                                         9-15

-------
       Additional research is needed to address these issues to reduce the uncertainty associated
with the potential cancer hazard of exposure to DE.
       The relevance of this hazard characterization to current ambient DE exposures hinges on
recognizing that the health effects data are derived from engine technologies and fuels that
existed in the past, and that some changes in the DE exhaust mixture have occurred and can be
expected in the future. Although decreases in amount and changes in composition of DE
emissions have occurred over time for on-road engines, a change is slow to manifest in the
environment because, for example, vehicular fleet turnover is slow and the change is slow to
dominate across an engine fleet. Available studies have not focused on the potential
toxicological  effect of the emission changes. There is no compelling evidence at present to show
that past  and present exhaust characteristics are so lexicologically dissimilar as to render the
current use of the assessment's findings outdated.

9.5.  DOSE-RESPONSE ASSESSMENT
       In assessments of estimated human health risks, human data from environmental
exposures are always preferred over animal data, if available, as their use obviates the need for
extrapolation across species, e.g., from animals to humans. However,  for most environmental
agents, available health effects information is generally limited to occupational exposures  in
studies of humans (e.g., workers) or high experimental exposures to laboratory animals. For the
agents with high-exposure data compared to environmental exposure levels of interest, dose-
response assessment is performed in two  steps: assessment of data in the observable range to
derive a point of departure (which usually is the lowest exposure or dose that induces some,
minimal, or no apparent effects), followed by extrapolation to lower exposures to the extent
necessary.  Extrapolation to low exposures is ideally based on the understanding of mode(s) of
toxic action of the agent which allows the development and use of a mode of action specific
exposure-response model. In the absence of sufficient data, default methods and models are
used to extrapolate to the lower exposure levels.
       For DE, there is sufficient evidence to conclude that acute or short-term inhalation
exposure at relatively high levels can cause irritant effects to the eye, nose, and throat,
respiratory  symptoms, and neurophysiological  symptoms such as headache, nausea, etc.,
however, no quantitative data are available to derive an estimate of human exposure that is not
likely to elicit irritant and inflammatory effects in humans.
       There is also sufficient evidence to support the conclusion that DE has the potential to
cause cancer and noncancer effects of the lung from long-term inhalation exposure. Chapters 6
and 8 provide dose-response analyses related to the noncancer and cancer hazards to humans,
respectively, from lifetime exposure to DE. A dose-response analysis  to estimate the expected

                                          9-16

-------
response at environmental exposure levels has less uncertainty the closer the animal test or
estimated human epidemiologic-related exposures are to the environmental levels of interest.
With increasing exposure margins (EM), and thus a greater range of extrapolation, the
uncertainty about the shape of the dose-response curve in the region of low-dose extrapolation
increases and the possibility of a zero risk cannot be ruled out.

9.5.1. Evaluation of Risk for Noncancer Health Effects
       As discussed previously (Section 9.4.2.1), the evidence for potential chronic noncancer
health effects of DE is based primarily  on findings from chronic animal inhalation studies
showing a spectrum of dose-dependent chronic inflammation and histopathological changes in
the lung in several animal species including rats, mice, hamsters, and monkeys. A limited
number of epidemiologic investigations of workers exposed to DE have not provided
consistently clear evidence of significant chronic respiratory effects associated with DE
exposure. On the other hand, the relatively large epidemiologic database for ambient PM shows
a clear relationship between respiratory effects and ambient fine PM that is partially composed
of DPM. The specific role of DPM or any other source-related constituent of ambient PM in
causing the observed respiratory effects has not been defined.
       The approach taken in this assessment to estimate a level of DE in the air to which
humans may be exposed throughout their lifetime without an appreciable risk of deleterious
effects is to derive a reference concentration (RfC) for DE based on the consistent data for
respiratory inflammation in the rat studies. This approach assumes that humans would respond
to DE similarly to the tested animals under similar exposure conditions.  An uncertainty of this
approach stems from the circumstance that animal studies have used high DE exposures, and the
animal results must be translated to humans as well as to lower exposure levels since the
potential chronic health effects of DE in humans at environmental exposure levels cannot be
ascertained from the available DE human data.
       It also is relevant to recognize that DPM is a component of ambient fine PM and that
there is a relative wealth of human effects data for ambient PM showing a similarity of certain
adverse health effects for DPM and  ambient fine PM. This allows one to reasonably expect that
the PM2 5 National Ambient Air Quality Standard (NAAQS) would provide a measure of
protection from DPM, reflecting DPM's current and approximate proportion to PM25. Ambient
PM2 5 has been shown to be statistically associated with increased mortality (especially among
people over 65 years of age with preexisting cardiopulmonary conditions) and morbidity, as
measured by increases in hospital admissions, respiratory  symptom rates, and decrements in lung
function with both long- and short-term changes in ambient PM2 5 concentrations.
                                          9-17

-------
9.5.1.1. Chronic Reference Concentrations for Diesel Exhaust
        An inhalation Reference Concentration (RfC) is based upon long-term data, i.e., chronic
exposure, and can be derived from either human or animal data. An RfC is correctly defined as
"an estimate of a continuous inhalation exposure to the human population, including sensitive
subgroups, with uncertainty spanning perhaps an order of magnitude, that is likely to be without
appreciable risks of deleterious noncancer effects during a lifetime."  The RfC methodology
assumes that there is an exposure threshold below which effects will  not occur.  The RfC is not a
bright line; rather, as the long-term human exposure increases above  the RfC, the margin of
protection decreases.
       With the absence of DE exposure-response data in humans, this assessment derives an
RfC for DE based on dose-response data from four chronic inhalation studies in rats (Mauderly
et al., 1987; Ishinishi et al.,  1988; Heinrich et al., 1995; Nikula  et al., 1995). All of these studies
used DPM (expressed as |ig/m3) as a measure of DE exposure.  The pulmonary effects, including
inflamation and histopathologic lesions, were considered to be the critical noncancer effects. As
shown in Table 6-2, the no-observable-adverse-effects levels (NOAELs), the lowest-observable-
adverse-effects levels (LOAELs), and the adverse effects levels (AELs) for lung inflammation
and histopathologic changes were identified for the first three studies. For the Nikula et al.
study, lower 95% confidence estimates of the concentrations of DPM associated with a 10%
incidence (BMCL10) of chronic pulmonary  inflammation and fibrosis were derived since
NOAEL's were not observed. For all four studies, human equivalent concentrations (HECs)
corresponding to the animal NOAEL, LOAEL, AEL, and BMCL10 were then computed using a
dosimetry model developed by Yu et al. (1991) as described in  Chapter 6, Section 6.5.2, and
Appendix A.  The dosimetry model accounts for species differences (rat to human) in respiratory
exchange rates, particle deposition efficiency, differences in particle  clearance rates at high and
low doses,  and transport of particles to lymph nodes. The purpose is to identify the highest HEC
value with no apparent effect, i.e., NOAE!^,..
       The highest NOAELj^ associated with no apparent effect is  144 |ig/m3 from the
Ishinishi  et al. (1988) study; this then becomes the point of departure for deriving an RfC. To
obtain the RfC, this point of departure was  divided by two types of uncertainty factors (UF): a
factor of 3 recognizes interspecies (i.e., rat to human) extrapolation uncertainties, and a factor of
10 reflects uncertainties about interindividual human variation in  sensitivity. An evaluation of
the interspecies extrapolation issues for dosimeteric and pharmacodynamic equivalence between
rats and humans showed that although some adjustments could  be accounted for, there remained
a residual uncertainty, and thus an uncertainty of 3 out of a possible factor of 10 is used.  In the
absence of mechanistic or specific data,  a default value of 10 is considered appropriate to
account for possible human variability in sensitivity, particularly for  children and people with

                                          9-18

-------
preexisting respiratory conditions. The spectrum of the population that may have a greater
susceptibility cannot be better characterized until there is additional knowledge about mode of
action. The resulting RfC for DE is 5 |lg/m3 of DPM.
       Overall, the confidence level in the RfC is considered medium in a range of low to high
confidence.  A principal uncertainty of the RfC analysis is the reliance on animal data to predict
human risk.  The critical effects, chronic inflammation, and pathologic changes, which are well
characterized in four animal species, are considered relevant to humans.  Collective evidence for
all poorly soluble particles, including DPM, indicates that the rat is the most sensitive laboratory
animal species tested to date. Although in general the rat is thought to be more sensitive to lung
injury than humans to poorly soluble particles (ILSI, 2000), it is not clear that this is the case
specifically for diesel. We must recall that DE is a mixture of not just carbon particles  but also
various organics, both on the particles and in gases. In addition, differences in particle
deposition, retention, and clearance mechanisms have been largely but perhaps not completely
addressed by the use of the rat-to-human dosimetry model.  The use of rat data is not likely to
grossly underestimate the human risk for pulmonary noncancer health effects. In terms of the
potential for other critical health effects, there is  growing evidence suggesting that DE  can
exacerbate allergenic effects to known sensitizers, while also evoking production of biochemical
markers typically associated with asthma.  Some work in this area indicates that humans may be
as sensitive as rats and mice to the immunologic  effects (Chapter 6, Section 6.3.4). This
database is currently lacking key exposure-response data, but may in the future provide an
alternative basis for RfC derivation.  It also should be noted that the ambient PM health effects
data show a broader array of adverse human health concerns (e.g., cardiovascular effects, as well
as acute exposure effects). With DPM being a ubiquitous component of ambient PM, there is an
uncertainty about the adequacy of the existing DE noncancer database to identify all of the
pertinent DE-caused noncancer health hazards.

9.5.1.2. Risks Based on Ambient PM25
       As discussed in Chapter 6 (Section 6.4), in 1997 EPA established an annual NAAQS for
PM25, at a level of 15 |ig/m3 to provide protection against adverse health effects associated with
both long- and short-term exposures to ambient fine PM.  DPM is a typical constituent of
ambient fine PM (generally about 10% of PM25 with some examples up to 36%).2  Given the
       2"A qualitative comparison of adverse effects of exposure to DPM and ambient fine PM
shows that the respiratory system is adversely affected in both cases, though a wider spectrum of
adverse effects has been identified for ambient fine PM. In contrast to the diesel PM database,
there is a wealth of human data for fine PM noncancer effects which indicates that the health
effects from fine PM do not have a discernable threshold at this time."
                                          9-19

-------
similarity of health concerns for respiratory inflammation and pulmonary health effects from
both DPM and fine particles, it is reasonable to expect that DPM contributes to some of the
health effects associated with PM2 5. Current knowledge is insufficient, however, to describe the
relative potencies of DPM and the other components of PM25 As long as the percentage of
DPM to total ambient PM25 remains in similar proportion, protective levels for PM25 would be
expected to offer a measure of protection from effects associated with DPM.

9.5.1.3. Conclusions
       This assessment estimates an exposure air level of DE (as measured by DPM) to which
humans may be exposed throughout their lifetime without experiencing any adverse noncancer
health effects.  The approach taken applies the RfC method using data  specific to DE to produce
an RfC of 5 |lg/m3 of DPM on the basis of four chronic inhalation studies of DE in rats and a
composite uncertainty factor of 30. In addition, this assessment also recognizes the relative
wealth of data regarding health effects associated with ambient PM and presumes that a health
protective level for PM25 also would be expected to provide a measure of protection from DPM,
a constituent part of PM25.  The PM25 standard of 15 |ig/m3 as an annual average thus is
expected to provide a measure of protection from DPM noncancer health effects, reflecting
DPM's current approximate proportion to PM25.

9.5.2.  Evaluation of Cancer Risks
       As discussed in Section 9.4.3, the combined weight of evidence indicates that DE has the
potential to pose a cancer hazard to humans at anticipated levels of environmental exposure.  The
target organ of DE-induced carcinogenicity is the lung.  Strong evidence exists for a causal
relationship between risk for lung cancer and occupational exposure to DE in certain
occupational workers such  as railroad workers, truck drivers, heavy-equipment operators (e.g.,
shipyard, diesel farm equipment, and construction), and transit workers. The evidence, however,
was less than sufficient to confidently characterize DE as carcinogenic to humans, and instead
the assessment concludes that DE is likely to be a human carcinogen. It also has been shown
unequivocally in several studies that DE can cause benign and malignant lung tumors in rats in a
dose-related manner following chronic inhalation exposure to high concentrations; however, this
response is  not thought applicable to predict a hazard to humans exposed at lower environmental
levels.  The mechanism(s) by which DE would induce lung cancer in humans has not been
established, but available data suggest that mutagenic and nonmutagenic modes of action are
possible. Hence, for estimating DE cancer risk at low environmental exposures, linear low-dose
extrapolation would be considered an appropriate default assumption, which is consistent with
EPA's science policy position that in the absence of an understanding of modes of carcinogenic

                                          9-20

-------
action, a nonthreshold effect is to be presumed (U.S. EPA, 1986, 1996a). This same assumption
has been used by other organizations/risk assessors who have previously used either linear risk
extrapolation models or mechanistically based models to estimate cancer risk from
environmental exposure to DE (e.g., WHO-IPCS, 1996; Cal EPA, 1998; also see Appendix C).
       Dose-response assessment is based on either human or animal data, although human data
are always preferred if available.  Several quantitative assessments have been conducted by
organizations and investigators on the basis of both occupational data and rat data (see Appendix
C). However, more recent evidence indicates that DE causes tumors in the rat lung via a mode
of action that involves impairment of lung clearance mechanisms (referred to as "lung overload
response") associated with high exposures. This lung overload response is not expected in
humans exposed to environmental levels (nor is it expected to occur in many occupational
exposures), and thus the rat lung tumor dose-response data are not considered suitable for
predicting human risk at low environmental exposures. Given that the rat data are not
appropriate for estimating cancer risk to humans, this assessment focuses on using the
occupational epidemiologic data for estimating environmental risk of DE to humans.
       Even though occupational data are considered most relevant for use in dose-response
assessment, uncertainties exist, including the following issues:

       •  the use of DPM (expressed as |ig/m3) as a surrogate dosimeter for DE exposure,
          given that the relative roles of various constituents in mediating carcinogenic effects
          and the mode of carcinogenic action are still unknown;
       •  the representativeness of occupational populations for the general population and
          vulnerable subgroups, including infants and children and individuals with preexisting
          diseases, particularly respiratory conditions;
       •  the lack of actual DE exposure data for workers in the available epidemiologic
          studies;
       •  possible confounders (smoking and asbestos exposure) that could contribute to the
          observed lung cancer risk in occupational  studies of DE if the control for these
          confounders is not adequate; and
       •  whether or not an exposure-response relationship for occupational lung cancer risk
          can be estimated for DE.

Chapter 8, Section 8.3, provides a discussion of these uncertainties, along with an evaluation of
the suitability of available occupational studies for a derivation of a cancer unit risk estimate for
DE. Unit risk is defined as the estimated upper-bound cancer risk at a specific exposure or dose
                                          9-21

-------
from a continuous average lifetime exposure of 70 years (in this case, cancer risk per |lg/m3 of
DPM).
       Among the occupational studies, the railroad worker studies (Garshick et al., 1987, 1988)
and the Teamsters Union truck driver studies (Steenland et al., 1990, 1998) are considered to
have the best available exposure data for possible use in establishing exposure-response
relationships and deriving a cancer unit risk. There have been different views on the suitability
of these studies for estimating environmental cancer risks (e.g., Cal EPA, 1998; HEI, 1995,
1999). Given the equivocal evidence for the presence or absence of an exposure-response
relationship for the study of railroad workers, and exposure uncertainties for the study of truck
drivers, it is judged that available data are too uncertain at this time for the development of a
confident quantitative dose-response analysis and subsequent derivation of cancer unit risk for
DE.
       In the absence of a cancer unit risk to assess population cancer risk, this assessment
provides a "perspective" about the possible magnitude of risk in the population from
environmental exposure to DE. One approach to estimating the possible magnitude of risk
involves  simply noting that risks to the general public would be low in comparison with
occupational risk if the differences in the lower environmental exposures compared to the higher
occupational exposures are large. If the differences are small, the environmental risks would
approach the workers' risk observed in studies of past occupational exposures.  A comparison of
environmental equivalent occupational and ambient environmental exposures showed that for
certain occupations, there is a potential for overlap between environmental exposure and the
estimated environmental equivalent occupational exposure, while in other cases the
environmental exposures could be up to about 100-fold lower than the occupational levels (see
Table 8-1).  For the exposure overlap case,  one can infer that the environmental risk could be the
same as, or approach, the risk magnitudes observed in the occupational studies. In the 100-fold
lower case, the environmental risk could be about 100-fold lower than the observed risk
magnitudes in  the occupational studies.  Risks to the general public are of potential concern
when a significant risk is seen in the occupational setting and the difference between
occupational and ambient exposure may overlap or is relatively small (within one to two orders
of magnitude).
       A second approach, which is related to the first approach but more quantitative, is to
estimate possible ranges of lung cancer risk from occupational exposures to DE, and then use a
proportional relationship of exposure differences (e.g., EMs) to scale the occupational risk to the
environmental exposure setting. Given the range of observed relative risks or odds ratios of lung
cancer in a number of occupational studies, a relative risk increase of 1.4 was selected as a
reasonable estimate of occupational risk for the purpose of this analysis. The relative risk of 1.4

                                          9-22

-------
means that the workers faced an extra risk that is 40% higher than the approximate 5%
background lifetime lung cancer risk in the U.S. population. Using the relationship [excess risk
= (relative risk-1) x background risk], 2% of these DE-exposed workers (i.e., 10"2 risk) would
have been at risk (and developed lung cancer) attributable to occupational exposure to DE.
       Using a nationwide average environmental exposure (0.8 |lg/m3 DPM), and assuming (a)
the excess lung cancer risk from occupational exposure is about 10"2; and (b) the past
occupational exposures were no higher than about 1,200 |lg/m3 (equivalent to an environmental
equivalent EM of 315, connoting a relatively large EM), the environmental cancer risk would
fall between 10"4 and 10"5.  The selection of 1,200 |lg/m3 is a very high value intentionally
selected to illustrate a high-end exposure boundary and thus a lower bounding of risk calculated
by this exploratory approach.  On the other hand, if occupational exposures for some groups
were lower, for example, closer to 100 |lg/m3 (equivalent to an environmental equivalent EM of
26, connoting a smaller EM), the environmental risk would be higher and approach 10"3. The
selection of 100 |ig/m3 is purposefully toward the lower end of the reported occupational
exposure range which spans 7-403 |ig/m3  in Table 8-1. The risk estimates are attended by
numerous uncertainties; their inclusion in this document does not constitute Agency
endorsement of their validity as a surrogate for cancer unit risk; the range of values is not useful
for estimating numbers of cancer cases; and the range of possible risk from environmental
exposures also could be lower and a zero risk cannot be ruled out.
       These types of exploratory analyses are not intended to be precise or provide a definitive
characterization of cancer risk but are useful in illustrating and  gauging the possible range of risk
based on applying reasonable judgment. The analyses provides a sense of where an upper limit
(or "upper bound") of the risk may be. These analyses are subject to uncertainties, particularly
the lack of actual exposure information for the occupational epidemiologic studies and the use of
public-health-conservative risk assessment assumptions. The possible risks also could be lower
and a zero risk cannot be ruled out because (a) some individuals in the population may have a
high tolerance to exposure from DE and therefore not be susceptible to cancer from
environmental exposure, and (b) although  not reported, there could be a threshold of exposure
below which there is no cancer risk.  Given these circumstances, we refer to this risk analysis as
a "perspective" on possible risks. Best scientific judgment guided the selection of assumptions
and other elements of this analysis which are deemed reasonable and appropriate for identifying
possible risks based on the information currently available. Further research is needed to more
accurately assess and characterize  environmental cancer risks from DE.
                                          9-23

-------
9.6.  SUMMARY AND CONCLUSIONS
       The available health effects data show that acute (short-term episodic exposure) and
chronic (long-term) exposure to DE can pose hazards to humans and that environmental
exposures, in some cases, may have a risk.
       At relatively high acute exposures, DE can cause acute irritation to the eye and upper
respiratory airways and symptoms of respiratory irritation which may be temporarily
debilitating.  Evidence also shows that DE has immunological toxicity that can induce allergic
responses (some of which are also typical of asthma) and/or exacerbate existing respiratory
allergies. While the hazard potential is important for these acute and short exposure-related
effects, quantitative dose-response estimates for these effects could not be developed because of
the lack of exposure-response information.
       It is concluded that long-term exposure to low levels of DE poses a hazard for chronic
inflammation and pathological changes in the human lung. A level of human lifetime exposure
thought to be without appreciable risk for lung damage  is estimated to be 5 |ig/m3 of DPM, this
being a calculated RfC value for DE.  Because DPM is  a constituent of ambient PM2 5 and there
is some similarity in potential adverse effects from DE and PM2 5, it is expected that a measure of
protection from health effects associated with DE is provided by the 1997 annual PM2 5 NAAQS,
set at a level of 15 |ig/m3.
       DE is considered to pose a human lung carcinogenicity hazard, which is expressed in a
weight-of-evidence conclusion that DE is judged to be a "probable" human carcinogen, or is
"likely to be carcinogenic in humans by inhalation" at environmental or higher exposure
conditions.  Because  of uncertainty in the available exposure-response data, a cancer unit
risk/cancer potency for DE has not been derived. One should note that the closeness of the high-
end environmental exposures and low-end estimates of occupational exposure suggest less
uncertainty in the extrapolation of hazard and possible risk to the environmental setting.
Exploratory  analyses using public health conservative assumptions provides a perspective on the
possible range of lung cancer risk from environmental exposure to DE. Best scientific judgment
guided the selection of assumptions and other elements  of this analysis which are deemed
reasonable and appropriate for identifying possible risks based on the information currently
available.  These analyses indicate that lifetime cancer risk may exceed 10"5 and could be as high
as 10"3 or nearly so, though considering the assumptions used and the uncertainties, lower risk is
possible and a zero risk cannot be ruled out. This range of values is attended by numerous
uncertainties, the inclusion of the range in this assessment does not constitute Agency
endorsement of their validity as surrogates for cancer unit risk values, and the range is not
suitable for estimating numbers of cancer cases. These  risk findings should not be viewed as a
definitive characterization of risk.
                                          9-24

-------
       Even though the evidence for potential human health hazards for DE is convincing and
persuasive, uncertainties exist because of the use of assumptions to bridge data and knowledge
gaps about human exposures to DE and the underlying mechanisms by which DE may cause the
observed toxicities in humans and animals. A notable uncertainty of this assessment is how the
physical and chemical nature of DE emissions has changed over the years because the
toxicological and epidemiologic observations are based on older engines and their emissions, yet
the desire is to focus on the potential health hazards related to exposure from present-day or
future emissions. There have been changes in the physical and chemical composition of some
DE emissions (on-road vehicle emissions) over time, though there is no definitive information to
show that the emission changes portend significant toxicological changes.  The mode(s) of action
for DE toxicity in humans is not understood, and hence knowledge is lacking  about the role of
exhaust mixture components in modulating the toxicity.  Taken together, these considerations
have lead to a judgment that the hazards identified from older technology-based exposures are
applicable to current-day exposures. As new and cleaner diesel engines replace a substantial
number of existing engines, the general applicability of the conclusions in this assessment will
need to be reevaluated.
       Other uncertainties include the assumptions that health effects observed at high doses
may be applicable to low doses, and that toxicologic findings in laboratory animals are
predictive of human responses. Also, the available data are  not sufficient to demonstrate the
absence or  presence of an exposure/dose-response threshold in humans for DE toxicity at
environmental exposures. Again, this is due in part to the lack of understanding of how DE may
cause adverse health effects in exposed humans and laboratory animals. Although there are
hypotheses about the specific mechanisms by which DE might cause cancer and other toxicities,
no specific biological pathways or specific constituents of DE have been firmly established as
responsible for low-dose effects. The assumptions used in this assessment, i.e., the presence of
a biological threshold for chronic respiratory effects based on cumulative dosage and the absence
of a threshold for lung cancer stemming from subtle and irreversible effects, are considered
prudent and reasonable default choices.
       The characterization of health hazards and risks contained in this document assumes that
the potential DE  health hazards are relevant for long-term exposures, up to and including
lifetime exposures, and would apply to a wide spectrum of individuals but not necessary those
that would  have significant differential susceptibility. There is no DE-specific information that
provides  direct insight into the question  of differential susceptibility within the general human
population  or vulnerable subgroups, for  example, children or the elderly. Although default
approaches to account for interindividual variation have been included in the derivation of the
noncancer effects RfC (i.e., use of an uncertainty factor of 10), this may or may not adequately

                                          9-25

-------
protect certain subgroups that could be more vulnerable.  Differential susceptibility to DPM
among individuals in the population would be due to differences in dosimetry (i.e., differences in
retained particle mass or number in the lung) and/or differences in respiratory system tissue
response sensitivity.  From the dosimetry perspective, we understand that age, gender, and
disease status can influence deposition in the lung and other areas of the respiratory tract (U.S.
EPA, 1996b, Section 10.7.7).  For example, given that DE chronic toxicity is focused on the
respiratory system, vulnerable subgroups might include those individuals who predispose their
lungs to increased particle retention (e.g., smoking, high particulate burdens from nondiesel
sources) or those having existing respiratory or lung inflammation, repeated respiratory
infections, or chronic bronchitis or asthma.  For children, there is also the hypothesis of possible
increased sensitivity to exposure, given the ongoing processes of development from birth to
maturation, of the respiratory  and immune systems.
       Despite the uncertainties regarding intraspecies variability, the default approach of using
an uncertainty factor of 10 in the derivation of the noncancer effects RfC to account for possible
interindividual variation in the toxic response to DE exposure is appropriate and reasonable
given the lack  of DE-specific  data.
       Variation in DE exposure is another source of uncertainty. Because of variation in
human activity patterns and their proximity to DE sources of emissions, different population
subgroups could potentially receive higher or lower exposure to DE. The highest exposed are
clearly occupational subgroups whose job brings them very close to DE sources, such as diesel
engine vehicle drivers and workers, diesel powered machinery operators,  some underground
miners, etc.  High exposures in the general population would be to those living very near or
having time outdoors in proximity to DE sources as well  as those engaged in activities that cause
high breathing rates where DE is present. Accordingly, where appropriate, analyses in this
assessment have included possible high-end DE exposures in addition to the lower nationwide
average exposure estimates.
       Lastly, this assessment considers only potential heath effects from exposures to DE
alone. DE exposure could be additive or synergistic to concurrent exposures to other air
pollutants. For example, there is evidence that DPM that has been altered by being in the
presence of ambient ozone significantly increases the rat  lung inflammatory effect compared to
DPM that was not subjected to ozone (Madden et al., 2000). This observation suggests a
hypothesis that inflammation-related noncancer hazards of airborne DPM may be worsened by
the increasing  presence of ozone in the ambient air.  Other concerns include the possible
impacts for children and adults on the exacerbation of existing allergens resulting from DE
exposure. However,  in the absence of more definitive data demonstrating interactive effects
                                          9-26

-------
from combined exposures to DE and other pollutants, it is not possible to further address these
issues at this time.
                               REFERENCES FOR CHAPTER 9

California Environmental Protection Agency-OEHHA (Cal EPA). (1998) Part B: Health risk assessment for diesel
exhaust, Public and Scientific Review Draft. February 1998.

Garshick, E; Schenker, MB; Munoz, A; et al. (1987) A case-control study of lung cancer and diesel exhaust
exposure in railroad workers.  Am Rev Respir Dis 135:1242-1248.

Garshick, E; Schenker, MB; Munoz, A; et al. (1988) A retrospective cohort study of lung cancer and diesel exhaust
exposure in railroad workers.  Am Rev Respir Dis 137:820-825.

Health Effects Institute (HEI). (1995) Diesel exhaust: a critical analysis of emissions, exposure, and health effects.
Cambridge, MA: Health Effects Institute.

HEI. (1999) Diesel emissions and lung cancer: epidemiology and quantitative risk assessment. A special report of
the Institute's Diesel Epidemiology Expert Panel. Cambridge, MA: Health Effects Institute.

Heinrich, U; Fuhst, R; Rittinghausen, S; et al. (1995) Chronic inhalation exposure of Wistar rats and two different
strains of mice to diesel engine exhaust, carbon black, and titanium dioxide. Inhal Toxicol 7:553-556.

International Agency for Research on Cancer (IARC). (1989) Diesel and gasoline engine exhausts and some
nitroarenes. IARC monographs on the evaluation of carcinogenic risks to humans: v. 46. Lyon, France: World
Health Organization; pp. 41-185.

International Life Sciences Institute (ILSI). (2000) ILSI Risk Science Institute workshop: the relevance of the rat
lung response to particle overload for human risk assessment. Gardner, DE, ed. Inhal Toxicol 12(1-2).

International Programme on Chemical Safety (IPCS), World Health Organization. (1996) Diesel fuel and exhaust
emissions. Environmental Health Criteria 171. Geneva: World Health Organization.

Ishinishi, N; Kuwabara, N; Takaki, Y; et al. (1988) Long-term inhalation experiments on diesel exhaust. In: Diesel
exhaust and health risks. Results of the HERP studies. Ibaraki, Japan: Research Committee for HERP Studies; pp.
11-84.

Madden, M; Richards, J; Dailey, L; et al. (2000) Effect of ozone on diesel exhaust toxicity in rat lung.  Toxicol Appl
Pharmacol 168:140-148.

Mauderly, JL; Jones, RK; Griffith, WC; et al.  (1987) Diesel exhaust is a pulmonary carcinogen in rats exposed
chronically by inhalation. Fundam Appl Toxicol 9:208-221.

National Toxicology Program (NTP). (2000) 9th report on carcinogens. Public Health Service, U.S. Department of
Health and Human Services, Research Triangle Park, NC. Available from: http://ntp-server.niehs.nih.gov.

Nikula, KJ; Snipes, MB; Barr, EB; et al. (1995) Comparative pulmonary toxicities and carcinogenicities of
chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl Toxicol 25:80-94.

Steenland, K; Silverman, DT; Hornung, RW. (1990) Case-control study of lung cancer and truck driving in the
Teamsters Union. Am J Public Health 80:670-674.
                                                 9-27

-------
Steenland, K; Deddens, J; Stayner, L. (1998) Diesel exhaust and lung cancer in the tracking industry: exposure-
response analyses and risk assessment. Am J Ind Med 34:220-228.

U.S. Environmental Protection Agency (EPA). (1986) Guidelines for carcinogen risk assessment. Federal Register
51(185):33992-34003.

U.S. EPA. (1996a) Proposed guidelines for carcinogen risk assessment. Office of Research and Development.
Federal Register 61(79):17960-18011. EPA/600/P-92/003C.

U.S. EPA. (1996b) Air quality criteria for paniculate matter.  National Center for Environmental Assessment;
Research Triangle Park, NC: EPA/600/P-95/001aF.

U.S. EPA. (1999) Guidelines for Carcinogen Risk Assessment. Review Draft. NCEA-F-0644, July. Risk
Assessment Forum, Washington, DC. http://www.epa.gov/ncea/raf/cancer.htm.

U.S. EPA. (2000) National air pollutant emission trends, 1900-1998. EPA-454/R-00-002, March.

World Health Organization (WHO-IPCS). (1996) Diesel fuel and exhaust emissions. Environmental Health Criteria
171, WHO-IPCS, Geneva, Switzerland.

Yu, CP; Yoon, KJ; Chen, YK. (1991) Retention modeling of DE particles in rats and humans. J Aerosol Med
4:79-115.
                                                 9-28

-------
              Appendix A
Calculation of Human Equivalent Continuous
     Exposure Concentrations (HECs)
                  A-l

-------
A.l.  INTRODUCTION
       As discussed in Chapter 3, the lung burden of diesel particulate matter (DPM) during
exposure is determined by both the amount and site of particle deposition in the lung and,
subsequently, by rates of translocation and clearance from the deposition sites. Mathematical
models have often been used to complement experimental studies in estimating the lung burdens
of inhaled particles in different species under different exposure conditions. This appendix
presents a mathematical model that simulates the deposition and clearance of DPM in the lungs
of rats and humans of Yu et al.(1991) also published as Yu and Yoon (1990).
       Diesel particles are aggregates formed from primary spheres 15-30 nm in diameter.  The
aggregates are irregularly shaped and range in size from a few molecular diameters to tens of
microns.  The mass median aerodynamic diameter (MMAD) of the aggregates is typically 0.2
|im and is poly disperse with a  geometric standard deviation of around 2.3. The organics
adsorbed onto the aggregates normally account for 10% to 30% of the particle mass. However,
the  exact size distribution of DPM and the specific composition of the adsorbed organics depend
upon many factors, including engine design, fuels used, engine operating conditions, and the
thermodynamic process of exhaust. The physical and chemical characteristics of DPM have
been reviewed extensively by Amann and Siegla (1982) and Schuetzle (1983).
       Four mechanisms deposit DPM within the respiratory tract during exposure:  impaction,
sedimentation, interception, and diffusion.  The contribution from each mechanism to deposition,
however, depends upon lung structure and size, the breathing condition of the subject, and
particle size distribution. Under normal breathing conditions, diffusion is the most dominant
mechanism and the other three mechanisms play minor roles.
       Once DPM is deposited in the respiratory tract,  both the carbonaceous core and the
adsorbed organics will be removed from the deposition sites by mechanical clearance, provided
by mucociliary transport in the ciliated conducting airways as well as macrophage phagocytosis
and migration in the nonciliated airways, and dissolution. As the carbonaceous core or soot of
DPM is insoluble, it is removed from the lung primarily by mechanical clearance, whereas the
adsorbed organics are removed principally by dissolution (Chapter 3).

A.2.  PARTICLE MODEL
       To develop a mathematical model that simulates the deposition and clearance of DPM in
the  lung,  an appropriate model for diesel particles must be introduced.  For the deposition study,
an equivalent sphere model developed by Yu and Xu (1987) was used to simulate the dynamics
and deposition of DPM in the respiratory tract by various mechanisms. For the clearance study,
a diesel particle is assumed to be composed of three different material components according to
their characteristic clearance rates: (1) a carbonaceous core of approximately 80% of the particle

                                          A-2

-------
mass; (2) absorbed organics of about 10% of particle mass, which are slowly cleared from the
lung; and (3) adsorbed organics quickly cleared from the lung, accounting for the remaining 10%
of particle mass.  The presence of two discrete organic phases in the particle model is suggested
by observations that the removal of particle-associated organics from the lung exhibits a biphasic
clearance curve (Sun et al., 1984; Bond et al., 1986), as discussed in Chapter 3.  This curve
represents two major kinetic clearance phenomena:  a fast- phase organic washout with a half-
time of a few hours, and a slow phase with a half-time that is a few hundred times longer.  The
detailed components involved in each phase are not known. It is possible that the fast phase
consists of organics that are leached out primarily by diffusion mechanisms while the slow phase
might include any or all of the following components: (a) organics that are "loosened" before
they are released, (b) organics that have become intercalated in the carbon core and whose
release is thus impeded, (c) organics that are associated for longer periods of time because of
hydrophobic interaction with other organic-phase materials, (d) organics that have been ingested
by macrophages and as a result effectively remain in the lung for a longer period of time because
of metabolism by the macrophage (metabolites formed may interact with other cellular
components), and (e) organics that have directly acted on cellular components, such as the
formation of covalent bonds with DNA and other biological macromolecules to  form adducts.
       The above distinction of the organic components is general and made to  account for the
biphasic clearance of DPM; it does not specifically imply the actual nature of the adsorbed
organics. For aerosols made of pure organics, such  as benzo(a)pyrene (BaP) and nitropyrene
(NP) in the same size range of DPM, Sun et al. (1984) and Bond et al. (1986) observed a nearly
monophasic clearance curve.  This might be explained by the absence of intercalative
phenomena (a) and of hydrophobic interaction imposed by a heterogeneous mixture of organics
(b). The measurement of a pure organic might also  neglect that quantity which has become
intracellularly (c) or covalently bound (d).

A.3.  COMPARTMENTAL LUNG MODEL
       The model of Yu et al. (1991) comprises three principal compartments involved in
deposition and clearance:  tracheobronchial (T or TB), alveolar (A), and lung-associated lymph
node (L), as shown in Figure A-l. The outside compartments blood (B) and GI  tract (G) and
nasopharyngeal or head (H) are also represented.  The alveolar compartment in the model is
obviously the most important for long-term retention studies. However, for short-term
consideration, retentions in other lung compartments may also be significant.  The presence of
these lung compartments and the two outside compartments in the model therefore provides a
complete description of all clearance processes involved.
                                          A-3

-------
       In Figure A-1, r%  r $ and r fare, respectively, the mass deposition rates of DE material
component i (i=l [core], 2 [slowly cleared organics], and 3 [rapidly cleared organics]) in the
head, tracheobronchial, and alveolar compartments; and A$ represents the transport rate of
material component i from any compartment X to any compartment Y.  Let the mass fraction of
material component i of a diesel particle be L  Then

                                   4°  =ftra>                                   (A-1)
                                    ? =f,rT,                                   (A-2)
                                                                                 (A-3)
where rH, rT, and rA are, respectively, the total mass deposition rates of DPM in the H, T, and
A compartments, determined from the equations:
                              rH = c(TV}(RF}(DF)H ,                              (A-4)
                              rT = c(TV)(RF)(DF)T ,                              (A-5)
                              rA  = c(TV)(RF)(DF)A  .                              (A-6)
       In Equations A-4 to A-6, c is the mass concentration of DPM in the air, TV is the tidal
volume, RF is the respiratory frequency, and (DF)H, (DF)T, and (DF)A are, respectively, the
deposition fractions of DPM in the H, T, and A compartments over a respiratory cycle.  The

                                         A-4

-------
values of (DF)H, (DF)T, and (DF)A, which vary with the particle size, breathing conditions, and
lung architecture, were determined from the deposition model of Yu and Xu (1987).
       The differential equations for mj^, the mass of material component i in compartment X as
a function of exposure time t, can be written as

Head (H)
                           dt
Tracheobronchial (T)

                     dm®
                           _  \l)    \ \l)   \l)   \ \l)  \l)    \ \l)   \l)                     ( A  Q\
                           	 / rp   '  /V A 'T'ffl A    /VT~'/~'//iI rp    /V'T'D//iI ^T  .                    \"^"^~ /
                      dt
Alveolar (A)
                      dt

Lymph nodes (L)
                             dt

Equation A-9 may also be written as
                           dm?      m  m    m   m
                              L   -- xSW0 - ^f  •                            (A-10)
                                     /(/,/(      / ,ri   /,                              ^     '
where

is the total clearance rate of material component i from the alveolar compartment.  In Equations
A-7 to A- 10, we have assumed vanishing material concentration in the blood compartment to
calculate diffusion transport.
       The total mass of the particle-associated organics in compartment X is the sum of m (j
and m ^the total mass of DPM in compartment X is equal to
                             nx = mf + mf + mf                            (A-13)

                                          A-5

-------
The lung burdens of diesel soot (core) and organics are defined, respectively, as
and
                       (2)+(3)     (2)     (2)      (3)      (3)                         , .  n ,,
                     mLung   = mT   + ™A   + mT   + ™A    •                      (A-15)
Because the clearance of diesel soot from compartment T is much faster than from compartment
A, m y< m y a short time after exposure, Equation A-14 leads to

       Solution to Equations A-7 to A-10 can be obtained once all the transport rates AjQ are
known.  When X^Y are constant, which is the case in linear kinetics, Equations A-7 to A-10 will
have a solution that increases with time at the beginning of exposure but eventually saturates and
reaches a steady-state value. This is the classical retention model developed by the International
Commission of Radiological Protection (ICRP,  1979). However, as discussed in Chapter 3, data
have shown that when rats are exposed to DPM at high concentration for a prolonged period,
long-termed clearance is impaired.  This is the so-called overload effect, observed also for other
insoluble particles.  The overload effect cannot be predicted by the classical ICRP model.
Soderholm (1981) and Strom et al. (1987, 1988) have proposed a model to simulate this effect by
adding a separate sequestering compartment in the alveolar region. In the present approach, a
single compartment for the alveolar region of the lung is used and the overload effect is
accounted for by a set of variable transport rates X^, X^D and ^ which are functions of mA. The
transport rates $ and X^L in Equations A-7 to A-10 can be determined directly from experimental
data on lung and lymph node burdens,  and ^ and ^ from Equation A-12.

A.4.  SOLUTIONS  TO KINETIC EQUATIONS
      Equation A-l 1 is a nonlinear differential equation  ofm^ with known function
For diesel soot, this equation becomes

                                         A-6

-------
                            dm(l}
                            U'"A      (1)    ~ (I),   N   (1)
                            —- =      ~
Because clearance of the particle-associated organics is much faster than diesel soot, m^and m^}
constitute only a very small fraction of the total particle mass (less than 1%) after a long
exposure, and we may consider A^as a function of w^alone. Equation A-17 is then reduced to a
differential equation with m^the only dependent variable.
       The general solution to Equation A-17 for constant r^at any time, t, can be obtained by
the separation of variables to give
1
                               V     dm™
       If r^is an arbitrary function oft, Equation A-17 needs to be solved numerically such as
by a Runge-Kutta method. Once m^is found, the other kinetic equations A-7 to A-10 for both
diesel soot and the particle-associated organics can be solved readily, as they are linear
equations. The solutions to these equations for constant r$, r$ and rf are given below:
Head (H)
                                                             0                    (A-19)
                             where   KJ/ = KJ/G + KJ/B                             (A-20)
                                           A-7

-------
Tracheobronchial (T)


         m®  = exp (-X(;} O |   ( 4°  + ^AT mA  ) exP  <$ t} dt + m®          (A-21)
                             Jo

                             where  X®  = tifG  + X®                             (A-22)
Lymph nodes (L)
                            «
                    = exp(-X   0     x|»exp(X   t ) dt + m)                (A-23)
       In Equations A-19 to A-23, #z j|, represents the value ofm£} at t = 0.
       In the sections to follow, the methods of determining r$, r$ and r$ or (DF)H, (DF)T, and
(DF)A  r^ r(D^l and r(D^ as well as the values ofAjy in the compartmental lung model are
presented.

A.5.  DETERMINATION OF DEPOSITION FRACTIONS
       The mathematical models for determining the deposition fractions of DPM in various
regions of the respiratory tract have been developed by Yu and Xu (1986, 1987) and are adopted
in this report. Yu and Xu consider DPM as a polydisperse aerosol with a specified mass median
aerodynamic diameter (MMAD) and geometrical standard deviation og. Each diesel particle is
represented by a cluster-shaped aggregate within a spherical envelope of diameter de.  The
envelope diameter de is related to the aerodynamic diameter of the particle by the relation
                                                                              (A-24)
where C, is the bulk density of the particle in g/cm3, Q, = 1 g/cm3; 9 is the packing density, which
is the ratio of the space actually occupied by primary particles in the envelope to the overall
envelope volume; and Cx is the slip factor given by the expression:
                                         A-8

-------
                                                      -
                   Cx = 1 +  2[1.257 + 0.4 exp  -(—-^ )]                    (A-25)
                              £/                        A
in which A = 8 x 10"6cm3 is the mean free path of air molecules at standard conditions.  In the
diesel particle model of Yu and Xu (1986), £ has a value of 1.5 g/cm3 and a (p value of 0.3 is
chosen based upon the best experimental estimates.  As a result, Equation A-24 gives de/da =
1.35.  In determining the deposition fraction of DPM, de is used for diffusion and interception
according to the particle model.

A.5.1. Deposition in the Head
       Particle deposition in the naso- or oropharyngeal region is referred to as head or
extrathoracic deposition.  The amount of particles that enters the lung depends upon the
breathing mode. Normally, more particles are collected via the nasal route than by the oral route
because of the nasal hairs and the more complex air passages of the nose.  Since the residence
time of diesel particles in the head region during inhalation is very small (about 0.1 s for human
adults at normal breathing), diffusional deposition is insignificant and the major deposition
mechanism is impaction.  The following empirical formulas derived by Yu et al. (1981) for
human adults are adopted for deposition prediction of DPM:
For mouth breathing:

                            (DF)H m =  0, for d}<  3000                            (A-26)
                       =  "L117 + °'324 \°§>(daQlf°r daQ > 300°               (A-27)
                                        ex =  ,                                   (A-28)

and for nose breathing:

               (DF)H m =  -0.014 + 0.023 \og(d2aQ), for d2aQ  < 337                (A-29)
                                          A-9

-------
              (DF)H m =  -0.959 + 0.397 log«£>), for daQ > 337                (A-30)

               (DF)H ex = 0.003  + 0.033  \og(d*Ql for d^Q < 215                (A-31)

              (DF)H ex = -0.851  + 0.399  \og(d*Ql for d^Q > 215               (A-32)

where (DF)H is the deposition efficiency in the head, the subscripts in and ex denote inspiration
and expiration, respectively, da is the particle aerodynamic diameter in  jim, and Q is the air
flowrate in cm3/sec.
       Formulas to calculate deposition of diesel particles in the head region of children are
derived from those for adults using the theory of similarity, which assumes that the air passage in
the head region is  geometrically similar for all ages and that the deposition process is
characterized by the Stokes number of the particle.  Thus, the set of empirical equations from
A-26 through A-32 are transformed into the following form:
For mouth breathing:

                          (DF)H, m  = °> f°r daQ * 300°                          (A-33)
and for nose breathing:
        (DF)H m =  -1.117 + 0.972 logK  + 0.324 log(Wa2£>), for d*Q > 3000        (A-34)
                                       , ex
                                          =  0.                                  (A-35)
       (DF)H m  =  - 0.014 + 0.690 log K + 0.023 \og(d^Q\ for d*Q < 337        (A-36)
       (DF)H m  = -0.959 +  1.191 log K + 0.397 log (d2aQ), for d*Q > 337       (A-37)
         (DF)H ex = 0.003 + 0.099 log K  + 0.033  \og(d^Q), for d%Q <215         (A-38)
                                         A-10

-------
where K is the ratio of the linear dimension of the air passages in the head region of adults to that
of children, which is assumed to be the same as the ratio of adult/child tracheal diameters.

         (DF)H ex  =  0.851 + 1.197 log K + 0.399 \og(d^Q\  for d%Q >215          (A-39)

       For rats, the following empirical equations are used for deposition prediction of DPM in
the nose:
         (DF)H, tn = (DF^H, ex = °-046  +  °-009  logW), for dQ < 13.33          (A-40)

A.5.2.  Deposition in the Tracheobronchial and  Alveolar Regions
       The deposition model adopted for DPM is the one previously developed for
monodisperse (Yu, 1978) and poly disperse spherical aerosols (Diu and Yu, 1983). In the model,

        (DF)H m  = (DF)H m =  -0.522 + 0.514 \og(d2aQ\ for dfe > 13.33          (A-41)
the branching airways are viewed as a chamber model shaped like a trumpet (Figure A-2). The
cross-sectional area of the chamber varies with airway depth, x, measured from the beginning of
the trachea. At the last portion of the trumpet, additional cross-sectional area is present to
account for the alveolar volume per unit length of the airways. Inhaled diesel particles that
escape capture in the head during inspiration will enter the trachea and subsequently the
bronchial airways (compartment T) and alveolar spaces (compartment A).
       Assuming that the airways expand and contract uniformly during breathing, the equation
for the conservation of particles takes the form:
                                  )-  + Q     =  - Qc^                         (A-42)
                                   dx       dx
where c is the mean particle concentration at a given x and time t; At and A2 are, respectively,
the summed cross-sectional area (or volume per unit length) of the airways and alveoli at rest; r|
is the particle uptake efficiency per unit length of the airway; P is an expansion factor, given by:

                                           V.
                                  P =  1  + —                                   (A-43)
                                           v
                                             i
                                         A-ll

-------
and Q is the air flow rate, varying with x and t according to the relation
                                 —  = 1  -  —                                  (A-44)
                                  Q           V,                                  {    }
                                  *z-o            I
where Q0 is the air flow rate at x = 0.  In Equations A-43 and A-44, V, is the volume of new air
in the lungs and Vx and Vc are, respectively, the accumulated airway volume from x = 0 to x, and
total airway volume at rest.
       Equation A-42 is solved using the method of characteristics with appropriate initial and
boundary conditions. The amount of particles deposited between location xl and x2 from time tt
to t2 can then be found from the expression



                                      'n
                               DF =11 QcT\dxdt                                 (A-45)
                                      J J
                                      *i xi

       For diesel particles, T|  is the sum of those due to the individual deposition mechanisms
described above, i.e.,
where r^, T|s, T|P, and T|D are, respectively, the deposition efficiencies per unit length of the
                             'H  = Tl/ +  Tls +  ^p + "^D                             (A-46)

airway due to impact!on, sedimentation, interception, and diffusion. On the basis of the particle
model described above, the expressions for r^, T|s, T|P, and T|D are obtained in the following form:
                                                                                  (A-47)
                                                 - e2/3  + sin'1 e1/3]                (A-48)
                                                                                  (A-49)
                                          A-12

-------
                   t]   = l[l-0.819exp(-14.63A) -
                         L
                        0.0976 exp(-89.22A) -
              0.0325 exp(-228A) -  0.0509 exp(-125A2/3)]

                     T!  = IAI/Z (1  - 0.444A1/2)
                                                                                (A-50)
                                                                               (A-51)
for Reynolds numbers of the flow smaller than 2000, and for Reynolds numbers greater than or
equal to 2000, where ST=d2au/(18flR) is the particle Stokes number, 0 = L/(8R), 6 =
3jMsL/(32uR), F= d/R, and A = DL/(4R2u).  In the above definitions u is the air velocity in the
airway; |i is the air viscosity; L and R are,  respectively, the length and radius of the airway; us =
Cad2J(18fJ,) is the particle settling velocity; and D = CJfTfSn/JdJ is the diffusion coefficient
with k  denoting the Boltzmann constant and T the absolute temperature.  In the deposition
model, it is also assumed that r^ and T|P = 0 for expiration, while T|D and T|s have the same
expressions for both inspiration and expiration.
       During the pause, only diffusion and sedimentation are present.  The combined
deposition efficiency in the airway, E, is equal to:
                      E=\  -
                                                - ED) .
(A-52)
where Ep and Es are, respectively, the deposition efficiencies due to the individual mechanisms
of diffusion and sedimentation over the pause period. The expression for ED and Es are given by
ED = 1 -   L,  — exP(- <'
           i  =  lai
V «,
T i pvn
Z-^ 9/ exP
i = la.
4T1/2
ti^
,.(1 ^ 4,
^ (i - 2^ — )
i = la2
                                                                                 (A-53)
where TD = DT/R2 in which T is the pause time and al3 a2, and CC3 are the first three roots of the
equation:
                                  J0(a)  =  0 .
in which J0 is the Bessel function of the zeroth order, and:
                                                                             (A-54)
                    Es =  1.1094-u5 - 0.1604-4/or 0 < TS < 1.
                                         A-13
                                                                             (A-55)

-------
and
                   Es = 1 -   .           .            .,
                                   for ts > 1,

where TS = usx/2R.
       The values of (DF)T and (DF)A over a breathing cycle are calculated by superimposing
DF for inspiration, deposition efficiency E during pause, and DF for expiration in the
tracheobronchial airways and alveolar space. It is assumed that the breathing cycle consists of a
constant flow inspiration, a pause, and a constant flow expiration, each with a respective
duration fraction of 0.435, 0.05, and 0.515 of a breathing period.

A.5.3.  Lung Models
       Lung architecture affects particle deposition in several ways: the linear dimension of the
airway is related to the distance the particle travels  before it contacts the airway surface; the air
flow velocity by which the particles are transported is determined by the cross-section of the
airway for a given volumetric flowrate; and flow characteristics in the airways are influenced by
the airway diameter and branching patterns.  Thus,  theoretical prediction of particle deposition
depends, to a large extent, on the  lung model chosen.

A.5.3.1. Lung Model for Rats
       Morphometric data on the lung airways of rats were reported by Schum and Yeh (1979).
Table A-l shows the lung model  data for Long Evans rats with a total lung capacity of
13.784 cm3. Application of this model to Fischer rats is accomplished by assuming that the rat
has the same lung structure regardless of its strain and that the total lung capacity is proportional
to the body weight. In addition, it is also assumed that the lung volume at rest is about 40% of
the total lung capacity and that any  linear dimension of the lung is proportional to the cubic root
of the lung volume.

A.5.3.2. Lung Model for Human Adults
       The lung model of mature human adults used in the deposition calculation of DPM is the
symmetric lung model developed by Weibel (1963). In Weibel's model, the airways are
assumed to be a dichotomous branching system with 24 generations. Beginning with the 18th
generation, increasing numbers of alveoli are present on the wall of the airways, and the last
three generations are completely aleveolated. Thus, the alveolar region in this model consists of

                                          A-14

-------
all the airways in the last seven generations.  Table A-2 presents the morphometric data of the
airways of Weibel's model adjusted to a total lung volume of 3000 cm3.

A.5.3.3. Lung Model for Children
       The lung model for children in the diesel study was developed by Yu and Xu (1987) on
the basis of available morphometric measurements. The model assumes a lung structure with
dichotomous branching of airways, and it matches Weibel's model for a subject when evaluated
at the age of 25 years, the age at which the lung is considered to be mature. The number and size
of airways as functions of age t (years) are determined by the following equations.

A.5.3.3.1.  Number of airways and alveoli. The number of airways N;(t) at generation i for age t
is given by
                       N,(f) = 2',
                           for 0  < 7  <20
                                   (A-57)
                        \N21(f)=N(tl
                                    < 221
                                                                                 (A-58)
                          -.21
                 N22(f)  = Nr(f) -221,
                 N23(t)  = 0,
                             for 221   221 +
                                                               -.22
                                                                (A-60)
where Nr(t) is the total number of airways in the last three airway generations. The empirical
equation for Nr which best fits the available data is
Thus, Nr(t) increases from approximately 1.5 million at birth to 15 million at 8 years of age and

                                                                                (A-61)
remains nearly constant thereafter. Equations A-58 to A-60 also imply that in the last three
            2.036 x 107(l-0.926e-aiY),  t < 8
            1.468 x io7                t > 8
                                         A-15

-------
generations, the airways in the subsequent generation begin to appear only when those in the
preceding generation have completed development.
       The number of alveoli as a function of age can be represented by the following equation
according to the observed data:

                       NA(f)  =  2.985 x  108(1 -0.9\9e-°A5t)                       (A-62)


       The number of alveoli distributed in the unciliated airways at the airway generation level
is determined by assuming that alveolization of airways takes place sequentially in a proximal
direction. For each generation, alveolization is considered to be complete when the number of
alveoli in that generation reaches  the number determined by Weibel's model.

A.5.3.3.2. Airway size. Four sets of data are used to determine airway size during postnatal
growth: (a) total lung volume as  a function of age; (b) airway size as given by Weibel's model;
(c) the growth pattern of the bronchial airways; and (d) variation in  alveolar size with age.  From
these data, it is found that the lung volume, LV(t) at age t, normalized to Weibel's model at 4800
cm3 for an adult (25 years old), follows the equation

                  LV(f)  =  0.959  x 105(1 - 0.998e -°-00V)  (cm3).                   (A-63)

       The growth patterns of the bronchial airways are determined by the following equations

                         Dt(t)  - Dm =  a,[H(t)  - #(25)],                          (A-64)
                          Lt(f) - Lw = p.[#(0  - H(25)],                           (A-65)
where D;(t) and L;(t) are, respectively, the airway diameter and length at generation i and age t,
Diw and Liw the corresponding values for Weibel's model, a; and p; are coefficients given by

                       a,. = 3.26 x  10-2exp[-1.183  (/ + l)a5]                       (A-66)
                       P,. =  1.05  x  1(T6 exp [10.1] (/ + l)-°-2]                       (A-67)
                                          A-16

-------
and H(t) is the body height, which varies with age t in the form
                    H(t) =  1.82 x  102(1 -  0.725e-°-140  (cm).
       For the growth patterns of the airways in the alveolar region, it is assumed that
                                                      (A-68)
D.
                          L

       D
       —?
                                              for 17 < i < 23
                                                      (A-69)
where Da is the diameter of an alveolus at age t, Daw = 0.0288 cm is the alveolar diameter for
adults in accordance with Weibel's model, and f(t) is a function determined from
16
(LV(t) - £ ^L
i = O4
23
( E -AXtyO)
/ = 174

>;2(0 Z.(Y)Af.(0}


36 a^^U}
                                                                               (A-70)
A.6.  TRANSPORT RATES
       The values of transport rates A|Y f°r rats have been derived from the experimental data of
clearance for diesel soot (Chan et al., 1981; Strom et al., 1987, 1988) and for the particle-
associated organics (Sun et al., 1984; Bond et al., 1986; Yu et al., 1991). These values are used
in the present model of lung burden calculation and are listed below:
                              gj = 1.73 (/ =  1,2,3)
                                                      (A-71)
               = A^l = 0.00018
                                                                                (A-72)
i (3)  _
                      = 12.55
                                         A-17
                                                                               (A-73)


                                                                               (A-74)

-------
                            gj = 0.693      (/  = 1,2,3)                           (A-75)
                             0.00068 [1  - exp(-0.046wj'62)]                       (A-76)
                                             / .    r* ^\
                                              l  = 2>3)                            (A-77)
                         "$T = 0.012 exp(-0.11mj'76) +
                                                                                 (A-78)
                       0.00068 exp(-0.046m]-62)  (/ = 1,2,3)
                         0.012 exp(-0.11mj'76)  + 0.00086
                                      = 0.012 exp(-0.11mj'76)
                         0.00068 exp(-0.046wj'62) + 15.7
                                                                                 (A-79)
                       '^AL '  '^AT '  '^AB = 0-012 exp(-0.11mA  ) +
                                                                                 (A-80)
                       0.00068  exp(-0.046maL62) +  0.0161
                                                                                 (A-81)
where ^ is the unit of day"1, and mA = m ^ is the particle burden (in mg) in the alveolar
compartment.
       Experimental data on the deposition and clearance of DPM in humans are not available.
To estimate the lung burden of DPM for human exposure, it is necessary to extrapolate the
transport rates ^ from rats to humans. For organics, it is assumed that the transport rates are the
same for rats and humans.  This assumption is based upon the observation of Schanker et al.
(1986) that the lung clearance of inhaled lipophilic compounds appears to depend only on their
lipid/water partition coefficients and is independent of species. In contrast, the transport rates of
diesel soot in humans should be different from those of rats, since the alveolar clearance rate, XA,

                                         A-18

-------
of insoluble particles at low lung burdens for human adults is approximately seven times that of
rats (Bailey et al., 1982).
       No data are available on the change of the alveolar clearance rate of insoluble particles in
humans due to excessive lung burdens. It is seen from Equation A-79 that X ^;;for rats can be
written in the form

                             >4  = a e\p(-bmA) + d                            (A-82)
where a, b, c, and d are constants. The right-hand side of Equation A-82 consists of two terms,
representing, respectively, macrophage-mediated mechanical clearance and clearance by
dissolution.  The first term depends upon the lung burden, whereas the second term does not.
To extrapolate this relationship to humans, we assume that the dissolution clearance term is
independent of species and that the mechanical clearance term for humans varies in the same
proportion as in rats under the same unit surface particulate dose. This assumption results in the
following expression forA ^in humans

                           41} = ^  e*p[-b(mA/S)c]  +  d                          (A-83)


where P is a constant derived from the human/rat ratio of the alveolar clearance rate at low lung
burdens and S is the ratio of the pulmonary surface area between humans and rats. Equation
A-83 implies that rats and humans have equivalent amounts of biological response in the lung to
the same specific surface dose of inhaled DPM.
       From the data of Bailey et al. (1982), a value of X <» = 0.00169 day'1 is obtained for
humans at low lung burdens leading to P = 14.4. A value for S of 148 is reported from the data
of the anatomical lung model of Schum and Yeh (1979) for rats and Weibel's model for human
adults. For humans less than  25 years old, the model assumes the same value for P, but S is
computed from the data of the lung model for young humans (Yu and Xu 1987).  The value of S
for different ages is shown in Table A-3.
       The equations for other transport rates that have a lung-burden-dependent component are
extrapolated from rats to  humans in a similar manner.  The following lists the values of/I j|
(in day"1) for humans used in the present model calculation:

                              ^ = 1.73 (/  =  1,2,3)                             (A-84)
                                         A-19

-------
      AHB = *TB = ^LB = ^AB = 0.00018                        (A-85)



      ^HB  = ^TB  = ^LB  =  ^AB  =  0.0129                        (A-86)



        HB =  TB =  LB =  AB = 12.55                         (A-87)



        Xg, = 0.693      (/  = 1,2,3)                          (A-88)




  = 0.00068 (1  - 0.0694  exp[-0.046(wA/S)L62]}                 (A-89)
        41  =   ^      0'  = 2, 3)                           (A-90)
 41 = 0.0694 (0.012 Qxp[-0.ll(mA/S)L76] +
                                                             (A-91)
 0.00068 exp[-0.046(m,/S)L76]}  (i = 1, 2, 3)
                                                             (A-92)
0.0694 (0.012 exp[-O.U(mA/S)L76]} + 0.00086
    0.0694(0.012 exp[-0.11(m/4)L76]  +
                                                             /A  «0x
                                                             (A'93)
   0.00068  exp[-0.046(wy5)L76]} + 0.016                      (A-94)

-------
A.7.  RESULTS
A.7.1. Simulation of Rat Experiments
       To test the accuracy of the model, simulation results are obtained on the retention of
DPM in the rat lung and compared with the data of lung burden and lymph node burden obtained
by Strom et al. (1988). A particle size of 0.19 jim MMAD and a standard geometric deviation,
Gg, of 2.3 (as used in Strom's experiment) are used in the calculation.
       The respiratory parameters for rats are based on their weight and calculated using the
following correlations of minute volume, respiratory frequency, and growth curve data.

                            Minute volume = 0.9W (cnrVmin)                      (A-95)

                         Respiratory frequency = 475W"03 (1/min)                   (A-96)

where W is the body weight (in grams) as determined from the equation

                         W = 5+537T/(100+T), for T>56 days                   (A-97)
in which T is the age of the rat measured in days.
       Equation A-95 was obtained from the data of Mauderly (1986) for rats ranging in age
from 3 mo to 2 years old; Equation A-96 was obtained from the data of Strom et al. (1988); and
Equation A-97 was determined from the best fit of the experimental deposition data. Figures A-
3 and A-4 show the calculated lung burden of diesel soot (m (^+ m $ and lymph node burden,
respectively, for the experiment by Strom et al. (1988) using animals exposed to DPM at
6 mg/m3 for 1, 3, 6, and 12 weeks; exposure in all cases was 7 days/week and 20 h daily.
The solid lines represent the calculated accumulation of particles during the continuous exposure
phase and the dashed lines indicate calculated post-exposure retention.  The agreement between
the calculated and the experimental data for both lung and lymph node burdens during and after
the exposure periods was very good.
       Comparison of the model calculation and the retention data of particle-associated BaP in
rats obtained by Sun et al. (1984) is shown  in Figure A-5.  The calculated retention is  shown by
the solid line. The experiment of Sun et al. consisted of a 30-min exposure to diesel particles
coated with [3H] benzo[a]pyrene ([3H] - BaP) at a concentration of 4 to 6 |ig/m3 of air and
followed by a post-exposure period of over 25 days.  The fast and slow phase of ([3H] - BaP)
clearance half-times were found to be 0.03  day and 18 days, respectively. These correspond to
4^ = 0.0385 day"1 and k% = 23.1  day"1 in our model, where X(^0 is the value of X|^ at  mA -  0.
Figure A-5 shows that the calculated retention is in excellent agreement with the experimental
data obtained by Sun et al. (1984).
                                         A-21

-------
A.7.2.  Predicted Burdens in Humans
       Selected results of lung burden predictions in humans are shown in Figures A-6 to A-9.
The particle conditions used in the calculation are 0.2 jim MMAD with og = 2.3, and the mass
fractions of the rapidly and slowly cleared organics are each 10% (j 1 = j2 = 0.1). Figures A-6
and A-7 show, respectively, the lung burdens per unit concentration of diesel soot and the
associated organics in human adults for different exposure patterns at two soot concentrations,
0.1 and 1 mg/m3.  The exposure patterns used in the calculation are (a) 24 h/day and 7 days
week; (b) 12 h/day and 7 days/week; and (c) 8 h/day and 5 days/week, simulating environmental
and occupational exposure conditions.  The results show that the lung burdens of both diesel soot
and the associated organics reached a steady-state value during exposure. Because of differences
in the amount of particle intake, the steady-state lung burdens per unit concentration were
highest for exposure pattern (a) and lowest for exposure pattern (b). Also,  increasing soot
concentration from 0.1 to 1 mg/m3 increased the lung burden per unit concentration.  However,
the increase was not noticeable for exposure pattern (c).  The dependence of lung burden on the
soot concentration is caused by the reduction of the alveolar clearance rate at high lung burdens
discussed above.
       Figures A-8 and A-9 show the effect of age on lung burden, where the lung burdens per
unit concentration per unit weight are plotted versus age. The data of lung weight at different
ages are those reported by Snyder (1975). The exposure pattern used in the calculation is
24 h/day and 7 days/week for a period of 1 year at the two soot concentrations, 0.1 and 1  mg/m3.
The results show that, on a unit lung weight basis, the lung burdens of both soot and organics are
functions of age, and the maximum lung burdens occur at approximately 5  years of age. Again,
for any given age, the  lung burden per unit concentration is slightly higher at 1 mg/m3 than at
0.1 mg/m3.

A.8. PARAMETRIC STUDY OF THE MODEL
       The deposition and clearance model of DPM in humans, presented above, consists of a
large number of parameters that characterize the size and composition of diesel particles, the
structure and dimension of the respiratory tract, the ventilation conditions of the subject, and the
clearance half-times of the diesel  soot and the particle-associated organics. Any single or
combined changes of these parameters from their normal values in the model would result in a
change in the predicted lung burden. A parametric study has been conducted to investigate the
effects  of each individual parameter on calculated lung burden in human adults.  The exposure
pattern chosen for this study is 24 h/day and 7 days/week for a period of 10 years at a constant
soot concentration of 0.1 mg/m3.  The following presents two important results from the
parametric study.

                                         A-22

-------
A.8.1.  Effect of Ventilation Conditions
       The changes in lung burden due to variations in tidal volume and respiratory frequency
are depicted in Figures A-10 and A-l 1. Increasing any one of these ventilation parameters
increased the lung burden, but the increase was much smaller with respect to respiratory
frequency than to tidal volume.  This small increase in lung burden was a result of the decrease
in deposition efficiency as respiratory frequency increased, despite a higher total amount of
DPM inhaled. The mode of breathing has only a minor effect on lung burden because
switching from nose breathing does not produce any appreciable change in the amount of
particle intake into the lung (Yu and Xu, 1987).  All lung burden results presented in this report
are for nose breathing.

A.8.2.  Effect of Transport Rates
       Transport rates have an obvious effect on the retention of DPM in the lung after
deposition.  Because we are mainly concerned with the long-term clearance of diesel soot and
the associated organics, only the effects of two transport rates, X ^;;and X(%, are studied.
Experimental  data of X ^ from various diesel studies in rats have shown that X ^ can vary by a
factor of two or higher. We use a multiple of 0.5 to 2 for the uncertainty in  X ^ and X %•* to
examine the effect on  lung burden. Figures A-12 and A-13 show respectively, the lung burden
results for diesel soot  and the associated organics versus the multiples of X %} and X ^ used in the
calculation. As expected, increasing the multiple of X y reduced the lung burden of diesel soot
with practically no change in the organics burden (Figure A-12), while just the  opposite occurred
when the multiple of X ^ was increased (Figure A-13).

A.9.  OPERATIONAL DERIVATION OF HUMAN EQUIVALENT
      CONCENTRATIONS (HECs)
       The model of Yu et al. (1991) is ordered into two parts; one part parameterized on the
physiology and anatomy of a 300 g rat and the other part parameterized on the physiology and
anatomy of a 25 year old human male. The sequence of steps taken to calculate the human
equivalent continuous concentrations (the HECs), outlined in Table A-4, were as follows:

          The exposure scenario of the rats was entered into the rat portion of the model  and the
          model ran to obtain the output of lung burden in mg DPM/ rat lung at the time  of the
          sacrifice of the rats.
       •   The output of mg DPM/ rat lung was normalized to mg DPM/ cm2 of rat lung tissue
          based on a total pulmonary surface area of 4090 cm2.
                                        A-23

-------
          The normalized rat lung burdens were used to calculate the corresponding lung
          burden based on the pulmonary surface area of 627,000 cm2.  This operation yielded
          mg DPM / lung of a 25 year old human male.
       •   Various air concentrations were run in an iterative fashion with the human portion of
          the model under a continuous exposure scenario of 24 h/day, 7d/wk for 70 years
          with ventilatory parameters set at 0.926 L for tidal volume and 15 breaths per minute
          as the respiratory frequency to yield a total daily pulmonary volume of 20 m3. This
          was continued until the output (mg DPM/lung) was matched to the mg DPM /human
          lung obtained from the normalized rat lung burden; the concentration from the model
          that matched this lung burden was termed the human equivalent continuous
          concentration, the HEC.  The human modeling runs did not consider the preadult
          status of airway and alveoli number discussed above but rather were ran for 1 to
          70 years with adult (25 years of age) parameters mentioned above.

       These HEC values address kinetic issues  of DPM deposition and retention in the lung by
humans.  As noted above, these values do not reflect the kinetic variability that may exist in the
human population exposed to DPM which includes men and women, young and old.  However,
the limited parametric analysis of the model clearly shows variability of those parameters most
determinative in humans (e.g., tidal volume, respiration rate, and rates of clearance of particles
from the airways) were mirrored in the corresponding output of the model (lung burden of
DPM). One interpretation of this parallel in parameter-output is that the variability in the
physiological characteristics of humans reflects the variability in the model such that, for
example, a small tidal volume would be reflected with a decreased lung burden of DPM.
Variability among humans of these key parameters such as tidal volume do vary but within an
order of magnitude. This would mean that the DPM dose received by different individuals in the
population from the same concentration would indeed vary within the extremes of these
determinative parameters.
                                         A-24

-------
                    Table A-l.  Lung model for rats at total lung capacity
Generation
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16b
17
18
19
21
22
25
24
Number of
airways
1
2
3
5
8
14
23
38
65
109
184
309
521
877
1,477
2,487
4,974
9,948
19,896
39,792
79,584
318,336
636,672
Length (cm)
2.680
0.715
0.400
0.176
0.208
0.117
0.114
0.130
0.099
0.091
0.096
0.073
0.075
0.060
0.055
0.035
0.029
0.025
0.022
0.020
0.019
0.017
0.017
Diameter (cm)
0.340
0.290
0.263
0.203
0.163
0.134
0.123
0.112
0.095
0.087
0.078
0.070
0.058
0.049
0.036
0.020
0.017
0.016
0.015
0.014
0.014
0.014
0.014
Accumulative
volumea (cm)
0.243
0.338
0.403
0.431
0.466
0.486
0.520
0.569
0.615
0.674
0.758
0.845
0.948
1.047
1.414
1.185
1.254
1.375
1.595
2.003
2.607
7.554
13.784
including the attached alveoli volume (number of alveoli = 3 x 107, alveolar diameter = 0.0086 cm).
bTerminal bronchioles.
                                             A-25

-------
Table A-2. Lung model by Weibel (1963) adjusted to 3000 cm3 lung volume
Generation
number
0
2
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16b
17
18
19
20
21
22
23
Number of
airways
1
2
4
8
16
32
64
128
256
512
1,024
2,048
4,096
8,192
16,384
32,768
65,536
131,072
262,144
524,283
1,048,579
2,097,152
4,194,304
8,388,608
Length (cm)
10.260
4.070
1.624
0.650
1.086
0.915
0.769
0.650
0.547
0.462
0.393
0.333
0.282
0.231
0.197
0.171
0.141
0.121
0.100
0.085
0.071
0.060
0.050
0.043
Diameter (cm)
1.539
1.043
0.710
0.479
0.385
0.299
0.239
0.197
0.159
0.132
0.111
0.093
0.081
0.070
0.063
0.056
0.051
0.046
0.043
0.040
0.038
0.037
0.035
0.035
Accumulative
volumea (cm)
19.06
25.63
28.63
29.50
31.69
33.75
35.94
38.38
41.13
44.38
48.25
53.00
59.13
66.25
77.13
90.69
109.25
139.31
190.60
288.16
512.94
925.04
1,694.16
3,000.00
                               A-26

-------
Table A-3. Ratio of pulmonary surface areas between humansand rats as a function
of human age
Age (year)
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
27
28
19
20
21
22
23
24
25
Surface area
4.99
17.3
27.6
36.7
44.7
51.9
58.5
64.6
70.4
76.0
81.4
86.6
91.6
96.4
101
106
110
115
119
123
128
132
136
140
144
148
                               A-27

-------
      Table A-4.  Human equivalent continuous concentrations (HECs) calculated with the model of Yu et al. (1991) from long-term
      repeated exposure rat studies of PPM exposure
c
6.85E-5
4.95E-3
1.09E-2
5.87E-5
2.45E-4
4.51E-3
9.75E-3
2.81E-4
3.16E-3
7.63E-3
1.58E-2
7.00E-3
1.86E-2
9.4E-4
8.4E-3
2.4E-2
mgDPM/
human lungc
43
3101
6825
37
153
2828
6115
176
1984
4786
9914
4391
11674
587
5274
14993
HEC
(mg/m3)'
0.038
1.375
3.05
0.032
0.128
1.25
2.75
0.144
0.883
2.15
4.4
1.95
5.1
0.33
2.35
6.7
        a These are entered into the program as hrs/day, days/week for the total number of weeks exposed and the last week of exposure before evaluation (as this would affect
        clearance). The parameters for the rat were based on a body weight which was set in the program at 300g.
        b These values were obtained with the rat portion of the model and are noted as lung burden, in mg DPM /lung of a 300 g rat, at the final week of the exposure scenario.  These
        outputs were then normalized to cm2 of the rat lung, at 4090 cm2 total (Xu and Yu, 1987).
        c Preparatory to using the human portion of the model, the mg DPM/cm2 value from above was used to project the mg DPM that would be present in the adult human lung
        based on a total lung surface area of 627,000 cm2 (Xu and Yu, 1987). Various air concentrations were then entered into the human model as 70 years continuous exposure
        scenarios and ran iteratively until the output (in mg DPM / lung at age 70) matched this mg DPM/human lung, i.e., the total lung burden. This matching air concentration is, by
        definition, the human equivalent continuous concentration (HEC).
        d weeks = (months of exposure) x 4.33.

-------
         B
Figure A-l. Compartmental model of DPM retention.
                                     A-29

-------
  Trachea
                    Summed Alveolar Cross Sectional Area
                    Summed Airway




             Cross Sectional Area A^x)
                                                            A2(x)
                                                                           Airway Length x
Figure A-2. Trumpet model of lung airways.
                                           A-30

-------
                                                   T
                                                         12 wk
                              -•a	
                                                  i
                                                         1 wk   i
13
                                   26            39
                                     Time, week
52
65
Figure A-3.  The experimental and predicted lung burdens of rats to DPM at a solid and
            dashed concentration of 0.6 mg/m3 for different exposure spans. Lines are,
            respectively, the predicted burdens during exposure and post-exposure.
            Particle characteristics and exposure pattern are explained in the text. The
            symbols represent the experimental data from Strom et al. (1988).
                                       A-31

-------
13
                                  26            39
                                    Time, week
52
65
Figure A-4. Experimental and predicted lymph node burdens of rats exposed to CEPs at a
           concentration of 6.0 mg/m3 for different exposure spans. The solid and
           dashed lines are, respectively, the predicted burdens during exposure and
           post-exposure. Particle characteristics and exposure pattern are explained in
           the text.  The symbols represent the experimental data from Strom et al.
           (1988).
                                      A-32

-------
     0.8
  CD
     0.2  -
                                10
   15
Time, day
20
25
Figure A-5.  Comparison between the calculated lung retention (solid line) and the
            experimental data obtained by Sun et al. (1984) for the particle-associated
            BaP in rats.
                                         A-33

-------
       700
   600


CO
 E

 75>

 £ 500
     I 400
     -i—i
     03
     CD
     O

     § 300

     O
     CD
     DO

     O)
     c
       200
       100
                                      1 mg/ m
                                      4             6

                                         Time, year
                                                                          10
Figure A-6.  Calculated lung burdens of diesel soot per unit exposure concentration in

            human adults exposed continuously to DPM at two different concentrations of

            0.1 and 1.0 mg/m3. Exposure patterns are (a) 24 h/day and 7 days/week,

            (b) 12 h/day and 7 days/week, and (c) 8 h/day and 5 days/week.
                                         A-34

-------
                                      1 mg/ m
   f~^  5
  n
   E
   ^)


      4
                                      0.1
(A)
   0)
   o
   o
   O
                                      0.1
(B)
   CD
   O)
                                      0.1
(C)
                                                                                10
                                        Time, year
Figure A-7.  Calculated lung burdens of the particle-associated organics per unit exposure
            concentration in human adults exposed continuously to DPM at two different
            concentrations of 0.1 and 1.0 mg/m3.  Exposure patterns are (a) 24 h/day and
            7 days/week, (b)  12 h/day and 7 days/week, and (c) 8 h/day and 5 days/week.
                                        A-35

-------
     0.4
   CD
   E
   ^§0.3

   g
   "ro

   §
   c
   ° n o
   Q 0.2
   CO
   CO
     0.1
   CD
   -D
   3
   00
   CO
                                             1 mg/m
                                     10
15
20
25
                                         Age, year
Figure A-8.  Calculated lung burdens of diesel soot per gram of lung per unit exposure
            concentration in humans of different ages exposed continuously for 1 year to
            DPM of two different concentrations of 0.1 and 1.0 mg/m3 for 7 days/week and
            24 h daily.
                                        A-36

-------
      0.01
      .008   -
   en
.o
'-I—'
CO
-I— '



I
o
o
   •s? .004
   O)
                                           1 mg/ m
      .006   -
   cu
   •
   CD

   cn
      .002
                                       10            15

                                           Age, year
                                                                20
25
Figure A-9. Calculated burdens of the particle-associated organics per gram of lung per

            unit exposure concentration in humans of different ages exposed continuously

            for 1 year to DPM of two different concentrations of 0.1 and 1.0 mg/m3 for

            7 days/week and 24 h daily.
                                        A-37

-------
    100
     80
     60
  D)
  E
  o
  o
     40
     20
               Soot
                                                                             1.5
                                                                               O)
                                                                               E
                                                                               CD
                                                                               O)
                                                                             0.5
       0.3
0.4               0.5
          Tidal Volume, Liter
0.6
0.7
Figure A-10.  Calculated lung burdens in human adults versus tidal volume in liters for
             exposure to DPM at 0.1 mg/m3 for 10 years at 7 days/week and 24 h daily.
             Parameters used in the calculation are: (a) MMAD=0.2 urn, og=2.3, /2=0.1,
             /3=0.1; (b) respiratory frequency = 14 min"1; and (c) lung volume = 3000 cm3.
                                       A-38

-------
     60
     50
                                  Soot
                                                                           1.4
                                                                           1.2
  O)
  •
  O
     40
     30
     20
     10
                                   Organics
                                                                            0.8
                                                                            0.6
                                                                           0.4
                                                                            0.2
                                                                                O)
                                                                                E
                                                        CD
                                                        O)
        10
12               14               16
     Respiratory Frequency, 1/min.
18
Figure A-ll.  Calculated lung burdens in human adults versus respiratory frequency in
             bpm for exposure to DPM at 0.1 mg/m3 for 10 years at 7 days/week and 24 h
             daily. Parameters used in the calculation are:  (a) MMAD=0.2 um, og=2.3,
             /2=0.1, /3=0.1; (b) tidal volume = 500 cm3, and (c) lung volume = 3200 cm3.
                                       A-39

-------
     120
     100
                                    2.5
      80
  O)
  E
  -5  60
  o
  CO
      40
                                       Soot
                                                                                O)
                                                                                E
                                    15--
                                    I .0  £-
                                        03

                                        O
      20
                                       Organics
                                    0.5
                               I
I
0.6       0.8        1       1.2
                       Multiple of;
                                                1.4
                1.6
1.8
Figure A-12.  Calculated lung burdens in human adults versus multiple of k ^ for
             exposure to DPM at 0.1 mg/m3 for 10 years at 7 days/week and 24 h daily.
             Parameters used in the calculation are:  (a) MMAD=0.2 urn, og=2.3, /2=0.1,
             /3=0.1; (b) tidal volume = 500 cm3, respiratory frequency = 14 min"1; and
             (c) lung volume = 3200 cm3.
                                        A-40

-------
    60
    50  =.
    40  -
  D)
  E
  "5 30
  o
    20  -
     10  -
           0.6      0.8
1       1.2      1.4
     Multiple of A,(A
1.6
Figure A-13.  Calculated lung burdens in human adults versus multiple of i % for exposure
             to DPM at 0.1 mg/m3 for 10 years at 7 days/week and 24 h daily. Parameters
             used in the calculation are: (a) MMAD=0.2 urn og=2.3, /2=0.1, /3=0.1;
             (b) tidal volume = 500 cm3, respiratory frequency = 14 min 1; and (c) lung
             volume = 3200 cm3.
                                        A-41

-------
                           REFERENCES FOR APPENDIX A

Amann, CA; Siegla, DC. (1982) Diesel particles ~ what are they and why. Aerosol Sci Technol 1:73-101.

Bailey, MR; Fry, FA; James, AC. (1982) The long-term clearance kinetics of insoluble particles from the human
lung. Ann Occup Hyg 26:273-289.

Bond, JA; Sun, JD; Medinsky, MA; et al. (1986) Deposition, metabolism and excretion of l-[14C]nitropyrene and
l-[14C]nitropyrene coated on diesel exhaust particles as influenced by exposure concentration. Toxicol Appl
Pharmacol85:102-117.

Chan, TL; Lee, PS; Hering, WE. (1981) Deposition and clearance of inhaled diesel exhaust particles in the
respiratory tract of Fisher rats. J Appl Toxicol 1:77-82.

Diu, CK; Yu, CP. (1983) Respiratory tract deposition of poly disperse aerosols in humans. Am Ind Hyg Assoc J
44:62-65.

Heinrich, U; Fuhst, R; Rittinghausen, S; et al. (1995) Chronic inhalation exposure of Wistar rats and two different
strains of mice to diesel engine exhaust, carbon black, and titanium dioxide. Inhal Toxicol 7:533-556.

ICRP. (1979) Limits for intakes of radionuclides by workers. Ann ICRP 2. Publication 30, part 1.

Ishinishi, N; Kuwabara, N; Takaki, Y; et al. (1988) Long-term inhalation experiments on diesel exhaust. In: Diesel
exhaust and health risks: results of the HERP studies. Tsukuba, Ibaraki, Japan: Japan Automobile Research
Institute, Inc., Research Committee for HERP Studies; pp. 11-84.

Mauderly, JL. 1986. Respiration of F344 rats in nose-only inhalation exposure tubes. J Appl Toxicol 6:25-30.

Mauderly, JL; Jones, RK; Griffith, WC; et al. (1987) Diesel exhaust is a pulmonary carcinogen in rats exposed
chronically  by inhalation. Fundam Appl Toxicol 9:208-221.

Nikula, KJ; Snipes, MB; Barr, EB; et al. (1995) Comparative pulmonary toxicities and carcinogenicities of
chronically  inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl Toxicol 25:80-94.

Schanker, LS; Mitchell, EW; Brown, RA. (1986) Species comparison of drug absorption from the lung after aerosol
inhalation or intratracheal injection. Drug Metab Dispos 14(l):79-88.

Scheutzle, D. (1983) Sampling of vehicle emissions for chemical analysis and biological testing. Environ Health
Perspect 47:65-80.

Schum, M;  Yeh, HC. (1979) Theoretical evaluation of aerosol deposition in anatomical models of mammalian lung
airways. Bull Math Biol 42:1-15.

Snyder, WS. (1975) Report of task group on reference man. Oxford, London: Pergamon Press, pp. 151-173.

Solderholm, SC. (1981) Compartmental analysis of diesel particle kinetics in the respiratory system of exposed
animals. Oral presentation at EPA Diesel Emissions Symposium, Raleigh, NC, October 5-7. In: Toxicological
effects of emissions from diesel engines (Lewtas J, ed.). New York: Elsevier, pp. 143-159.

Strom, KA; Chan,  TL; Johnson, JT. (1987) Pulmonary retention of inhaled submicron particles in rats: diesel
exhaust exposures and lung retention model. Research Publication GMR-5718. Warren, MI: General Motors
Research Laboratories.

Strom, KA; Chan,  TL; Johnson, JT. (1988) Inhaled particles VI. Dodgson, J; McCallum, RI; Bailey, MR; et al., eds.
London: Pergamon Press, pp. 645-658.
                                                A-42

-------
Sun, JD; Woff, RK; Kanapilly, GM; et al. (1984) Lung retention and metabolic fate of inhaled benzo(a)pyrene
associated with diesel exhaust particles. Toxicol Appl Pharmacol 73:48-59.

Weibel, ER. (1963) Morphometry of the human lung. Berlin: Springer-Verlag.

Xu, GB; Yu, CP. (1987) Desposition of diesel exhaust particles in mammalian lungs: a comparison between rodents
and man. Aerosol Sci Tech 7:117-123.

Yu, CP. (1978) Exact analysis of aerosol deposition during steady breathing. Powder Technol 21:55-62.

Yu, CP; Xu, GB. (1986) Predictive models for deposition of diesel exhaust participates in human and rat lungs.
Aerosol Sci Technol 5:337-347.

Yu, CP; Xu, GB. (1987) Predicted deposition of diesel particles in young humans. J Aerosol Sci 18:419-429.

Yu, CP; Yoon, KJ. (1990) Retention modeling of diesel exhaust particles in rats and humans. Res Rep Health Eff
Inst40:l-33.

Yu, CP; Diu, CK; Soong, TT. (1981) Statistical analysis of aerosol deposition in nose and mouth. Am Ind Hyg
Assoc 142:726-733.

Yu, CP; Yoon, KJ; Chen,YK. (1991) Retention modeling of diesel exhaust particles in rats and humans. J. Aerosol
Med. 4(2): 79-115.
                                                A-43

-------
           Appendix B

Benchmark Concentration Analysis of
           Diesel Data
                B-l

-------
B-l.  INTRODUCTION TO BENCHMARK
     The benchmark dose or benchmark concentration approach, hereafter referred to as the
BMC approach, is an alternate to the N/LOAEL option for deriving effect levels.  The BMC is
currently undergoing extensive consideration by the Agency with promulgation of software and
guidelines for application of this methodology (U.S. EPA, 2000). The BMC approach involves
fitting a dose-response function to dose and effect information from a single study to derive the
best fit of those data. This "best fif'is statistically termed the maximum likelihood estimate but
is referred to in the benchmark terminology as the BMC curve. The curve defining the
corresponding lower 95% confidence limit of this "best fif'estimate is termed the BMCL curve.
This BMCL curve is used to predict the dose that will result in a level of response that is defined
a priori as the benchmark response "x", BMCLX. In the analyses below, for example, the
benchmark response for a 10% increase in incidence1 of chronic inflammation is defined as a
BMCL10; the corresponding 10% increase as determined from the BMC curve would be termed
the BMC10. This BMCL10 would be derived by first using the data and the programs to determine
the BMC and BMCL curves.  The concentration corresponding to a 10% increase in incidence
would then be  determined directly from the BMCL. The BMCL10 then would be used as the
representative value for the effect level or point of departure in the dose-response assessment.
     The latest version of the Agency Benchmark Dose Software (BMDS Version 1.2; U.S.
EPA, 2000) was used to analyze data on chronic inflammation and pulmonary histopathology
present in the chronic studies that were amenable to benchmark analysis.  At this time, the
Agency BMDS offers sixteen different models total that are appropriate for the analysis of
dichotomous data (gamma, logistic, probit, Weibull, log-logistic, multistage, log-probit,
quantal-linear, quantal-quadratic), continuous data (linear, polynomial,  power, Hill) and nested
developmental toxicology data (NLogistic, NCTR, Rai & Van Ryzin).  Results from all models
include a reiteration of the model formula and model run options chosen by the user,
goodness-of-fit information, a graphical presentation for visual inspection and the concentration
estimate for the response at the designated BMCLX, as well as the corresponding BMCX. More
details on the modeling results are described and presented in the analysis on dichotomous data
following.
     The U.S. EPA benchmark dose (BMD/C) methods guidance has not been finalized at this
time to provide definitive procedures and criteria (U.S. EPA 1995). Therefore, in this document
provisional criteria for minimum data to perform a benchmark analysis are designated such that
(1) complete quantitative information on the response of interest should be available (e.g.,
       For increases in incidence "extra risk" is used which is response incidence (inc) normalized to the
background (BG) incidence; response - BG/l-BG.

                                          B-2

-------
incidence as number affected / total, means with variability) and that (2) at least two exposure
levels with responses that differ from those of the controls are provided, and (3) a benchmark
response of 10% is employed such that outcomes are BMCL10s.  A response of 10% is at or near
the limit of sensitivity in most long-term bioassays as determined from both the typical number
of animals used in bioassays and a low spontaneous background rate (e.g., 0.1%) for a given
effect (Haseman, 1984; Haseman et al.,  1989).

B-2.  DIESEL DATA FOR BENCHMARK ANALYSIS
     Using the criteria set forth in Section B-l and the information about the critical effects that
have been identified (pulmonary inflammation, pulmonary histopathology including indicators
of fibrotic changes such as increases in alveolar-capillary wall thickness) the following rat
chronic studies identified in Chapter 6 were analyzed for information suitable for BMC analysis:
Ishinishi et al. (1986, 1988), Mauderly et al. (1987a,b; 1988); Heinrich et al. (1986, 1995), and
Nikulaetal. (1995).
     Results from this analysis yielded  only a few data sets from a single study, that of Nikula
et al. (1995), that could be used for BMC analysis. The basis for not including data from the
other studies varied. Information on pulmonary histopathology in the studies of Ishinishi et al.
(1986, 1988), for example, was supplied only in narrative form with no quantitative information
given.  A similar situation was found for those reports of the ITRI study; Wolff et al. (1987)
reports on clearance alterations due to DPM exposure; Henderson et al. (1988) does give
information on hydroxyproline but only in graphical form; the 1988 study of Mauderly et al.
deals with pulmonary function as a function of DPM lung loading; the 1987a reference of
Mauderly et al. discusses tumor prevalence only and the Mauderly 1987b reference reports on
diesel exhaust in developing lung to a single exposure concentration of DPM with no dose-
response information available. Those reports on the General Motor study contain extensive
information relating not to the critical effects, but mostly to precursors of inflammation such  as
levels of polymorphonuclear neutrophils and lymphocytes in bronchoalveolar lavage from DPM
exposed rats (Strom, 1984) and guinea pigs (Barnhart et al., 1981) as well as information on
collagen biosynthesis (Misiorowski et al.,  1980) all of which is presented in graphical rather than
tabular form amenable for benchmark analysis.  The information on noncancer histopathology
reported by Heinrich et al. (1995) is in text form only and this author's 1986 study deals
primarily with clearance and mortality.  Nikula et al. (1995), however, do present extensive
quantitative dose-response information (incidence / dichotomous data) on several  measures of
the critical  effect including chronic inflamation (presence of focal aggregates of neutrophils),
focal fibrosis with epithelial hyperplasia (nodular fibrosis rimmed by hyperplasia), and septal
fibrosis (interstitial fibrosis within alveolar septa) although the study had but 2 exposure

                                          B-3

-------
concentrations both of which are different from the controls, a minimal number on which
benchmark analysis should be performed.

B-3.  BENCHMARK ANALYSIS OF DIESEL DATA
     These data from Nikula et al. (1995) were extracted, HEC concentrations calculated using
the model of Yu et al. (1991; Appendix A), and analyzed using all 9 applicable models for
dichotomous data. Because the benchmark models were ran with the HEC, general from the
model of Yu et al. (1991), the BMCL10s are also HECs. The results and data are presented in
Table B-l. Results were evaluated based on the nature of the data set, visual inspection of the
graphical output, and on the goodness-of-fit parameters,  including p values and the AIC. When
p values were generated for model fits, values for p that were less than 0.1 were  considered to
reflect a minimal fit to the data and were disqualified from further consideration. However, the
small set of only 3 data  points was often matched by the  number of parameters fitted in several
of the models such that  the outcome of the model exactly fit the data and thus no p value is
generated; these model fits are often referred to as being overparameterized, and are indicated as
"NA" in Table B-l.  Values for p that were less than 0.1  were considered to reflect a minimal fit
to the data. The AIC (Akaike Information Coefficient; Akaike, 1973; Stone,  1998) is a
parameter generated for the models in U.S. EPA (2000) that allows for a general comparison
among models run on the same data set.  The AIC is defined as -2 log L + 2 p where log L is the
log likelihood of the fitted model, and p is the number of parameters estimated; smaller values
indicate better fits.
     The overall results of this mathematical analysis is reasonable in a biologically mechanistic
sense in that chronic inflammation is more prevalent and apparently occurs at lower
concentrations (i.e., has lower BMCL10 values) than does focal fibrosis.  The information on
septal fibrosis were not  interpretable as the data were not amenable (no or zero background and
then total incidence) to any meaningful benchmark or other dose-response analysis. The most
sensitive  endpoint, chronic inflammation, is therefore the most sensitive benchmark
concentration followed by focal fibrosis.
     The choice for the most appropriate BMCL10 from  among the various modeled values for
chronic inflammation requires analysis of both the statistical and graphical outputs of the data.
The shape of the dose-response curve from information given in Chapter 6 (Table 6-2) gives
evidence of considerable "S" character, e.g., several low HECs without any reported effects up
to about 0.2 mg/m3.  The shape of the dose-response curves generated by  several of the models,
including gamma-hit, Weibull, multistage, and quantal linear were all a uniformly upward
sloping arc from the origin (graphs not shown) with minimal evidence of any "S" character, a
shape not concordant with the data array in Table 6.2. Models that did generate  curves with "S"

                                          B-4

-------
character included log-logistic, logistic, probit, quantal-quadratic, and log-probit.  Because of
their concordance with this independent data array on dose-response, the latter outputs are
further analyzes.
     The results for both chronic inflammation and focal fibrosis for those models with outputs
having appreciable "S" character suggest that females may be more sensitive than males for
these endpoints as the incidences are higher and the BMCL10 values are generally lower for
females than for males.  However, the model fits of the BMCL10s to the chronic inflammation
data segregated by  sex were generally inadequate as judged from the p values (most being far
less than  0.1) or from visual inspection of the fits to the data, several of which (e.g., log-logistic
and log-probit) were lacking any appreciable "S" character. However, combining female and
male data improved data fitting as judged by the increased p values to where nearly all were
>0.1 and  to where the visual fits were  concordant with the independent information on dose-
response. Too,  most of the combined  BMCL10s were either intermediate between the female and
male values or somewhat closer to the female values such that the combined BMCL10 values
were not  much different from the females BMCL10s.
     From among the combined male and female model outputs in Table B-l, the logistic,
probit, and  quantal  quadratic results were all excluded based on the high AIC value relative to
the log-logistic  and log-probit results.  The log-logistic results were excluded based on the shape
of the lower portion of the dose-response curve which was upward sloping near the origin (graph
not shown) and not as concordant with the independent dose-response information in Table 6-2
as was the fit of the log-probit model (Figure B-l). This leaves the fit of the log-probit model as
being most reflective of the information in Table 6-2. The BMCL10 of the log-probit curve at
0.37 mg/m3 remains and, by elimination, appears to be the most defensible choice from among
the BMCL10s arrayed in Table B-l.  Figure B-l shows the graphical representation of the  log-
probit model fit to the data and the origin of the BMCL10.  This graph also shows the relationship
of the BMCL10 of 0.37 mg/m3 to the variability that exists around the control value and that the
value of 0.37 mg/m3 is not far removed from the outer range of this variability.  The log-probit
BMCL10 for focal fibrosis (combined) of 1.3 mg/m3 noted as being representative of this lesion
from the BMC analysis in Table B-l.
     Characterization of this benchmark value indicates that it may not be a suitable candidate
for use as a point of departure for development of a dose-response assessment such as the RfC.
An attribute of the benchmark method is that the response (such as the 10% as used here) is near
the range of the actual experimental values,  such that extrapolation is not far below the observed
experimental  range. However, due to  the paucity of data points overall and lack of any values
below an HEC of nearly  2 mg/m3 in the Nikula et al.  (1995) study, the extrapolation of this BMC
to the 10% response level is considerable, the BMLC10 of 0.37 mg/m3 being > 5-fold below the

                                          B-5

-------
nearest observed value of 1.95 mg/m3. Also, the high experimental exposures used in this study
are in the range of those resulting in pulmonary overload conditions in rats and therefore in the
range of the model assumptions of Yu et al. (1991) about this phenomenon in humans for
calculation of the HECs (Chapter 3).  The BMCL10 of 0.37 mg/m3 is considerably greater than
other NOAELs in the DPM data base of 0.144 mg/m3 and 0.128 mg/m3 (Table 6-2 in Chapter 6),
possibly indicating that these NOAELs represent actual incidence levels that are considerably
less thanlO%; from the same log-probit model the corresponding BMCL05 was 0.21  mg/m3
(near the range of these NOAELs) and the corresponding BMCL01 was 0.07 mg/m3  (below the
range of these NOAELs). These limitations on this BMCL10 make it a less than optimal
candidate for consideration as a point of departure in the development of dose-response
assessments.

B-4. SUMMARY
     The recently developed EPA Benchmark dose software (U.S. EPA, 2000) and preliminary
guidance was utilized to analyze diesel data by the benchmark approach. Data from only one of
the array of principal studies identified elsewhere (Chapter 6) was found to contain data
amenable to benchmark analysis.  The data from this study, that of Nikula et al. (1995) on
pulmonary inflammation and histopathology, was extracted and analyzed as dichotomous data
using all available models and designating a 10% response level such that BMCL10s were
calculated; as the models were ran with  HECs, the BMCL10s were also HECs.
     The analysis resulted in an array of BMCL10s from 3 different effects in two sexes (both
separate and combined) with 9 different models. These BMCL10s were each considered from a
perspective of biological relevance, known dose-response character, and from the individual  fit
to the data by the models from statistical parameters and visual judgments. The BMCL10 that
emerged after the above considerations was 0.37 mg/m3 for the combined male plus female
incidence of chronic active pulmonary inflammation. A BMCL10 of 1.3 mg/m3 for pulmonary
focal fibrosis was also noted in this analysis. Characterization of these benchmark values
indicates that neither may be a suitable candidate for use as a point of departure in development
of a dose-response assessment such as the RfC but that they are concordant with other
quantitative dose-response aspects of the DPM database.
                                         B-6

-------
td
      Table B-l. BMC analysis of pathology incidence data in male and female F344 rats from the study of Nikula et al. (1995)
      using the different models available from U. S. EPA benchmark dose project (U.S. EPA, 2000) for dichotomous data based on
      10% extra risk (i.e., a 10% increase relative to a total that has been adjusted for background) and no threshold term. The
      concentrations used in the analysis are human continuous equivalent concentrations (HECs) obtained from the interspecies
      extrapolation model of Yu et al. (1991). The table listings include the BMCL10 (the benchmark response level of 10%
      obtained from the lower 95% limit of the benchmark curve in mg/m3), the BMC10 (the corresponding estimate at 10%
      response from the best fit benchmark curve, also in mg/m3), P = goodness-of-fit values. NA indicates a G-O-F value was not
      available, usually due to the lack of degrees of freedom. AIC =  Akaike Information Coefficient (see U.S. EPA, 2000 and
      below) which may be used for model comparison on the same data set.
Effect (from Table 5
and 6, p 86, Nikula
et al., 1995)
Chronic active
inflammation >18 mos,
grades 1-3, male +
female combined
Chronic active
inflammation >18 mos,
grades 1-3 in males
Chronic active
inflammation >18 mos,
grades 1-3 in females
Focal fibrosis with
epithelial hyperplasia,
grades 1-4 in males and
females combined
Focal fibrosis with
epithelial hyperplasia,
grades 1-4 in males
Focal fibrosis with
epithelial hyperplasia,
grades 1-4 in females
Septal fibrosis,
>18 mos, grades 1-4 in
males
Septal fibrosis,
>18 mos, grades 1- 4 in
females
Inc @ Inc @
Inc@ 1.95 mg/m3 5.1 mg/m3
0 mg/m3 HEC HEC
5/177 59/162 118/174
1/86 19/81 54/85
4/91 40/81 64/89
0/177 18/162 63/174
0/86 5/81 19/85
0/91 13/81 44/89
1/86 79/81 83/85
2/91 75/81 87/89
BMCL10
(BMC10)
log-logistic
0.32(0.64)
P=NA
AIC= 483
0.67(1.16)
P=NA
AIC=217
0.18(0.26)
P=NA
AIC= 257
1.25(1.8)
P= 1.000
AIC= 345
1.72(2.7)
P= 1.00
AIC= 132
0.80(1.4)
P= 1.00
AIC= 199
.003(.008)
P=0.35
AIC= 53
0.009 (.05)
P=NA
AIC= 87
BMCL10
(BMC10)
log-probit
0.37(.70)
P=NA
AIC = 483
0.74(1.22)
P = NA
AIC = 217
.016(30)
P = NA
AIC = 257
1.3(1.8)
P = 1.000
AIC = 345
1.6(2.7)
P = 1.000
AIC = 132
0.87(1.47)
P = 1.000
AIC = 199
(failed)
(failed)
BMCL10
(BMC10)
multi-stage
0.43(.49)
P= 0.982
AIC= 481
0.56(.95)
undefined
AIC=217
0.33(.40)
P= 0.173
AIC= 257
1.21(1.8)
P= 1.000
AIC= 345
1.79(2.8)
undefined
AIC= 134
0.77
P=0.99
AIC= 199
0.07(.08)
P= 0.000
AIC= 65
0.08(.10)
P= 0.003
AIC= 91
BMCL10
(BMC10) -
Weibull
0.43(.49)
P= 0.982
AIC=481
.56(1.04)
P=NA
AIC= 216
0.33(.40)
P= 0.173
AIC= 257
1.21(1.8)
P= 1.000
AIC= 345
1.79(2.8)
P= 1.00
AIC= 132
0.77(1.4)
P=1.0
AIC=199
0.07(.08)
P= 0.000
AIC= 65
0.08(.10)
P= 0.000
AIC= 91
BMCL10
(BMC10) -
gamma
0.43(.49)
P=0.98
AIC= 480
.56(1.09)
P=NA
AIC= 217
0.33(.40)
P=0.17
AIC= 257
1.21(1.8)
P= 1.0
AIC= 345
1.79(2.75
P=1.0
AIC= 132
0.71(1.4)
P= 1.00
AIC= 199
0.07(.08)
P= 0.000
AIC= 65
0.08(.10)
P= 0.003
AIC= 91
BMCL10
(BMC10) -
quantal
linear
0.43(.49)
P= .982
AIC=481
0.50(.61)
P=0.15
AIC= 216
0.33(.40)
P= 0.173
AIC= 257
1.1(1.3)
P= 0.363
AIC= 345
1.7(2.4)
P= 0.70
AIC= 131
0.71(.88)
P= 0.445
AIC= 198
0.07(.08)
P= 0.000
AIC= 65
0.08(.10)
P= 0.003
AIC= 91
BMCL10
(BMC10) -
probit
1.06(1.19)
P= 0.000
AIC= 499
1.31(1.55)
P=0.05
AIC=219
0.83(.96)
P= 0.0001
AIC= 272
2.32(2.61)
P= 0.013
AIC= 353
2.98(3.5)
P= 0.199
AIC= 134
1.76
P= 0.037
AIC= 205
0.29(.37)
P= 0.000
AIC=114
0.32(.40)
P= 0.000
AIC=131
BMCL10
(BMC10)-
logistic
1.12(1.26)
P=0.000
AIC= 502
0.67(1.16)
P=NA
AIC=217
0.85(1.0)
P= 0.000
AIC= 273
2.50(2.8)
P= 0.006
AIC= 356
3.17(3.69)
P= 0.153
AIC= 135
1.89(2.2)
P= 0.02
AIC= 207
0.32(.44)
P= 0.000
AIC= 86
0.34(.45)
P= 0.000
AIC= 109
BMCL10
(BMC10)
quantal
quadratic
1.34(1.45)
P= 0.000
AIC = 505
1.42(1.57)
P= 0.055
AIC = 218
1.21(1.35)
P= 0.000
AIC = 279
2.14(2.34)
P= 0.091
AIC = 347
2.68(3.1)
P=0.552
AIC= 131
1.7(1.9)
P=0.21
AIC = 200
0.42(0.47)
P= 0.000
AIC = 100
0.46(.51)
P= 0.000
AIC =119

-------
      0.8-

      0.7-

      0.6-

  I  o.5H

  &
  <  0.4-
   C
  -.§  0.3-
   O
   CO
  LL  0.2-

      0.1-

       0 -
	Best Estimate (log-probit fit)
	Lower Bound of Best Estimate
            BMCL
          BMC
                             10
               0
                        2           3
                       Concentration
Figure B-l.  Benchmark concentration analysis (log-probit) of chronic pulmonary
            inflammation in rats exposed to DPM from Nikula et al. (1995). BMCL10,
            the lower confidence estimate of the concentration of DPM associated with
            a 10% incidence (extra risk); BMC10, the corresponding estimate from the
            best (log-probit) fit. (0) data with 95% error bounds.

-------
                            REFERENCES FOR APPENDIX B

Akaike, H. (1973) Information theory and an extension of the maximum likelihood principle. In: Proceedings of the
Second International Symposium on Information Theory, B.N. Petrov and F. Csaki, eds. Akademiai Kiado,
Budapest, pp. 267-281

Barnhart, MI; Chen, S-T; Salley, SO; et al. (1981) Ultrastructure and morphometry of the alveolar lung of guinea
pigs chronically exposed to diesel engine exhaust: six months' experience. J Appl Toxicol 1:88-103.

Haseman, JK. (1984)  Statistical issues in the design, analysis, analysis and interpretation of animal carcinogenicity
studies. Environ Health Persp 58: 385-392.

Haseman, JK; Huff, JE; Rao, GN and Eustis, SL. (1989) Sources of variability in rodent carcinogenicity studies.
Fund Appl Toxicol 12: 793-804.

Heinrich, U; Muhle, H; Takenaka, S; et al. (1986) Chronic effects on the respiratory tract of hamsters, mice, and rats
after long-term inhalation of high concentrations of filtered and unfiltered diesel engine emissions. J Appl Toxicol
6:383-395.

Heinrich, U; Fuhst, R; Rittinghausen, S; et al. (1995) Chronic inhalation exposure of Wistar rats and two strains of
mice to diesel engine exhaust, carbon black, and titanium dioxide. Inhal Toxicol 7:553-556.

Henderson, RF; Pickrell, JA; Jones, RK; et al. (1988) Response of rodents to inhaled diluted diesel exhaust:
biochemical and cytological changes in bronchoalveolar lavage fluid and in lung tissue. Fundam Appl Toxicol
11:546-567.

Ishinishi, N; Kuwabara, N; Nagase, S; et al. (1986) Long-term inhalation studies on effects of exhaust from heavy
and light duty diesel engines on F344 rats. In: Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. Carcinogenic
and mutagenic effects of diesel engine exhaust: proceedings of the international satellite symposium on
lexicological effects of emissions from diesel engines; July; Tsukuba Science City, Japan. (Developments in
toxicology and environmental science: v. 13.) Amsterdam: Elsevier Science Publishers BV; pp. 329-348

Ishinishi, N; Kuwabara, N; Takaki, Y; et al. (1988) Long-term inhalation experiments on diesel exhaust. In: Diesel
exhaust and health risks: results of the HERP studies. Tsukuba, Ibaraki, Japan: Japan Automobile Research
Institute, Inc., Research Committee for HERP Studies; pp. 11-84.

Mauderly, JL; Jones, RK; Griffith, WC; et al. (1987a) Diesel exhaust is a pulmonary carcinogen in rats exposed
chronically by inhalation. Fundam Appl Toxicol 9:208-221.

Mauderly, JL; Bice, DE; Carpenter, RL; et al. (1987b) Effects of inhaled nitrogen dioxide and diesel exhaust on
developing lung. Cambridge, MA: Health Effects Institute; research report no. 8.

Misiorowski, RL; Strom, KA; Vostal, JJ; et al. (1980) Lung biochemistry of rats chronically exposed to diesel
particulates. In:  Pepelko, WE; Danner, RM; Clarke, NA, eds. Health effects of diesel engine emissions: proceedings
of an international symposium; December 1979. Cincinnati, OH: U.S. Environmental Protection Agency, Health
Effects Research Laboratory; pp. 465-480; EPA report no. EPA-600/9-80-057a. Available from: NTIS, Springfield,
VA;PB81-173809.

Nikula, KJ; Snipes, MB; Barr, EB; et al. (1995) Comparative pulmonary toxicities and carcinogenicities of
chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl Toxicol 25:80-94.

Stone, M. (1998) Akaike's Criteria. In: Encyclopedia of Biostatistics, Armitage, P. and Colton, T., eds. Wiley,
New York.
                                                  B-9

-------
Strom, KA. (1984) Response of pulmonary cellular defenses to the inhalation of high concentrations of diesel
exhaust. J Toxicol Environ Health 13:919-944.

U.S. Environmental Protection Agency (U.S. EPA). (1995) The use of the benchmark dose approach in health risk
assessment. Washington, DC: Office of Research and Development, Risk Assessment Forum United States
Environmental Protection Agency.

U.S. EPA. (2000) Benchmark dose software version 1.2. Washington, DC: National Center for Environmental
Assessment, United States Environmental Protection Agency. Available: http://www.epa.gov/ncea/bmds.htm
[2000, June?].

Wolff, RK; Henderson, RF; Snipes, MB; et al. (1987) Alterations in particle accumulation and clearance in lungs of
rats chronically exposed to diesel exhaust. Fundam Appl Toxicol 9:154-166.

Yu, CP; Yoon, K.J.; Chen, Y.K. (1991) Retention modeling of diesel exhaust particles in rats and humans.
J. Aerosol Res. 4 (2): 79-115.
                                                B-10

-------
              Appendix C

A Summary Review of Cancer Dose-response
        Analyses on Diesel Exhaust
                   c-i

-------
C.I. INTRODUCTION
       Several individuals and organizations have previously conducted dose-response
assessments to estimate quantitatively the cancer risk from exposures to DE. Estimations were
performed on the basis of either epidemiologic and/or experimental data.  As concluded in
Section 8.5, EPA finds that available epidemiologic data are too uncertain to confidently derive a
unit risk estimate for DE-induced lung cancer, and that rat data are not suitable for estimating
human risk. Nevertheless, a review of historical dose-response evaluations is provided here as
background information. This information is not intended to constitute endorsement or a
recommendation for use in quantitative risk assessment.
       Early analyses to quantitatively assess the carcinogenicity of DE were hindered by a lack
of positive epidemiologic studies and long-term animal studies. One means of overcoming these
obstacles was the use of comparative potency methods based on combined epidemiologic and
experimental data.  By the late 1980s, the availability of dose-response data from animal
bioassays and epidemiologic studies provided an opportunity for the derivation of both animal
and human data-based estimates, although considerable uncertainties were generally
acknowledged by the authors of these assessments.

C.2. COMPARATIVE POTENCY METHODS
       In this method, the potency of diesel particulate matter (DPM) extract is compared with
other combustion or pyrolysis products for which epidemiology-based unit risk estimates have
been developed. Comparisons are made using short-term tests such as skin painting, mutations,
and mammalian cell transformation.  The ratio of the  potency of DPM extract to each of these
agents is then multiplied by their individual unit risk estimates to obtain the unit risk for DE. If
epidemiology-based estimates from more than one pollutant are used,  the derived potencies are
generally averaged to obtain an overall mean. Major  uncertainties of this method include the
assumptions that (1) the cancer potency of DE can be determined on the basis of the relative
effectiveness of the organic fraction alone; (2) the relative potency in short-term tests is an
accurate predictor of lung cancer potency; and (3) DPM extracts are similar in chemical
composition and proportion as combustion or pyrolysis products.
       In the  study by Albert et al. (1983), epidemiology-based unit cancer risk estimates for
coke oven emissions, cigarette smoke condensate, and roofing tar were used.  Samples of DPM
were collected from three light-duty engines (a Nissan 220 C, an Oldsmobile 350, and a
Volkswagen turbocharged Rabbit), all run on a highway fuel economy test cycle, and from a
heavy-duty engine  (Caterpillar 3304) run under steady-state, low-load conditions.  The DPM
extracts were tested in a variety of assays.  Dose/concentration-dependent increases in response
were obtained for the four assays listed below:

                                          C-2

-------
       •   Ames Salmonella typhimurium (TA98) reverse mutation,
       •   Gene mutation in L5178Y mouse lymphoma cells,
       •   Sencar mouse skin tumor initiation test, and
       •   Viral enhancement of chemical transformation in Syrian hamster embryo cells.
       Only the first three assays were used to develop comparative potency estimates because
of variability of responses in the enhancement of the viral transformation assay.  The in vitro
studies were carried out both in the presence and absence of metabolic activators. The potency,
defined as the slope of the dose-response curve,  was measured for each sample in each short-
term assay.
       The skin tumor initiation test was positive for all the engines tested except the Caterpillar
engine.  Only the Nissan engine, however, resulted in strong dose-response data.  Because skin
tumor initiation was considered to be the most biologically relevant test, it was used to derive
potency estimates for the Nissan engine. An estimate for the Nissan engine was then derived by
multiplying the epidemiology-based potency estimates for each of the three agents (coke oven
emissions, roofing tar, and cigarette smoke condensate) by the ratios of their potencies in the
skin
tumor initiation test to that of the Nissan diesel engine. According to this method, three 95%
upper-bound estimates of lifetime cancer risk per microgram per cubic meter of extractable
organic matter were derived for the Nissan diesel, based on potency comparisons with each of
the three agents. These values are: coke oven emissions, 2.6 x 10"4; roofing tar, 5.2 x 10"4; and
cigarette smoke condensate, 5.4 x 10"4. The average of the three equals 4.4 x 10"4.
       The potency of the other diesel emission samples was not estimated directly because of
the weak response in the skin tumor initiation test. Instead, their potency relative to the Nissan
engine was estimated as the arithmetic mean of their potency relative to the Nissan in the
Salmonella assay in strain TA98, the sister chromatid exchange assay in Chinese hamster ovary
cells, and the mutation assay in mouse lymphoma cells.  The estimated lifetime cancer risk per
microgram per cubic meter of extractable organic matter for extracts from these engines are as
follows: Volkswagen, 1.3 x lO'4; Oldsmobile, 1.2  x  10'4; and Caterpillar,  6.6 x 10'6.
       Harris (1983) developed comparative potency estimates for the same four engines used
by Albert et al. (1983) but used only two epidemiology-based potency estimates:  those for coke
oven emissions and for roofing tar. He employed preliminary data from three of the same  assays
used by Albert et al. (1983): the Sencar  mouse skin tumor initiation assay, enhancement of viral
transformation in Syrian hamster embryo cells, and the L5178 mouse lymphoma test.  The DE
cancer potency estimates were then derived by multiplying the epidemiology-based  cancer
potency estimates for both coke oven emissions  and roofing tar by the ratio of their potencies
compared with DPM extract in each of the three bioassays. Harris (1983) derived an overall

                                          C-3

-------
mean relative risk value of 3.5 x 10"5per |ig/m3 for the three light-duty engines with a 95% upper
confidence limit of 2.5 x 10"4.  Individual mean values for each engine were not reported.
       McClellan (1986), Cuddihy et al. (1981, 1984), and Cuddihy and McClellan (1983)
estimated a risk of about 7.0 x 10"5 per |ig/m3 DPM using a comparative potency method similar
to those reported in the preceding paragraph. The database was similar to that used by Albert et
al. (1983) and Harris (1983).

C.3. EPIDEMIOLOGY-BASED ESTIMATION OF CANCER RISK
       The first lung cancer risk estimates based on epidemiologic data were derived by Harris
(1983). He assessed the risk of exposure to DE using data from the London Transport Worker
Study reported by Waller (1981). Five groups of employees from the London Transport
Authority (LTA) were used: bus garage engineers, bus drivers, bus conductors, engineers in
central works, and motormen and guards. The first group was considered to  have received the
highest exposure; the next two, intermediate; and the last two groups, none.  When cancer death
rates for the high-exposure group were compared with those of London males, there was no
increase in the observed-to-expected (O/E) ratios.  The author, in fact, considered the results to
be negative.  However, because the low rate of lung cancer in all the LTA exposure groups may
have been the result of a "healthy worker" effect, Harris (1983) compared the exposed groups
with internal controls.  He merged the three exposed groups and compared them with the two
groups considered to be unexposed.  An adjustment was made for the estimated greater exposure
levels of garage engineers compared with bus drivers and conductors. Using this method, the
relative risk of the exposed groups was greater than 1 but was statistically significant only for
garage engineers exposed from 1950 to 1960.  In that case, the O/E ratio was 29% greater than
the presumed unexposed controls.
       Harris (1983) identified a variety of uncertainties relative to potency assessment based on
this study. These included:
       •  small  unobserved differences in smoking incidences among groups, which could have
          a significant effect on lung cancer rates;
       •  uncertainty about the magnitude of exposure in the exposed groups;
       •  uncertainty regarding the extent of change in exposure conditions over time;
       •  random effects arising from the stochastic nature of the cancer incidence; and
       •  uncertainty in the mathematical specification of the model.
       Taking the uncertainties into account, he derived a maximum likelihood excess relative
risk estimate of 1.23 x 10"4, with a 95% upper confidence limit of 5 x 10"4 per |ig/m3 DPM per
year.
                                          C-4

-------
       McClellan et al. (1989) reported risk estimates based on the Garshick et al. (1987) case-
control study in which lung cancer in railroad workers was evaluated. Using a logistic
regression, the expected relative risk of lung cancer death was estimated to rise 0.016 per year of
exposure to DE.  Adjustments were made to convert to continuous exposure (168 vs. 40 hours)
for 70 years. Because exposure levels could not be defined exactly, two sets of calculations were
made, assuming inhaled DPM concentrations of either 500 or 125 |ig/m3 DPM. The number of
excess cancer deaths per year in the United  States was estimated to be 3,800 (95% C.I. 400-7400
when an exposure of 125 |ig/m3 was used, and 950 (95% C.I. 100-1,900) when 500 |ig/m3 DPM
was used.
       The California EPA (Cal-EPA,  1998) derived unit risk estimates for lung cancer based
upon the Garshick et al. (1987) case-control study and the Garshick et al. (1988) cohort study of
U.S. railroad workers. A variety of exposure patterns were considered, characterized by two
components: the average exposure concentration for the workers as measured by Woskie et al.
(1988) and the extent of change in exposure from 1959 to 1980. The lowest lifetime risk
estimate derived was  1.3 x  10"4 per |ig/m3 and the highest was 2.4 x 10"3 per |ig/m3.  The
geometric mean was 6 x 10"4 per  |ig/m3.
       Steenland et al. (1998) estimated lung cancer risk of truck drivers on the basis of a case-
control study of decedents in the Teamsters Union (Steenland et al., 1990). Retrospective
exposure estimates were made starting with a set of 1990 exposure  measurements for different
job categories and then retrospectively estimating from 1982 to about 1950 using various factors,
including diesel vehicle miles traveled and engine emission rates per mile.  The 1990 job
category estimates came from  an extensive industrial hygiene survey of elemental carbon (EC)
exposures in the trucking industry by Zaebst et al. (1991).  Lifetime (through age 75) excess risk
of lung cancer death for male truck drivers was calculated with the  aid of a cumulative exposure
model.  Assuming a most likely emissions scenario of 4.5 g/mile in 1970, and a 45-year exposure
to 5 |ig/m3 of EC beginning at age 20 and ending at age 65, the estimated excess lung cancer risk
was determined to be 1.6% (95%  CI 0.4%-3.1%). Using the same data base, Stayner et al (1998)
presented an estimate of excess lifetime risk of 4.5E-4 for a worker exposed to 1 |ig/m3 of DE
for 45 years.

C.4. ANIMAL BIOASSAY-BASED CANCER POTENCY  ESTIMATES
       With the availability of chronic cancer bioassays, a considerable number of potency
estimates were derived using lung tumor induction in rats. A high degree of uncertainty exists in
the use of the rat data to predict human risk. Major uncertainties include: (1) differences in
particle deposition patterns between rats and humans, (2) differences in sensitivity between rats
and humans to the carcinogenic action of DE, and (3) extrapolation of rat lung tumor responses

                                          C-5

-------
at high concentrations to ambient concentrations without a clear understanding of the mode of
action of DE. It is now widely recognized that the rat lung tumor response associated with any
insoluble particles at high concentrations is mediated by a particle-overload mechanism (ILSI,
2000),  suggesting that rat data for DE are not suitable for estimating human risk at low
environmental concentrations.
       The first risk estimate was reported by Albert and Chen (1986), based on the chronic rat
bioassay conducted by Mauderly et al. (1987). Using a multistage model and assuming
equivalent deposition efficiency in humans and rats, they derived a 95% upper confidence limit
of 1.6 x 10"5 for lifetime risk of exposure to 1 |ig/m3. Pott and Heinrich (1987) also used a linear
model and data reported by Brightwell et al.  (1989), Heinrich et al. (1986), and Mauderly et al.
(1987).  They reported risk estimates ranging from 6 x 10"5 to 12 x  10"5 per |ig/m3.  Smith and
Stayner (1990), using time-to-tumor models based on the data of Mauderly et al. (1987), derived
point (MLE) estimates ranging from 1.0 x  10"4 to 2.1 x 10"4 per |ig/m3 after converting from
occupational to environmental exposure scenario.
       Pepelko and  Chen (1993) developed unit risk estimates based on the data of Brightwell  et
al. (1989), Ishinishi  et al. (1986), and Mauderly et al. (1987) using a detailed dosimetry model to
extrapolate dose to humans and a linearized multistage (LMS) model. Taking the geometric
mean of individual estimates from the three bioassays, they derived unit risk estimates of 1.4 x
10"5 per |ig/m3 when dose was based on carbon particulate matter per unit lung surface area
rather than whole DPM, and 1.2 x 10"4 per |ig/m3 when based on lung burden per unit body
weight.
       Hattis and Silver (1994) derived a maximum likelihood estimate for occupational
exposure of 5.2 x 10"5 per |ig/m3 based on lung burden and bioassay data reported by Mauderly
et al. (1987) and use of a five-stage Armitage-Doll low-dose extrapolation model. California
EPA (CAL-EPA, 1998) derived a geometric  mean estimate of 6 x 10"5per |ig/m3 from five
bioassays using an LMS model.
        To demonstrate the possible influence of particle effects as well  as particle-associated
organics, an additional modeling approach was conducted by Chen and Oberdorster (1996).
Employing a biologically based two-stage  model and using malignant tumor data from Mauderly
et al. (1987), the upper-bound risk estimate for exposure to 1  |ig/m3 was estimated to be
1.7 x 10"5. This estimate is virtually identical to that using the LMS model, assuming
nonthreshold effect of particles.  If a threshold of particle effect is assumed, however, the
estimated risk decreases about fivefold.  The results also show that the mechanism of DE-
induced lung tumor at high exposure concentrations may differ from that at low exposure
concentrations, with the organics and particles playing primary roles of tumorigenesis,
respectively, at low and high concentrations. Overall, the potency estimates on the basis of

                                          C-6

-------
animal bioassays are in the range of 10"6  to 10"4per 1 |ig/m3 of DPM.

       Valberg and Crouch (1999) conducted a meta-analysis of rat bioassays by pooling

together data of low-dose groups from different bioassays.  There are eight bioassays used in the

meta-analysis; half of them had duration of 24 months, and the remaining studies had duration of

30 months or more. Animals with continuous lifetime exposure of less than 600 |ig/m3 of DE

were included in the analysis.  Continuous lifetime exposure is calculated by protracting actual

DE exposure to 30 months (24 hours per day, 7 days per week). The researchers concluded that

exposure of rats to DE at concentrations not associated with lung overload is consistent with no

tumorigenic effect.


                             REFERENCES FOR APPENDIX C

Albert, RE; Chen, C. (1986) U.S. EPA diesel studies on inhalation hazards. In: Carcinogenicity and mutagenicity of
diesel engine exhaust. Ishinishi, N; Koizumi, A; McClellan, R; et al., eds. Amsterdam: Elsevier, pp. 411-419.

Albert, RE; Lewtas, J; Nesnow, S; et al. (1983) Comparative potency method for cancer risk assessment: application
to diesel paniculate emissions. Risk Anal 3:101-117.

Brightwell, J; Fouillet, X; Cassano-Zoppi, AL; et al. (1989) Tumours of the respiratory tract in rats and hamsters
following chronic inhalation of engine exhaust emissions. J Appl Toxicol 9:23-31.

California Environmental Protection Agency. (CAL-EPA, OEHHA) (1998) Health risk assessment for DE.  Public
and Scientific Review Draft.

Chen, CW; Oberdorster, G. (1996) Selection of models for assessing dose-response relationships for particle-
induced lung cancer. Inhal Toxicol 8(suppl):259-278.

Cuddihy, RG; McClellan, RO. (1983) Evaluating lung cancer risks from exposures to diesel engine exhaust. Risk
Anal 3:119-124.

Cuddihy, RG; Griffith, WC; Clark, CR; et al. (1981) Potential health and environmental effects of light-duty diesel
vehicles. Report No. LMF-89. Albuquerque, NM: Inhalation Toxicology Research Institute.

Cuddihy, RG; Griffith, WC; McClellan, RO. (1984) Health risks from light-duty diesel vehicles. Environ Sci
Technol 18:14-21a.

Garshick, E; Schenker, MB; Munoz, A; et al. (1987) A case-control study of lung cancer and DE exposure in
railroad workers. Am Rev Respir Dis 135:1242-1248.

Garshick, E; Schenker, MB; Munoz, A; et al. (1988) A retrospective cohort study of lung cancer and DE exposure
in railroad workers. Am Rev Respir Dis 137:820-825.

Harris, JE. (1983) Diesel emissions and lung cancer. Risk Anal 3:83-100.

Hattis, D; Silver, K. (1994) Use of mechanistic data in occupational health risk assessment: the example of diesel
particulates. In: Chemical risk assessment and occupational health. Smith, MC; Christiani, DC; Kelsey, KT, eds.
Westport, CT: Auburn House, pp. 167-178.
                                               C-7

-------
Heinrich, U; Muhle, H; Takenaka, S; et al. (1986) Chronic effects on the respiratory tract of hamsters, mice and rats
after long-term inhalation of high concentration of filtered and unfiltered diesel engine emissions. J Appl Physiol
6:383-395.

Ishinishi, N; Kuwabara, N; Nagase, S; et al. (1986) Long-term inhalation studies on effects of exhaust from heavy
and light duty diesel engines on F344 rats. In: Carcinogenic and mutagenic effects of diesel engine exhaust.
Ishinishi, N; Koizumi, A; McClellan, RO; et al., eds. Amsterdam: Elsevier, pp. 329-348.

Mauderly, JL; Jones, RK; Griffith, WC; et al. (1987) DE is a pulmonary carcinogen in rats exposed chronically by
inhalation. Fundam Appl Toxicol 9:208-221.

McClellan, RO. (1986) Health effects of DE: a case study in risk assessment. Am Ind Hyg Assoc J 47:1-13.

McClellan, RO; Cuddihy, RG; Griffith, WC; et al. (1989) Integrating diverse data sets to assess the risks of airborne
pollutants. In: Assessment of inhalation hazards. Mohr, U, ed. New York: Springer Verlag, pp. 3-22.

Pepelko, WE; Chen, C. (1993) Quantitative assessment of cancer risk from exposure to diesel engine emissions.
Regul Toxicol Pharmacol 17:52-65.

Pott, F; Heinrich, U. (1987) Dieselmotorabgas und Lungenk auf die Gefahrdung des Menschen.  In: Umwelthygiene,
vol. 19. Med. Institut. f. Umwelthygiene, Annual Report 1986/87. Dusseldorf, F.R.G., pp. 130-167.

Smith, RA; Stayner, L. (1990) An exploratory assessment of the risk of lung cancer associated with exposure to DE
based on a study with rats. Final report. Division of Standards Development and Technology Transfer; Cincinnati,
OH: NIOSH.

Stayner, L; Ddankovic, D; Smith, R; et al. (1998) Predicted lung cancer risk among miners exposed to diesel
exhaust particles. Am J Ind Med 34:207-219.

Steenland, NK;  Silverman, DT; Hornung, RW. (1990) Case-control study of lung cancer and truck driving in the
Teamsters Union. Am J Publ Health 80:670-674.

Steenland, K; Deddens, J; Stayner, L. (1998) DE and lung cancer in the trucking industry: exposure-response
analysis and risk assessment. Am J Ind Med 34:220-228.

Valberg, PA; Crouch, EA. (1999) Meta-analysis of rat lung tumors from lifetime inhalation of diesel exhaust.
Environ Health Perspect 107(9):693-699.

Waller, RE. (1981) Trends in lung cancer in London in relation to exposure to diesel fumes. Environ Int 5:479-483.

Woskie, SR; Smith, TJ; Hammond, SK; et al. (1988) Estimation of the DE exposures of railroad workers: II.
National and historical exposures. Am J Ind Med 13:395-404.

Zaebst, D; Clapp, D; Blade, L; et al.  (1991) Quantitative determination of trucking industry workers' exposures to
diesel particles.  Am Ind Hyg Assoc  J 52:529-541.

-------
^^^^|j^^^                                 Please make all necessary changes on the below label,               PRESORTED STANDARD
                                                  detach or copy, and return to the address in the upper left-              Di~iCTAr^c » cccc DA in
                                                  hand corner.                                               KUt> I Aljt   httb KAIU
United States                                                                                                     EPA
Environmpntal Protection Anpnrv/ORD          if you do not wish to receive these reports CHECK HERE n;                PERMIT No. G-35
tnvironi         roieciion Mgency/UKU          detach, or copy this cover, and return to the address in the
National Center for                               upper left-hand corner.
    Environmental Assessment
Washington, DC 20460


Official Business
Penalty for Private Use
$300

EPA/600/8-90/057F
May 2002

-------