DRAFT                                                   EPA/600/8-90/057C
DO NOT CITE OR QUOTE                                 ,       February 1998
                                                          SAB Review Draft
                Health Assessment Document
                       for Diesel Emissions
                                NOTICE

     THIS DOCUMENT IS A PRELIMINARY DRAFT. It has not been formally
     released by the Environmental Protection Agency and should not at this stage
     be construed to represent Agency policy. It is being circulated for comment on its
     technical accuracy and policy implications.
           National Center for Environmental Assessment-Washington Office
                      Office of Research and Development
                     U.S. Environmental Protection Agency
                             Washington, DC

-------
                                 DISCLAIMER

      This document is a draft for review purposes only and does not constitute U.S.
Environmental Protection Agency policy. Mention of trade names or commercial products does
not constitute endorsement or recommendation for use.
2/1/98                                 i        DRAFT-DO NOT CITE OR QUOTE

-------
                                  CONTENTS
j


 LIST OF TABLES	,. .•. xi

 LIST OF FIGURES	,	,	... xvi

 PREFACE 	-....-	 Xvii

 AUTHORS, REVIEWERS, AND CONTRIBUTORS	xviii

 ACKNOWLEDGMENTS	..'	xx

 1.  INTRODUCTION	1-1

 2.  DIESEL EMISSIONS, TRANSPORT, AND TRANSFORMATION	.2-1
    2.1. INTRODUCTION.	2-1
    2.2. OVERVIEW OF DIESEL POLLUTANTS AND POLLUTANT
        FORMATION	....	2-4
        2.2.1. Gas-Phase Pollutant Emissions	2-4
              2.2.1.1. Oxides-of-Nitrogen Formation	 2-4
              2.2.1.2. Hydrocarbons and Carbon Monoxide Formation	2-5
        2,2.2. Particle Formation and Emission	2-6
        2.2.3. Gas-to-Particle Conversion			 2-7
              2.2.3.1. Condensation of Organic Matter ,	2-7
              2.2.3.2. Oxidation of Sulfur Oxides	 2-7
        •2.2:4. Nitroarene Formation	2-8
    2.3. PRIMARY DIESEL EMISSIONS	2-8
        2.3.1. Gaseous Emissions	2-10
        2.3.2. Particulate Emissions	2-13
              •2.3.2.1. Diesel Particulate Matter	...'..,.	*... 2-13
              2.3.2.2. Particulate-Phase Inorganics	2-15
              2.3.2.3. Particulate-Phase Organic Compounds	2-15
        2.3.3. Gaseous-/Particulate-Phase Emission Partitioning of
              Polycyclic Aromatic Hydrocarbons	 2-23
    2.4. ATMOSPHERIC TRANSFORMATIONS OF PRIMARY DIESEL
        EMISSIONS	2-25
        2.4.1. Long-Range Transport and Fate of Primary Diesel Emissions	2-25
        2.4.2. Chemical Transformations	:	2-28
              2.4.2.1. Gas-Phase Reactions	2-28
              2.4.2.2. Particulate-Phase Reactions		..... 2-40
        2.4.3. Physical Removal Processes	 2-44
              2.4.3.1. Dry Deposition  		2-44
              2.4.3.2. Wet Deposition	2-44
    2.5. ATMOSPHERIC CONCENTRATIONS OF PRIMARY DIESEL
        EMISSIONS AND THEIR TRANSFORMATION PRODUCTS	 2-45


 2/1/98                    ,            iii       DRAFT-DO NOT CITE OR QUOTE

-------
                             CONTENTS (continued)
        2.5.1,  Volatile Organic Compounds Attributable to Traffic	2-47
        2.5.2.  Polycyclic Aromatic Hydrocarbons	2-47
        2.5.3.  Nitroarene Concentrations in Ambient Air ....	2-51
   2.6.  EXPOSURE PERSPECTIVE	..	2-56
   2.7.  SUMMARY	 2-58
   2.8.  REFERENCES	 2-59

4.  DOSIMETRIC FACTORS	.44
   4.1.  INTRODUCTION	...:...	4-1
   4.2.  REGIONAL DEPOSITION OF INHALED PARTICLES	4-2
        4.2.1.  Physical Processes, Physiological/Anatomical  Considerations,
              and Particle Characteristics	4-2
        4.2.2.  Species Variability in Regional Dose	4-4
   4.3.  RESPIRATORY TRACT CLEARANCE RATES	4-6
        4.3.1.  Tracheobronchial Clearance	4-6
        4.3.2,  Clearance From the Alveolar Region	 4-9
              4.3.2.1. Alveolar Clearance in Humans	4-9
              4.3.2.2. Alveolar Clearance in Animals	4-10
              4.3.2.3. Lung Burden and Pulmonary Overload Resulting iii
                    Impaired Clearance	.	4-13
        4.3.3.  Role of AMs in the Clearance of Particulate Matter	4-16
              4.3.3.1. AM-Mediated Clearance of Particulate Matter	4-16
              4.3.3.2. Translocations of Particles to Extraalveolar
                    Macrophage Compartment Sites	4-18.
              4.3.3.3. Potential Mechanisms for an AM Sequestration
                    Compartment for Particles During. Particle Overload	4-20
              4.3.3.4. Physiologically Based Models for Pulmonary
                    Retention and Clearance of Insoluble Particles	4-21
   4.4.  BIOAVAILABILITY OF ORGANIC CONSTITUENTS PRESENT
        ON DIESEL EXHAUST PARTICLES	 4-22
        4.4.1.  Laboratory Animal Studies	4-23
        4,4.2.  Extraction of Diesel Particle-Associated Organics in
              Biological Fluids			 4-25
        4.4.3.  Extraction of Diesel Particle-Associated Organics by
              Alveolar Lung Cells and Other Cell Types	4-27
        4.4.4.  Bioavailability of Adsorbed Compounds as a Function of Particle
              Clearance Rates and Extraction Rates of Adsorbed Compounds ......... 4-28
   4.5.  CONSIDERATIONS FOR DOSIMETRY MODELING	4-30
   4.6.  SUMMARY	 4-31
   4.7.  REFERENCES	4-33

5.  NONCANCER HEALTH EFFECTS OF DIESEL EXHAUST	  5-1
   5.1.  HEALTH EFFECTS OF WHOLE EXHAUST	5-1

2/1/9.8                                  iv        DRAFT-DO NOT CITE OR QUOTE

-------
                           CONTENTS (continued)
       5.1.1.  Human Data	  5-1
             5.1.1.1. Short-Term Exposures	,,	  5-1
             5,1.1,2. Long-Term Exposures	,	  5.7
       5.1.2.  Laboratory Animal Studies	  5-17
             5,1.2.1. Acute Exposures	  5-17
             5.1,2,2. Short-Term and Subchronic Exposures	,	  5-19
             5.1.2.3. Chronic Exposures	  5-24
   5.2. COMPARISON OF HEALTH EFFECTS OF FILTERED AND
       UNFILTERED DIESEL EXHAUST	....'.....	  5-71
   5.3. INTERACTIVE EFFECTS OF DIESEL EXHAUST  	  5-75
   5.4. COMPARATIVE RESPONSIVENESS AMONG SPECIES TO THE
       PULMONARY EFFECTS OF DIESEL EXHAUST	  5-77
   5.5. DOSE-RATE AND PARTICIPATE CAUSATIVE ISSUES  .,..	  5-78
   5.6. SUMMARY AND DISCUSSION	  5-81
       5.6.1,  Effects of Diesel Exhaust on Humans	  5-81
       5.6,2.  Effects of Diesel Exhaust on Laboratory Animals	  5-83
             5.6.2.1. Effects on Survival and Growth	  5-83
             5.6.2.2. Effects on Pulmonary Function  	  5-83
             5.6.2.3. Histopathological and Histochemical Effects	,	  5-84
             5.6.2.4. Effects on Defense Mechanisms	  5-84
             5.6.2.5. Neurological and Behavioral Effects	  5-85
             5.6.2.6. Other Noncancerous Effects	  5-86
       5.6.3. Comparison of Filtered and Unfiltered Diesel Exhaust..;	  5-86
       5.6.4. Interactive Effects of Diesel Exhaust	  5-86
       5.6.5. Summary  		.  5-86
   5.7. REFERENCES	  5-87

6.  QUALITATIVE AND QUANTITATIVE ASSESSMENT OF NONCANCER
   HEALTH EFFECTS-DERIVATION OF THE INHALATION
   REFERENCE CONCENTRATION	.. 6-1
   6.1. INTRODUCTION	....	6-1
   6.2, DETERMINATION OF CRITICAL TARGET SITE	 6-2
   6.3. APPROACH FOR DERIVATION OF THE INHALATION
       REFERENCE CONCENTRATION	.. 6-3
   6.4. THE PRINCIPAL STUDIES FOR INHALATION REFERENCE
       CONCENTRATION DERIVATION	6-5
   6.5. SUPPORTING STUDIES FOR INHALATION REFERENCE
       CONCENTRATION DERIVATION	 6-8
       6.5.1.  Respiratory Tract Effects in Species Other Than the Rat	6-12
   6.6. DERIVATION OF THE INHALATION REFERENCE
       CONCENTRATION	;....... 6-14
       6.6.1.  Application of the Benchmark Dose Approach to
2/1/98                               v       DRAFT-DO NOT CITE OR QUOTE

-------
                           CONTENTS (continued)
             Derivation of the RfC	6-16
   6.7. SUMMARY	.,	6-26
   6.8. REFERENCES	 6-27

7.  CARCINOGENICITY OF DIESEL EMISSIONS IN LABORATORY ANIMALS .... 7-1
   7.1. INTRODUCTION	7-1
   7.2. INHALATION STUDIES	7-2
       7.2,1.  Rat Studies	•	7-2
       7.2.2.  Mouse Studies			:	 7-19
       7.2.3.  Hamster Studies	7-23
       7.2.4.  Monkey Studies	7-25
   7.3. LUNG IMPLANTATION OR INTRATRACHEAL
       INSTILLATION STUDIES	7-25
       7.3.1.  Rat Studies	7-25
       7.3.2.  Syrian Hamster Studies	7-28
       7.3.3.  Mouse Studies		7-30
   7.4. SUBCUTANEOUS AND INTRAPERJTONEAL INJECTION STUDIES  	7-31
       7.4.1.  Mouse Studies	,		..".	,	7-31
   7.5. DERMAL STUDIES  . ,	7-33
       7.5.1.  Mouse Studies .		'.	7-33
   7.6. SUMMARY AND CONCLUSIONS OF LABORATORY ANIMAL
       CARCINOGENICITY STUDIES	7-38
   7.7. REFERENCES	 7-45

8.  EPIDEMIOLOGIC STUDIES OF THE CARCINOGENICITY OF
   EXPOSURE TO DIESEL EMISSIONS	,	8-1
   8.1. INTRODUCTION	,	8-1
   8.2. COHORT STUDIES	8-2
       8.2.1.  Waller (1981): Trends in Lung Cancer in London in
             Relation to Exposure to Diesel Fumes ,	 8-2
       8.2.2.  Howe et al. (1983):  Cancer Mortality (1965 to 1977) in
             Relation to Diesel Fume and Coal Exposure in a Cohort of
             Retired Railroad Workers	8-3
       8.2.3,  Rushton et al, (1983):  Epidemiological Survey of Maintenance
             Workers in the London Transport Executive
             Bus Garages and Chiswick Works	8-5
       8.2.4.  Wong et al. (1985): Mortality Among Members of a Heavy
             Construction Equipment Operators Union With Potential
             Exposure to Diesel Exhaust Emissions	8-6
       8.2.5.  Edlingetal. (1987): Mortality Among Personnel Exposed
             to Diesel Exhaust	8-10
2/1/98                                vi       DRAFT-DO NOT CITE OR QUOTE

-------
                              CONTENTS (continued)
        8.2.6.  Boffetta and Stellman( 1988): Diesel Exhaust Exposure and
               Mortality Among Males in the American Cancer Society
               Prospective Study	..'..!	8-11
        8.2.7.  Garshicketal. (1988): A Retrospective Cohort Study of Lung
               Cancer and Diesel Exhaust Exposure in Railroad Workers	8-13
        8.2.8.  Gustavssonetal. (1990): Lung Cancer and Exposure to
               Diesel Exhaust Among Bus Garage Workers	8-16
        8.2.9.  Hansen(1993):  A Follow-Up Study on the Mortality of
               Truck Drivers	".	8-17
   8.3.  CASE-CONTROL STUDIES OF LUNG CANCER	8-22
        8.3.1.  Williams et al. (1977): Associations of Cancer Site and Type
               With Occupation and Industry From the Third National
               Cancer Survey Interview	  8-22
        8.3.2.  Hall and Wynder (1984): A Case-Control Study of Diesel   '
               Exhaust Exposure and lung Cancer	8-23
        8,3.3.  Damber and Larsson( 1987):  Occupation and Male Lung
               Cancer, a Case-Control Study in Northern Sweden	8-25
        8.3.4.  Lerchen et al. (1987):  Lung Cancer and Occupation in New Mexico	8-26
        8.3.5.  Garshicketal. (1987): A Case-Control Study of Lung Cancer
               and Diesel Exhaust Exposure in Railroad Workers	8-28
        8.3.6.  Benhamou et al. (1988):  Occupational Risk Factors of Lung
               Cancer in a French Case-Control Study	,	8-31
        8.3.7.  Hayes etal. (1989): Lung Cancer in Motor
               Exhaust-Related Occupations	'•-...,	  8-32
        8.3.8.  Steenland et al. (1990): A Case-Control Study of Lung
               Cancer and Truck Driving in the Teamsters Union	8-34
        8.3.9.  Boffetta et al. (1990):  Case-Control Study on Occupational
               Exposure to Diesel Exhaust and Lung Cancer Risk	8-35
        8.3.10. Emmelin et al. (1993): Diesel Exhaust Exposure and Smoking:
               A Case-Referent Study of Lung Cancer Among
               Swedish Dock Workers	 8-36
   8.4.  CASE-CONTROL STUDIES OF BLADDER CANCER	8-38
        8.4.1.  Howeetal, (1980):  Tobacco  Use, Occupation, Coffee,
               Various Nutrients, and Bladder Cancer	 8-38
        8.4.2.  Wynder etal. (1985): A Case-Control Study of Diesel
               Exhaust Exposure and Bladder Cancer	8-44
        8.4.3.  Hoar and Hoover (1985): Truck Driving and Bladder
               Cancer Mortality in Rural New England	8-46
        8.4.4.  Steenland etal. (1987): A Case-Control Study of Bladder
               Cancer Using City Directories as a Source of Occupational  Data	,. 8-47
        8.4.5.  Iscovichetal. (1987):  Tobacco Smoking, Occupational
               Exposure, and Bladder Cancer in Argentina	8-49


2/1/98                                   vii        DRAFT-DO NOT CITE OR QUOTE

-------
                          CONTENTS (continued)
       8.4.6. Iyer et al, (1990): Diesel Exhaust Exposure and Bladder Cancer Risk	8-51
       8.4.7. Steineck et al. (1990): Increased Risk of Urothelial Cancer
            in Stockholm From 1985 to 1987, After Exposure to
            Benzene and Exhausts .,	8-53
   8.5.  DISCUSSION AND SUMMARY	8-54
       8.5.1. The Cohort Mortality Studies	8-60
       8.5.2. Case-Control Studies of Lung Cancer	8-63
       8.5.3. Case-Control Studies of Bladder Cancer .,	•	 8-64
       8.5.4. Relevant Methodologic Issues	."	8-65
       8.5.5. Criteria of Causal Inference	8-66
   8.6.  REFERENCES	8-70

9.  MUTAGENICITY	9-1
   9.1.  GENE MUTATIONS	9-1
   9.2.  CHROMOSOME EFFECTS	9-4
   9.3.  OTHER GENOTOXIC EFFECTS	.9-5
   9.4.  SUMMARY	.9-5
   9.5.  REFERENCES	 9-7

10. METABOLISM AND MECHANISM OF ACTION IN DIESEL
   EMISSION-INDUCED CARCINOGENESIS	10-1
   10:1.  PARTICLE-INDUCED CARCINOGENIC RESPONSE	 10-1
   10.2.  METABOLISM AND MECHANISM OF ACTION OF ORGANIC
        CARCINOGENIC COMPONENTS OF DIESEL EXHAUST  	10-4
        10.2.1.  Metabolism and Disposition of B[a]P Relative to Diesel Exhaust	 10-6
        10.2.2.  Metabolism and Disposition of Nitroarenes	10-10
        10.2.3.  Formation of Reactive Oxygen Species From Organic
              Constituents of Diesel Exhaust and Their Involvement
              in the Induction of Lung Cancer	10-12
   10.3.  POTENTIAL INVOLVEMENT OF PULMONARY LEUKOCYTES
        IN THE DEVELOPMENT OF LUNG TUMORS	 10-13
   10.4.  DNA ADDUCT CONSIDERATIONS	 10-17
   10.5.  SUMMARY OF METABOLISM AND MECHANISM OF ACTION OF
        CARCINOGENIC COMPONENTS OF DIESEL EXHAUST  	10-22
   10.6.  REFERENCES	10-24

11. QUALITATIVE AND QUANTITATIVE EVALUATION OF THE
   CARCINOGENICITY OF DIESEL ENGINE EMISSIONS	11-1
   11.1.  INTRODUCTION	.;....:	11-1
   112.  QUALITATIVE EVALUATION OF THE CARCINOGENIC RISK
        OF EXPOSURE TO DIESEL EXHAUST	 11-1
2/1/98                             viii      DRAFT-DO NOT CITE OR QUOTE

-------
                             CONTENTS (continued)
         11.2.1.  Hazard Evaluation Based on EPA 1986 Guidelines for
                Carcinogen Risk Assessment	11-1
         11,2.2.  Hazard Evaluation Based on EPA Proposed Guidelines for
                Carcinogen Risk Assessment	11-3
   11.3.  QUANTITATIVE ESTIMATION OF THE CARCINOGENIC RISK
         OF EXPOSURE TO DIESEL EXHAUST	11-4
         11.3.1.  Review of Previous Quantitative Risk Estimates .,	 11-4
         11.3.2.  Current Approaches to Quantitation of Human Cancer Risk
                From Exposure to Diesel Exhaust	,...'	11-10
                11.3.2.1.  Dose-Response Estimates Using Human
                        Epidemiologic Data	11-10
                11.3.2.2.  Dose-Response Estimates Based on Animal
                        Bioassay Data	,	,	 11-11
                11.3.2.3.  Dose-Response Estimation Using Benzo(a)pyrene
                        as a Dosimeter	,	;	11-17
         11.3.3.  Strengths and Weaknesses of Differing Approaches for
                Quantitating Cancer Risk	11-17
                11.3.3.1.  Dose-Response Estimates Based on Human
                        Epidemiologic Data	.....	 11-17
                11.3.3.2.  Dose-Response Estimate-Based Laboratory
                        Animal Bioassays	11-19
                11.3.3.3.  Dose-Response Estimates Based on the
                        Biomarker Approach	11-21
                11.3.3.4.  Dose-Response Estimates Based on the
                        Comparative Potency Approach	11-21
   11.4.  SUMMARY AND RECOMMENDATIONS	11-22
   11.5.  REFERENCES	11-23

12. HEALTH RISK CHARACTERIZATION FOR DIESEL ENGINE EMISSIONS	12-1
   12.1.  INTRODUCTION	 12-1
         12.1.1.  Scope	 12-1
         12.1.2.  What Is Diesel Exhaust in a Risk Assessment Context?	12-1
   12.2.  HAZARD ASSESSMENT  ....,	12-3
         12.2.1.  Hazard Assessment for Health Effects Other Than Cancer	 12-3
                12.2.1.1.  Effects From Acute Exposure	12-3
                12.2.1.2.  Effects From Short-Term and Chronic Exposure	12-5
         12.2.2.  Toxicity Mode of Action		....... 12-7
         12.2.3.  Carcinogenic Hazard Assessment	 12-9
                12.2.3.1.  Human Evidence	12-9
                12.2.3.2.  Animal Evidence	 12-11
         12.2.4.  Weight-of-Evidence Summary	,	12-13
   12.3.  DOSE-RESPONSE ASSESSMENT	12-14
         12.3.1.  Considerations in Modeling the Dose-Response Relationship	..12-14

2/1/98                                  ix        DRAFT-DO NOT CITE OR QUOTE

-------
                           CONTENTS (continued)
        12.3.2. Dose-Response Assessment for Health Effects Other Than Cancer  .... 12-15
              12.3.2.1. Derivation of an Inhalation Reference Concentration	12-16
        12.3.3. Dose-Response Assessment for Carcinogenic Effects	12-17
              12.3.3.1. High-to Low-Dose Extrapolation and
                      Mode-of-Action Considerations 	12-17
              12.3.3.2. Estimating the Cancer Risk of Diesel Exhaust	12-19
              12.3.3.3. Bounds for Cancer Risk and Margin of Exposure	12-25
        12.3.4. Susceptible Subgroups	'. .•	12-27
        12.3.5. Other Comprehensive Diesel Exhaust Health Assessments
              in the United States  	..'..'	12-29
   12.4. EXPOSURE PERSPECTIVE		12-29
   12.5. DIESEL EXHAUST HEALTH RISKS-A PLAIN-LANGUAGE
        OVERVIEW OF KEY INFORMATION	12-31
   12.6. REFERENCES	 12-38

APPENDIX A: EXPERIMENTAL PROTOCOL AND COMPOSITION OF
             EXPOSURE ATMOSPHERES	 A-l
APPENDIX B: ALTERNATIVE MODEL FOR DIESEL CANCER
             RISK ASSESSMENT	B-l
APPENDIX C: MODELS FOR CALCULATING LUNG BURDENS	C-l
2/1/98                                x        DRAFT-DO NOT CITE OR QUOTE

-------
                                  LIST OF TABLES
2-1,    Major components of diesel engine emissions and their known
       atmospheric transformation products	'	, 2-3

2-2,    Some nitroarenes identified in vehicle exhaust	2-9

2-3.    Emission rates of volatile organic compounds (VOC) from
       diesel and gasoline engines	.,...,,;	..2-11,

2-4.    Particulate matter emission rates and their distribution between total
       carbon (TC) and organic carbon (OC) for heavy-duty diesel (HDD),
       light-duty diesel (LDD), and gasoline engines	 2-15

2-5.    Particle phase and gaseous phase emissions from diesel exhaust,
       and their atmospheric reaction products  	,	2-16

2-6.    Summary of composition and emission rates (in milligrams per kilometer)
       of airborne paniculate matter from on-road vehicles, Tuscarora Mountain
       Tunnel 1977 experiment	,	2-17

2-7.    Classes of organic compounds identified in particulate-phase
       combustion emissions	2-18

2-8.    Polycyclic aromatic hydrocarbons identified and quantified in
       extracts of diesel particles	 2-20

2-9.    Emission rates of particle-bound polycyclic aromatic hydrocarbons (PAHs)
       (in milligrams per kilometer) from HDD, LDD, and gasoline engines	2-21

2-10.   Concentrations of nitro-polycyclic aromatic hydrocarbons identified
       in a light-duty diesel (LDD) particulate extract			2-22

2-11.   Vapor pressures at 25 °C for a series of polycyclic aromatic
       hydrocarbons (PAHs)	. 2-24

2-12.   Calculated atmospheric lifetimes for gas-phase reactions of
       selected compounds present in automotive emissions with
       atmospherically important reactive species	 2-29

2-13.   Summary of the nitroarenes produced from the gas-phase hydroxyl (OH)
       radical-initiated and dinitrogen pentoxide (N2O5) reactions and electrophilic
       nitration of polycyclic aromatic hydrocarbons (PAHs)	2-34
2/1/98                                     xi        DRAFT-DO NOT CITE OR QUOTE

-------
                             LIST OF TABLES (continued)
2-14,   Mean particle, gas, and overall scavenging ratios for neutral
       organic compounds	,	2-46

2-15.   Concentrations of individual hydrocarbons and aldehydes measured
       in the Raleigh, NC, roadside study	 2-48

2-16.   Particle-and vapor-phase polycyclic aromatic hydrocarbon
       concentrations for Baltimore Harbor Tunnel samples	2-50

2-17.   Average ambient concentrations of polycyclic aromatic hydrocarbons
       measured in Glendora, CA	 2-51

2-18.   Maximum concentrations of nitrofluoranthene (NF) and nitropyrene (NP)
       isomers observed at three South Coast Air Basin sampling sites	2-52

2.19.   Estimated annual ambient concentrations of diesel exhaust particulate
       matter(ug/m3)	 2-57

4-1.    Predicted doses of inhaled diesel exhaust particles per minute
       based on total lung volume (M), total airway surface area (Mi), or
       surface area in alveolar region (M2)	 4-5

5-1.    Human studies of exposure to diesel exhaust	 5-12

5-2.    Short-term effects of diesel exhaust on laboratory animals	5-20

5-3.    Effects of chronic exposures to diesel exhaust on survival and
       growth of laboratory animals	 5-25
                                                         *
5-4.    Effects of chronic exposures to diesel exhaust on organ weights and
       organ-to-body-weight ratios	 5-27

5-5.    Effects of diesel exhaust on pulmonary function of laboratory animals	5-31

5-6.    Histopathoiogical effects of diesel exhaust in the lungs of
       laboratory animals	:	5-35

5-7.    Effects of exposure to diesel exhaust on the pulmonary defense
       mechanisms of laboratory animals	 5-46

5-8.    Effects of exposures to diesel exhaust on the immune system of
       laboratory animals	5-56


2/1/98                                    xii        DRAFT-DO NOT CITE OR QUOTE

-------
                            LIST OF TABLES (continued)
5-9.    Effects of exposure to diesel exhaust on the liver of laboratory animals	5-58

5-10.  Effects of exposure to diesel exhaust on the hematological and
       cardiovascular systems of laboratory animals	 5-59

5-11.  Effects of chronic exposures to diesel exhaust on serum chemistry
       of laboratory animals	5-62

5-12.  Effects of chronic exposures to diesel exhaust on microsomal '
       enzymes of laboratory animals 	5-64

5-13.  Effects of chronic exposures to diesel exhaust on behavior and
       neurophysiology	 5-68

5-14.  Effects of chronic exposures to diesel exhaust on reproduction and
       development in laboratory animals	5-70

5-15.  Composition of exposure atmospheres in studies comparing
       unfiltered and filtered diesel exhaust	5-72

6-1.    Human equivalent continuous concentrations from the principal studies	 6-9

6-2.    Definition of benchmark response levels for endpoints important
       in the diesel exhaust BMC analysis	6-19

6-3.    Results of benchmark concentration analyses using the
       polynomial model and data from the ITRI study	 6-20

6-4.    Results of benchmark concentration analyses using the
       polynomial model and data from the HERP study	6-22

6-5.    Results of benchmark concentration analyses using the
       polynomial model and data from the GM and Nikula studies	 6-23

6-6.    Comparison of benchmark concentration calculated using the
       polynomial and Weibull models for selected endpoints	6-25

7-1.    Summary of animal carcinogenicity studies	 7-3

7-2.    Tumor incidence and survival time of rats treated with
       fractions from diesel exhaust condensate (35 rats/poup)	 7-26
2/1/98                           -         xiii       DRAFT-DO NOT CITE OR QUOTE

-------
                             LIST OF TABLES (continued)
7-3.    Tumor incidences in rats following intratracheal instillation of
       diesel exhaust particles (DPM), extracted DPM, carbon black (CB),
       benzo[a]pyrene (BaP), or particles plus BaP	7-29

7-4.    Tumorigenic effects of dermal application of acetone extracts
       of diesel particulate matter (DPM)	7-34

7-5.    Dermal tumorigenic and carcinogenic effects of various
       emission extracts	7-37

7-6.    Cumulative (concentration * time) exposure data for rats exposed
       to whole diesel exhaust	7-41

8-1.    Epidemiologic studies of the health effects of exposure to diesel exhaust:
       cohort mortality studies	8-19

8-2.    Epidemiologic studies of the health effects of exposure to diesel exhaust:
       case-control studies of lung cancer	8-39

8-3.    Epidemiologic studies of the health effects of exposure to diesel exhaust:
       case-control studies of bladder cancer	8-55

11-1.   Estimated 95% upper confidence limits of the lifetime risk of cancer
       from inhalation of 1 |ag/m3 diesel particulate matter (DPM)	11-9-

11-2.   Incidence of lung tumors in Fischer 344 rats (males and  females
       combined) exposed to heavy-duty engine exhaust  	11-13

11-3.   Incidence of lung tumors in Fischer 344 rats (males and  females
       combined) exposed to heavy-duty engine exhaust	11-13
                                                                              \
11-4.   Incidence of lung tumors in Fischer 344 rats (males and  females
       combined) exposed to diesel engine exhaust	11-14

11-5.   Unit risk estimates per |ag/m3 of diesel exhaust based on the
       Yuetal. (1991)  dosimetry model	11-15

11 -6.   Relative importance of particles and organics at different exposure
       concentrations assuming either threshold or nonthreshold effects
       of particles	11-16
2/1/9.8                                    xiv       DRAFT-DO NOT CITE OR QUOTE

-------
                          LIST OF TABLES (continued)
12-1,  Particle-phase and gaseous-phase emissions from diesel exhaust
      and their atmospheric reaction products	 12-4
2/1/98                    .             xv       DRAFT-DO NOT CITE OR QUOTE

-------
                                  LIST OF FIGURES


2-1,    Typical size distribution of diesel exhaust particles	  2-14

2-2.    Vapor-/particle-phase polycyclic aromatic hydrocarbon distribution in
       samples collected in Baltimore Harbor Tunnel	  2-25

2-3.    Diesel-derived pollutants: emission-to-deposition atmospheric cycle	2-26

2-4.    Mass chromatograms of the molecular ion of the nitrofluoranthenes (NF) and
       nitropyrenes (NP) formed from the gas-phase reaction of fluoranthene and
       pyrene with the OH radicals and present in the ambient particulate sample
       collected at Torrance, CA	2-54

2-5.    Mass chromatograms of the molecular ion of the nitrofluoranthenes
       and nitropyrenes present in ambient particulate samples collected in
       Torrance, CA, and Clarernont, CA  	,:..  3-55

4-1.    Deposition distribution patterns of inhaled diesel exhaust particles
       in the airways of different species	4-5

4-2.    Clearance of insoluble particles deposited in tracheobronchial
       and alveolar regions	.4-8

4-3.    Short-term thoracic clearance of inhaled particles as determined
       by model prediction and experimental measurement	4-9

4-4.    Physiology-oriented compartmental kinetics model of retention of
       inhaled insoluble particles in rats	  4-15

7-1.    Cumulative exposure data for rats exposed to whole diesel exhaust	7-43

8-1.    Lung cancer and exposure to diesel exhaust in railroad workers	8-68

8-2.    Lung cancer and exposure to diesel exhaust in truck drivers	8-68
2/1/98        ,                            xvi        DRAFT-DO NOT CITE OR QUOTE

-------
                                      PREFACE

       This draft health assessment was prepared by the National Center for Environmental
Assessment-Washington Office, which is a risk assessing program in the EPA's Office of
Research and Development, This assessment has been prepared for EPA's Office of Mobile
Sources which has mandates to consider the health hazards associated with diesel engine use in
vehicular transportation. As diesel exhaust emissions also affect air toxics and ambient
particulate matter, other EPA air programs also have an interest in this assessment. An earlier
draft of this assessment was released for public comment in December 1994, and the Agency's
Clean Air Scientific Advisory Committee (CASAC) met in public session in May 1995 to review
the draft. This February 1998 draft builds on the 1994-1995 history.
       The scientific literature search for this assessment is generally current through December
1997 for the health chapters and is generally current through early 1995 for the background
chapters dealing with emission characterization. The health assessment document series is not
intended to be encyclopedic, rather the focus is on key studies and thus, the health assessment
series is less comprehensive than some other types of assessments prepared by EPA.
       This February 1998 version of the assessment will be reviewed a second time by CASAC
and then EPA will finalize the document.
2/1/98                                   xvii        DRAFT-DO NOT CITE OR QUOTE

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

Dr. Ronald Bradow
Atmospheric Research and Exposure
 Assessment Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC

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

Dr. Kenney Crump
Clement International Corporation
Ruston, LA

Dr. Daniel Guth
National Center for Environmental
 Assessment
U.S. Environmental Protection Agency
Research Triangle Park, NC

Dr. John Johnson
Michigan Technological University
Houghton, MI

Dr. Apama Koppikar
National Center for Environmental
 Assessment
U.S. Environmental Protection Agency
Washington, DC

Ms. Tarnmie Lambert
Clement International Corporation
Ruston, LA

Dr. Brace Lehnert
Pulmonary Biology-Toxicology Section
Los Alamos National Laboratory
Los Alamos, NM
     Dr. Samuel Lestz
     State College, PA

     Dr. Kumar Menon
     Pikesville, MD

     Dr. Giinter OberdSrster
     Department of B iophysics
     University of Rochester Medical Center
     Rochester, NY

     Dr. Dennis Opresko
     Biomedical and Environmental Information
      Analysis
     Health and Safety Research Division
     Oak Ridge National Laboratory
     Oak Ridge, TN

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

     Mr. Chris Rambin
     Clement International Corporation
     Ruston, LA

     Dr. Larry Valcovic
     National Center for Environmental
      Assessment
     U.S. Environmental Protection Agency
     Washington,  DC

     Mr. Michael Walsh
     Arlington, VA

     Dr. Ronald K. Wolff
     Lilly Research Laboratories
     Greenfield, IN
2/1/98
xvm
DRAFT-DO NOT CITE OR QUOTE

-------
            AUTHORS, REVIEWERS, AND CONTRIBUTORS (continued)
Dr. Robert A. Young
Biomedical and Environmental Information
  Analysis
Health Sciences Research Division
Oak Ridge National Laboratory
Oak Ridge, TN

Dr. Chia-Ping Yu
State University of New York at Buffalo
Department of Mechanical and Aerospace
  Engineering
Buffalo, NY

Project Managers

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

Contributors and Reviewers
     Dr. Barbara Zielinska
     Desert Research Institute
     Energy and Environmental Engineering
       Center
     Reno, NV
     Dr. William Pepelko
     National Center for Environmental
      Assessment
     U.S. Environmental Protection Agency
     Washington, DC
      The following individuals reviewed the current and/or an earlier draft of this document
and participated in a peer review workshop on July 18 and 19,1990.
Dr. Roy Albert
University of Cincinnati
Cincinnati, OH

Dr. James Bond
Chemical Industries Institute of Toxicology
Research Triangle Park, NC

Dr. Glen Cass
California Institute of Technology
Pasadena, CA

Dr. Eric Garshik
Harvard Medical School
Channing Laboratory
Boston, MA
     Dr. Judith Graham
     National Exposure Research Laboratory
     U.S. Environmental Protection Agency
     Research Triangle Park, NC

     Dr. Uwe Heinrich
     Department of Environmental Hygiene
     Fraunhofer Institute
     Hanover, Germany

     Dr. Joellen Lewtas
     National Health and Environmental
      Effects .Research Laboratory
     Research Triangle Park, NC
2/1/98
xix
DRAFT-DO NOT CITE OR QUOTE

-------
            AUTHORS, REVIEWERS, AND CONTRIBUTORS (continued)
Dr. Joe Mauderly
Lovelace Inhalation Research Institute
Albuquerque, NM

Dr. Roger McClellan
Chemical Industries Institute of Toxicology
Research Triangle Park, NC

Dr. Fred Miller
Duke University Medical Center
Durham, NC

Dr. Otto Raabe
University of California
Davis, CA

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

Dr." Herbert Rosenkranz
Department of Environmental Sciences
School of Medicine
Case Western University
Cleveland, OH

Dr. Irving Salmeen
Ford Motor Company
Scientific Research Laboratory
Dearborn, MI

Dr. Andrew Sivak
Cambridge, MA
    Dr. Jeanette Wiltse
    National Center for Environmental
      Assessment
    U.S. Environmental Protection Agency
    Washington, DC

    Dr. Thomas Smith
    University of Massachusetts Medical Center
    Worcester, MA '

    Dr. Frank Speizer
    Charming Laboratory
    Boston, MA

    Dr. Leslie Stayner
    National Institute  for Occupational Safety
      and Health - Taft Laboratories
    Cincinnati, OH

    Dr. Werner Stdber
    Chemical Industries Institute of Toxicology
    Research Triangle Park, NC

    Dr. Jaroslav Vostal
    EHAC International
    Bloomfield Hills,  MI

    Dr. James Walker
    National Center for Environmental
      Assessment
    U.S. Environmental Protection Agency
    Washington, DC
2/1/98
xx
DRAFT-DO NOT CITE OR QUOTE

-------
                              ACKNOWLEDGMENTS
 Word Processing Support

 Ms. Debra Butler
' The CDM Group, Inc.
 Chevy Chase, MD

 Ms. Gienda Johnson
 Biomedical and Environmental Information
  Analysis
 Health Sciences Research Division
 Oak Ridge National Laboratory
 Oak Ridge, TN

 Document Production

 Ms. Theresa Konoza
 National Center for Environmental
  Assessment
 U.S. Environmental Protection Agency
 Washington, DC

 Ms. Kay Marshall
 Mr. Eric Sorensen
 The CDM Group, Inc.
 Chevy Chase, MD

 Ms. Marianne Barrier
 Mr. John Barton
 Ms. Sheila Lassiter
 Ms. Wendy Lloyd
 Ms. Edie Smith
 ManTech Environmental Technology, Inc.
 Research Triangle Park, NC
     References

     Mr. Douglas Fennell
     Environmental Criteria and Assessment
      Office
     U.S. Environmental Protection Agency
     Research Triangle Park, NC

     Ms. Catherine Carter
     Ms. Blythe Hatcher
     Information Organizers, Inc.
     Research Triangle Park, NC

     Reprographics

     Mr. Richard Wilson
     Environmental Criteria and Assessment
      Office
     U.S. Environmental Protection Agency
     Research Triangle Park, NC
 2/1/98
xxi
DRAFT-DO NOT CITE OR QUOTE

-------
                                         1. INTRODUCTION

  1            Diesel engines are an important part of transportation and industry throughout the
  2      industrialized world, and their use may expand because of increased pressures for fuel efficiency.
  3      For example, in addition to their widespread use in heavy-duty trucks and in many nonroad
  4      applications (such as construction and agriculture), there is increased use of diesel engines in
  5      light/medium duty trucks. Also, new technologies, such as hybrid vehicles powered by advanced
  6      diesel engines, and the potential for increased use of light-duty diesel engines in the sport utility
  7      market could fuel this expansion. In both the United States and Europe (where diesel passenger
  8      vehicles are much more common than in the United States), decisions are being made about how
  9      to regulate diesel engine and diesel fuel emissions; these decisions need to be based on sound
 10      scientific foundations.
 11            EPA's Office of Mobile Sources (OMS) has traditionally regulated diesel engine
 12      emissions through Clean Air Act (CAA) provisions, based on diesel engines' contribution to
 13      nonattainment of the PM-10 and ozone National Ambient Air Quality Standards (NAAQS).  In
 14      particular, diesel engines are a significant source of particulate and nitrogen oxide (NOX)
 15      emissions.  Since the passage of the first CAA in 1977, great strides have been made in
 16      controlling particulate matter and NOX emissions (including NO and NO2, both of which
 17      contribute to ozone formation) from highway and nonroad diesel engines. Heavy-duty urban bus
. 18      diesel engine particulate certification standards decreased by approximately 90% between 1988
 19      and 1996, and heavy-duty highway diesel engine particulate certification standards decreased by
 20      approximately 80% in the same period. NOX emissions standards have also been reduced,
 21      resulting in significant control of these emissions over the years. Manufacturers have reduced
 22      particulate NOX emissions by a variety of control technologies, such as altered combustion
 23      chamber design, better fuel injection technology such as high-pressure injection, and oxidation-
 24      type catalysts in some cases. EPA has also implemented initial regulations controlling emissions
 25      (NOX) from nonroad diesel engines. In addition, EPA has regulated diesel fuel quality by
 26      limiting the sulfur content of diesel fuel, thereby reducing particulate emissions.
 27            New authority granted to OMS under the CAA Amendments of 1990 allows for further
 28      regulation of nonroad diesel engines.  Particulate and NOX emission standards have been or will
 29      be established for diesel engines used to power nonroad equipment such as construction
 30      equipment, marine vessels, and locomotives.
 31            As mentioned previously, EPA has taken action to control diesel engine particulate matter
 32      emissions in order to facilitate attainment with the PM-10 NAAQS as a direct result of the
 33      adoption of a new NAAQS for particulate matter less than 2.5 um in diameter. On the basis of

         2/1/98                                     1-1        DRAFT-DO NOT CITE OR QUOTE

-------
 1      current health research, it is this fraction of the PM that is implicated in both the cancer and
 2      noncancer effects observed from diesel participate emissions exposure. PM 2.5 data indicate that
 3      diesel engines contribute significantly to the national particulate emission inventories.  Further
 4'      -regulation of diesel engine particulate emissions will require sound scientific evidence from the
 5      NAAQS effort as well as continuing research related to particulate effects on public health.
 6             EPA has never prepared a comprehensive health assessment document focused on the
 7      possible health hazards from direct exposure to diesel exhaust emissions.  Piecemeal health
 8      assessment work was done at various times in the 1980s, as was EPA-sponsored health research,
 9      with most of the attention focused on the possible carcinogenicity of DE exposure. The purpose
10      of this assessment is to capture the key data relating to toxicity and to assess that body of
11      evidence in order to produce a risk characterization describing the human health hazards of DE
12      exposure. Companion comprehensive characterizations of human exposure, diesel engine
13      emission factors, postemission DE pollutant transformation, and transport in the ambient air are
14      not included in the assessment, though some overview information is provided so that the health
15      assessment findings can be placed in a context with related issues.
16             In Chapters 2 through 11 of this assessment, key human and animal toxicity topics, as
17      well as overview topics, are reviewed and conclusions drawn as appropriate.  Chapter 12
18      integrates the various findings about the potential for health hazards and risk and additionally
19      provides a risk characterization.
20             The health assessment findings from this assessment will be advisory to EPA's Office of
21      Mobile Sources by providing health impact information that can be used to characterize residual
22      health hazards and risk.
        2/1/98                                    1-2        DRAFT-DO NOT CITE OR QUOTE

-------
                  2.  DIESEL EMISSIONS, TRANSPORT, AND TRANSFORMATION

  1      2.1.  INTRODUCTION
  2'            This chapter reviews a number of background topics that are valuable for perspective and
  3      context as one considers the health data (in later chapters) for diesel exhaust exposure. The
  4      contents of this chapter are not intended to necessarily reflect the most current literature, and thus
  5      should not be considered comprehensive and authoritative in that respect.
  6             The diesel engine was patented in 1892 by Rudolf Diesel, who conceived it as a prime
  7      mover that would provide much improved fuel efficiency compared with spark-ignition engines.
  8      To the present day, the diesel engine's high efficiency remains its strongest selling point. In the
  9      United States, the diesel engine is used mainly in trucks, buses, agricultural and other off-
10      highway equipment, locomotives, ships, and many stationary applications.
11             The chief advantages of the diesel engine over the gasoline engine are its fuel economy
12      and durability. Diesel engines, however, emit more oxides of nitrogen (NOJ and carbonaceous
13.    particulate matter than do gasoline engines.  Over the past decade, modifications of diesel engine
14      components have substantially reduced gaseous and particle emissions (Hammerle et al., 1994).
15             The diesel engine compresses air to high pressure and temperature. Fuel, when injected
16      into this compressed air, autoignites, releasing its chemical energy,  and the resulting combustion
17      gases expand, doing work on the piston,, before being exhausted to the atmosphere. Power output
18      is controlled by the amount of injected fuel rather than by throttling the air intake. Compared to
19      its spark-ignited (SI) counterpart, the diesel engine's superior efficiency derives from a higher
20      compression ratio and no part-load throttling.  Because of its poorer air utilization, a diesel
21      engine requires a larger piston displacement for the same power output as a comparable SI
22      engine. To ensure structural integrity for prolonged reliable operation at the higher peak
23      pressures brought about by a higher compression ratio and autoignition, the structure of a diesel
24      engine generally  is more massive than its SI counterpart.   ,
25             Diesel engines may be broadly identified as being either two- or four-stroke cycle,
26      injected directly or indirectly, and naturally aspirated or supercharged. They also are classified
27      according to service requirements such as light-duty (LD) or heavy-duty (HD) automotive, small
28  •    or large industrial, and rail or marine engines.
29             All diesel engines use hydraulic fuel injection in one form or another. The fuel system
30      must meet four main objectives if a diesel engine is to function properly over its entire operating
31      range:  (1) meter the correct quantity of fuel, (2) distribute the metered fuel to the correct
32      cylinder, (3) inject the metered fuel at the correct time, and (4) inject the fuel so that it is
33      atomized and  mixes well with the in-cylinder air. The first two objectives are functions of a
34      well-designed injection pump, and the last two are mostly functions of the injection nozzle.  As a
        2/1/98                      .               2-1        DRAFT-DO NOT CITE OR QUOTE

-------
 1      part of the effort to obtain lower exhaust emissions without diminishing fuel efficiency, fuel
 2      injection systems are moving toward the use of electronics for more flexible control than is
 3   •   available with purely mechanical systems,
 4            Both the fuel and the lubricants that are used to service diesel engines are highly finished
 5     • petroleum-based products combined with chemical additives.  Diesel fuel oil is a mixture of
 6      many different hydrocarbon molecules from about C? to about C35, with a boiling range from
 7      roughly  350°F to 650 °F.  Many of the fuel oil properties, such as its specific energy content,
 8      ignition  quality, and specific gravity, are related to its hydrocarbon composition. Therefore, fuel
 9      and lubricant composition affects many aspects of engine performance, including economy and
10      exhaust  emissions. For example, a decrease of fuel aromatic content, sulfur, and volatility
11      usually leads to a reduction of regulated emissions (Ullman, 1989). The four stages of the
12      combustion process in the diesel engine are:
13
14            1.   Ignition delay period:  the elapsed time from the start of injection until the start of
15                 combustion.  This is the time required to atomize fuel, evaporate droplets, and mix
16                 vapor with air and  for the necessary preflame reactions to occur. The ignition delay
17                 period is really two inseparable overlapping delay periods, physical delay and
18                 chemical delay.
19
20            2.   Uncontrolled burning period:  during this stage the fuel that has passed entirely
21                 through the first stage autoignites and burns in a premixed fashion and then
22                 diffusion takes over control of the burning (Lyn, 1963; Kahn, 1970).  During this
23                 stage a high rate of pressure rise and noise associated with diesel knock occur.
24
25            3.   Controlled burning period:  during this stage the fuel burns as it is injected in what
26                 is essentially a diffusion-controlled process.  The burning rate and the rate at which
27                 energy is released in this stage are lower than during the second stage.
28
29            4.   Afterburning period: the elapsed time from the end of fuel injection until the end  of
30                 combustion.  This stage is characterized by the diffusion mode giving way to the
31                 premixed mode of combustion.
32
33            Diesel emissions are derived from the complete and incomplete combustion of fuel and
34      lubricating oil; they are a mixture of gases and low molecular weight (MW) carbon particles.
35      High MW organic compounds are adsorbed on the particles.  Table 2-1 lists the major diesel
36      combustion emissions and their atmospheric reaction products.
37            Combustion of fuel in the diesel engine results in the formation of a complex mixture of
38      gaseous and particulate exhaust. Because of concerns over possible health effects  associated
39      with diesel particulate emissions, measurements have been made to characterize chemically in
40      detail the exhausts from light-duty diesel (LDD) and, to a lesser extent, heavy-duty diesel (HDD)


        2/1/98                                     2-2        DRAFT-DO NOT CITE OR QUOTE

-------
        Table 2-1,  Major components of diesel engine emissions and their known
        atmospheric transformation products
Emission component
Atmospheric reaction products
A. Vapor-phase emissions*
Carbon dioxide
Carbon monoxide
Oxides of nitrogen
Sulfur dioxide
Hydrocarbons
Aikanes (sC,8)
Alkenes (sC4)
(e.g., 1 ,3-butadiene)
Aldehydes
Formaldehyde
Higher aldehydes
(e.g., acrolein)
Monocyclic aromatic compounds
(e.g., benzene, toluene)
PAHs (s4 rings)6
(e.g., phenanthrene, fluoranthene)
Nitro-PAHs (2 and 3 rings) .
(e.g., nitronaphthalenes)
—
—
Nitric acid, ozone
Sulfuric acid
Aldehydes, alkyl nitrates, ketones
Aldehydes, ketones
Carbon monoxide, hydroperoxyl radicals
Peroxyacyl nitrates
Hydroxylated and hydroxylated-nitro derivatives6
Nitro-PAHs (s4 rings)1*
Quinones and hydroxylated-nitro derivatives
B. Particle-phase emissions
Elemental carbon
Inorganic sulfate
Hydrocarbons (C14-C35)
PAHs (*4 rings)
(e.g., pyrene, benzo[a]pyrene)
Nitro-PAHs (a 3 rings)
(e.g., nitropyrenes)
—
—
Little information; possibly aldehydes, ketones, and
alkyl nitrates
Nitro-PAHs (s4 rings)4
Nitro-PAH lactones
Hydroxylated-nitro derivatives
"Unless otherwise stated, the impact results from both the emissions components and the atmospheric reaction products.
'Some reaction products expected to partition into the particle phase.
"PAHs containing four rings are usually present in both the vapor and particle phases.
^Nitro-PAHs with more than two rings' will partition into the particle phase.

Source: HEI, 1995.
2/1/98
2-3
DRAFT-DO NOT CITE OR QUOTE

-------
 1      engines. Most of these measurements are of primary pollutants, that is, gases and particulate
 2      matter emitted directly into the air from their sources,
 3            The primary pollutants are subject to dispersion and transport and, at the same time, to
 4      chemical and physical transformations into secondary pollutants; the time scales of these
 5      atmospheric transformations and physical loss processes vary widely .^ Atmospheric lifetimes
 6      range from <1 min for some highly reactive organic compounds to months for other much more
 7      inert constituents of direct emissions.  Thus, to assess the environmental effects of diesel
 8      emissions, it is necessary to determine the chemical and physical changes that primary diesel
 9      emissions undergo during their transport through the atmosphere.
10            Combustion of fuel in the diesel engine results in the formation of a complex mixture of
11      gaseous and particulate exhaust. Because of concerns over possible health effects associated
12      with diesel particulate emissions, measurements have been made to characterize chemically in
13      detail the exhausts from light-duty diesel (LDD) and, to a lesser extent, heavy-duty diesel (HDD)
14      engines. Most of these measurements are of primary pollutants, that is, gases and particulate
15      matter emitted directly into the air from their sources.
16            The primary pollutants are subject to dispersion and transport and, at the same time, to
17      chemical and physical transformations into secondary pollutants; the time scales of these
18      atmospheric transformations and physical loss processes vary widely.  Atmospheric lifetimes
19      range from <1 min for some highly reactive organic compounds to months for other much more
20      inert constituents of direct emissions.  Thus, to assess all of the environmental effects of diesel
21      emissions, it is necessary to determine the chemical and physical changes that primary diesel
22      emissions undergo during their transport through the atmosphere.
23
24      2.2.  OVERVIEW OF DIESEL POLLUTANTS AND POLLUTANT FORMATION
25      2.2.1. Gas-Phase Pollutant Emissions
26      2.2.1.1.  Oxides-of-Nitrogen Formation
27            Because the diesel combustion process is very complex and involves burning of fuel
28      droplets, it has proven difficult to predict pollutant concentrations or emission rates
29      quantitatively. In SI gasoline engines, NO emission can be quantitatively explained by adiabatic
30      compression of the initially burned mixture  (nearest the spark plug) by the  combustion pressure
31      developed in the later stages of combustion. Thus, the originally burned gases are raised to a
32      much higher temperature than that achieved in the flame by the subsequent compression.  Shahed
33      (1985) has reviewed the work on this subject for diesel engines. He reports that, qualitatively,
34      NO formation in diesel engines cannot be explained by this phenomenon. The time-temperature
35      history of the burning droplets seems to determine the extent of NO formation.  Yu and Shahed
36  ,    (1981) have defined relevant engine operating parameters that control NO emissions for heavy-
        2/1/98                                    2-4       DRAFT-DO NOT CITE OR QUOTE

-------
  1      duty diesel engines.  For instance, retarding injection or recirculating exhaust gases reduces NO
  2      formation and emission at the expense of increasing soot formation and hydrocarbons, all other .
  3      factors being equivalent. Wu and Peterson (1986) studied NO formation kinetics in an IDI
  4      passenger-car diesel engine over a wide range of operating conditions. These authors found that
  5      a variable-temperature model accounting for the average gas temperature at the time of droplet
  6      burning explained the observed NO considerably better than a constant-temperature (peak-cycle-
  7      temperature) model. Global NO formation rates in their study suggested that the NO must be
  8      formed in the vicinity of the droplet flame zone.
  9             Lipkea et al. (1987) and Lipkea and DeJoode (1987) have constructed a successful engine
10      model that adequately explains NO formation.  Process parameters that control NO formation
11      include fuel jet momentum flux, in-cyUnder air density at the start of injection, swirl cross-flow
12      momentum flux, and in-cylinder temperature. Air system design characteristics that change air
13      density and temperature can result in constant work but decreasing NOX.  On the other hand,
14      decreasing cylinder temperatures also result in increased hydrocarbon emissions (Uyehara, 1987;
15      Gill, 1988). The net result is a tradeoff between NOX and paniculate material at high
16      temperatures and between NOX and hydrocarbons at low operating temperatures.
17
18      2.2.1.2. Hydrocarbons and Carbon Monoxide Formation
19        ,     Small quantities of gaseous unburned hydrocarbons (HCs) and carbon monoxide (CO) are
20      emitted from diesel engines, but less than those emitted by comparable SI engines. Myers and
21      Uyehara (1947) have explained the observed CO on the basis of locally rich combustion. During
22      the ignition delay period, especially small amounts of fuel vaporize from the initial droplets. The
23      gas-phase reactions of this material are responsible ultimately for its  ignition and, thus, the
24      ignition of the droplets. However, this premixed patch is likely to be locally quite rich, even
25      though the overall fuel-air mixture has  considerable excess air. Therefore, CO is formed in
26      concentrations of 2,000 ppm or even slightly more in diesel exhaust.  By comparison, typical
27      gasoline engines might have exhaust CO concentrations of 10,000 to 20,000 ppm.
28            These locally rich combustion processes are also responsible  for a small release of low
29      molecular weigh! hydrocarbons, principally methane, ethylene, and acetylene, in diesel exhaust.
30      About 10% of these materials are C, to C4 combustion-derived compounds. The bulk of the
31      emission in the gas phase is diesel fuel  in the C10 to C25 molecular weight range; these materials
32      account for 70% to 80% of the HCs emitted. The balance of the material, including partiele-
33      bound HC, is in the same molecular weight range as lubricating oil.  Similar findings have been
34      reported with HD diesel engines, both of the two-stroke and four-stroke cycle types.  Therefore,
35      both fuel and lubricant can supply organic matter to diesel engine exhaust HCs in the gaseous
36      and paniculate states.
        2/1/98                                    2-5        DRAFT-DO NOT CITE OR QUOTE

-------
 1            Hampton et al, (1983) reported a gas chromatography-mass spectrometry (GC-MS) study
 2      of heavy HCs in a Pennsylvania Turnpike roadway tunnel that has varying amounts of diesel and
 3      gasoline passenger-car traffic. There were characteristic differences in gaseous HC content of the
 4      tunnel gases between the diesel and gasoline vehicles. Typically, diesel traffic was characterized
 5      by substantial quantities of aliphatic hydrocarbons with lesser amounts of long-chain substituted
 6      monoaromatics. Gasoline traffic was dominated by methyl- and ethyl-benzene emissions, based
 7      on the porous polymer trapped samples used in this study.  These results are in general agreement
 8      with the Black and High (1979) conclusion that diesel HC emissions are primarily fuel-derived.
 9
10      2.2.2.  Particle Formation and Emission
11            The chemical mechanism that accounts for carbon formation in diesel combustion is not
                                                                                           /
12      completely established; the major weight of scientific opinion seems to support some role for
13      intermediate formation of poly cyclic aromatic hydrocarbons in the process.  Diffusion flames,
14      whether rich or lean, usually form some carbon. However, carbon is consumed on the lean side
15      by reaction with hydroxyl (OH) radicals (Fenimore and Jones, 1967), and only when the OH
16      radical population is reduced by other reactions (with fuel hydrocarbons, for example) is carbon
17      found  to be a major combustion product.  Thus, carbon is normally a stable combustion product
18      only of rich flames.
19        "    Beck and Uyehara (1988) have pointed out that there is a very good linear correlation
20      between CO emission rate and carbon emission rate from HD engines. Once formed, both
21      substances are difficult to remove, requiring highly energetic OH radicals for reaction. These
22      authors argue that carbon formation normally takes place over a rather narrow temperature range
23      and  that the maximum rate of carbon formation is found at temperatures around 2,250 K, At
24      temperatures of 2,400 K and above, carbon is burned out, and at temperatures of 1,900 K and
25      below, it is never formed. However, if attempts are made to raise temperature and thus burn out
26      the carbon, high NO results.  If flame zone temperatures are lowered by adding water, alcohol, or
27      exhaust gas, HC emissions are increased (Ball, 1987; Kadota and Henein, 1981). Beck and
28      Uyehara have described a qualitative model of droplet combustion in which the burning rate of
29      fuel droplets is controlled by boiling rate. It is shown that an ideal.condition involves high-
30  •    pressure, high-velocity injection with a minimum ignition-delay period.  Carbon can be
31      minimized by reducing rich gaseous combustion and NO can be controlled by delayed-injection
32      timing.  Under these conditions, the fuel cetane number might become a critical parameter
33      controlling the ignition delay and the region of uncontrolled combustion. In practice, Ullman
34      (1989) has shown that reducing aromatic content and, hence, decreasing the ignition delay of
35      diesel  fuel is  beneficial in the control of paniculate material and of NOX emissions in three HD
36      engines built to meet the  1988 California or 1991 Federal emissions standards,  Tosaka et al.
        2/1/98                    .              2-6        DRAFT-DO NOT CITE OR QUOTE

-------
  1      (1989) have studied the effect of foe! aromatics in promoting diesel carbon formation.  These
  2      authors have found an aliphatic radical-benzene condensation process that apparently accounts
  3      for the additional amounts of carbon that result in the diesel combustion of aromatic fuels,
  4
  5      2.2.3.  Gas-to-Particle Conversion
  6      2.2.3.1. Condensation of Organic Matter
  7             Generally, the formation of carbon particles is thought to involve growth of particles by
  8      polymerization of gaseous intermediates at the surface of small particles (Kadota and Henien,
  9      1981; Plee et al., 1981). Thus, the growth and agglomeration of carbon particles are probably
10      gas-to-particle processes.  Ross etal. (1982) studied the properties of diesel particles obtained
11      from an engine operated on high-purity dodecane. In this case, the fuel was too volatile to have
12      an impact on particle composition.  These authors found that the carbon particles contained an
13      HC film, which must have been condensed from the gas phase.  The carbon had rather low
14      specific surface area, about 0.5 m2/g, which could be materially increased by high-temperature
15      treatment.
16             Heats of sorption of a variety of HCs were determined by a GC technique. It was found
17      that the absorptivity of gaseous organic compounds on these particles was controlled by Henry's
18      law of absorption in the organic surface film. The presorbed organic layer was several layers
19      thick, and this material essentially acted as an organic droplet, dissolving materials from the gas
20      phase. Thus, the heat of sorption was adequately explained by the heat of vaporization of the
21      organics (Ross etal., 1982).
22         ..
23      2.2.3.2. Oxidation of Sulfur Oxides
24             Studies of diesel particle composition have produced some information about the fate of
25      fuel sulfur.  In the earliest studies, sulfate was found to be a significant component of diesel
26      particles (Hare et al., 1976).  Generally, the sulfate found in particles accounted for only about
27      2% of the fuel sulfur charged, the balance  being emitted as  sulfur dioxide (SO2).
28             Sulfate emission rates have been measured using both engine and chassis dynamometer
29      test procedures.  Hare et al. (1976) measured composite particle emission rates on the older
30      Federal compliance test for HD engines and found that with overall emission rates from 0.3 to
31      1.0 g/kW h (0.4 to 1.3 g/bhp h) of particle mass, sulfate emissions were relatively constant at
32      about 0.02 g/kW h (0.03 g/bhp h). Dietzmann et al. (1980) measured sulfate emission rates from
33      a number of HD vehicles driven over simulated urban driving schedules and found emission rates
34      from about 0.03 to 0.05 g/km for overall particle emission rates from 0.5 to 1.6 g/km. Thus, with
35      previous engines, sulfate was a significant but small (2% to 3%) component of diesel particle
36      mass. With newer engine designs and particle emission rates characteristically below 0.25 g/bhp
        2/1/98                                   2-7        DRAFT-DO NOT CITE OR QUOTE

-------
 1      h, this emission rate amounts to 10% to 15% of the emitted particle mass, a very important
 2      portion of the allowable limit. Currently, no means of reducing this sulfate formation is available
 3"     other than reducing the sulfur concentration of diesei fuel.
 •4            Earlier in this document, the evidence regarding the oxidation of nitrogen to nitric oxide
 5      was documented.  Dietzmann et al. (1980) have reported analysis of HD diesei vehicle exhaust
 6      for N02 during simulated urban driving. These authors have reported NO2 concentrations from
 7      about 1 to 5 ppm, accounting for 2% to 5% of the NO emitted. Harris et al. (1987) have
 8      measured NO2 and HNO3 in air-diluted diesei exhaust. In the exhaust from HD engines, the NO2
 9      concentration was from 1 to 30 ppm, whereas the HNO3 ranged from.0.08 to 0.8 ppm.
10      Generally, HNO3 accounted for a few percentage points of the NO2, which in turn accounted for a
11      few percentage points of the NO.
12
13      2.2.4. Nitroarene Formation
14            The soluble extract from diesel-generated particulate material was shown to cause
1.5      mutations when subjected to a bacterial assay (Huisingh et al., 1978).  It was not long before
16      evidence began to point toward the class of compounds consisting of nitrated polycyclic aromatic
17      hydrocarbons (PAHs) that have come to be called nitroarenes (Pederson and Siak, 1981; Newton
18      et at, 1982). Hare and Bradow (1979) reported a summary of the EPA findings to that date on
19      the formation of semipolar mutagens in both HD and LD diesei engines.  Dietzmann et al. (1980,
20      1981) reported emissions of mutagenic material from a variety of HD trucks operated over
21      transient driving cycles, and Gibbs et al. (1980) reported the emission of bacterial mutagens from
22      a large number of in-use diesei passenger cars.
23            Very-high-resolution organic analytical procedures have been applied to diesei exhaust
24      samples (Liberti et al., 1984; Schuetde and Frazier, 1986; Schuetzle and Perez, 1983).
25      Generally, a variety of nitrated polycyclic aromatic compounds has been found, which accounts
26      for a substantial portion of the mutagenicity found. However, not all the bacterial mutagenicity
27      has been identified in this way, and the identity of the remainder of the mutagenic compounds
28      remains unknown. The nitrated aromatics thus far identified in diesei exhaust were the subject of
29      review in the International Agency for Research on Cancer (1989) monograph on diesei exhaust
30      (Table 2-2).                      '                                              "     .
31
32      2.3. PRIMARY  DIESEL EMISSIONS
33            Detailed chemical characterization of diesei engine emissions was performed mostly in
34      the late 1970s  and early 1980s. Since that time substantial changes have occurred in engine and
35      emission control technologies, as well  as in chemical analysis methodology. It is likely that
36      emissions from currently manufactured diesei vehicles are not the same as those measured and
        2/1/98                                    2-8        DRAFT-DO NOT CITE OR QUOTE

-------
        Table 2-2.  Some nitroarenes identified in vehicle exhaust3
  1,3-Dihydroxynitropyrene-l
  2.5-Dinitrofluorene
  2,7-Dinitrofluorene
  2,7- D in itro-9-fluorenone
  1,3-Dinitropyrene
  1,6-Dinitropyrene
  1,8-Dinitropyrene
         9-Methylcarbazole
         1-Nitro-3-acetoxypyrene
        . 9-Nitroanthracene
         2-Nitroanthracene or -phenanthrene
         x-Nitroanthracene or -phenanthrene (two isomers)b
         6-Nitrobenzo[a]pyrene
         x-Nitrobenzoquino lineb
         2-Nitrobiphenyl
         3-Nitrobiphenyl
         4-Nitrobiphenyl
         1-Nitrochrysene
         x-Nitrodibenzothiophene (two isomers)b
         x-Nitro-y,z-dimethylanthraeene or -phenanthrene (five isomers)b'
         I-Nitrofluoranthene
         3-Nitrofluoranthene
         7-Nitrofluoranthene
         8-Nitrofluoranthene
         2-Nitrofluorene
         3-Nitro-9-fluorenone
  10-Nitro-l-methylanthracene  .of -phenanthrene
  lO-Nitro-9-mtethylanthracene  or -phenanthrene
         x-Nitro-y-methylanthracene or -phenanthreneb
         1 -Nitro-2-methy Inaphthalene
         3-Nitro-1 -methylpyrene
         6-Nitro-l-methylpyrene
         8-Nitro-1 -methy Ipyrene
         1 -Nitronaphthalene
         2-Nitronaphthalene
         2-N itrophenanthrene
         1-Nitropyrene
         5-Nitroquinoline
         8-Nitroquinoline
         x-Nitroterphenylb
         x-Nitro-y,z,z'-trimethylanthracene or -phenanthrene (six Isomers)b
 	 x-Nitrotrimethybiaphthalene (three isomers)b	
"From International Agency for Research on Cancer, 1989,
bx, y, z, and z' imply position is unknown.
2/1/98                             .              2-9         DRAFT-DO NOT CITE OR QUOTE

-------
 1      reported earlier. When possible, the latest data were used; however, the data presented in this
 2      chapter should not be considered fully representative of emissions from the wide range of diesel
 3      engines currently used or those that may have occurred in the past.
 4
 5      2.3.1.  Gaseous Emissions
 6             Diesel passenger-car and light-truck emissions of CO and total gaseous hydrocarbons
 7      (THC) are considerably lower than those of gasoline vehicles. For HDD vehicles, the CO and
 8      THC emission rates are somewhat lower than, but comparable with, those of HD gasoline
 9      vehicles, but NOX emissions are many times those of average traffic. .In addition to these
10      regulated pollutants, diesel exhausts also contain some sulfur dioxide because of the presence of
11      sulfur in the diesel fuel. Following combustion, approximately 98% of the sulfur is emitted as
12      S02 and 2% as particulate sulfate (Pierson et al., 1978,1979; Truex et al., 1980), Most of the
13      sulfate is in the form of sulfuric acid (H2SO4) (Truex et al., 1980).
14             The atmospheric concentration of nitric acid (HNO3) from LDD exhausts has been
15      reported to be negligible in comparison with that from other anthropogenic sources (Okamoto et
16      al., 1983; Harris et al., 1987). A range of concentrations from =100 ppbv (=250 ug/m3) to =800
17      ppbv (=2 mg/m3) (Harris et al., 1987) and an emission rate of = 1.3 mg/km were reported
18      (Okamoto et al., 1983).
19             A small amount of ammonia was also detected in diesel engine exhausts (Pierson and
20      Brachaczek, 1983 a). The highest value for NH3 (25 mg/km) was reported for HDDs; =4 mg/km
21      was reported for LDD, and ~ 10 mg/km and =5 mg/km for gasoline-powered vehicles, with and
22      without catalyst, respectively (Pierson and Brachaczek, 1983a).  The emission rates of total
23      aliphatic amines were reported to be below the detection limit of 0.04 to 0.3 mg/km for gasoline-
24      powered vehicles and 0.08 to 0.7 mg/km for heavy-duty trucks.  It was concluded that motor
25      vehicles are an insignificant source of atmospheric NH3 and that amines emitted from motor
26      vehicles cannot give rise to carcinogenic nitrosoamines in the amount said to exist in ambient
27     samples (Pierson and Brachaczek, 1983a).
28             In addition, low concentrations of phenols have been reported in HDD and LDD
29      emissions (Hare and Baines, 1979; Hare and Bradow,T979). Aliphatic carboxylic acids (mainly
30     formic, acetic, propionic, and benzoic acids) were also reported in vehicle exhausts (Kawamura
31      etal., 1985; Roggeetal., 1993).
32            Table 2-3 compares the emission rates of some representative alkanes, alkenes, aromatic
33     hydrocarbons, and aldehydes from HDD and LDD engines and gasoline engines with and without
34     a catalytic converter.  Data on catalyst-equipped gasoline vehicles were obtained with a chassis
35     dynamometer and are averaged from 46 in-use passenger cars, 1975 to 1982 models, selected to
36     be representative of vehicles actually driven by the U.S. public (Sigsby et al., 1987).  Table 2-3
       2/1/98                                   2-10       DRAFT-DO NOT CITE OR QUOTE

-------
        Table 2-3,  Emission rates of volatile organic compounds (VOC) from
        cliesel and gasoline engines


VOC fe/mi
Diesel
VOC
THC
Methane
Ethylene
Acetylene
Propylene
n-Pentane
ijo-Pentane
«-Decane
«-Dodecane
Benzene
Toluene
Xylenes
Ethyl benzene
Naphthalene
Formaldehyde
Acetaldehyde
Acrolein
Benzaldehyde
Total aldehyde
HDD
3.65 (2.28)d
NA
NA
NA
NA
NA
NA
0.01(0.007)'
0.027 (0.0 17)f
0.024 (0.015)d
0.01 (0.007)f
0.006 (0.004)d
0.005 (0.003)f .
0.0 1 (0.007/
NA-
NA
0.053 (0.033)"
NA
' NA
LDD«
0.23(0.14)'
0.01 (0.008)
0.04 (0.03)
' NA
0.01 (0.008)
NA
. NA
NA
NA
0.02(0.015)'
0.006 (0.004)=
0.002(0.001)'
0.001 (0.0006)'
0.003 (0.002)e
0.02(0.01)
0.007 (0.004)*
0.01 (0.006)
NA
0.03(O.Q2)e
fe/kml)

Gasoline
Catalyst"
1.8(1.2)
0.26(0.16)
0.14(0,09)
0.04 (0.02)
0.04(0.03)
0.03 (0.02)
0.07(0.04)
0.003
0.003
0.06 (0.04)
0.1(0.07)
0.08 (0.05)
0.02(0.01)
NA
0.025(0.015)
0.01(0.007)
0.002(0.001)
0.003 (0.002)
0.04 (0.03)
Noncatalysf
5.4 (3.4)
0.27(0.17)
0.3 (0.2)
0.26(0.16)
0.15(0.09)
0.09 (0.06)
0.27(0,17)
(0.0016)*
(0.002)h
0.31(0.19)
0.7 (0.45)
0.96 (0.6)
0.21 (0.13)
NA
0,06 (0.04)'
NA
NA '
NA
NA
HDD = Heavy-duty diesel.
LDD = Light-duty diesel.
NA = Data not available.

"From National Research Council (1982), except as indicated.
'From Sigsby et al. (1987).
'From Bailey et al. (1990), except as indicated.
••From Westerhoim et al. {1991).
'From Smith (1989) and Smith and Paskind (1989); four-cycle FTP test, 1986 Mercedes Benz.
'From Hampton et al. (1983), data from Allegheny Mountain Tunnel. .
"From Hampton ct al. (1983), no differentiation between vehicles with and without catalyst.
''From Hampton et al. (1983).
'From Schuetzle and Frazier{ 1986),
2/1/98
2-11
DRAFT-DO NOT CITE OR QUOTE

-------
 1.     shows the data from the Federal Test Procedure (FTP) only (which attempts to simulate a typical
 2      urban driving pattern with average speed of =20 mph), although two other driving cycles
 3   •'  (Crowded Urban Expressway and the New York City cycle) are reported in the original
 4      publication.
 5     .       The emission data on noncatalyst gasoline vehicles shown in Table 2-3 are averaged from
 6      25 in-use passenger cars, representing late 1980s vehicles in intensive use in the United Kingdom
 7      (U.K.) (Bailey et al.,  1990). The vehicles were driven "as received" and fueled by leaded
 8      premium-grade gasoline, obtained locally from a single source.  They were driven on five routes
 9      chosen to cover the normal range of U.K. driving speeds and conditions, and the exhaust samples
10      were taken using a miniaturized constant-volume exhaust gas sampler.  To allow direct
11      comparison, the urban roadway data (=13,5 mph average speed) are given in Table 2-3.
12      However, it has to be pointed out that hydrocarbon emission rates are highly dependent on
13      driving speeds; in general, THC emission rate expressed in grams per traveled distance decreases
14      as the driving speed increases, but the individual hydrocarbons display various patterns, which
15      relate to their origin.  The gasoline components (hydrocarbons with carbon number C a 4) are
16      present in highest proportion at low speeds, whereas at higher speeds these components are more
17      efficiently used and the proportion of combustion-derived products increases.
18            The data on LDD emissions are from the National Research Council report (1982) and
19      from studies (Smith,  1989; Smith and Paskind, 1989) concerning the evaluation of particle trap
20      efficiencies for two diesel passenger cars (a 1986 Mercedes Benz and a Volkswagen prototype
21      Jetta). However, little has been published recently; most of the available data are from the late
22      1970s and early  1980s,                                                               .
23            Quantitative data on emissions from heavy-duty vehicles are relatively sparse and are
24      generally expressed in terms of grams per unit work performed by HDD engines.  Most of the
25      data on HDD vehicle emission rates expressed in terms of grams per traveled distance  were
26      obtained from tunnel field experiments, in particular from the Allegheny and Tuscarora
27      Mountain Tunnel experiments (Hampton et al., 1982,1983); Unfortunately, because of the
28      sampling method selected, no quantitative data on hydrocarbons with carbon number C < 8 could
29      be obtained.
30            Recently, exhaust emissions from a HDD truck (Scania 143H, equipped with a turbo-
31      charged Scania DSC  1403 diesel engine) were characterized chemically and tested for
32      mutagenicity during transient driving conditions on a chassis dynamometer (Westerholm et al.,
33      1991). Table 2-3 gives selected data from this study.
34            As can be seen from Table 2-3, the emissions of gaseous organic compounds from diesel
35      engines and spark-ignition engines are qualitatively similar (e.g., similar chemical components
36      are present in both exhausts), although there are significant quantitative differences. In theory,
        2/1/98                                    2-12        DRAFT-DO NOT CITE OR QUOTE

-------
  1      new spark-ignition vehicles equipped with catalytic converters emit almost no reactive
  2      hydrocarbons in their exhausts. However, catalyst deterioration over the lifetime of the vehicle
  3      and evaporative and refueling emissions will result in an increase in the amount of reactive
  4      material released.
  5      •       Table 2-3 lists the emission rates for exhaust pipe emissions only.  Currently, fuel
  6      evaporation (e.g., from fuel lines and carburetors) accounts for 30% to 60% of the total
  7      hydrocarbon emissions from passenger gasoline vehicles with and without catalytic converters
  8      (International Agency for Research on Cancer, 1989). Under ambient conditions, the vapor
  9      pressure of most diesel fuels in current use is so low that emissions resulting from evaporation
10      are not significant.  However, for both diesel and gasoline engine emissions, methane, ethane,
11      ethylene, acetylene, propane, and propylene originate strictly from tailpipe emissions, as
12      evidenced by the fact that the lowest molecular weight components of gasoline are normally
13      hydrocarbons with  carbon number C4.
14             Differences in the quantity of emitted material between vehicles of the same category are
15      very significant, and such differences arise from many factors of engine design, fuel control,
16      engine conditions, and the general condition of the vehicle at the time of test. The differences
17      between test parameters (e.g., speed, cold or hot start, fuel composition, dynamometer or road
18      measurements, etc.) make the comparison of the data given in Table 2-3 more uncertain.
19      However, it is clear from these data that the emission profile of gaseous organic compounds is
20      different for diesel and spark-ignition vehicles; the aromatic hydrocarbons and low molecular
21      weight alkanes (C10) are more characteristic of diesel emissions (Hampton et al, 1983; Carey
23      and Cohen, 1980).
24
25      2.3.2. Particulate  Emissions
26      2.3.2.1. Diesel Particulate Matter
27             Diesel exhaust particles are aggregates of spherical primary particles, and 75% to 95% of
28      the paniculate mass is in the accumulation mode centered about an aerodynamic diameter of 0.2
29      urn. Figure 2-1 shows a typical size distribution of diesel exhaust particles. The size distribution
30      is important, because transport of the particles in the atmosphere and deposition in the human
31      respiratory tract depend essentially on aerodynamic diameter (see Chapter 4 for discussion  of
32      deposition).
33             Diesel particulate matter is generally defined as any material that is collected on a
34      filtering medium at a temperature of 52 ° C or less after dilution of the raw exhaust.  In general,  ,
35      diesel engines produce more particulate emissions than do gasoline engines with or without
36      catalytic converters. The main constituent of diesel particles is carbon, which accounts for =80%
        2/1/98                                   2-13        DRAFT-DO NOT CITE OR QUOTE

-------
                                                    Accumulation Mode
                                                    75 to 95% of
                                                    Particulate Mass
                  Nuclei Mode
                  Oto 10% of
                  Particulate Mass
                                   Coarse Particle Mode
                                   5 to 15% of
                                   Particulate Mass
                   0.01
0.03       0.1        0.3         1.0
      Particle  Diameter ((am)
                 3.0
       Figure 2-1. Typical size distribution of diesel exhaust particles.

       Source: National Research Council (1982).
 1     of total particle mass. Approximately 70% of this so-called total carbon (TC) occurs in the form
 2     of elemental carbon (EC); the rest is in the form of organic compounds and is called organic
 3     carbon (OC). Table 2-4 compares the emission rates of particulate matter and its distribution
 4     between TC and OC for HDD, LDD, and gasoline engines. Data on HDD, LDD, and gasoline
 5     engines without catalytic converters are mostly from the survey of 13 HDD vehicles, 19 LDD
 6     vehicles, and 22 spark-ignition vehicles hi intensive use hi Sydney, Australia (Williams et al.,
 7     1989a,b).  The vehicles were tested on a dynamometer using several test procedures, but Table 2-
 8     4 lists the data from the ADR 37 cycle (for spark-ignition and LDD vehicles) and from the
 9  •   modified ADR 36 cycle (for HDD vehicles; multimode steady-state procedure). In addition, data
10     from a heavy-duty truck (Scania 143H), tested on a chassis dynamometer using transient driving
11     conditions (Westerholm et al., 1991), are also given hi Table 2-4.
12            Because the dynamometer studies are not fully representative of the road conditions, the
13     data obtained from the field experiments hi two highway tunnels, the Allegheny and Tuscarora
14     Mountain Tunnels of the Pennsylvania Turnpike (Pierson and Brachaczek, 1983b;  Szkarlat and
15     Japar, 1983), are also given hi Table 2-4 for comparison. Table 2-5 lists the particle-phase and
       2/1/98
                 2-14
DRAFT-DO NOT CITE OR QUOTE

-------
              Table 2-4. Partkulate matter emission rates and their distribution between
              total carbon (TC) and organic carbon (OC) for heavy-duty diesel (HDD),
              light-duty diesel (LDD), and gasoline engines
Emission
Paniculate matter in
g/mi (g/km)
TC (% w/w)
OC(%ofTC)
HDD
3.2 (2f
1.4(0.87)d
0,99 (0.62)e
78a
58a
30d
LDD
0.6 (0,37)a
80a
.36a
Noncatalyst
0.1 (0.07)b
0,04
3lb
87*
Catalyst
0.02(0,01)°
(0.025)d
NA
67f
NA
        aFrom Williams et al. (1989b).
        bFrom Williams et al. (1989a).
        cFrom Schuetzle and Frazier (1986).
        dFrom Pierson and Brachaczek(1983b).
        eFrom Westerholm et al. (1991).
        %rom Szkarlat and Japar (1983).
        NA = Not available.
 1      gaseous-phase emissions of diesel exhaust, along with their atmospheric reaction products
 2      (discussed in Section 2.4).
•3                  •
 4      2.3.2.2. Particulate-Phase Inorganics
 5            Organic and elemental carbon account for =80% of the total particulate matter mass.  The
 6      remaining 20% is composed of sulfate (mainly H2S04)  (Pierson and Brachaczek, 1983b) and
 7      some inorganic additives and adventitious components of fuel and motor oil. Table 2-6 gives the
 8      average compositions of inorganic constituents of airborne particulate matter associated with
 9      vehicles on the road (from Pierson and Brachaczek, 1983b). All airborne constituents of
10      particulate, matter associated with vehicle traffic (other than atmospheric transformation products
11      of primary emissions) are included, whether emitted from the exhaust or not (e.g., originated
12      from tire wear debris and soil dust).
13
14      2.3.2.3. Particulate-Phase Organic Compounds
                                                                                          n
15            Carbonaceous, diesel-emitted particles have high specifie.surface areas of 30 to 50 m /g
16      (Frey and Corn, 1967). Because of this high surface area, diesel particles are capable of
17      adsorbing relatively large quantities of organic material originating from unburned fuel and
18      lubricating oil and from pyrosynthesis occurring during combustion of fuel (see Section 2.1).

        2/1/98                                    2-15       DRAFT-DO NOT CITE OR QUOTE

-------
              Table 2-5. Particle phase and gaseous phase emissions from diesel exhaust,
              and their atmospheric reaction products
                      Emission component
            Atmospheric reaction products
        Particle-phase emissions
        Elemental carbon
        Inorganic sulfate
        Hydrocarbons (Cu-CK)

        PAHs (>4 rings) (e.g., pyrene, benzo[a]pyrene)
        Nitro-PAHs (s3 rings) (e.g., nitropyrenes)
        Gaseous-phase emissions
        Carbon dioxide
        Carbon monoxide
        Oxides of nitrogen
        Sulfur dioxide
        Hydrocarbons
         Alkanes (sC,g)
         Alkenes (sC4) (e.g., 1,3-butadiene)
        Aldehydes
         Formaldehyde
         Higher aldehydes (e.g., acrolein)
        Monocyclic aromatic compounds (e.g., benzene,
        toluene)                            .
        PAHs (s4 rings) (e.g., phenanthrene, fluoranthene)
        Nitro-PAHs (2 and 3 rings) (e.g., nitronaphthalenes)
   Little information; possibly aldehydes, ketones, and
   alkyl nitrates
   Nitro-PAHs (>4 rings); nitro-PAH lactones
   Hydroxylated-nitro derivatives
   Nitric acid, ozone
   Sulfuric acid

   Aldehydes, alkyl nitrates, ketones
   Aldehydes, ketones

   Carbon monoxide, hydroperoxyl radicals
   Peroxyacyl nitrates
   Hydroxylated and hydroxylated-nitro derivatives

   Nitro-PAHs (s4 rings)
   Quinones and hydroxylated-nitro derivatives
       Source: Health Effects Institute (1995).

1  .    After removed of extractable organic material, the surface area of diesel particles increases up to
2      90 m2/g (Pierson and Braehaczek, 1976).
3             The extractable fraction of diesel particles is typically in the range of 20% to 30%, but it
4      may be as high as 90% (Williams et at, 1989b) depending on vehicle type and operating
5      conditions.  In general, if a diesel engine is running under low load, the incomplete combustion
6      results in a relatively low particle concentration and a higher proportion of organic-associated
7      particles (Dutcher et ah, 1984).  In addition, recent progress in in-cylinder paniculate matter
       2/1/98
2-16
DRAFT-DO NOT CITE OR QUOTE

-------
               Table 2-6. Summary of composition and emission rates (in milligrams per
               kilometer) of airborne participate matter from on-road vehicles, Tuscarora
               Mountain Tunnel 1977 experiment
           Constituent
                 Gasoline*
         (mg/km [% of total mass])
                 Dieselb
        (mg/km [% of total mass])
        H
        B  ,
        C
        N
        Nac
        Mgc
        Alc
        Sid
        pe
        S[S04-2]
        Clf
        Kd
        Cac
        Tid
        Mnc'f
        Fec
        Cu
        Zne
        Brf
        Baf
        Pbf
 5 ±4(10 ±6)
 0.04 ±0.6(0.07 ±0.11)
 34 + 21 ($7 ± 42)   •
 1.1  ±0.8 (2 ±2)
 0.09 ± 0.37 (0,2 ± 0.7)
 0.7  ± 0,3 (1.3 ± 0.6)
 0.2  ± 0.5 (0.3 ± 0.9)
 0.5  ±0.7(1.0 ± 1.3)
 0.07 ±0.06(0.13 ±0.11)
 0.4[3,4 ± 0.9] (0.9[7 ± 3])
 0.8  ±0.4(1.6 ±0.8)
 0.17 ±0.08(0.3 ±0.2)
 1.3  ±0.3(2.5 ±0.7)
 0.006 ± 0.01 (0,01 ± 0.02)
 0.08 ±0.01 (0,16 ±0,025)
- 0.32 ± 0.32 (0.6 ± 0.6)
 0.04 ± 0.02 (0.07 ± 0.03)
 0.04 ± 0.04 (0.08 ± 0.08)
 5.75 ±0.45(11.2 ±0,9)
 0.03 ± 0.01 (0.07 ± 0.02) -
 12.4 ± 1.6(24)	
 47 ± 11 (5 ± 1)
 1.14 ±0.16(0.13 ±0.02)
 725  ± 117(84 ± 14)
 16 ±2 (1,9 ±0.3)
 6.6 ±  1.0. (6.8 ±0.1)
 8 ±  1(0.9 ±0.15)
 8.5 ±  1(1.0. ±0.2)
 14 ±2 (1.6 ±0.2)
 1.3 ±0,2(0.15 ±0.02)
 23[42 ± 5] (2.7[4.9 ± 0.9])
 0 (0)
 1.5 ±0.2 (0.17 ±0.03)
 5.8 ±  1.4(0.7 ±0.2)
 0.12 ± 0.03 (0.014 ± 0.004)
 0.34 ± 0.04 (0.04 ± 0.004)
 5.0 ±0.9 (0.6 ±0.1)
 0.22 ± 0.09 (0.025 ± 0.01)
 1.4 ±0.1 (0.16 ±0.1)
 0(0)
 0.66 ± 0.03 (0.08 ± 0.033)
 11.5 ±3(1.3 ±0.3)
       aMostly passenger cars, no distinction between catalytic and noncatalytic vehicles.
       ''Mostly heavy-duty diesel trucks, average weight -30 ton.
       cPartiaMy attributable to soil dust.
       dWholly attributable to soil dust.
       'Attributable to motor oil.
       Attributable to fuel additives.

       Source: Pierson and Brachaczek, 1983b.
1      control has been most effective in reducing the elemental carbon fraction of the particulate
2      matter, so that the organic carbon fraction now accounts for a much larger share than it had
3      previously.
4              The extractable portion of total carbon, although commonly used as a measure of organic
5      compound content, is not totally equivalent to the OC fraction, as measured by the thermal-
6      optical carbon analysis technique (Japar et al., 1984). The average ratio of OC to extractable
7      mass was shown to be 0,70 ± 0.05, when toluene/propanol-1 mixture was used as an extraction
       2/1/98
                            2-17
DRAFT-DO NOT CITE OR QUOTE

-------
  1      solvent, and this ratio was probably the result of the presence of both oxygenated organic
  2      compounds and inorganic sulfates in the extracted mass.
  3
  4      2.3.2.3.1. Extraction and fractionation techniques,  A variety of solvents and extraction
  5      techniques have been used in the past for the separation of organic compounds from diesel
  6      particles (Levsen, 1988). Although the reports on the extraction efficiencies are in part
  7      contradictory, it appears that Soxhlet extraction and the binary solvent system composed of
  8      aromatic solvent and alcohol gave the best recovery of PAHs, as determined by 14C-B[a]P
  9      (benzo[a]pyrene) spiking experiments (Schuetzle and Perez, 1981).  Direct chemical analysis of
 10      the entire extractable fraction of diesel paniculate matter is not generally possible because a large
 11      number of compounds of different polarity are present.  The separation of diesel particulate
 12      organic matter (POM) into various fractions according to chemical functionalities is a necessary
 13      preliminary step to chemical identification of individual compounds. Open-column liquid
 14      chromatography (LC) and liquid-liquid separation procedures have been the most widely used
 15      fractionation methods (Lee and Schuetzle, 1983).  Open-column LC is very often followed by
 16      normal-phase high-performance liquid chromatography (HPLC) if the identification of less
 17      abundant components is required.
 18
 19      2.3.2.3.2. Chemical composition.  Table 2-7 lists the general classes of extractable organic
•20      compounds identified in POM from combustion emissions, including diesel emissions.
 21            Liquid chromatography methods usually divide the complex environmental mixtures of
 22      organic compounds into nonpolar, moderately polar, and polar fractions.  This separation is
               Table 2-7. Classes of organic compounds identified in particulate-phase
               combustion emissions

                           Hydrocarbons
                           Derivatives* of hydrocarbons
                           Polycyclic aromatic hydrocarbons (PAHs)
                           Derivatives of PAH
                           Multifunctional derivatives of PAH
                           Heterocyclic compounds
                           Derivatives of heterocyclics
         	Multifunctional derivatives of heterocyclic compounds 	

         "Derivatives include acids, alcohols, aldehydes, esters, ketones, nitrates, and sulfonates.
         Source: Adapted from Schuetzle (1988).
         2/1/98                                     2-18       DRAFT-DO NOT CITE OR QUOTE

-------
  1      achieved by using specific solvents (or solvent mixtures) for the elution of compounds from
  2      chromatographic columns. Schuetzle and co-workers (1985) proposed that standard chemical
  3      compounds deselected for establishing reference points for the fractionation of diesel POM into
  4      nonpolar, moderately polar, and polar fractions by normal-phase HPLC.  They proposed that the
  5      elution of 1 -nitronaphthalene would define the end of the elution of nonpolar compounds and the
  6      beginning of the moderately polar fractions. In a similar way,  1,6-pyrene quinone would define
  7      the end of elution of moderately polar compounds and the beginning of the polar region.
  8            For diesel engine emissions, -51% of the extracted organic mass is contained in the
  9      nonpolar fraction (Schuetzle, 1983). About 90% of this fraction consists of aliphatic
.10      hydrocarbons from approximately C14 to about C40, with a carbon number maximum at C22 to C26
11      (Black and High, 1979; Pierson et al., 1983).  Polycyclic aromatic hydrocarbons and alkyl-
12      substituted PAH account for the remainder of the nonpolar mass.
13            The moderately polar fraction (=9% w/w of extract) consists mainly of oxygenated PAH
14      species and nitrated PAHs. The polar fraction (=32% w/w of extract) is composed mainly of
15      carboxylic and dicarboxylie acids of PAH, hydroxy-PAH, hydroxynitro-PAH, nitrated
16      N-containing heterocyclic compounds, etc. (Schuetzle, 1983; Schuetzle et al., 1985).
17            Limited recovery studies have shown that there is little  depadation or loss of diesel  POM
18      on the HPLC column.  More than 90% of the mass and 70% to 100% of the Ames
19      S. typhimurium-active material injected onto the column have been recovered (Schuetzle et al.,
20      1985).
21
22      2.3.2.3.3. Polycyclic aromatic hydrocarbons.  Particle-bound PAHs and their derivatives
23      (mainly nitrated PAH) attracted considerable attention relatively early because of their mutagenic
24      and, in some cases, carcinogenic properties (see, for example, National Research Council, 1982).
25      The most widely used methods of PAH analysis included thin layer chromatography (TLC),
26      capillary gas chromatography (GC), gas chromatography/mass spectrometry (GC/MS), and
27      HPLC with ultraviolet or fluorescence  detection (Levsen, 1988). Table 2-8 lists the PAHs and
28      thioarenes identified and quantified by GC/MS in three LDD particulate matter extracts (Tong
29      and Karasek, 1984). Data listed in this table reveal the presence of a large number of alkyl
30      derivatives of PAH, which are sometimes more abundant than the parent PAH.
31            Table 2-9 compares the emission rates of several representative PAHs from HDD, LDD,
32      and gasoline (with and without catalytic converter) engines.       .
33             .                                       '
34      2.3.2.3.4. Nitratedpofycyclic aromatic hydrocarbons. Nitro-PAHs (nitroarenes) have been
35      shown to be present in diesel particulate extracts, though in much lower concentration than the
36 ,     parent PAHs (Schuetzle et al., 1981,1982; Paputa-Peck et al., 1983).  Because many nitroarenes
        2/1/98                                   2-19       DRAFT-DO NOT CITE OR QUOTE

-------
        Table 2-8, Polycyclic aromatic hydrocarbons identified and quantified in
        extracts of diesel particles
Compound*'"
' Aeenaphthylene
Trimethy (naphthalene
Fluorene
Dimethylbiphenyl
C4-Naphthalene
Trimethylbiphenyi
Dibenzothiophene
Phenanthrene
Anthracene
Methyldibenzothiophene
Methylphenanthrene
Methylanthracene
Ethylphenanthrene
4H-Cyclopenta[de/]phenanthrene
Ethyldibenzothiophene
2-Phenylnaphthalene
Dimethyl(phenanthrene/anthracene)
Fluoranthene
Benzo[
-------
               Table 2-9. Emission rates of particle-bound polycyclic aromatic hydrocarbons
               (PAHs) (in milligrams per kilometer) from HDD, LDD, and gasoline engines
Gasoline cars
PAH
Pyrene
Fluoranthene
Benzo[a]pyrene

Benzq|e]pyrene
HDD
17.6 (ll)a
27.2 (17)a
<0.1 (0.06)a
2.3(l,4)e
0.24(.15)a
LDD
66 (42)b
50 (31)b
1 (0.6)d
NDb
3(1.9)d
Noncatalyst
45 (28)c
32(20)c
3.2 (2)c

4.8 (3)c
Catalyst
7 (4,4)d
5(3.1)d
0.4 (0.25)d

. 0.4 (0.25)d
        aFrom Westerhoim et al. (1991).
        bFrom Smith (1989), four-cycle FTP test; 1986 Mercedes Benz.
        'From Alsberg et al. (1985).
        dFrom Schuetzle and Frazier (1986).
        'From Dietzman et al. (1980), averaged value for four different engines.
        ND = None detected.
  1      are potent direct-acting (e.g., without metabolic activation) mutagens in the Ames assay using S,
  2      typhimurium strains (Rosenkranz and Mermelstein, 1983), the analysis of nitro-PAHs in diesel
  3      POM attracted considerable attention in the early 1980s,
  4             Numerous nitro-PAHs were identified in LDD participate extracts using capillary GC
  5      with thermionic nitrogen-phosphorus detector (NPD) (Paputa-Peck et al., 1983). Positive isomer
  6      identification for 16 nitro-PAHs has been made utilizing the GC retention times of authentic
  7      standards and low- and high-resolution mass spectra as identification criteria. These include 1-
  8      nitropyrene; 2-methyl-l-nitronaphthalene; 4rnitrobiphenyl; 2-nitrofluorene; 9-nitroanthracene; 9-
  9      methyl- 10-nitroanthracene; 2-nitroanthracene; 2-nitrophenanthrene; l-methyl-9-nitroanthracene;
10      l-methyl-3-nitropyrene; l-methyl-6-nitropyrene; l-methyl-8-nitropyrene; 1,3-, 1,6-, and 1,8-
11      dinitropyrene; and 6-nitrobenzo[a]pyrene.  In addition, two nitrated heterocyclic compounds
12      were identified, 5- and 8-nitroquinoline.  Forty-five additional nitro-PAHs were tentatively
*l 3      identified in this diesel particulate extract (Paputa-Peck et al., 1983).
14             The concentration of nitro-PAHs adsorbed on diesel particles varies substantially from
15      sample to sample. Usually 1-nitropyrene is the predominant component, and concentrations
16      ranging from *7 to = 165 |ig/g of particles are reported (Levsen, 1988).
17             Table 2-10 gives the approximate concentrations of several more abundant nitro-PAHs in
18      LDD particulate extracts.
19
20      2.3.2.3.5. Oxygenated polycyclic aromatic hydrocarbons.  The moderately polar fraction of
21      diesel particulate extract contains a variety of oxy-PAHs (in particular, aldehydes, ketones,
22      quinones, and acid anhydrides) in much higher amounts than nitro-PAHs, which elute in the
        2/1/98
2-21
DRAFT-DO NOT CITE OR QUOTE

-------
               Table 2-10.  Concentrations of nitro-polycyclic aromatic hydrocarbons identified
               in a light-duty diesel (LDD) particulate extract

                                                                       Concentration
        Nitro-PAHa	.	   fcg/g of particles)	
        4-Nitrobiphenyl                                                      2.2
        2-Nitrofluorene                                                     .1.8
        2-Nitroanthracene                                                     4.4
        9-Nitroanthracene                                                     1.2
        9-Nitrophenanthrene                                                   1.0
        3-Nitrophenanthrene                                                   4.1
        2-Methyl-l-nitroanthracene                                             8.3
        1-Nitrofluoranthene                                                   1.8
        7-Nitrofluoranthene                                                   0.7
        3-Nitrofluoranthene                                                   4.4
        8-Nitrofluoranthene                                                   0.8
        1-Nitropyrene                                                       18.9; 75b
        6-Nitrobenzo[a]pyrene                                                 2.5
        l,3-Dinitropyreneb                                                    0.30
        l,6-Dinitropyreneb                                                    0.40
        l,8-Dinitropyreneb    -                                                0.53
        2,7-Dinitrofluorenec                                                   4.2; 6.0
        2,7-Dinitro-9-fluorenonec	8.6; 3.0	'

        "From Campbell and Lee (1984) unless noted otherwise. Concentrations recalculated from fig/g of extract to
        fig/g.of particles using a value of 44% for extractable material (w/w).
        "From Paputa-Peck et al. (1983).
        Trom Schuetzle (1983).
 1      same fraction.  Oxy-PAHs are nonmutagenic or very weakly mutagenic; this explains the
 2      relatively low interest in this group of compounds. The most detailed study of oxy-PAHs was
 3      published by Schuetzle et al. (1981), who identified more than 100 compounds.  A large number
 4      of oxy-PAHs in the molecular weight range of 182 to 272 were also identified by Tong et al.
 5      (1984). The main components identified by Tong and co-workers in particulate matter extracts
 6      of three LDD cars (Volkswagens) were 9-fluorenone, anthraquinone, 4H-
 7      cyclopehta|We/|phenanthrene-4-one, 9-phenanthrene aldehyde, benzo|We]anthracene-7-one, and
 8      benzo[cd]pyrene-6-one. These components are present at concentrations of 30 to 300 ug/g
 9      particles.  Some of these oxy-PAHs are formed during sampling (Levsen, 1988).
10
11      2.3.2.3.6. Polar polycylic aromatic hydrocarbon derivatives. According to Schuetzle et al.
12      (1985), although 65% to 75% of the directly acting mutagenicity (as tested by Ames S.
13      typhimurium assay) for LDD particulate extracts is associated with the fraction of moderate

        2/1/98                                      2-22        DRAFT-DO  NOT CITE OR QUOTE

-------
  1      polarity, more than 65% of the mutagenic activity for HDD particulate extract is concentrated in
  2      the most polar fraction. However, because of the serious analytical difficulties, only preliminary
  3      data exist on the identification of compounds that are responsible for the mutagenic activity of
  4      this fraction (so-called "polar mutagens"). Schuetzle and co-workers (1985) employed the
  5      concept of "bioassay directed chemical analysis" (see Section 2.6) for the isolation and
  6      identification of polar PAH derivatives from the extracts of HDD particulate matter (National
  7      Institute of Standards and Technology [NISTJ standard reference material SRM 1650). Several
  8      hydroxynitro-PAHs, hydroxy-PAHs, and nitrated heterocyclic compounds were tentatively
  9      identified in the polar fraction. It has to be noted, however, that NIST SRM 1650 was not
10      intended to be representative of HDD engines but was a material made available to investigators
11      for the purpose of methods development.
12            In another study (Bayona et al., 1988), the polar HPLC fractions of the same NIST SRM
13      1650 were analyzed by fused silica capillary GC with low- and high-resolution mass MS, using
14      electron impact (El) and negative ion chemical ionization (NICI) techniques.  In addition, direet-
15      probe El and NICI-MS analyses were performed. More than 80 poly cyclic  aromatic compounds
16      (PAC) belonging to several different chemical classes (anhydrides, carboxaldehydes, diazaarenes,
17      cyclic imides, hydroxynitro-PAH, nitroaza-PAC, nitrolactones, and quinones) were tentatively.
18      identified. Ten were positively identified by comparison of retention times with authentic
19      standards. Among them, phenazine and phthalic anhydride were positively identified for the first
20      time in diesel exhaust particles.  In addition, cyclic imides and their alkylated derivatives were
21      tentatively identified.
22
23      2.3.3, Gaseous-/Partieulate-Piiase Emission Partitioning of Polycyclic Aromatic
24            Hydrocarbons
25            The distribution of the emissions between the gaseous and particulate phases is
26      determined by the vapor pressure of the  individual species, by the amount and type of the
27      particulate matter present (adsorption surface available), and by the temperature (Ligocki and
28      Pankow, 1989). Table 2-11 gives the vapor pressures at 25 °C of some representative PAHs
29      ranging from naphthalene to benzofojpyrene.
30            The factor of = 107 in the range of vapor pressures is reflected in the  fact that, at
31      equilibrium at ambient temperature, naphthalene exists almost entirely in the gas phase, whereas
32      B[a]P, other five-ring PAHs, and higher-ring PAHs are predominantly adsorbed on particles,
33      The intermediate three- and four-ring PAHs are distributed between- the two phases.
34            However, the vapor pressures of these intermediate PAHs can be significantly reduced by
35      their adsorption on various types of surfaces. Because of this phenomenon, the amount and type

        2/1/98                                   2-23       DRAFT-DO NOT CITE OR QUOTE

-------
              Table 2-11. Vapor pressures at 25°C for a series of polycyclic aromatic
              hydrocarbons'1 (PAHs)

                                                                  Vapor pressure
       	PAH	            at298K(torr)
                Naphthalene                                           8.0 x 10'2
                Acenaphthylene                                        6.7 X 10~3
                Acenaphthene                                          2.2 x 10"3
                Fluorene                                             6.0 x 10A
                Phenanthrene                                          1.2'x 10"
                Anthracene                                           6,0 x 10'6
                Fluoranthrene                                          9.2 x 10*
                Pyrene                                               4.5 x 10~*
                Benzo[a] anthracene                                     2.1 x 10"7
                Benzo[a]pyrene                                        5.6 x 10"9
                Chryseneb                                            6.4 X IP'9

       'Sonnefeld et al, (1983), except as indicated.
       "Yamasaki tt al. (1984).
 1      of participate matter present play an important role, together with temperature, in the vapor-
 2      particle partitioning of semivolatile organic compounds (SOC).
 3    '        The measurements of gas/particulate phase distribution are often accomplished by using a
 4      high-volume filter followed by an adsorbent such as polyurethane foam (PUF), Tenax, or XAD-2 •
 5      (Cautreels and Van Cauwenberghe,  1978; Thrane and Mikalsen, 1981; Yamasaki et al., 1982).
 6      However, the pressure drop behind a high-volume filter or cascade impactof contributes to
 7      volatilization of the three- to five-ring PAHs, to a degree reflecting their vapor pressures.  The
 8      magnitude of this "blow-off artifact depends on a number of factors, including sampling
 9      temperature and the volume of air sampled (Van Vaeck et'al, 1984; Coutant et al., 1988).
10      Despite these problems from volatilization, measurements with the high-volume filters followed
11      by a solid adsorbent have provided most estimates of vapor-particle partitioning of SOC in
12      ambient air, as well as insights into the factors influencing SOC adsorption onto aerosols.  -
13            Average distributions of PAH between high-volume filter and PUF plugs (positioned
14      downstream of the filter) in samples collected in a heavily traveled roadway tunnel (Baltimore
15      Harbor Tunnel) are shown in Figure 2-2. As discussed in the preceding text, the "blow-off* from
16      the filter precludes detailed quantitative interpretation; However, it' can be seen from this figure
17      that significant fractions of phenanthrene, anthracene, and their alkylated derivatives, along with
18      fluoranthene and pyrene, exist in the gas phase. No PAHs less volatile than pyrene were

        2/1/98                                    2-24       DRAFT-DO NOT CITE OR QUOTE

-------
                                            Particle
                    Vapor Phase
               Phenanthrene
                 Anthracene
        3-MethyIphenanthrene
       .,  2-Methylanthracene
        1-Methylphenanthrene
         C2-phenanthrene(1)
         C2-phenanthrene(2)
         C2-phenanthrene(3)
         C2-phenanthrene(4)
         C2-phenanthrene(5)
         C2-phenanthrene(6)
               Fluoranthrene
                    Pyrene
                          0%
25%           50%           75%
     Percent in Participate Phase
                             100
       Figure 2-2.  Vapor/particle phase polycyclic aromatic hydrocarbon distribution in samples
                   collected in Baltimore Harbor Tunnel.
       Source: Benneretal. (1989).
 1     observed in any of the PUF samples.  Comparison of the observed vapor-to-particle PAH ratios
 2     and those calculated based on the relationship derived by Yamasaki and co-workers (1982)
 3     generally agreed within a factor of 2 (Benner et aL, 1989).
 4                                  •       •
 5     2.4.  ATMOSPHERIC TRANSFORMATIONS OF PRIMARY DIESEL EMISSIONS
 6     2.4.1. Long-Range Transport and Fate of Primary Diesel Emissions
 7           Once released into the atmosphere, primary diesel emissions (or any other direct
 8     emissions) are subject to dispersion and transport and, at the same time, to various physical and
 9     chemical processes that determine their ultimate environmental fate. The role of the atmosphere
10     may be compared in some way with that of a giant chemical reactor in which materials of varying
11     reactivity are mixed together, subjected to chemical and/or physical processes, and finally
12     removed (Schroeder and Lane, 1988).  The main features of the atmospheric cycle for primary
       2/1/98
      2-25
DRAFT-DO NOT CITE OR QUOTE

-------
 1      diesel emissions, beginning with emission and ending with deposition to the Earth's surface, are
 2      shown in Figure 2-3.
 3            Initial mixing describes the physical processes that act on pollutants immediately after
 4      their release from an emission source.  The dilution of diesel exhaust under roadway conditions
 5    •  is an important factor to consider; whereas a dilution factor of 10 is typical of many dilution
 6      tunnels used in dynamometer studies of automobile exhaust, a dilution factor of = 103 is more
 7      realistic under roadway conditions.  This discrepancy leads to slightly different particle size
 8      distributions under real driving conditions than those predicted from laboratory data (Kittelson
 9      and Dolan, 1980); for example, because of slower coagulation processes, more particles in the
10      Aitken nuclei range (<0.08 urn diameter) may be expected under typical roadway conditions.
11            Diffusion and transport processes occur simultaneously in the  atmosphere and account for
12      the dispersion of emissions.  The actual distance traveled by gaseous- and particulate-phase
                                     Chemical
                                   Transformation
                                                          Ambient Air
                                                          Concentration
                                                  Wet
                                             Transformation
                Initial
               .Mixing
Transport and
  Diffusion
                                                                                 Scavenging
           '     Diesel
               Emissions
        Figure 2-3.  Diesel-derived pollutants: emission-to-deposition atmospheric cycle.

        Source: Schroeder and Lane (1988).      .     .
        2/1/98
                2-26
DRAFT-DO NOT CITE OR QUOTE

-------
  1      pollutants depends on the amount of time a specific pollutant resides in the atmosphere and is
  2      available for dispersion (Sehroeder and Lane, 1988). As primary diesel emissions are a very
  3      complex mixture containing thousands of organic and inorganic constituents in the gas and
  4      particulate phases that have different chemical reactivities, they are removed by dry and wet
  5      deposition processes at different rates. The more reactive compounds with short lifetimes will be
  6      removed from the atmosphere relatively quickly, whereas more stable pollutants can be
  7      transported over greater distances.  Clearly, a knowledge of the atmospheric loss processes and
  8      lifetimes for automotive emissions is important, because these lifetimes determine the geographic
  9      extent of the influence of these emissions.
10            Anthropogenic pollutants can travel through the atmosphere over long distances. In
11      particular, the long-range transport of SO2 and its transformation to SOJ have been studied
12"     extensively (Galloway and Whelpdale, 1980; Lowenthal and Rahn, 1985). The organic
13      pollutants, particularly those adsorbed on carbonaceous particles, are also subjected to long-range
14      transport. Organic compounds, such as PAHs, adsorbed on diesel particulate matter are generally
15      more resistant to atmospheric reactions than those in the gas phase. In addition, particles of
16      smaller diameter (<1 um), such as diesel particulate matter (Figure 2-1), are removed less
17      efficiently than larger particles by wet and dry deposition and thus have longer atmospheric
18      residence times.
19            It has been reported (Laflamme and Hites, 1978; Hites et al., 1980) that PAHs and their
20      alkyl homologs are distributed in sediments throughout the world and that the PAH patterns are
21      similar to each other and to air particulate matter for most of the locations studied. Furthermore,
22      the quantities of PAHs increase with proximity to urban areas. This suggests anthropogenic
23      combustion sources and long-range atmospheric transport of PAHs.
24            Evidence for long-range transport of PAHs was also reported from the recent
25      measurements of PAHs in Siskiwit Lake, located on a wilderness island in northern Lake
26      Superior (McVeety and Hites, 1988). Because of the lake's remote location, PAHs found in this
27      lake are likely to have originated from atmospheric transport.
28            Earlier studies by Bj0rseth and co-workers (Bj0rseth and Olufsen, 1983) showed that
29      PAHs are transported from Great Britain and the European continent to remote locations in
30 "     Norway and Sweden. The specific sources of PAH emissions could not be identified, however.
31      The authors speculated that combustion engines were not the major sources, because the amounts
32      of B[a]P found in the samples collected in Norway and Sweden were higher than could be
33      accounted for from gasoline and diesel fuel consumption in .Great Britain and because the
34      coronene/B[a]P ratios in the samples were lower than those usually found in gasoline and diesel
35      exhaust. However, because of the lack of PAH profiles specifically for combustion engines (or,

        2/1/98                                   2-27      DRAFT-DO NOT CITE OR QUOTE

-------
 1      as a matter of fact, any other specific tracer), the relative contribution of gasoline and diesel
 2      vehicle exhausts to long-range transport of organics could not be determined.
 3
 4      2.4.2.  Chemical Transformations
 5      2,4.2.1. Gas-Phase Reactions
 6            The following chemical processes contribute to the removal of gas-phase compounds
 7      from the atmosphere (Atkinson, 1988):
 8            •  Photolysis during daylight hours;
 Q                   ,
10            •  Reaction with hydroxyl (OH) radicals during daylight hours;
11
12            •  Reaction with ozone (O3) during  daytime and nighttime;
13
14            *  Reaction with hydroperoxyl (HO2) radicals, typically during late daytime and early
15               nighttime hours;
16
17            *  Reaction with gaseous nitrate (NO3) radicals during nighttime hours;
18
19            •  Reaction with dinitrogen pentoxide (N2O5) during nighttime hours;
20
21            *  Reaction with NO2 during daytime and nighttime hours; and
22                                     .
23            •  Reaction with gaseous nitric acid (HNO3) and other species such as nitrous (HNO2)
24               acid and sulfuric acid (H2SO4).
25                                                    ,
26            It has been shown (Atkinson et al,, 1990) that the N2O5 reactions with PAHs proceed by
27      initial NO3 addition to form an NO3-PAH adduct, which either dissociates back to reactants or
28      reacts exclusively with NO2 to form nitroarenes and other products. Because under atmospheric
29      conditions, where N2O5, NO3 radicals, and NO2 are in equilibrium, these reactions are kinetically
30      equivalent to a reaction with N2O5 with an effective N2O5 reaction rate constant, we will further
31      refer to these reactions as N2O5 reactions.
32            The reactive gaseous species, such-as OH radicals, NO3 radicals, HO2 radicals, and ozone,
33      are present in the atmosphere either during the daytime (OH radicals) or nighttime (N2O5 and
34      NO3 radicals) hours or both time periods (ozone, NO2). For the routes of formations of these
35      species and their concentrations in the troposphere, see Finlayson-Pitts and Pitts (1986).
36            Table 2-12 gives the calculated atmospheric lifetimes for some selected compounds
37      present in automotive gas-phase emissions as the result of known tropospheric chemical removal
38      reactions (Atkinson, 1988). These lifetimes (i.e., the time for the compound to decay to 1/e or
        2/1/98                                    2-28       DRAFT-DO NOT CITE OR QUOTE

-------
        Table 2-12.  Calculated atmospheric lifetimes for gas-phase reactions of selected
        compounds present in automotive emissions with atmospherically important
        reactive species
Atmospheric lifetime resulting from reaction with
Compound
N02
NO
HNO3
SO,
NH3
Propane
rc-Butane
«-Octane
Ethylene
Propylene
Acetylene
Formaldehyde
Acetaldehyde
Benzaldehyde
Acrolein
Formic acid
Benzene
Toluene
m-Xylene
Phenol
Naphthalene
2-Methylnaphthalene
2,3-Dimethylnaphthalene
Acenaphthene
Acenaphthylene
Phenanthrene
Anthracene •
Fluoranthene*
Pyrene*
OH"
2 days
4 days
180 days
26 days
140 days
19 days
9 days
3 days
3 days
llh
30 days
3 days
1 day
2 days
1 day
50 days
18 days
4 days
llh
10 h
1 day
5h
4h
2h
2h
9h
•2h
6h
6h
0,"
12 h
1 min
—
> 200 years
—
> 7,000 years
> 4,500 years
—
9 days
1.5 days
6 years
>2 x 104 years
> 7 years
- _
60 days
—
600 years
300 years
75 years
—
> 80 days
> 40 days
>40 days
> 30 days
~50 min
—
—
—
—
NO/ HO/ hv'
1 h 2 h 2 min
3 min 20 min —
— —
>4 x 10" years . > 600 years —
— — —
— — —
9 years — —
3 years — —
3 years — —
15 days — —
> 14 years — — '
210 days 23 days 4 h
50 days — 60 h
60 days — —
— — — .
— — —
> 16 years — —
9 years — —
2 years — —
20 min — —
	 r . 	 	
	 r 	 	
	 r 	 • 	
=3h - —
13 min — —
— — ' —
— —
	 r 	 	
r 	
"For 12-h average concentration of OH radical of 1 x itf molecule/cm'.
bFor 24-h average 0, concentration of 7 x 10" molecule/crrf.
'For 12-h average NOj concentration of 2 x 108 molecule/cnf.
dFor 12-h average HQ; concentration of itf molecule/cm1,
Tor solar zenith angle of 0°.
'Lifetimes due to gas-phase reactions with a 12-h average concentration of 1^0, of 2 x 10'" molecule/cm5 are:
 naphthalene, =80 days; 2-methylnaphthalene,  =35 days; 2,3-dimEthyinaphthalene, =20 days; fluoranthene,
 -64 days; and pyrene, -20 days.
•Lifetimes calculated from kinetic data given in Atkinson et al. (1990),

Source;  Atkinson (1988) unless noted otherwise.
2/1/98
2-29
DRAFT-DO NOT CITE OR QUOTE

-------
 1      37% of its original concentration) are calculated from the corresponding measured reaction rate
 2      constants and the average ambient concentration of the tropospheric species involved.
 3            Although the individual rate constants are known to a reasonable degree of accuracy (in
 4      general, to within a factor of 2), the tropospheric concentrations of these key reactive species are
 5      much more uncertain. For example,  the ambient concentrations of OH radicals at any given time
 6      and/or location are uncertain to a factor of at least 5, and more likely 10 (Atkinson, 1988).  The
 7      tropospheric diurnally and annually averaged OH radical concentrations are more certain, to
 8      possibly a factor of 2. For this reason,  the calculated lifetimes listed in Table 2-12 are
 9      approximate only and are valid for those reactive species concentrations listed in the table
10      footnotes. However, these data permit estimation of the contribution of each of these
11      atmospheric reactions to the overall rates of removal of most pollutants from the atmosphere.
12            As can be seen from Table 2-12, the major atmospheric loss process for most of the
13      automotive emission constituents listed is by daytime reaction with OH radicals. For some
14      pollutants photolysis, reactions with  ozone, and reactions with NO3 radicals during nighttime
15      hours are also important removal routes.
16            The atmospheric lifetimes do not take into consideration the potential chemical or
17      biological importance of the products of these various reactions.  For example, the reaction of
18      gas-phase PAHs with N2O5 appears to be of minor significance as a PAH loss process but,  as will
19      be discussed in subsequent sections,  is more important as a route of formation of mutagenic
20      nitro-PAHs.
21
22      2.4.2.1.1. Reactions of nitrogen oxides. Only the major atmospheric reactions of NOX are
23      considered here; for detailed discussion of the chemistry of these important species, see
24      Finlayson-Pitts and Pitts (1986).
25             Oxides of nitrogen emitted by diesel engines include mainly NO, with  lesser amounts of
26      NO2. Nitric oxide is easily oxidized to NO2 in reactions with HO2 radicals and alkylperoxy
27      radicals (the reaction of NO with O2  is too slow at typical ambient concentrations of NO).  In
28      addition, NO reacts rapidly (Table 2-3) with ozone via Reaction 1:
29                                       NO + 03-NO2 + O2                                (1)
30      Nitrogen dioxide is photolyzed rapidly at wavelengths of <430 nm:
31                                       NO2 + hv - NO + O(3P)                              (2)
32      The oxygen atom produced in this reaction reacts with O2, forming ozone:
33                                          O2 + 0(3P)-O3                                  (3)
34             The photolysis of NO2 is the  only known significant anthropogenic source of ozone in the
35      ambient air and is produced via Reactions 2 and 3.  With this series of reactions, NO, NO2, and
36      O3 are in a photostationary state;
        2/1/98                                    2-30       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
1.1
12
                             NO > O3
with
                                  [OJ
k2[N02]
kJNOJ
where k, and k, are the rate constants for Reactions 1 and 2, respectively, and brackets signify
concentrations. This photostationary state is strongly affected by NO-to-NO2 conversions caused
by reactions involving organic compounds.
      The important atmospheric reactions of NO2 also include formation of N03 radicals and
N205:
                                 NO2 + O3 - NO3 + 02
                                              M
                                 NO, + NO,
                                         N,O5
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
with N2O5 being in equilibrium with NO2 and NO3 radicals.
      The other important atmospheric reactions of NO and NO2 include nitrous and nitric acid
formation, respectively, by reaction with OH radicals:
                         NO + OH-HNO2
                         NO2 + OH - HNO3

2.4.2.1.2. Reactions of sulfur dioxide. Reaction with OH radical is the dominant SO2
atmospheric gas-phase reaction process (Stockwell and Calvert, 1983):
                         SO2 + OH - HSO3
followed by the formation of HO2 radicals and H2S04:
                         HSO + 0 - H0 + S03
                                                  H2O
                                             HS0
      Because SO2 is soluble in water, it undergoes scavenging by fog, cloud water, and
raindrops.  In aqueous systems, SO2 is readily oxidized to sulfate (Calvert and Stockwell, 1983).
       2/1/98
                                       2-31
         DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
1 \J
19
20
21
22
23
24
25
26
27
28
29

30
31
32
33

34
35
36-
2.4.2.1.3. Reactions of alkanes. Only a brief overview of the most important atmospheric


reactions of alkanes is presented here; for detailed discussion, consult Finlayson-Pitts and Pitts
(1986) and Atkinson (1988, 1990).

Under atmospheric conditions, alkanes react with OH radicals during the daytime and
with NO3 radicals during the nighttime:
RH + OH-R' + H2O
RH + NO3 - R' + HNO3

Alkyl radical R' reacts with O2, forming an alkylperoxy radical:

R- + O2^RO2*
which, under polluted urban atmospheric conditions, reacts predominantly with NO by two
pathways: (a) oxidation of NO to NO2 and formation of alkoxy radical (RO'), the only









significant path for the smaller (
-------
 1            For the simplest alkoxy radicals, reaction with O2 (pathway a) is predominant. For larger
 2      radicals, however, isomerization (c) and decomposition (b) may become significant, the relative
 3      importance of these two pathways depending on the structure of the radical. In all three cases,
 4      free radicals are produced, which then will carry on the chain reactions. The first-generation
 5      products include aldehydes, ketones, and alkyl nitrates, which can react further under
 6      atmospheric conditions.
 7
 8      2.4.2.1.4. Reactions ofaikene. Lower molecular weight alkenes, such as ethylene, propylene,
 9      and isomeric butenes, are present in exhaust from gasoline and (although in lower amount) diesel
10      engines (see Table 2-2),  Gas-phase alkenes are removed from the troposphere by reaction with
11      OH radicals, NO3 radicals, and O3 (Finlayson-Pitts and Pitts, 1986; Atkinson, 1988; Atkinson
12      and Carter, 1984). Reactions with OH radicals are rapid (see Table 2-13) and proceed by  OH
13      radical addition to the double bond.
14            Because the OH-olefin adduct  is essentially an alkyl radical, it reacts further in a manner
15      similar to alkyl radicals formed from the reaction of alkanes with OH radicals (e.g., by the
16      addition of O2 followed by reaction with NO). For example, for radical (a):
17       .
18                                               NO           O
19
20                        CH3CHCH2OH -i- O2 —
                                                            CH,CHCH,OH
21                                               N02                                        .
22
23      Table 2-3).  Similar to the OH radical reaction, this reaction proceeds through NO3 radical
24      addition to the double bond, followed by reaction with O2 (Finlayson-Pitts and Pitts, 1986;
25      Atkinson, 1988).  Carbonyl compounds are formed as major products, but minor products
26      (possibly dinitrates) are not well defined (Atkinson, 1988).
27            The reactions of alkenes with ozone compete with the daytime OH radical reaction (Table
28      2-3). These reactions proceed by addition to the double bond, followed by rapid decomposition
29      of so-called "ozonide" or "molozonide" into a carbonyl compound and an energy-rich biradical:
        2/1/98                     :              2-33       DRAFT-DO NOT CITE OR QUOTE

-------
Table 2-13, Summary of the nitroarenes produced from the gas-phase hydroxyl (OH) radical-initiated and
dinitrogen pentoxide (N2O5) reactions and electrophilie nitration of poly cyclic aromatic hydrocarbons (PAHs)








to
U)
-fc.


O
H
I
D
O
2!
O
H
D
3
O
90
O
w
PAH
Naphthalene
1 -Methy (naphthalene

2-Methylnaphthalene

Acenaphthylene
Acenaphthene

Biphenyl
Phenanthrene


Anthracene
Fluoranthene
Pyrene
Acephenanthrylene

"Concurrent NO3 radical



Position of nitration
OH
1- (0,3%); 2- (0.3%)
5->4->6->3->7-=2^8-
Total yield (=0.4%)
5->6-=7-=4-=8^>3->l-
Totai yield (=0.2%)
4- (2%)
5->3->4-
Total yield (=0.2%)
3- (5%)
Two isomers (not 9-nitrophenanthrene)
Total yield (iO.1%)

Total yield (=0.2%)
2- (3%); 7- (-0.15%); 8- (=0.15%)
2- (=0.25%); 4- (=0.045%)
Two isomers (not 4- or 5-nitro-)
Total yield (=0.1%)

(yield) in reaction with
N2OS Position of electrophilie nitration
If i *JO/ \, ''I /*70/ "i 1 ""si")
~ 1 1 / /Of. At" ( i/O I I ~"-*^4£,**'
3->5->4->8-:=6->7->2- 4->2->5->8->7-=3->6-
Totalyield(«30%)
4->l-=5->8-=3-=7-=6- ' l^>8->4->6->5->3->7-
Total yield (=30%)
None observed 1 -
4- (40%); 3- (=2%); 5- (=2%)a 3-; 5-

No reaction observed 2-; 4-
Four isomers (including 9->3-;2-;l-
9-nitrophenanthrene)
Total yield (<1%)
l-;2- 9-
Total yield (<2%)
2-(=25-30%) 3->8->7->l-
4-; 2- 1-
Total yield (<1%)
None observed 4-; 5-

reaction will dominate over N2O5 reaction in ambient air.

•





-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
 R'
   \
                  R,
     c = c
0,
                                                R,
\
/
c  -  c
                                                             \
                              R,R2COO -i- R3R4C=0
                                R,R2C=O + R3R4COO
The importance of pathways (a) and (b) is assumed to be equal (Atkinson and Carter, 1984).
       The major uncertainty concerns the atmospheric fate of the energy-rich intermediate, the
so-called "Criegee biradical." It is believed that some of the Criegee intermediates contain
sufficient excess energy to spontaneously decompose. The remaining "thermalized" species are
expected to react in the atmosphere with NO, NO2, S02, H2O vapor, CO, and carbonyl
compounds, although the rate constants and mechanisms of these reactions are uncertain
(Finlayson-Pitts and Pitts, 1986).

2.4i2.1.5. Reactions of oxygen-containing organics. Diesel and gasoline vehicle exhausts
contain aldehydes and carboxylic acids. The major loss processes for aldehydes involve
photolysis  and reaction with OH radicals (see Table 2-12), Photolysis is an important loss
process for formaldehyde, forming H atom and HCO radical in the first step:
                                               H + HCO

                           HCHO + hv-
26
27
28
29
30
31
The rapid reaction of H atom and HCO radical with 02 produces HO2 radicals:
                   HCO + O2 - HO2 + CO
The higher aldehydes also photqdissociate, ultimately yielding HO2 radicals.
       2/1/98
                                       2-35
                    DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
      The OH radical reaction with formaldehyde yields CO and HO2:
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
                  HCHO + OH - H,O + HCO
                                                I O,
                                               HO2 + CO
      For higher aldehydes, the RCO radicals initially formed add O2 to yield acylperoxy
radical, which can react further with NO (pathway a) and NO2 (pathway b):
                    RCHO.+ OH'-» RCO" + H2O
                                        02
                                   RC(O)OO-
                    NO
              RC(O)O- + NO2
                           It
NO,
                   R- + CO2
                          RC(O)OONO2
                        (peroxyacyl nitrates)
Acetaldehyde forms peroxyacetyl nitrate (PAN), which has been shown to be a direct-acting
mutagen toward Ames S. typhimurium strain TA100 (Kleindienst et al., 1985) and is phytotoxic.
      Benzaldehyde, the simplest aromatic aldehyde, forms peroxybenzoyl nitrate by the series
of reactions analogous to pathway (b) above and nitrophenols as a result of the reaction with NO,
analogous to pathway (a) above:
        CJL- + O,
CH,OO-
                                        NO
                                        NO,
                                              C6H50-
                                           HO,
                                                  NO,
                                            C6H5OH + O2
  NO2C6H4OH
  (o- and p-)
       2/1/98
                                      2-36
                         DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
                                   NO,
                   C6H5C(0)OO- —
                                             C6H5C(0)OON02
                                   NO  '
                                             CH,- + CO,
                                   NO,
      Carboxylic acids react with OH radicals under atmospheric conditions (Table 2-3). For
formic acid, hydrogen atoms are produced (Atkinson, 1988):
                          HC(0)OH + OH - H2O + CO2 + H
For the higher carboxylic acids, reaction products are currently unknown.

2.4.2.1.6. Reactions of monocyclic aromatic compounds. The monocyclic aromatic
compounds are removed from the atmosphere solely by reactions with OH radicals (see Table 2-
3). These reactions proceed by two pathways (Finlayson-Pitts and Pitts, 1986; Atkinson, 1988):
      (a)   a major pathway by OH radical addition to the aromatic ring, and

      (b)   a minor pathway by H atom abstraction, either from the aromatic ring (in the case of
           benzene) or from an alkyl group (in the case of alkyl-substituted aromatics). This
           latter pathway leads to the formation of aromatic aldehydes and is analogous to OH
           radical/reaction of alkanes.
For example, for toluene:

                                      C^  H                                   ,
                                           -OH
                                        -90%
                                        10%
                                                 (+ lesser amount
                                                 of p-isomer)
                                                       NO
                                                               'CH>°
                                                                                  HO,
                                                       NO
       ,2/1/98
                                       2-37
DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
SO"
31
32
33
34
35
36
       The products arising from the OH radical addition pathway (a) are not well known.
Reaction with O2, again occurring by two pathways, is expected to predominate:
                    CH3
                          H
                           OH
                                                               HO2
Pathway (a) yields phenolic compounds and, for toluene, accounts for -20% of the overall
reaction yield (Atkinson, 1988). The major reactions involve ring cleavage (opening), leading to
a variety of bifunctional products (Finlayson-Pitts and Pitts, 1986).
       For phenolic compounds, in addition to OH radical reaction (proceeding mainly by initial
OH radical addition to the ring), the NO3 radical reaction that yields nitrophenols appears to be
important (Atkinson, 1988):
NO
                                  HNO
+ o-isomer
2.4.2.1.7. Reactions ofpolycyclic aromatic compounds. As discussed in Section 2.3.4, two- to
four-ring PAHs emitted from diesel and spark-ignition engines are distributed between gas and
particle phases. For those PAHs present in the gas phase, the reaction with OH radical is
predominant, leading to atmospheric lifetimes of a few hours or less (see Table 2-12). The
nighttime gas-phase reaction withN2O5 is-of minor significance as a PAH loss process but (as
will be discussed in the following text) may be important as a formation route of mutagenic
nitro-PAH. In addition, for the PAH-containing cyclopenta-fused ring, such as acenaphthene,
acenaphthylene, and acephenanthrylene, the NO3 radical reaction can be an important gas-phase
loss process during nighttime hours.
       Relatively few product data are available concerning these gas-phase reactions.  It has
recently 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 at al., 1986, 1989; Atkinson et al.,
       .2/1/98
                                         2-38
                                  DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
 1987, 1990; Zielinska et al., 1988, 1989a). The postulated reaction pathway involves initial OH
 radical addition to the most reactive ring position; for example, C-3 position for fluoranthene
 (Pitts et al., 1985a):
                                8   9
                 OH
                                                               H   OH
followed by NO2 addition in the C-2 position. Subsequent elimination of water results in
2-nitrofluoranthene formation:
                   NO,
         H OH
               -ao
HOH
       The analogous reaction sequence for pyrene produces 2-nitropyrene (2-NP) (Pitts et al,,
1985a), In contrast, the electrophilic nitration reaction of fluoranthene or pyrene involving the
NO, ion produces mainly 3-nitrofluoranthene (3-NF) from fluoranthene and 1-NP from pyrene.
       The gas-phase reactions of N2O5 with naphthalene,  1- and 2-methylnaphthalene, .
acenaphthene, phenanthrene, anthracene, fluoranthene, and pyrene yield, 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-NF is produced from both OH radical and N2O5 gas-phase fluoranthene reactions, but
the reaction with N2O5 produces a much higher yield. The postulated reaction pathway involves
initial NO3 radical addition to the most reactive position, followed by NO2 addition to the
neighboring position and elimination of HNO3 (Zielinska et al., 1986):
        2/1/98
                                        2-39
               DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
                                              NO,
                                      H  ONO,
                                                                            NO,
       Table 2-13 summarizes the nitroarene product data for OH radical-initiated and N2O5 gas-
phase reactions with several PAHs studied to date in environmental chambers (Arey et al., 1989;
Zielinska et al., 1990).  Section 2.5.3 will discuss the fact that, generally, the same nitro-PAH
isomers that are formed from OH radical and N2O5 reactions are observed in ambient air samples.

2.4.2.2. Particulate-Phase Reactions
       Organic compounds present in diesel exhaust are partitioned into the particulate phase
under atmospheric conditions. The following chemical processes are likely to contribute to the
degradation of these compounds in the troposphere (Atkinson, 1988):
       •   Photolysis,

       «   Reaction with O3,

       •   Reaction with N2O5 during nighttime hours,

       *   Reaction with NO2 during nighttime and daytime hours,

       «   Reaction with H,O2, and

       •   Reaction with HNO3, HNO2, and H2SO4.
       However, the atmospheric lifetimes of particle-bound organic compounds are not well
known, partly because (1) these chemical processes depend on the nature of the substrate (e.g.,
Behymer and Hites, 1985,1988) and (2) many of the laboratory studies have been done using
atmospherically unrealistic adsorbents, such as glass-fiber and Teflon-coated glass-fiber filters,
        2/1/98
                                        2-40
DRAFT-DO NOT CITE OR QUOTE

-------
  1      silica gel, and alumina. The extrapolation of sometimes contradictory results reported by
  2      different laboratories to atmospherically realistic conditions presents major problems.
  3             The atmospheric fate of particle-bound PAHs has received much attention since their
  4      potential toxicity was first observed. In their recent publication, Behymer and Hites (1988)
  5      define two opposite schools of thought on this subject. One says that particle-bound PAHs
  6      degrade quickly in the atmosphere with lifetimes as short as a few hours (e.g., Kamens et al.,
  7      1988; Nielsen, 1988; Behymer and Hites, 1988). The other says that PAHs degrade slowly, if at
  8      all, in the atmosphere and eventually deposit on soil or water. The latter conclusion is supported
  9      by the studies of marine and lacustrine sediments (the ultimate environmental sinks of PAHs)
10      that have shown that the relative abundances of PAHs, even at the most remote locations, are
11      similar to those in combustion sources and in air particulate matter (Laflamme and Hites, 1978;
12      Hites et al., 1980; McVeety and Hites, 1988).
13
14      2.4.2.2.1. Photooxidation of particulate polycyclic aromatic hydrocarbons. Laboratory studies
15      of photolysis of PAHs adsorbed on 18 different fly ashes, carbon black, silica gel, and alumina
16      (Behymer and Hites, 1985,1988) and several coal stack ashes (Yokley et at, 1986;  Dunstan et
17      al., 1989) showed that the extent of photodegradation of PAHs depended very much on the
18      nature of the substrate to which they are adsorbed.  The dominant factor in the stabilization of
19      PAHs adsorbed on fly ash was the color of the fly ash, which is related to the amount of black
20      carbon present. It appeared that PAHs were stabilized if the black carbon content of the fly ash
21      was greater than =5%. On black substrates, half-lives of PAHs studied were on the  order of
22      several days (Behymer and Hites, 1988).
23             Similar conclusions were reached from studies of photolysis of PAHs adsorbed on coal
24      stack ashes (Yokley et al.,  1986; Dunstan et al., 1989).  The relative quantity of carbon in coal
25      ash was the main factor determining the extent of photochemical degradation of pyrene and
26      benzo[a]pyrene adsorbed on the surface. In addition, in coal ashes that contained a relatively
27      large quantity of iron, the magnetic particles played a minor role in stabilizing adsorbed pyrene
28      toward photodegradation (Dunstan et al., 1989).
29            On the other hand, the environmental chamber studies of Kamens and co-workers (1988)
30      on the daytime decay of PAH  present on residential wood smoke particles and on gasoline
31      internal combustion emission particles showed PAH half-lives on the order of 1 h at moderate
32      humidities and temperatures.  At very low-angle sunlight, very low water-vapor concentration, or
33      very low temperatures, PAH daytime half-lives increased to a period of days.
34            Atmospheric studies by Nielsen (1988), carried out in rural areas during the winter and
35      early spring when ambient temperatures and concentrations of NO2 and O3 were low, showed
36 -     evidence for atmospheric decay of more reactive PAHs, such as benzo[a]pyrene and
        2/1/98                                    2-41        DRAFT-DO NOT CITE OR QUOTE

-------
 1      cyclopenteno[c 8-» 7-^ 1-NF.
24      However, no 2-NF (the sole isomer formed from gas-phase N2O5 reaction) was produced from
25      this adsorbed-phase reaction (Pitts et al., 1985c). It was speculated that N2O5 becomes ionized
26      on the filter surface prior to the reaction with fluoranthene, but the resulting NO2 ion is not "free"
27      nitronium ion, that is, not completely dissociated (Zielinska et al., 1986).
28            Based on these laboratory studies,  it has been proposed that some nitro-PAHs detected in
29      ambient particles may be formed from the reaction  of the parent PAH with gaseous copollutants
30      in the atmosphere, during the collection of particulate matter, or both (Pitts et al., 1978,1985a;
31      Jager and Hanus, 1980; Brorstrom et al., 1983). However, the extrapolation of the data obtained
32      under laboratory conditions to the ambient atmosphere requires several major assumptions.
33      These include, for example, the assumptions that substrate effect, PAH concentration,  the
34      presence of copollutants, relative humidity, etc., have no major impact on PAH nitration
35      reactivities. If these assumptions are valid, the available data indicate that the  nitration of
36 •     particle-bound PAH with NO2/HNO3 and N2O5 is probably not significant under atmospheric
        2/1/98                                   2-42       DRAFT-DO NOT CITE OR QUOTE

-------
 1      conditions (Pitts et al,, 1985c). However, this may not always be the case in air sheds that have
 2      high NO2 and nighttime N2O5 concentrations.
 3          .  The formation of nitro-PAHs during sampling may be an important problem for diesel
 4      participate matter collection because of the presence of NO2 and HNO3. However, Schuetzle
 5      (1983) concluded that the artifact formation of 1-NP during dilution tube sampling accounts for
 6      less than 10% to 20% of the total 1-NP present in diesel particles if the sampling time is less than
 7      23 min (one FTP cycle) and if the sampling temperature is not higher than 43 °C.
 8            The formation of nitroarenes during ambient high-volume sampling conditions has been
 9      reported to be minimal, at least for the most abundant nitropyrene and nitrofluoranthene isomers
10      (AreyetaL, 1988).
11
12      2.4.2.2.3. Ozonolysis of particulate polycyclic aromatic hydrocarbons. Numerous laboratory
13      studies have shown that PAHs deposited on combustion-generated fine particles and on model
14      substrates undergo  reaction with O3 (e.g., Katz et al., 1979; Pitts et al., 1980,1986; Van Vaeck
15      and Van Cauwenberghe, 1984; Finlayson-Pitts and Pitts, 1986),  The dark reaction of several
16      PAHs deposited on model substrates toward O3 has been shown to be  relatively fast under
17      simulated atmospheric conditions (Katz et al., 1979; Pitts et al., 1980,1986). Half-lives of the
18      order of one to several hours were reported for the more reactive PAHs, such as B[a]P,
19      anthracene, and benz[a]anthracene (Katz et al., 1979).
20            The reaction of PAH deposited on diesel particles with 1.5 ppm O3 under high-volume
21      sampling conditions has been shown to be relatively fast, and half-lives of the order of 0.5 to  1 h
22      have been reported for most PAHs studied (Van Vaeck and Vari Cauwenberghe, 1984). The
23      most reactive PAHs include B[a]P, perylene, benz[a]anthracene, cyclopenta[c
-------
 1      2.4.3.  Physical Removal Processes
 2      2.4.3.1. Dry Deposition
 3            Dry deposition is the removal of particles and gases from the atmosphere through the
 4      delivery of mass to the surface by nonprecipitation atmospheric processes and the subsequent
 5      physical attachment to, or chemical reaction with, surfaces such as vegetation, soil, water, or the
 6      built*environment (Dolske and Gatz, 1985). It should be noted that the surface itself may be wet
 7      or dry; the term "dry deposition" refers to the mechanism of transport to the surface, not to the
 8      nature of the surface itself.  Dry deposition plays an important role as a removal mechanism of
 9      pollutant in the absence of precipitation. Even in remote locations such as Siskiwit Lake, located
10      on a wilderness island in northern Lake Superior, the dry deposition of aerosol was found to
11      exceed the wet removal mechanism by an average ratio of 9:1 (McVeety and Kites, 1988),
12            For particles, the deposition velocities depend on the particle size, exhibiting a minimum
13      for particles of mean diameter of =0.1 to ~\ um.
14
15      2.4.3.2. Wet Deposition
16            Wet deposition encompasses all processes by which airborne pollutants are transported to
17      the Earth's surface in aqueous form (i.e., in rain, snow, or fog). The mechanisms of wet removal
18      from the atmosphere may be very different for particle-associated compounds and for gas-phase
19      compounds. However, because many organic compounds are partitioned between the aerosol
20      and vapor phase, processes of both gas and particle scavenging may be important for a given
21      compound (Ligocki et al., 1985a,b; Bidleman, 1988). When there is no exchange of material
22      between the particulate and dissolved phases in the rain, the total scavenging of a given
23      compound can be expressed as (Pankow et al., 1984):
24
25                                      W = Wg (1 -) +  WpcJ>
26                                                         .
27      where W  is the overall scavenging ratio:
28                                     '    w     [rain, total]
29                                               [air, total]
30      Wg is the  gas scavenging ratio:
31                                      w  -  [ra*n»  dissolved]
32                                        g       [air,  gas]
33      Wp is the  particle scavenging ratio:
34                                            [rain, particulate]
35                                        p     [air, particulate]
36      and (j> is the fraction of the atmospheric concentration that is associated with particles.
        2/1/98                                   2-44       DRAFT-DO NOT CITE OR QUOTE

-------
 1            Particle scavenging is a complex process that depends on the meteorological conditions in
 2      the cloud as well as the size and chemical composition of the aerosol particles. The simplest
 3      model for in-cloud particle scavenging involves nucleation scavenging followed by coalescence
 4      or growth of the cloud droplet into raindrops.
 5    •        In cold clouds, ice crystals grow by vapor accretion and by collection of supercooled
 6      droplets (riming). Scavenging ratios may be considerably lower than 106 under these conditions.
 7      In the case of below-cloud scavenging, Wp values have been estimated to be 103 to 105 for 0.01 to
 8      1.0 um particles (Slinn et al., 1978). From these data, one may expect to observe overall particle
 9      scavenging ratios in the range of 103 to  106.
10            Ligocki and co-workers (Ligocki et al., 1985a,b) measured gas- and particle-scavenging
11      ratios for a number of organic compounds, including PAHs and their derivatives.  Table 2-14
12      gives mean gas, particle, and overall scavenging ratios for measured neutral organic compounds.
13      It can be seen from this table that particle scavenging ratios range from 102 to 10s, whereas gas
14      scavenging ratios range from 22 to 105.  Gas scavenging dominates over particle scavenging for
15      compounds of lower molecular weights (MW <252 for PAHs).  Particle scavenging dominates
16      for the alkanes, which are essentially insoluble in water.
17            The complexity of liquid-phase inorganic acid formation from gaseous precursors and the
18      problems of acid rain and acid fog are beyond the scope of this chapter and are not discussed here
19      (see Finlayson-Pitts and Pitts [1986] for more information).
20
21      2.5.  ATMOSPHERIC CONCENTRATIONS OF PRIMARY DIESEL EMISSIONS AND
22           THEIR TRANSFORMATION PRODUCTS
23            Most of the data collected on vehicle emissions are, from laboratory studies that used
24      dynamometer/dilution tube measurements. The relevance of these measurements to the
25      atmosphere is always a question, because emissions from vehicles on the road have much higher
26      dilution ratios (»103 vs. 10), are collected at lower temperatures, are composed of a large number
27      of individual vehicle exhausts, have usually experienced longer residence times (seconds to days
28      versus =5 s) before collection or measurement, and, as discussed in Section 2.4, have the
29      opportunity to interact with ambient air  pollutants (including exhausts of other vehicles and
30      vehicle types),
31            Pierson and co-workers (Pierson et al., 1983; Salmeen et al., 1985) conducted field
32      experiments in the Allegheny Mountain Tunnel of the Pennsylvania Turnpike to address this
33      problem. They found that the diesel-produced particulate matter at tunnels was, in general, very
34      similar to that encountered in dilution-tube studies with respect to total particulate matter
        2/1/98                                   2-45       DRAFT-DO NOT CITE OR QUOTE

-------
       Table 2-14. Mean particle, gas, and overall scavenging ratios for neutral
       organic compounds3
Compound
Toluene
1 ,2,4-Trimethylbenzene
Ethylbenzene
m+p-Xy\ene

-------
  1      emission rates, percentage extractables, hydrocarbon molecular weight distribution, HPLC
  2      profiles, particle size distribution, elemental compositions, and extract mutagenicities. However,
  3      these findings did not preclude the possibility of substantial differences in detailed chemical
  4      compositions.  Indeed, the concentration of 1-NP in the extract of particulate samples collected in
  5      the Allegheny Mountain Tunnel was reported to be lower than would be predicted on the basis of
  6      laboratory dilution tube measurements either for diesel or spark-ignition vehicles (Gorse et al.s
  7      1983).
  8             Some data on organic compound concentrations in air sheds heavily affected by motor
  9      vehicle emissions (tunnels, roadsides, etc.) are reviewed in the following text.
10
11      2.5.1.  Volatile Organic Compounds Attributable to Traffic
12             Individual volatile hydrocarbons and aldehydes were measured along a section of U.S.
13      Highway 70 near Raleigh, NC (Zweidinger et al., 1988).  Traffic volume during sampling was
14      determined by visual counting (~ 1,050 ± 10% vehicles per hour in each direction) and was
15      classified into four groups: (1) light-duty, including gasoline and diesel vehicles; (2) heavy-duty
16      gasoline; (3) HDD; and (4) motorcycles.  Typical distributions were 91.5%, 3.2%, 5.1%, and
17      0.2%, respectively.
18             Table 2-15 lists the mean concentrations from four roadsides for selected hydrocarbons
19      and aldehydes, expressed in ppb C and as a percentage contribution of individual hydrocarbons
20      and aldehydes to total nonmethane hydrocarbons (TNMHC) and total aldehydes, respectively.
21             The roadside VOC distribution was compared with dynamometer/dilution tube results on
22      in-use vehicles, which were weighted in an attempt to reflect the same model year distribution as
23      observed on the roadway (Sigsby et al., 1987; see also Table 2-3). The two sets of data were
24      similar in that the different driving cycles, like the different sampling sites, generally show no
25      significant differences in the distribution of hydrocarbons or aldehydes on a percentage of total
26      basis. There were, however, differences observed between the sets of data, particularly for the
27      contribution of combustion products (i.e., hydrocarbons below C4 and aldehydes).  For the
28      roadside study, ethylene, formaldehyde, and acetaldehyde were lower, whereas acetylene was
29      higher than in "the dynamometer study. However, noncatalyst vehicles, which constituted  15% of
30 '     all light-duty vehicles in the roadside study, were not included in the dynamometer study,  nor
31      were LDD and HDD vehicles and tracks.
32
33      2.5.2.  Polycyclic Aromatic Hydrocarbons
34             Particulate and vapor phase samples were collected from the traffic passing through the
35      Baltimore Harbor Tunnel and analyzed for PAHs and related compounds (Benner et al., 1989).
36      High-volume air samplers equipped with Teflon filters backed by PUF plugs were used for
        2/1/98                                    2-47      DRAFT-DO NOT CITE OR QUOTE

-------
       Table 2-15.  Concentrations of individual hydrocarbons and aldehydes measured
       in the Raleigh, NC, roadside study"
Compound
Hydrocarbons
Ethane
Ethylene
Acetylene
Propane
Propylene
n-Butane
1-Butene
n-Pentane
j'so-Pentane
Methylcyclopentane
Methylcyclohexane
n-Decane
Benzene
Toluene
m- and p-Xylenes
o-Xylene
Ethylbenzene
TNMHC"
Total paraffins
Total olefins
Total aromatics
Total unidentified NMHC
Aldehydes
Formaldehyde
Acetaldehyde
Acrolein
Benzaldehyde
Total aldehydes
Concentration (ppbC)

16.30
64.30
50.90
7.90
22.60
15.80
5.70
25.40
53.00
10.40
4.70
3.00
29.00
59:30
53.10
12.70
12.00
900.00
369.20
164.50
252.00
63.60

6.74
3.00
1.20
2.31
16.38
Percent contribution11'0

1.81
7.15
5.65
0.88
2.51
1.75
0.64
2.82
5.89
1.15
0.53
0.33
3.23
6.60
5.90
1.41
1.33
100.00
41.00
18.20
28.00
7.10

1.05
18.40
7.30
3.88
100.00
'From Zweidinger et al. (1988).
'Percent based on ppbC.
'Percent contribution of individual hydrocarbons to TNMHC and of individual aldehydes to total aldehydes.
JTNMHC = Total nonmethane hydrocarbons.
2/1/98                                      2-48        DRAFT-DO NOT CITE OR QUOTE

-------
  1      sample collection. There was no breakdown of traffic into numbers of diesel- and gasoline-
  2      fueled vehicles,
  3             The range of particle-phase PAH concentrations and the mean particle- and vapor-phase
  4      PAH concentrations for 48 samples collected in the tunnel are tabulated in Table 2-16.  The
  5      ratios of mean particle-phase PAH concentrations to that of B[e]P, which is considered to be a
  6      nonreactive PAH, are also given in this table.
  7             As can be seen from Table 2-16, alkyl-substituted phenanthrenes in the tunnel samples
  8      had relatively high concentrations compared with those of the parent compound. This suggests a
  9      significant contribution from diesel vehicle emissions (particularly diesel-fueled trucks) because
10      extracts of diesel particulate matter are known to have significant concentrations of methyl and
11      dimethylphenanthrenes (see Table 2-8 and Yu and Kites, 1981).
12             Factor analysis was applied to the tunnel data in an attempt to identify factors associated
13      with different types of vehicles; two factors were obtained. The alkylated phenanthrenes loaded
14      significantly on factor 1, suggesting the diesel vehicles as the source of these  compounds.
15      Several of the higher-molecular-weight PAHs loaded onto factor 2, which may be associated
16      with the contribution of gasoline-fueled emissions in the tunnel.
17             Ambient air sampling for PAHs was also  conducted during a summertime photochemical
18      air pollution episode in Glendora, CA, at a site situated less than 1 km from the heavily traveled
19      1-210 freeway and generally downwind of Los Angeles; therefore, the site was affected by motor
20      vehicle emissions (Atkinson et al., 1988). Samples were collected by means of high-volume
21      samplers equipped with Teflon-impregnated glass-fiber filters backed by PUF plugs. Table 2-17
22      shows the average (from three daytime and three  nighttime samples) concentrations of PAH
23      measured and the ratios of these concentrations to that of B[e]P. Unfortunately, no alkylated
24      phenanthrenes were measured.
25             As can be seen from the comparison of Tables 2-16 and 2-17, the concentrations of all
26      PAHs measured in Glendora were much lower than those measured in the tunnel, as would be
27      expected. However, the ratios of the concentrations of particle-bound PAH to those of B[e]P
28      were also different for the two sites, usually much lower for the Glendora site (except for higher
29      molecular weight PAHs, indeno[l,2,3-o/]pyrene, benzo[g/zz']perylene, and coronene). This may
30      indicate either contributions from sources other than motor vehicles in the Glendora study or
31      PAH photochemical transformations occurring on particles prior to or during  high-volume
32      sample collections (or both).  The generally higher PAH concentrations for nighttime versus
33      daytime sampling periods (Atkinson et al, 1988) seem to support the latter possibility.  However,
34      the influence of meteorology cannot be excluded. This conclusion is also consistent with high
35      levels of photochemical pollutants observed in Glendora; for example, the daily maxima of O3

        2/1/98                                   2-49       DRAFT-DO NOT CITE OR QUOTE

-------
        Table 2-16.  Particle- and vapor-phase polycyclic aromatic hydrocarbon
        concentrations for Baltimore Harbor Tunnel samples"

Compound
Phenanthrene
Anthracene
3-Methylphenanthrene
2-Methylphenanthrene
2-Methylanthracene
9-and 4-Methyliphenanthrene and
4H-cyclopenta[de/|-phenanthrene
1 -Methylphenanthrene
2 , 6-Dimethylphenanthrene
2 , 7-Dimethylphenanthrene
1,3-, 2,10-, 3,9-, and 3,10-
Dimethyl- and phenanthrene
1,6- and 2,9-Dimethylphenanthrene
1 , 7-Dimethylphenanthrene
2,3-Dimethylphenanthrene
Fluoranthene
Pyrene
Benzo [ghi] fluoranthene
Cyclopenta[oflpyrene
Benz[a]anthracene
Chrysene/triphenylene
Benzofluoranthenes[£,_/, +k]
Benzo[e]pyrene
Benzo[a]pyrene
lndeno[l,2, 3-oflpyrene
Benzo [ghi] pery lene
Coronene

Range,
particles
4.3-56
0.6-12
. 3.9-58
5.3-74
0.6-12
4.7-50

2.6-43
4.7-62
3.4-38
9.5-119

4.5-63
3.9-41
3.5-41
6.4-69
9.7-76
3.2-26 .
7.6-65
1.9-29
2.9-47
2.2-44
1.5-19
1.3-26
0.3-15
1.8-18
1.0-10
Concentration
Mean,
particles
18.0
2.9
13.9
19.0
3.0
12.9

9.8
14.0
9.2
26.0

14.0
10.2
9.3
20.0
27.0
9.6
20.0
7.6
12.0
10.6
5.0
5.8
4.6
8.0
4.7
(ng/m3)
Mean,
vaporb
132
18
70
—
5.3
71

43
30
16
61

27
20
16
16
26
NDd
ND
ND
ND
ND
ND
ND
ND
ND
ND

Ratioc
to B[e]P
4.3
0.6
3.3
4.6
0.7
3.0

2.3 .
3.4
2.2
6.3

3.3
2.4
2.2
4.5
6.3
2.1
4.6
1.5
2.4
2.1
1.0
1.1
0.9
1.6
0.9
 "From Benner et al. (1989).
- bMean concentrations of PAH collected on PUF plug (calculated from data given in Table III of Benner
 etal., 1989).
 'Mean ratios to paniculate phase B[e]P.
 JNone detected.
 2/1/98
2-50
DRAFT-DO NOT CITE OR QUOTE

-------
               Table 2-17. Average ambient concentrations of poly cyclic aromatic hydrocarbons
               measured in Glendora, CAa
PAH
Phenanthrene0
Anthracene0
Fluoranthene
Pyrene
Benzoff hi] -fluoranthene
Cy clopenta[ofl -py rene
Benz[a]anthracene
Chrysene/Triphenylene
Benzofluoranthenes [bj + k]
Benzo[e] pyrene
Benzo[a]pyrene
Indenofjf , 2, 3-af|-pyrene
Benzo[£fe]perylene
Coronene
Total concentration (ng/m3)
20.0
1.0
5.6 (0.26)d
4,l(Q.35)d
0.26
0.09
0.2
1,0
1.6
0.94
0.33
1.6
3.8
2.8
Ratio to B[e]Pb
—
—
0.27
0.37
0.28
0.1
0.22
1.1
1.7
1.0
0.35
1.7
4.0
3.0
        "From Atkinson et al. (1988).
        ""Ratios of particle-phase PAH to particle-phase B[e]P.
        cPhenanthrene and anthracene were not present on filters, only on PUF plugs.
        dFluoranthene and pyrene are distributed between gas and paniculate phases; numbers in parentheses represent
        particle concentrations.
 1      concentrations (which always occurred between 1400 and 1700 hours, Pacific standard time)
 2      ranged from 160 to 240 ppb throughout the entire 9 days of the study.
 3
 4      2.5.3. Nitroarene Concentrations in Ambient Air
 5             Diesel paniculate matter contains a variety of nitroarenes, with 1-NP being the most
 6      abundant among identified nitro-PAHs. The concentration of 1-NP was measured in the extract
 7      of particulate samples collected at the Allegheny Mountain Tunnel on the Pennsylvania Turnpike
 8      (Gorse et al., 1983). This concentration was 2.1 ppm and i5 ppm (by mass) of the extractable
 9      material from diesel and spark-ignition vehicle particulate matter, respectively. These values are
10      much lower than would be predicted on the basis of laboratory dilution tunnel measurements for
11      either diesel or spark-ignition engines.
12             Several nitroarene measurements were conducted in air sheds heavily affected by motor
13      vehicle emissions (Arey et al., 1987; Atkinson et at, 1988; Zielinska et al., 1989a,b; Ciccioli et
14      al., 1989). For example, ambient particulate matter samples were collected at three sites
15      (Claremont, Torrance, and Glendora) in the Los Angeles Basin; the Claremont and Glendora
16      sites are = 30 km and =20 km northeast, respectively, and the Torrance site is -20 km southwest

        2/1/98                                    2-51       DRAFT-DO NOT CITE OR QUOTE

-------
 1      of downtown Los Angeles (Arey et al, 1987; Atkinson et al., 1988; Zielinska et al., 1989a,b).
 2      The sampling was conducted during two summertime periods (Claremont, September 1985, and
 3'     Glendora, August 1986) and one wintertime period (Torrance, January and February 1986),
 4      Table 2-18 lists the maximum concentrations of nitropyrene and nitrofluoranthene isomers
 5      observed at these three sites during the daytime and nighttime sampling periods.

 6             As can be seen from Table 2-18,1 -NP, the most abundant nitroarene emitted from diesel

 7      engines, is not the most abundant nitroarene observed in ambient particulate matter collected at

 8      three sites heavily affected by motor vehicle emissions. Of the two nitropyrene isomers present,

 9      2-NP, the main nitropyrene isomer formed from the gas-phase OH  radical-initiated reaction with
10      pyrene (see Section 2.4.2.1), is sometimes more abundant. The 2-NF was always the most
11      abundant nitroarene observed in ambient particulate matter collected at these three sites (Ciccioli
12      et al., 1989), and this nitrofluoranthene isomer is not present in diesel and gasoline vehicle
13      emissions. The 2-NF is the only nitroarene produced from the gas-phase OH radical-initiated
               Table 2-18. Maximum concentrations of nitrofluoranthene (NF) and
               nitropyrene (NP) isomers observed at three South Coast
               Air Basic sampling sites
Concentration (pg/m3) at
Nitroarene, collection period
2-NF, day
2-NF, night
3-NF, day
3-NF, night
8-NF, day
8-NF, night .
1-NP, day
1-NP, night.
2-NP, day
2-NP, night-
Claremont">b
40
1,700
3
=3
2
' . 2
3
10
1
8
GlendoraM
350
2,000
NDf
ND
3
4
15
15
14
32
Torrance**
410
750
=3
70
8
50.
60
50
50
60
        'From Zielinska et al. (I989b).
        'Daytime sample collected from 1200 to 1800 hours and nighttime sample from 1800 to 2400 hours on
        September 13, 1985.
        'From Atkinson et al. (1988),
        JDaytime sample collected from 0800 to 2000 hours on August 20, 1986, and nighttime sample from 2000 to
        0800 hours on August 20 and 21, 1986.
        "Daytime sample collected from 0500 to 1700 hours on January 28. 1986, and nighttime sample from 1700 to
        0500 hours on January 27 and 28, 1986.
        'ND = None detected.


        2/1/98                                    2-52        DRAFT-DO NOT CITE OR QUOTE

-------
  1      N2O5 reactions with fluoranthene (see Sections 2,4.2.1 and 2.4.2.2), whereas mainly 3-NF and
  2      lesser amounts of 1-, 7-, and 8-nitroisomers are present in diesel participate matter and are
  3      produced from the electrophilic nitration reactions of fluoranthene. Figure 2-4 compares the
  4      nitroarenes formed from the OH radical-initiated reaction of fluoranthene and pyrene in an
  5      environmental chamber (upper trace) with the ambient samples collected at Torrance (lower
  6      trace).  It is very unlikely that N205 could have been present during the nighttime winter
  7      collections in Torrance, given the high level of NO present at sunset.  More likely, a relatively
 . 8      high level of OH radicals was present because of the measured high concentration of HNO2,
  9      which photolyzes to yield OH radicals.  This suggests that all isomers observed in Figure 2-4
 10      (lower trace), with the exception of 1-NP, are the product of the OH radical-initiated reactions of
 11      the parent PAH.  Direct emissions may account for the 1-NP (and 3-NF) observed at relatively
 12      low levels in these ambient samples. (See Zielinska et al. [1989b] for full discussion of all the
 13      MW 247 nitroarenes observed in ambient particles.)
 14             Although the reaction with OH radicals is the major atmospheric loss process for gas-
 15      phase fluoranthene and pyrene (Table 2-12), evidence for atmospheric formation of 2-NF from
 16      N2O5 reaction with fluoranthene has also been reported (Zielmska et al., 1989b). Because the 2-
 17      NF/2-NP yield ratio for N2O5 reactions, observed from environmental chamber experiments, is,
 18      > 100, compared to  ~ 10 for the OH radical reaction (Table 2-4), the high 2-NF/2-NP
 19      concentration ratio in ambient samples suggests a contribution from the N2O5 reaction with
20      fluoranthene.  Figure 2-5  shows a comparison of a wintertime sample collected in Torrance
21      (upper trace) with a summertime sample collected in Claremont (lower trace). The 2-NF/2-NP
22      ratio reached =200 for the summer night sample.  The N2O5 concentration was calculated to be 5
23      ppb for this night, which supports the suggested formation route of 2-NF via reaction with N2O5
24      (Zielinska etal, 1989b).
25             The evidence presented in the preceding text, as well as the observation that 2-NF has
26      been the most abundant MW 247 nitroarene in ambient samples collected worldwide  (Ramdahl
27      et al., 1986), strongly suggests that the atmospheric formation from the parent PAH, not the
28      direct automotive emissions, is the major source of these nitroarenes in ambient air. However,
29      under certain sampling conditions, when ambient particulate matter is collected  very close to
30      emission sources, the MW 247 nitroarene profile may be different. For example, in urban
31      samples collected during wintertime rush hours at a central square in Rome, Italy, at a height of
32      1.5 m above street level, 2-NF and 2-NP were not observed (Ciccioli et al.,  1989).
        2/1/98                                    2-53       DRAFT-DO NOT CITE OR QUOTE

-------
^
o>
o
       1OO
                                 2-NF
                m/z 247
                       Products of the
                       OH  Radical Reaction
                       with Fluoranthene
                       and P y re n e
                        7-NF
                           A
                                                     2-NP
                   8-NF
                     A
             23
              25
              26
              27
       1OO
       2-NF
               m/z 247
                                                  Ambient Sample
                                                    2-NP
                                               1-NP
                         7-NF
                 8-NF
                 ^    t\
              21
22
23
24
                                                                    25
Figure 2-4.   Mass chromatograms of the molecular ion of the nitrofluoranthenes (NF)
            and nitropyrenes (NP) formed from the gas-phase reaction of fluoranthene
            and pyrene with the OH radicals (top) and present in the ambient participate
            sample collected at Torrance, CA (bottom).

Source: Arey et al. (1989).
2/1/98
           2-54
       DRAFT-DO NOT CITE OR QUOTE

-------
     100
             m/z 247
                                2-NF         Winter  Night
 o>
 o
 i=
 CO
 CO

 o>
                    7-NF
             m/z 247
3-NF
       8-N F
                 2-NP
2-NF        Summer Night
                                        1-NP
                                                 2-NP

                      Relative Retention  Time
Figure 2-5.   Mass chromatograms of the molecular ion of the nitrofluoranthenes (NF)

           and nitropyrenes (NP) present in ambient particulate samples collected in

           Torrance, CA (top), and Claremont, CA (bottom).
2/1/98
  2-55      DRAFT-DO NOT CITE OR QUOTE

-------
 1      2.6. EXPOSURE PERSPECTIVE
 2            Diesel emissions are complex mixtures containing thousands of organic and inorganic
 3      constituents in both gas and particulate phases and with differing chemical reactivities. After
 4      entering the atmosphere, they are transported and transformed according to their distinctive
 5    •  characteristics, undergoing physical and chemical changes that may form secondary pollutants
 6      more harmful than their predecessors. Thus, a knowledge of diesel emissions at or near their
 7      sources is not sufficient to fully assess the impact of these emissions on human health and
 8      welfare. However, data on how diesel exhaust contributes to exposure levels for these secondary
 9      pollutants are currently lacking.
10            Determining the amount of diesel exhaust present in the ambient air is also complicated
11      by the difficulty of distinguishing organic compounds and particles that originate in diesel
12      engines from those that originate in gasoline engines or come from other sources. This, too,
13      cannot be accomplished at present because of the lack of a sufficient research base.
14            Nonoccupational exposure to diesel exhaust is worldwide in urban areas, with lesser
15      exposure in rural areas. Certain working populations are also exposed to higher levels of diesel
16      exhaust than the rest of the population. The level of exposure will differ within geographic areas
17      based on the number and types of diesel engines in the area, as well as atmospheric patterns of
18      dispersal and the location of the individual relative to the emission sources.
19            While a detailed exposure assessment for DE has not been conducted as part of this study,
20      the following exposure data are provided to  give some context for the hazard assessment and
21      dose-response analysis.
22            Estimates of annual average concentrations of particulate matter in the ambient air,
23      published in Chapter 9 of EPA's Motor Vehicle-Related Air Toxics Study (U.S. EPA, 1993),
24      may be used to generate a crude estimate of the concentration of particulates from diesel exhaust
25      in the ambient air. The Air Toxics study used two approaches to generate exposure estimates.  In
26      the first one, DPM (diesel particulate matter) national fleet average emissions factors for 1988
27      were multiplied by the urban and rural grams-per-meter conversion  factors obtained from EPA's
28      hazardous air pollution exposure model (HAP-EM: 1988). Based on this approach, the total
29      concentration of particulates from diesel exhaust in ambient air in urban areas for 1995 was
30      estimated as 2 |ig/m3; the concentration in rural areas as  0.6 |ig/m3; and the nationwide average
31      concentration as 1.1 |ig/m3.
32             In a second method, using ambient monitoring data, total suspended particulate matter
33      (TSP) for 1990 was determined to equal 48  |ig/m3. Approximately 5% of total particulate matter
34      is associated with diesel exhaust.  Multiplication of the total by the fraction contributed by diesel
35      exhaust, and adjusting for time spent indoors, results in an integrated estimate of 1.5 |ig/m3 (U.S.
36      EPA, 1993).
        2/1/98                                   2-56        DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
       Exposure estimates for more highly polluted locations are somewhat greater.  Estimated
mean concentrations of DPM for Los Angeles were reported to be 2.7 ug/m3 (Sienicki and Mago,
1992).  McClellan at al. (1986) estimated concentrations on urban freeways and street canyons to
be as great as 15  ug/m3. These exposure estimates are summarized in Table 2-19.
       Recent Cal-EPA (1996) studies show winter period estimates in three California locations
for diesel PM10 of 4 to 22 ug/m3. A broader Cal-EPA analysis shows average ambient outdoor
diesel PM,0 to range from 0.2 to 3.6 |ig/m3 across 14 California air basins with a population-
weighted average of 3.2 |ig/m3. Concentrations in occupational settings may be higher.
       Potentially, diesel exhaust may contribute to levels of a very hazardous pollutant, dioxin.
Only scant information is available to quantify this, though some study data are available.
Dioxin concentrations were measured in a Baltimore tunnel used only by heavy trucks
(unpublished). Measurements were also made of the exhaust from a heavy-duty track during
highway travel.  Based on these measurements, estimated dioxin production from all truck
diesels was estimated to be only 29 gm versus 5,000 gm from all combustion sources in the
United States. It is therefore concluded  that DE contribution to dioxin levels in the United States
is insignificant.
       The changing composition of DE, (i.e., older engines vs. newer technology ones, heavy-
duty vs. light-duty, and engines run under varied operating conditions) gives rise to questions
about how the health data and the risk assessment findings in this report, which are based on pre-
1998 engines, can be applied to present-day engine exhaust emissions and the resulting ambient
              Table 2-19. Estimated annual ambient concentrations of diesel exhaust participate
              matter (jig/m3)
Year
1986
1990
1995
2000
2010
U.S. EPA
(1993)
Method 1
Rural
.
1.1
0.6
0.4
0.2
U.S. EPA
(1993)
Method 1
Urban

2.0
1.2
0.7
0.4
U.S. EPA
(1993)
Method 1
National

1.8
1.1
0.6
0.4
U.S. EPA
(1993)
Method 2
National

1.5



Sienicki
and Mago
(1992)
Los Angeles

2.7



McClellan
(1986)
Highly
Exposed ,
15.0




        2/1/98
                                         2-57
DRAFT-DO NOT CITE OR QUOTE

-------
 1      exposures. This is a complex question that is not rigorously addressed in this version of the
 2      assessment. It is clear that newer technology engines will have somewhat different emission
 3      composition (i.e., perhaps reduced NOX with increased fine particles), not to mention the
 4      emission controls, which would reduce certain exhaust components, presumably larger particles.
 5      Since particle mass is the surrogate dosimeter used to correlate toxicity with exposure and public
 6      health impact, as the particle mass is changed by virtue of new technology or controls, so might
 7      the applicability of the health assessment findings in this report. Further analysis of emission
 8      changes may be a desirable research pursuit.
 9                   •    •
10      2.7. SUMMARY
11            Major research programs were carried out in the late 1970s and early  1980s to ascertain
12      the physical and chemical characteristics of emissions from diesel engines and the biological
13      effects of these emissions. New control technologies are being introduced into currently
14      manufactured diesel vehicles, and the effect of these changes on diesel emissions is likely to be
15      visible in the future. Diesel vehicles manufactured in the late 1970s and early 1980s are still on
16      the road and, in this sense, data collected from that period are still valid.
17            However, many of these data were collected using laboratory dynamometers with selected
18      new vehicles or vehicles well tuned to manufacturers' specifications. The well-controlled
19      conditions of the dynamometer tests have many benefits but do not necessarily represent vehicle
20      emissions under real on-road conditions, and the small number of vehicles tested in the
21      laboratory is not truly representative of the distribution within the on-road vehicle fleet.
22      Although several roadway and tunnel emission measurements were performed in the past, the
23      database on mobile-source emission rates necessary to assess the role of vehicle emissions in air
24      pollution problems is  still not sufficient.  More measurements carried out under realistic on-road
25      conditions are necessary, in particular for gaseous and particulate-phase organic compounds
26      present in vehicle emissions.
27            Once released into the atmosphere, diesel emissions are subject to dispersion and
28      transport and, at the same time, to chemical and physical transformation into secondary
29      pollutants, which may be more harmful than their precursors.  Thus, a knowledge of diesel
30 '    emissions at or near their sources is no longer sufficient to assess fully the impact of these
31      emissions on human health and welfare.  The understanding of physical and chemical changes
32      that primary diesel emissions undergo during their transport through the atmosphere is equally
33      important. As a result of the past two decades of laboratory and ambient experiments and
34      computer modeling, a comprehensive set of data now exists concerning the atmospheric loss
35      processes and transformation of automotive emissions, but knowledge concerning the products of
36      these chemical transformations is still limited. Study is required to determine the products from
        2/1/98                                   2-58       DRAFT-DO NOT CITE OR QUOTE

-------
  1       the OH radical-initiated reactions of the aromatic and aliphatic hydrocarbons from automobile

  2       emissions. The atmospheric transformation products of PAHs and their oxygen-, sulfur-, and

  3       nitrogen-containing analogs require study in the gaseous and adsorbed phases. In particular, the

  4       reactions occurring in adsorbed phases on atmospherically relevant surfaces are poorly

  5       understood and require further study.  In addition, gas-to-particle conversion processes and the

  6       chemical processes that lead to aerosol formation should be further investigated.

  7              The quantitation of the contribution of diesel emissions to total ambient aerosol mass

  8       concentration is not possible without developing a specific profile for diesel emissions, a

  9       "fingerprint" that may be used in receptor source apportionment models. The data indicate that it

10       may be possible to use PAHs and/or alkylated PAHs, alkanes, and possibly certain unique

11       compounds to assist in distinguishing between diesel and other pollutant sources. However, the

12       available data are not adequate for use in receptor modeling, and study is required to determine

13       the profile of diesel emissions by using sampling and analytical methods appropriate to receptor

14       modeling.

15

16       2.8. REFERENCES

17
18       Alkidas, AC. (1988) Effects of injector-tip configuration on the performance and emissions of an uncooled diesel.
19       Presented at: International fuels and lubricants meeting and exposition; October; Portland, OR. Warrendale, PA:
20       Society of Automotive Engineers, Inc.; SAE technical paper no. 881613.
21
22       Alsberg, T; Stenberg, U; Westerholm, R; Strandell, M; Rannug, U; Sundvall, A; Romert, L; Bernson, V; Pettersson,
23       B; Toftgard, R; Franzen, B; Jansson, M; Gustafsson, JA; EgebSck, ICE; Tejle, Q. (1985) Chemical and biological
24       characterization of organic material from gasoline exhaust particles. Environ Sci Technol 19:43-50.
25
26       Arey, J; Zielinska, B; Atkinson, R; Winer, AM; Ramdahl, T; Pitts, JN, Jr. (1986) The formation of nitro-PAH from
27       the gas-phase reactions of fluoranthene and pyrene with the OH radical in the presence of NOX. Atmos Environ
28       20:2339-2345.
29
30       Arey, J; Zielinska, B; Atkinson, R; Winer, AM. (1987) Polycyclic aromatic hydrocarbon and nitroarene
31       concentrations in ambient air during a wintertime high-NOx episode in the Los Angeles basin. Atmos Environ
32       21:1437-1444.
33
34       Arey, J; Zielinska, B; Atkinson, R; Winer, AM. (1988) Formation of nitroarenes during ambient high-volume
35       sampling. Environ Sci Technol 22:457-462.
36
37       Arey, J; Zielinska, B; Atkinson, R; Aschmann, SM.  (1989) Nitroarene products from the gas-phase reactions of
38       volatile polycyclic aromatic hydrocarbons with OH radical and N2O5. Int J Chem Kinet 21:775-799.
39
40       Atkinson, R. (1988) Atmospheric transformations of automotive emissions. In: Watson, AY; Bates, RR; Kennedy, D,
41       eds. Air pollution, the automobile, and public health. Washington, DC: National Academy Press; pp. 99-132.
42
43       Atkinson, R. (1990) Gas-phase tropospheric chemistry of organic compounds: a review. Atmos Environ Part A 24:1-
44       41.
45
        2/1/98                                       2-59        DRAFT-DO NOT CITE OR QUOTE

-------
 1       Atkinson, R; Carter, WPL. (1984) Kinetics and mechanisms of the gas-phase reactions of ozone with organic
 2       compounds under atmospheric conditions. Chem Rev 84:437-470.
 3
 4       Atkinson, R; Arey, J; Zielinska, B; Aschmann, SM. (1987) Kinetics and products of the gas-phase reactions of OH
 5       radicals and N2O5 with naphthalene and biphenyl. Environ Sci Technol 21:1014-1022.
 6
 7       Atkinson, R; Arey, J; Winer, AM; Zielinska, B. (1988) A survey of ambient concentrations of selected polycyclic
 8       aromatic hydrocarbons (PAH) at various locations in California [final report]. Sacramento, CA: California Air
 9       Resources Board; contract no. A5-185-32.
10
11       Atkinson, R; Arey, J; Zielinska, B; Aschmann, SM. (1990) Kinetics and nitro-products of the gas-phase OH and NO3
12       radical-initiated reactions of naphthalene-d8, fluoranthene-d10 and pyrene. Int J Chem Kinet 22:999-1014.
13
14       Bailey, JC; Schmidl, B; Williams, ML. (1990) Speciated hydrocarbon emissions from vehicles operated over the
15       normal speed range on the road. Atmos Environ Part A 24:43-52.
16
17       Ball, DJ. (1987) Particulate carbon emissions and diesel vehicles. In: Vehicle emissions and their impact on
18       European air quality: proceedings of the Institution of Mechanical Engineers international conference; November;
19       London, United Kingdom. London, Mechanical Engineering Publications Limited, pp. 83-88.
20
21       Balzotti, A; Cornetti, GM; Pidello, F; Signer, M; Scorsone, V. (1990) Italian city buses with particulates traps.
22       Presented at: International congress and exposition; February-March; Detroit, MI. Warrendale, PA: Society of
23       Automotive Engineers, Inc.; SAE technical paper no. 900114.
24
25       Bayona, JM; Markides, KE; Lee, ML. (1988) Characterization of polar polycyclic aromatic compounds in a heavy-
26       duty diesel exhaust particulate by capillary column gas chromatography and high-resolution mass spectrometry.
27       Environ Sci Technol 22:1440-1447.
28
29       Beck, NJ; Uyehara, OA. (1987) Factors that affect BSFC and emissions for diesel engines: Part II. Experimental
30       confirmation of concepts presented in part I. Presented at: SAE International Congress and Exposition; February;
31       Detroit, MI. Warrendale, PA: Society of Automotive Engineers, Inc.
32
33       Behymer, TD; Kites, RA. (1985) Photolysis of polycyclic aromatic hydrocarbons adsorbed on simulated atmospheric
34       particulates. Environ Sci Technol 19:1004-1006!
35
36       Behymer, TD; Kites, RA. (1988) Photolysis of polycyclic aromatic hydrocarbons adsorbed on fly ash. Environ Sci
37       Technol22:1311-1319.
38
39       Benner, BA, Jr; Gordon, GE; Wise, SA. (1989) Mobile sources of atmospheric polycyclic aromatic hydrocarbons: a
40       roadway tunnel study. Environ Sci Technol 23:1269-1278.
41
42       Bidleman, TF. (1988) Atmospheric processes. Environ Sci Technol 22:361-367.
43
44       Bj0rseth, A; Olufsen, BS. (1983) Long-range transport of polycyclic aromatic hydrocarbons. In: Bjeirseth, A, ed.
45       Handbook of polycyclic aromatic hydrocarbons. New York: Marcel Dekker, Inc.; pp. 507-524.
46
47       Black, F; High, L. (1979) Methodology for determining particulate and gaseous diesel hydrocarbon emissions.
48       Warrendale, PA: Society of Automotive Engineers, Inc.; SAE technical paper no. 790442.
49
50      Black, F; Braddock, J; Bradow, RL; Ingalls, M. (1985) Highway motor vehicles as sources of atmospheric particles.
51       Environ Int 11:1-32.
52
53       Boone, PM; Macias, ES. (1987) Methyl alkanes in atmospheric aerosols. Environ Sci Technol 21:903-909.
54
55       Bradow, RL. (1980) Diesel particle emissions. Bull NY Acad Med 56:797-811.


         2/1/98                                          2-60        DRAFT-DO NOT CITE OR QUOTE

-------
  1      Bradow, RL. (1982) Diesel particle and organic emissions: engine simulation, sampling, and artifacts. In: Lewtas, J.,
  2      ed. Toxicological effects of emissions from diesel engines: proceedings of the Environmental Protection Agency
  3      1981 diesel emissions symposium; October 1981; Raleigh, NC. (Developments in toxicology and environmental
  4      science: v. 10). New York: Elsevier Biomedical,pp. 33-47.
  5
  6      Brorstrom, E; Grennfelt, P; Lindskog, A. (1983) The effect of nitrogen dioxide and ozone on the decomposition of
  7      particle-associated polycyclic aromatic hydrocarbons during sampling from the atmosphere. Atmos Environ 17:601-
  8      605.
  9
 10      Broukhiyan, EMH; Lestz, SS. (1981) Ethanol fumigation of a light duty diesel engine. In: SP-503 alternate fuels for
 11      diesel engines. Warrendale, PA: Society of Automotive Engineers, pp.  15-25; SAE technical paper no. 811209.
 12
 13      California Environmental Protection Agency (Cal-EPA), (1996) Health risk assessment for diesel exhaust, public
 14      and scientific panel review draft; Office of Environmental Health Hazard Assessment, California Environmental
 15      Protection Agency.
 16
 17      Campbell, RM; Lee, ML. (1984) Capillary column gas chromatographic determination of nitro polycyclic aromatic
 18      compounds in paniculate extracts. Anal Chem 56:1026-1030.-
 19
 20      Carey, P; Cohen, J. (1980) Comparison of gas phase hydrocarbon emissions from light-duty vehicles equipped with
 21      diesel engines. U.S. Environmental Protection Agency technical report no, CTAB/PA/80-5.
 22
 23      Cass, OR. (1990) Risk assessment of diesel exhaust: 1990  and beyond. Presented at a workshop, March, Los
 24      Angeles, CA. Sacramento, CA: California Air Resources Board.
 25                                                     .
 26      Cass, GR; McRae, GJ. (1983) Source-receptor reconciliation of routine air monitoring data for trace metals: an
 27      emission inventory assisted approach. Environ Sci Technol 17:129-139.
 28
 29      Cautreels, W; Van Cauwenberghe, K. (1978) Experiments on the distribution of organic pollutants between airborne
 30      particulate matter and  the corresponding gas phase. Atmos Environ 12:1133-1141.
 31
 32      Ciccioli, AP; Cecinato, A; Brancaleoni, E; Draisci, R; Liberti, A. (1989) Evaluation of nitrated polycyclic aromatic
 33      hydrocarbons in anthropogenic emission and air samples: a possible means of detecting reactions pf carbonaceous
 34      particles in the atmosphere: Aerosol Sci Technol 10:296-310.
 35
 36      Clark, DH; Johnson, DM; Swedberg, SE. (1984) Modem engine oils. Warrendale, PA: Society of Automotive
 37      Engineers, Inc.; SAE technical paper no. 841207; SAE SP-603.
 38
 39      Coutant, RW; Brown,  L; Chuang, JC; Riggin, RM; Lewis, RG. (1988) Phase distribution and artifact formation in
40      ambient air sampling for polynuclear aromatic hydrocarbons. Atmos Environ 22:403-409.
41
42      Daisey, JM; Cheney, JL; Lioy, PJ. (1986) Profiles of organic particulate emissions from air pollution sources: status
43      and needs for receptor source apportionment modeling. J Air Pollut Control Assoc 36:17-33.
44
45      Dietzmann,  HE; Pamess, MA; Bradow, RL. (1980) Emissions from trucks by chassis version of 1983 transient
46      procedure. Presented at:  Society of Automotive Engineers  fuels and lubricants meeting; October; Baltimore, MD,
47      Warrendale, PA: Society of Automotive Engineers, Inc., paper no. 801371.
48
 49      Dietzmann,  HE; Parness, MA; Bradow, RL. (1981) Emissions from gasoline and diesel delivery tracks by chassis
 50      transient cycle. Presented at: ASME energy-sources technology conference and exhibition; January; Houston, TX.
 51      New York: American  Society of Mechanical Engineers, ASME paper no. 81-DGP-6.
 52
 53  •    Dolske, DA; Gatz, DF. (1985) A field intercomparison of methods for the measurement of particle and gas dry
 54      deposition. J Geophys Res 90:2076-2084.
 55                          ,


         2/1/98                                          2-61        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Dunstan, TDJ; Mauldin, RF; Jinxian, Z; Hipps, AD; Wehry, EL; Mamantov, G, (1989) Adsorption and
  2       photodegradation of pyrene on magnetic, carbonaceous, and mineral subtractions of coal stack ash. Environ Sci
  3       Technol 23:303-308,
  4
  5       Dutcher, JS; Sun, JD; Lopez, JA; Wolf, I; Wolff, RK; McClellan, RO. (1984) Generation and characterization of
  6       radiolabeled diesel exhaust. Am Ind Hyg Assoc J 45:491-498.
  7
  8       Fenimore, CP; Jones, GW. (1967) Oxidation of soot by hydroxyl radicals. J Phys Chem 71:593-597,
  9
10       Finlayson-Pitts, BJ; Pitts, JN, Jr. (1986) Atmospheric chemistry: fundamentals and experimental techniques. New
11       York: John Wiley & Sons.
12
13       Frey, JW; Corn, M. (1967) Diesel exhaust participates. Nature (London) 216:615-6 US.
14
15       Galloway, JN; Whelpdale, DM. (1980) An atmospheric sulfur budget for eastern North America. Atmos Environ
16       14:409-417.
17
18       Garrigues, P; Parlanti, E; Radke, M; Bellocq, J; Willsch, H; Ewald, M. (1987) Identification of alkylphenanthrenes
19       in shale oil and coal by liquid and capillary gas chromatography and high-resolution spectrofluorimetry (Shpol'skii
20       effect). JChromatogr 395:217-228.
21
22       Gibbs, RE; Hyde, JD; Byer, SM. (1980) Characterization of particulate emissions from in-use diesel vehicles.
23       Presented at:  Fuels and lubricants; October; Baltimore, MD. Warrendale, PA: Society of Automotive Engineers, Inc.,
24       SAE technical paper no. 801372.
25
26       Gibson, TL.  (1983) Sources of direct-acting nitroarene mutagens in airborne particulate matter. Mutat Res
27       122:115-121,
28
29       Gibson, TL; Ricci, AI; Williams, RL. (1981) Measurement of polynuclear aromat Society of Automotive Engineers,
30       Inc., SAE technical paper no. 880350.
31                   .              :
32       Gorse, RA, Jr.; Salmeen, IT; Clark, CR. (1982) Effects of filter loading and filter type on the mutagenicity and
33       composition of diesel exhaust particulate extracts. Atmos Environ 16:1523-1528.
34
35       Gorse, RA, Jr; Riley, TL;  Ferris, FC; Pero, AM; Skewes, LM. (1983) 1-Nitropyrene concentration and bacterial
36       mutagenicity in on-road vehicle particulate emissions. Environ Sci Technol 17:198-202,
37
38       Gray, HA, (1986) Control of atmospheric fine primary carbon particle concentration [PhD thesis]. Pasadena, CA:
39       California Institute of Technology; EQL report no. 23.
40
41       Hammerle, RH; Schuetzle, D; Adams, W. (1994) A perspective on the potential development of environmentally
42       acceptable light-duty diesel engines. Environ Health Perspect (Suppl) 102(4):25-30.
43
44       Hampton. CV; Pierson, WR; Harvey, TM; Updegrove, WS; Marano, RS. (1982) Hydrocarbon gases emitted  from
45       vehicles on the road. 1. A qualitative gas chromatography/mass spectrometry survey. Environ Sci Technol 16:287-
46       298.
47
48       Hampton, CV; Pierson, WR; Schuetzle, D; Harvey, TM. (1983) Hydrocarbon gases emitted from vehicles on the
49       road. 2. Determination of emission rates from diesel and spark-ignition vehicles. Environ Sci Technol 17:699-708.
50                   -
51       Hardenberg,  HO; Daudel, HL; Erdmannsdorfer, HJ. (1987) Experiences in the development of ceramic fiber coil
52       particulate traps. In: Diesel particulates: an update, SP-702. Warrendale, PA:  Society of Automotive Engineers, SAE
53       technical paper no. 870015.
54
         2/1/98                                         2-62        DRAFT-DO NOT CITE OR QUOTE

-------
  1    .  Hare, CT; Bradow, RL. (1979) Characterization of heavy-duty gaseous and participate emissions and effects of fuel
  2      composition. Warrendale, PA: Society of Automotive Engineers, Inc., SAE technical paper no. 790490.
  3
  4      Hare, CT; Baines, TM. (1979) Characterization of particulate and gaseous emissions from two diesel automobiles as
 ,5      functions of fuel and driving cycle. Presented at: Automotive engineering congress and exposition; February-March;
  6      Detroit, ML Warrendale, PA: Society of Automotive Engineers; SAE technical paper no. 790424.
  7
  8      Hare, CT; Springer, KJ; Bradow, RL. (1976) Fuel and additive effects on diesel particulate development and
  9      demonstration of methodology. Presented at: Automotive engineering congress and exposition; February; Detroit,
10      MI. Warrendale, PA: Society of Automotive Engineers, Inc., SAE technical paper no.  760130.
11
12      Harris, GW; Mackay, GI; Iguchi, T; Schiff, HI; Schuetzle, D. (1987) Measurement of NO2 and HNO3 in, diesel
13      exhaust gas by tunable diode laser absorption spectrometry. Environ Sci Technol 21:299-304.
14
15      Health Effects Institute. (1995) Diesel exhaust: a critical analysis of emissions, exposure, and health effects.
16      Cambridge, MA:  Health Effects Institute.
17
18    •  Henein, NA; Bolt, JA. (1967) Ignition lag in diesel engines. Warrendale, PA: Society of Automotive Engineers, Inc.,
19      SAE technical paper no. 670007.
20
21    •  Herr, JD; Yergey, JA; Dukovich, M; Tejada, S; Risby, TH; Lestz, SS. (1983) The role of nitrogen in the observed
22      microbial mutagenic activity for diesel engine combustion. Warrendale, PA: Society for Automotive Engineers, SAE
23      technical paper no. 820467.
24
25      Kites, RA; Laflamme, RE; Windsor, JG, Jr. (1980) Polycyclic aromatic hydrocarbons in marine/aquatic sediments:
26      their ubiquity. In: Petrakis, L; Weiss, FT, eds. Petroleum in the marine environment [proceedings of a symposium at
27      the 176th meeting of the American Chemical Society; September 1978; Miami Beach, FL]. (Comstock, MJS ed.
28      Advances in chemistry series: v. 185).  Washington, DC: American Chemical Society; pp. 289-311.
29
30      Horvath, H; Kreiner, I; Norek, C; Preining, O; Georgi, B. (1988) Diesel emissions in Vienna. Atmos Environ
31      22:1255-1269.
32
33      Huisingh, J; Bradow, R; lungers, R; Claxton, L; Zweidinger, R; Tejada, S; Bumgamer, J; Duffield,  F; Waters, M;
34      Simmon, VF; Hare, C; Rodriguez, C; Snow, L. (1978) Application of bioassay to the characterization of diesel
35      particle emissions. In: Waters, MD; Nesnow, S; Huisingh, JL; Sandhu, SS; Claxton, L, eds. Application of short-
36      term bioassays in the fractionation and analysis of complex environmental mixtures: proceedings of a symposium;
37      February; Williamsburg, VA. (Hollaender, A; Probstein, F; Welch, BL, eds. Environmental science research: v. 15).
38      New York: Plenum Press, pp. 383-418.
39
40      International Agency for Research on Cancer. (1989) Diesel and gasoline engine exhausts and some nitroarenes.
41      (IARC monographs on the evaluation of carcinogenic risks to humans: v. 46). Lyon, France: World Health
42      Organization.
43
44      J§ger, J; Hanus", V. (1980) Reaction of solid carrier-adsorbed polycyclic aromatic hydrocarbons with gaseous low-
45      concentrated nitrogen dioxide. J Hyg Epidemiol Microbiol Immunol 24:1 -15.
46                                                                                                  .
47      Japar, SM; Szkarlat, AC; Gorse, RA, Jr; Heyerdahl, EK; Johnson, RL; Rau, JA; Huntzicker, JJ. (1984) Comparison
48      of solvent extraction and thermal-optical carbon analysis methods: application to diesel vehicle exhaust aerosol
49      Environ Sci Technol 18:231-234.
50
51      Jensen, TE; Kites, RA. (1983) Aromatic diesel emissions as a function of engine-conditions. Anal Chem 55:594-599,
52
53      Kadota, T; Henein, NA. (1981) Time-resolved soot particulates in diesel spray combustion. In: Siegla, DC; Smith,
54      GW, eds. Particulate carbon: formation during combustion [proceedings of an international symposium; October
55      1981; Warren, MI], New York: Plenum Press, pp.  391-421.


         2/1/98                                         2-63        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Kagami, M; Akasaka, Y; Date, K; Maeda, T. (1984) The influence of fuel properties on the performance of Japanese
  2       automotive diesels. Warrendale, PA: Society of Automotive Engineers, Inc.; SAE technical paper no. 841082, SAE
  3       SP-581.
  4
 .5       Kahn, IM. (1970) Formation and combustion of carbon in a diesel engine. In: Diesel engine combustion symposium;
  6       April. London: Institute of Mechanical Engineers.
  7
  8       Kamens, RM; Quo, Z; Fulcher, JN; Bell, DA. (1988) Influence of humidity, sunlight, and temperature on the daytime
  9       decay of polyaromatic hydrocarbons on atmospheric soot particles. Environ Sci Technol 22:103-108.
10
11       Katz, M; Chan, C; Tosine, H; Sakuma, T. (1979) Relative rates of photochemical and biological oxidation (in vitro)
12   .    of polycyclic aromatic hydrocarbons. In: Jones, PW; Leber, P, eds. Polynuclear aromatic hydrocarbons: third
13       international symposium on chemistry and biology—carcinogenesis and mutagenesis; October 1978; Columbus, OH.
14       Ann Arbor, MI: Arm Arbor Science Publishers, Inc.; pp.  171-189.
15
16       Kawamura, K; Ng, L-L; Kaplan, IR. (1985)  Determination of organic acids (C,-C10) in the atmosphere, motor
17       exhausts and engine oil. Environ Sci Technol 19:1092-1096.
18
19       Kittelson, DB; Dolan, DF. (1980) Diesel exhaust aerosols. In: Willeke, K, ed. Generation of aerosols and facilities
20       for exposure experiments: [technical papers  presented at the symposium on aerosol generation and exposure
21       facilities]; April 1979; Honolulu, HI. Ann Arbor, MI: Ann Arbor Science; pp. 337-359.
22
23       Kleindienst, TE; Shepson, PB; Edney, EO; Claxton, LD.  (1985) Peroxyacetyl nitrate: measurement of its mutagenic
24       activity using  the Salmonella/mammalian microsome reversion assay. Mutat Res 157:123-128.
25
26       Laflamme, RE; Hites, RA. (1978) The global distribution of polycyclic aromatic hydrocarbons in recent sediments.
27       Geochim Cosmochim Acta 42:89-310.
28
29       Lee, FS-C; Schuetzle, D. (1983) Sampling, extraction,  and analysis of polycyclic aromatic hydrocarbons from
30       internal combustion engines. In: Bjerseth, A, ed. Handbook of polycyclic aromatic hydrocarbons. New York: Marcel
31       Dekker, Inc.; pp. 27-94.
32
33       Lee, FSC; Pierson, WR; Ezike, J. (1980) The problem  of PAH degradation during filter collection of airborne
34       particulates—an evaluation of several commonly used  filter media. In: Bjerseth, A; Dennis, AJ, eds. Polynuclear
35       aromatic hydrocarbons: chemistry and biological effects; proceedings of the 4th international symposium; Columbus,
36       OH. Columbus, OH: Battelle Press, pp. 543-563.
37
38       Levsen, K. (1988) The analysis of diesel particulate. Fresenius Z Anal Chem 331:467-478.
39
40       Liberti, A; Ciccioli, P; Cecinato, A; Brancaleoni, E; Di Palo, C. (1984) Determination of nitrated-polyaromatic
41       hydrocarbons  (nitro-PAHs) in environmental samples by high resolution chromatographic techniques. J High Resolut
42       Chromatogr Commun 7:389-397.
43
44       Ligocki, MP; Pankow, JF. (1989) Measurements of the gas/particle distributions of atmospheric organic compounds.
45       Environ Sci Technol 23:75-83.
46
47       Ligocki, MP; Leuenberger, C; Pankow, JF. (1985a) Trace organic compounds in rain—II. Gas scavenging of neutral
48       organic compounds. Atmos Environ 19:1609-1617.
49
50       Ligocki, MP; Leuenberger, C; Pankow, J. (1985b) Trace organic compounds hi rain—HI. Particle scavenging of
51       neutral organic compounds. Atmos Environ  19:1619-1629.
52
53       Lipkea, WH; DeJoode, AD. (1987) A model of a direct injection diesel combustion system for use hi cycle
54       simulation and optimization studies. Presented at: International congress and exposition; February; Detroit, MI.
55       Warrendale, PA: Society of Automotive  Engineers, Inc.,  SAE technical paper no. 870573.


         2/1/98                                         2-64        DRAFT-DO NOT CITE OR QUOTE

-------
  1      Lipkea, WH; DeJoode, AD; Christenson, SR. (1987) The relationship between nitric oxide and work as influenced
  2      by engine operating conditions and combustion system parameters for a direct injection diesel engine. Presented at:
  3      International congress and exposition; February; Detroit, MI. Warrendale, PA: Society of Automotive Engineers,
  4   '   Inc., SAE technical paper no. 870269.
  5
  6      Lofroth, G. (1981) Comparison of the mutagenic activity in carbon paniculate matter and in diesel and gasoline
  7    ,  engine exhaust. In: Waters, MD; Sandhu, SS; Huisingh, JL; Claxton, L; Nesnow, S, eds. Short-term bioassays in the
  8      analysis of complex environmental mixtures II: proceedings of the second symposium on the application of short-
  9      term bioassays in the fractionation and analysis of complex environmental mixtures; March 1980; Williamsburg,
10      VA. (Hollaender, A; Welch, BL; Probstein, RF, eds. Environmental science research series: v. 22). New York:
11      Plenum Press; pp. 319-336.
12
13      Lowenthal, DH; Rahn, KA. (1985) Regional sources of pollution aerosol at Barrow, "Alaska during winter 1979-80
14      as deduced from elemental tracers. Atmos Environ 19:2011 -2024.
15
16      Lyn, WT. (1960-61) Calculation of the effect of rate of heat release on the shape of cylinder pressure diagram and
17      cycle efficiency. Proc Automob Div Inst Mech Eng (1),
18
19      Lyn, WT. (1963) Study of burning rate and nature of combustion in a diesel engine. In: Ninth international
20      symposium on combustion. New York: Academic Press, Inc.
21
22      Lyn, WT; Valdmanis, E. (1966-67) The effect of physical factors on ignition delay. Proc Inst Mech Eng Part 2A (1).
23
24      McCann, J; Ames, BN. (1977) The Salmonella/mkrosome mutagenicity test: predictive value for animal
25      eareinogenicity. In: Hiatt, HH; Watson, JD; Winsten, JA, eds. Origins of human cancer: book C, human risk
26      assessment. (Cold Spring Harbor conferences on cell proliferation: v. 4). New York:  Cold Spring Harbor Laboratory,
27      pp.  1431-1450.
28
29      McClellan, RO. (1986) Health effects of diesel exhaust: a case study in risk assessment. Am Ind Hyg Assoc J 47:1-
30      13.
31
32      McCoy, EC; Rosenkranz, HS; Mermelstein, R. (1981) Evidence for the existence of a family of bacterial
33      nitroreductases capable of activating nitrated polycyclics to mutagens. Environ Mutagen 3:421-427.
34  .
35      McVeety, BD; Hites, RA. (1988) Atmospheric deposition of polycyclic aromatic hydrocarbons to water surfaces: a
36      mass balance approach. Atmos  Environ 22:511-536.
37
38      Miguel, AH; Pereira, PAP. (1989) Benzo(k)fiuoranthene, benzo(ghi)perylene, and  indeno(l,2,3-cd)pyrene: new
39      tracers of automotive emissions in receptor modeling. Aerosol Sci Technol 10:292-295.
40
41      Myers, PS; Uyehara, OA. (1947) Flame temperature measurements—electronic solution of temperature equations.
42  .    SAE Q Trans 1:592-611.
43
44      National Research Council, (1982) Diesel cars: benefits, risks and public policy, final report of the Diesel Impacts
45      Study Committee. Washington, DC: National Academy Press.
46                                                         .
47      Newton, DL; Erickson, MD; Tomer, KB; Pellazzari, ED; Gentry, P; Zweidinger, RB. (1982) Identification of
48      nitroaromatics in diesel exhaust particulate using gas chromatography/negative ion chemical ionization mass
49      spectromerry and other techniques. Environ Sci Technol 16:206-213.
50
51      Nielsen, T. (1988) The decay of benzo[a]pyrene and cyclopenteno(c
-------
  1       polycyclic organic species. (NATO advanced science institutes series C, mathematical and physical sciences: v.
  2       112), Dordrecht, The Netherlands: D, Reidel, pp. 197-210.
  3
  4       Obert, EF. (1973) Internality of diesel particulate matter a sampling artifact? Environ Sci Technol 17:621-624.
  5
  6       Okamoto, WK; Gorse, RA, Jr; Pierson, WR. (1983) Nitric acid in diesel exhaust. J Air Pollut Control Assoc 33:
  7       1098-1100.
  8
  9       Pankow, JF; Isabelle, LM; Asher, WE. (1984) Trace organic compounds in rain. 1. Sampler design and analysis by
10       adsorption/thermal desorption (ATD). Environ Sci Technol 18:310-318.
11
12       Paputa-Peck, MC; Marano, RS; Schuetzle, D; Riley, TL; Hampton, CV; Prater, TJ; Skewes, LM; Jensen, TE;
13       Ruehle, PH; Bosch, LC; Duncan, WP. (1983) Determination of nitrated polynuclear aromatic hydrocarbons to
14       particulate extracts by capillary column gas chromatography with nitrogen selective detection. Anal Chero 55: 1946-
15       1954.
16
17       Pierson, WR; Brachaczek, WW. (1976) Particulate matter associated with vehicles on the road. Warrendale, PA:
18       Society of Automotive Engineers; SAE technical paper no. 760039. SAE Trans  85:209-227.
19
20    .   Pierson, WR; Brachaczek, WW. (1983a) Emissions of ammonia and amines from vehicles on the road. Environ Sci
21       Technol 17:757-760.
22
23       Pierson, WR; Braehaezek, WW. (1983b) Particulate matter associated with vehicles on the road. II. Aerosol Sci
24       Technol 2:1-40.
25
26       Pierson, WR; Brachaczek, WW; Hammerle, RH; McKee, DE; Butler, JW. (1978) Sulfate emissions from vehicles on
27       the road. J Air Pollut Control Assoc 28:123-132.
28
29.      Pierson, WR; Brachaczek, WW; McKee, DE. (1979) Sulfate emissions from catalyst-equipped automobiles on the
30       highway. J Air Pollut Control Assoc 29:255-257.
31
32       Pierson, WR; Gorse,  RA, Jr; Szkarlat, AC; Brachaczek, WW; Japar, SM; Lee, FS-C; Zweidinger, RB; Claxton, LD.
33       (1983) Mutagenicity  and chemical characteristics of carbonaceous particulate matter from vehicles on the road.
34       Environ Sci Technol  17:31-44.
35
36       Pitts, JN, Jr; Grosjean, D; Mischke, TM; Simmon, VF; Poole, D. (1977) Mutagenie activity of airborne particulate
37       organic pollutants. Toxicol Lett 1:65-70,
38
39       Pitts, JN, Jr; Van Cauwenberghe, KA; Grosjean, D; Schmid, JP; Fitz, DR; Belser, WL, Jr; Knudson, GB; Hynds,
40       PM. (1978) Atmospheric reactions of polycyclic aromatic hydrocarbons: facile.formation of mutagenic nitro
41       derivatives. Science (Washington, DC) 202:515-519.
42
43       Pitts, IN, Jr; Lokensgard, DM; Ripley, PS; Van Cauwenberghe, KA; Van Vaeck, L; Shaffer,  SD; Thill, AJ; Belser,
44       WL, Jr. (1980) "Atmospheric" epoxidation of benzo[a]pyrene by ozone: formation of the metabolite
45       benzo[a]pyrene-4,5-oxide. Science (Washington, DC) 210:1347-1349.
46
47       Pitts, JN, Jr; Sweetman, JA; Zielinska, B; Winer, AM; Atkinson, R. (1985a) Determination of 2-nitrofluoranthene
48       and 2-nitropyrene in  ambient particulate organic matter: evidence for atmospheric reactions. Atmos Environ
49       19:1601-1608.
50
51       Pitts, JN, Jr; Zielinska, B; Sweetman, JA; Atkinson, R; Winer, AM. (1985b) Reaction of adsorbed pyrene and
52       perylene with gaseous N2O5 under simulated atmospheric conditions. Atmos Environ 19:911-915.
53
         2/1/98                                          2-66        DRAFT-DO NOT CITE OR QUOTE

-------
  1      Pitts, JN, Jr; Sweetman, JA; Zielinska, B; Atkinson, R; Winer, AM; Harger, WP. (1985c) Formation of nitroarenes
  2      from the reaction of polycyclic aromatic hydrocarbons with dinitrogen pentoxide. Environ Sci Technol 19:1115-1121.
  3
  4      Pitts, JN, Jr; Paur, H-R; Zielinska, B; Arey, J; Winer, AM; Ramdahl, T; Mejia, V. (1986) Factors influencing the
  5      reactivity of polycyclic aromatic hydrocarbons adsorbed on filters and ambient POM with ozone, Chemosphere
  6      15:675-685.
  7             .
  8      Ramdahl, T; Zielinska, B; Arey, J; Atkinson, R; Winer, A; Pitts, JN, Jr. (1986) Ubiquitous occurrence of 2-
  9      nitrofluoranthene and 2-nitropyrene in air. Nature (London) 231:425-427.
10
11      Rogge, W; Mazurek, MA; Hildemann, LM; Cass, GR. (1993) Quantification of urban organic aerosols at a
12      molecular level: identification, abundance and seasonal variation, Atmos Env 27Arl309-1330.
13                                                                               •                   •      .
14      Rosenkranz, HS; Mermelstein, R. (1983) Mutagenicity and genotoxicity of nitroarenes: all nitro-containing
15      chemicals were not created equal. Mutat Res 114:217-267.
16
17      Ross, MM; Risby, TH; Steele, WA; Lestz, SS; Yashin, RE. (1982) Physicochemical properties of diesel paniculate
18      matter. Colloids Surf 5:17-31.
19
20      Salmeen, IT; Pero, AM; Zator, R; Schuetzle, D; Riley, TL. (1984) Ames assay chromatograms and the identification
21      of mutagens in diesel particle extracts. Environ Sci Technol 18:375-382.
22
23      Salmeen, IT; Gorse, RA, Jr; Pierson, WR. (1985) Ames assay chromatograms of extracts of diesel exhaust particles
24      from heavy-duty trucks on the road and from passenger cars on a dynamometer. Environ Sci Technol 19:270-273.
25
26      Schroeder, WH; Lane, DA. (1988) The fate of toxic airborne pollutants: the atmosphere plays a major role in their   .
27      transformation and transport. Environ Sci Technol 22:240-246,
28
29      Schuetzle, D. (1983) Sampling of vehicle emissions for chemical analysis and biological testing. Environ Health
30      Perspect 47:65-80.
31
32      Schuetzle, D. (1988) Organic tracers for source apportionment of combustion emissions to  ambient air particulates.
33      In: proceedings of the sixth symposium on environmental analytical chemistry and source apportionment in ambient
34      and indoor atmospheres; June; Provo, UT.
35
36      Schuetzle, D; Perez, JM. (1981) A CRC cooperative comparison of extraction and HPLC techniques for diesel
37      particulate emissions. Presented at: 74th annual meeting of the Air Pollution Control Association; June; Philadelphia,
38      PA. Pittsburgh, PA: Air Pollution Control Association; paper no. 81-56.4,
39
40      Schuetzle, D; Perez, JM. (1983) Factors influencing the emissions of nitrated-polynuclear aromatic hydrocarbons
41      (nitro-PAH) from  diesel engines. J Air Pollut Control Assoc 33:751-755.
42
43      Schuetzle, D; Jensen, TE. (1985) Analysis of nitrated polycyclic aromatic hydrocarbons by mass spectrometry. In:
44      White, C. M., ed% Nitrated polycyclic aromatic hydrocarbons, Heidelberg, Germany: A. Huthing Verlag, pp. 121-
45      167.
46 •
47      Schuetzle, D; Frazier, JA. (1986) Factors influencing the emission of vapor and particulate  phase components from
48      diesel engines,  hi: Ishinishi, N; Koizumi, A; McClellan, RO; StOber, W, eds. Carcinogenic and mutagenic effects of
49      diesel engine exhaust: proceedings of the international satellite symposium on lexicological effects of emissions
50      from diesel engines; July; Tsukuba Science City, Japan. (Developments in toxicology and environmental science: v.
51      13). Amsterdam: Elsevier Science Publishers B. V., pp. 41-63.
52
53      Schuetzle, D; Lewtas, J. (1986) Bioassay-directed chemical analysis in environmental research. Anal Chem 58:
54      1060A-1076A.
55


         2/1/98                                           2-67        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Schuetzle, D; Lee, FS-C; Prater, TJ; Tejada, SB. (1981) The identification of polynuclear aromatic hydrocarbon
  2       (PAH) derivatives in mutagenic fractions of diesel participate extracts. Int J Environ Anal Chem 9:93-144.
  3
  4       Schuetzle, D; Riley, TL; Prater, TJ; Harvey, TM; Hunt, DF. (1982) Analysis of nitrated polycyclic aromatic
  5       hydrocarbons in diesel particulates. Anal Chem 54:265-271.
  6
  7       Schuetzle, D; Jensen, TE; Ball, JC. (1985) Polar PAH derivatives in extracts of particulates: biological
  8       characterization and techniques for chemical analysis. Environ Int 11:169-181.
  9
10       Schulz, B. (1988) The United States engine markets in perspective. Diesel Prog N Am 53(6):8,10,12.
11
12       Shahed, SM. (1985) The role of fuel-air mixing in diesel combustion and emissions. ASME international conference
13       on flows in internal combustion engines III, v. 28, pp. 159-164.
14
15       Siak, J; Chan, TL; Gibson, TL; Wolff, GT. (1985) Contribution to bacterial mutagenicity from nitro-PAH
16       compounds in ambient aerosols. Atmos Environ 19:369-376.
17
18       Sienicki, EJ; Mago, RS. (1992) Re-evaluation of diesel particulate matter inventories. In: Toxic  air pollutants from
19       mobile sources: proceedings of an international specialty conference. Pittsburgh, PA: Air and Waste Management
20       Association,  pp. 151-164.
21
22       Siegla, DC; Smith, GW, eds. (1980) Particulate carbon: formation during combustion; proceedings of an
23       international symposium; October 1980; Warren, MI. New York: Plenum Press.
24
25       Sigsby, JE, Jr; Tejada, S; Ray, W; Lang, JM; Duncan, JW. (1987) Volatile organic compound emissions from 46 in-
26       use passenger cars. Environ Sci Technol 21:466-475.
27
28       Simoneit, BRT. (1984) Organic matter of the troposphere—III. Characterization and sources of petroleum and
29       pyrogenic residues in aerosols over the western United States. Atmos Environ 18:51-67.
30
31       Slinn, WON; Hasse,  L; Hicks, BB; Hogan, AW; Lai, D; Liss, PS; Munnich, KO;  Sehmel, GA; Vittori, O. (1978)
32       Some aspects of the transfer of atmospheric trace constituents past the air-sea interface. Atmos Environ 12:  2055-
33       2087.
34
35       Smith, LR. (1989) Characterization of exhaust emissions from trap-equipped light-duty diesels [final report].
36       Sacramento,  CA: California Air Resources Board; contract no. A5-159-32.
37
38       Smith, LR; Paskind, J. (1989) Characterization of exhaust emissions from trap-equipped light-duty diesels. Presented
39       at: Passenger car meeting and exposition; September; Dearborn, MI. Warrendale, PA: Society of Automotive
40       Engineers, Inc.; SAE technical  paper no. 891972.
41
42       Sonnefeld, WJ; Zoller, WH; May, WE. (1983) Dynamic coupled-column liquid chromatographic determination of
43       ambient temperature vapor pressures of polynuclear aromatic hydrocarbons. Anal Chem 55:75-280.
44
45       Springer, KJ. (1988) Diesel lube oils—4th dimension of diesel particulate control. In: Springer,  KJ, ed. Engine
46  •     emission technology for the 1990's; proceedings of the internal combustion engine division technical conference;
47       October; San Antonio, TX. New York: American Society of Mechanical Engineers, pp. 45-51.
48
49       Springer, KJ. (1989) Low emission diesel fuel for 1991—1994. In: Chrisman, B; Serve, JV, eds. Advances  in engine
50       emissions control technology ICE, v. 5. New York; American Society of Mechanical Engineers.
51
52       Stockwell, WR; Calvert, JG. (1983) The mechanism of the HO-SO2 reaction. Atmos Environ 17:2231-2235.
53
54       Sweetman, JA; Zielinska, B; Atkinson, R; Ramdahl, T; Winer, AM; Pitts, JN, Jr.  (1986) A possible formation
55       pathway for the 2-nitrofluoranthene observed in ambient particulate organic matter. Atmos Environ 20:235-238.


         2/1/98                                         2-68        DRAFT-DO NOT CITE OR QUOTE

-------
•  1      Szkarlat, AC; Japar, SM. (1983) Optical and chemical properties of particle emissions from on-road vehicles, J Air
  2      Pollut Control Assoc 33:592-597.
  3
  4      Talcott, R; Wei, E, (1977) Airborne mutagens bioassayed in Salmonella typhimurium, J Natl Cancer Inst 58:449-
  5     -451.
  6
  7      Thrane, KE; Mikalsen, A. (1981) High-volume sampling of airborne polycyclic aromatic hydrocarbons using glass
  8      fibre filters and polyurethane foam. Atmos Environ 15:909-918.
  9
10      Tokiwa, H; Takeyoshi, H; Morita, K; Takahashi, K; Saruta, N; Ohnishi, Y. (1976) Detection of mutagenic activity in
11      urban air pollutants. Mutat Res 38:351.
12
13      Tokiwa, H; Morita, K; Takeyoshi, H; Takahashi, K; Ohnishi, Y. (1977) Detection of mutagenic activity in
14      paniculate air pollutants. Mutat Res 48:237-248.
15
16      Tokiwa, H; Kitamori, S; Takahashi, K; Ohnishi, Y. (1980) Mutagenic and chemical assay of extracts of airborne
17      particulates. Mutat Res 77:99-108.
18
19      Tokiwa, H; Kitamori, S; Nakagawa, R; Horikawa, K; Matamala, L. (1983) Demonstration of a powerful mutagenic
20      dinitropyrene in airborne particulate matter. Mutat Res 121:107-116.
21
22      Tong, HY; Karasek, FW. (1984) Quantitation of polycyclic aromatic hydrocarbons in diesel exhaust particulate
23      matter by high-performance liquid chromatography fractionation and high-resolution gas chromatography. Anal
24      Chem 56:2129-2134.
25
26      Tong, HY; Sweetman, JA; Karasek, FW; Jellum, E; Thorsrad, AK. (1984) Quantitative analysis of polycyciic
27      aromatic compounds in diesel exhaust particulate extracts by combined chromatographic techniques. J Chromatogr
28      312:183-202.
29
30      Tosaka, S; Fujiwara, Y; Murayama, T. (1989) The effect of fuel properties on diesel engine exhaust particulate
31      formation. Presented at: International congress and exposition; February-March; Detroit, MI. Warrendale, PA:
32      Society of Automotive Engineers, Inc., SAE technical paper no, 890421,
33                                                                   '
34      Truex, TJ; Pierson, WR; McKee, DE; Shelef, M; Baker, RE. (1980) Effects of barium fuel additive and fuel sulfur
35      level on diesel particulate emissions. Environ Sci Technol 14:1121-1124.
36
37      Ullman, TL. (1989) Investigation of the effects of fuel composition on heavy-duty diesel engine emissions. Presented
38      at: International fuels and lubricants meeting and exposition; September; Baltimore, MD. Warrendale, PA: Society
39      of Automotive Engineers, Inc., SAE technical paper no, 892072.
40                 •
41      U.S. Environmental Protection Agency. (1985) Compilation of air pollutant emission factors, volume I: stationary
42      point and area sources; volume II: mobiie sources. 4th ed. Research Triangle Park, NC: Office of Air Quality
43      Planning and Standards; EPA report no. AP-42-ED-4-VOL-1 and AP-42-ED-4-VOL-2. Available from: NTIS,
44      Springfield, VA; PB86-124906 and PB87-205266.
45
46      U.S. Environmental Protection Agency. (1986) Program to calculate size specific particulate emissions for mobile
47      sources. Ann Arbor, MI: Emission Control Technology Division; EPA report no. EPA/SW/MT-86/OI1. Available
48      from: NTIS, Springfield, VA; PB86-179702/XAB.
49
50      U.S. Environmental Protection Agency. (1989) User's guide to MOBILE4 (Mobile Source Emissions Factor Model).
51      Ann Arbor, MI: Office of Mobile Sources; EPA  report no, EPA-AA-TEB-89-01, Available from: NTIS, Springfield,
52      VA;PB89-164271.
53
         2/1/98                                         2-69        DRAFT-DO NOT CITE OR QUOTE

-------
 1       U.S. Environmental Protection Agency. (1993) Motor vehicle-related air toxic study. Ann Arbor, MI: Office of
 2       Mobile Sources. EPA/420/R-93/005. Available from National Technical Information Service, Springfield, VA,
 3       PB93-182590/XAB.
 4
 5      • Uyehara, OA. (1987) Factors that affect BSFC and emissions for diesel engines: part 1—presentation of concepts,
 6       Presented at: International congress and exposition; February; Detroit, MI, Warrendale, PA: Society of Automotive
 7       Engineers, Inc., SAE technical paper no, 870343.
 8
 9       Van Vaeck, L; Van Cauwenberghe, K. (1984) Conversion of polycyclic aromatic hydrocarbons on diesel paniculate
10       matter upon exposure to ppm levels of ozone. Atmos Environ 18:323-328.
11
12       Van Vaeck, L; Van Cauwenberghe, K; Janssens, J. (1984) The gas-particle distribution of organic aerosol
13       constituents: measurements of the volatilisation artefact in hi-vol cascade impactor sampling. Atmos Environ 18:417-
14       430.
15
16       Wade, WR; Jones, CM. (1984) Current and future light duty diesel engines and their fuels. Presented at:   •
17       International congress and exposition; February-March; Detroit^ MI. Warrendale, PA: Society of Automotive
18       Engineers, Inc., SAE technical paper no. 840105.
19
20       Warner-Selph, MA; Dietzmann, HE. (1984) Characterization of heavy-duty motor vehicle emissions under transient
21       driving conditions. Research Triangle Park, NC: U.S. Environmental Protection Agency, contract no. 68-02-3722.
22
23       Watson, JG; Robinson, NF; Chow, JC; Henry, RC; Kim, BM; Pace, TG; Meyer, EL; Nguyen, Q. (1990) The
24       USEPA/DRI chemical mass balance receptor model, CMB 7,0. Environ Software 5:38-49,
25
26       Westerholm, RN; Almen, J; Li, H; Rannug, JU; Egeback, K-E; Gragg, K, (1991) Chemical and biological
27       characterization of paniculate-, semivolatile-, and gas-phase-associated compounds  in diluted heavy-duty diesel
28       exhausts: a comparison of three different semivolatile-phase samplers; Environ Sci Technol 25:332-338.
29
30       White, CM. (1983) Determination of polycyclic aromatic hydrocarbons in coal-derived materials. In: Bj0rseth, A,
31       ed. Handbook of polycyclic aromatic hydrocarbons. New York,  NY: Marcel Dekker, Inc.; pp. 525-616.
32
33       White, CM. (1985) Analysis of nitrated polycyclic aromatic hydrocarbons by gas chromatography.  In: White, C. M.,
34       ed. Nitrated polycyclic aromatic hydrocarbons. Heidelberg, Germany: A. Huthing Verlag, pp. 1-86,
35
36       Williams, DJ; Milne, JW; Roberts, DB; Kimberlee, MC. (1989a) Paniculate emissions from *in-use' motor
37       vehicles—I. Spark ignition vehicles. Atmos Environ 23:2639-2645,
38
39       Williams, DJ; Milne, JW; Quigley, SM; Roberts, DB; Kimberlee, MC. (1989b) Paniculate emissions from 'in-use'
40       motor vehicles—II. Diesel vehicles. Atmos Environ 23:2647-2661.
41
42       Wu, KJ; Peterson, RC. (1986) Correlation of nitric oxide emission from a diesel engine with measured temperature
43       and burning rate. Presented at: International fuels and lubricants meeting and exposition; October; Philadelphia, PA.
44       Warrendale, PA: Society of Automotive Engineers, Inc., SAE technical paper no. 861566.
45
46     •  Yamasaki, H; Kuwata, K; Miyamoto, H. (1982) Effects of ambient temperature on aspects of airborne polycyclic
47       aromatic hydrocarbons. Environ Sci Technol 16:189-194.
48                                          -
49       Yamasaki, H; Kuwata, K; Kuge, Y. (1984) Determination of vapor pressure of polycyclic aromatic hydrocarbons in
50       the supercooled liquid phase and their adsorption on airborne paniculate matter. Nippon Kagaku Kaishi 8:1324-
51       1329.
52
53       Yokley, RA; Garrison, AA; Wehry, EL; Mamantov, G. (1986) Photochemical transformation of pyrene and
54       benzo[a]pyrene vapor-deposited on eight coal stack ashes. Environ Sci Technol 20:86T90.
55


         2/1/98                                         2-70         DRAFT-DO NOT CITE OR QUOTE

-------
  1       Yost, DM; Ryan, TW. Ill; Owens, EC. (1985) Fuel effects on combustion in a two-stroke diesel engine. Presented at:
  2       International fuels and lubricants meeting and exposition; October; Tulsa, OK. Warrendale, PA: Society of
  3       Automotive Engineers, Inc., SAE technical paper no. 852104.
  4
  5       Yu, M-L; Hites, RA. (1981) Identification of organic compounds on diesel engine soot. Anal Chem 53:951-954.
  6
  7     .  Yu, RC; Shahed, SM. (1981) Effects of injection timing and exhaust gas recirculation on emissions from a D.I.
  8       diesel engine. Presented at: Fuels and lubricants meeting; October; Tulsa, OK. Warrendale, PA: Society of
  9       Automotive Engineers, Inc., SAE technical paper no. 811234.
10
11       Zielinska, B; Arey, J; Atkinson, R; Ramdahl, T; Winer, AM; Pitts, JN, Jr. (1986) Reaction of dinitrogen pentoxide  .
12       with fluoranthene. J Am Chem Soc 108:4126-4132.
13
14       Zielinska, B; Arey, J; Atkinson, R; McElroy, PA. (1988) Nitration of acephenanthrylene under simulated
15       atmospheric conditions and in solution and the presence of nitroacephenanthrylene(s) in ambient particles. Environ
16       SciTechnol 22:1044-1048.
17
18       Zielinska, B; Arey, J; Atkinson, R; McElroy, PA. (1989a) Formation of methylnitronaphthalenes from the gas-phase
19       reactions of 1- and 2-methylnaphthalene with OH radicals and N2O5 and their occurrence in ambient air. Environ Sci
20       Technol 23:723-729.
21
22  '     Zielinska, B; Arey, J; Atkinson, R; Winer, AM. (1989b) The nitroarenes of molecular weight 247 in ambient
23       paniculate samples collected hi southern California. Atmbs Environ 23:223-229.
24
25       Zielinska, RB; Sigsby, JE, Jr; Tejada, SB; Stump, FD. (1990) The atmospheric formation of nitroarenes and their
26       occurence in ambient air. In: Proceedings of the fourth international conference on N-substituted aryl compounds:
27       occurrence, metabolism and biological impact of nitroarenes; July 1989; Cleveland, OH.
28
29       Zweidinger, RB; Sigsby, JE, Jr; Tejada, SB; Stump, FD; Dropkin, DL; Ray, WD; Duncan, JW. (1988) Detailed
30       hydrocarbon and aldehyde mobile source emissions from roadway studies. Environ Sci Technol 22:956-962.
         2/1/98                                          2-71         DRAFT-DO NOT CITE OR QUOTE

-------
                                       CHAPTERS

1     NOTE TO READERS
2           Because Chapter 3 recently has merged with Chapter 2, time did not permit renumbering
3     of the remaining chapters for this draft. Chapters 4 through 12 will be renumbered in the next
4     version.
      2/1/98                                3-1       DRAFT-DO NOT CITE OR QUOTE

-------
                                     4. DOSIMETRIC FACTORS

 1      4.1. INTRODUCTION
 2            Diesel engine emissions consist of a complex mixture of gases, vapors, and particles
 3      made up of a carbon core with a great variety of organic agents adsorbed to the surface.  The
 4      tumorigenic response observed in rats exposed to diesel exhaust is the result of pulmonary
 5      overloading with diesel soot particles. Data from studies in rats have shown that exposure
 6      duration and particle surface area and/or particle mass are relevant dosimetric factors. These
 7      same studies have shown that particle-adsorbed organics may not be instrumental in the
 8      tumorigenic response in rats. Whether similar tumorigenic mechanisms in humans exposed at
 9      low doses are critical dosimetric factors in estimating human risk to diesel exhaust is uncertain.
10      Although assessment of dose-response relationships may permit more advanced extrapolations
11      from high experimental exposure concentrations to ambient levels and from animal test species
12      to humans, the question of mechanistic similarities in a tumorigenic response between rats and
13      humans remains unanswered.
14            A review of animal carcinogenicity studies (Chapter 7) revealed that the gaseous phase
15      alone failed to induce increases in lung tumors in any of the long-term studies in rats or mice.
16      Because of the very limited positive data for this fraction and because the potential carcinogens
17      likely to be present in this fraction (formaldehyde and acetaldehyde) induce upper respiratory
18      tract tumors, which were not seen in the whole-exhaust studies, the gaseous phase is not
19      considered separately in determining carcinogenic risk of diesel exhaust.  Noncancer endpoints
20      examined in these studies (Chapter 5) also were more affected by the whole exhaust compared to
21      the gas phase of the exhaust. Because the tumorigenic effects observed in rats are likely to be a
22      function of particle overload, the dosimetry of particles will be emphasized in this chapter.  The
23      dosimetry of particle-adsorbed organics, however, will be included because they may be of
24      greater importance for humans exposed at low concentrations.
25            With a single exception (Iwai et al., 1986), the tumors reported in the diesel exhaust
26      inhalation studies reviewed in Chapter 7 all occurred in the lungs.  However, these findings are
27      questionable because of the relatively small number of animals used and the fact that the
28      lymphomas reported by Iwai et al. (1986) were not observed in other studies where rats were
29      similarly exposed.  Because particle overload carcinogenesis occurs under pulmonary overload
30      conditions, dosimetric considerations are limited to  the lung.
31            The insoluble carbon core of the exhaust particle and its persistence in the lung are now
32      recognized as critical factors in the particle overload phenomenon that apparently leads to the
33      tumorigenic response observed in rats.  Furthermore, total particle surface area appears to be an

        2/1/98      .                              4-1        DRAFT-DO NOT CITE OR QUOTE

-------
 1      important determinant in the particle overload phenomenon. The contribution of extractable
 2      organics has been shown to be inconsequential in this response (this subject is discussed in more
 3      detail in Chapter 10). It is, however, important to note that dose determination for inhaled
 4      compounds remains problematic among toxicologists (Dahl et al., 1991).
 5             The dosimetric aspects considered will include deposition in the conducting airways and
 6      alveolar regions, normal particle clearance mechanisms and rates in both regions, clearance rates
 7      during lung overload, elution of organics from the particles, particle transport to extraalveolar
 8      sites, and the interrelationships of these factors in determining the target organ dose.
 9
10      4.2. REGIONAL DEPOSITION OF INHALED PARTICLES
11             The regional deposition of particulate matter in the respiratory tract is dependent on the
12      interaction of a number of factors, including respiratory tract anatomy (airway dimensions and
13      branching configurations), ventilatory characteristics (breathing mode and rate, ventilatory
14      volumes and capacities), physical processes (diffusion, sedimentation, impaction, and
15      interception), and the physicoche.mical characteristics (particle size, shape, and density) of the
16      inhaled particles. Regional deposition of partieulate material is usually expressed as deposition
17      fraction of the total particles or mass inhaled and may be represented by the ratio of the particles
18      or mass deposited in a specific region  to the number or mass of particles inspired.  The factors
19      affecting deposition in these various regions and their importance in understanding the fate of
20      inhaled diesel exhaust particulate matter are discussed in the following sections. It is beyond the
21      scope of this document to present a  comprehensive  account of the complexities of respiratory
22      mechanics, physiology, and toxicology. Where appropriate, the reader is referred to publications
23      that provide a more in-depth treatment of these topics (Weibel, 1963; Brain and Mensah, 1983;
24      Raabe et al., 1988; Stober et al., 1993).
25
26      4.2.1.  Physical Processes, Physiological/Anatomical Considerations, and Particle
27             Characteristics
28             Deposition of particles may  occur through several processes or combinations thereof,
29      including diffusion, sedimentation (gravitational settling), interception, electrostatic
30      precipitation, and impaction.  It is important to appreciate that these processes are not necessarily
31      independent but may, in some instances, interact with one another such that total deposition in
32      the respiratory tract resulting from these processes may be less than the calculated  probabilities
33      for deposition by the individual processes (Raabe, 1982).  Depending on the particle size and
34      mass, varying degrees of deposition may occur in the nasopharyngeal, tracheobronchial, and
35  •    alveolar regions of the respiratory tract.

        2/1/98                                      4-2        DRAFT-DO NOT CITE OR QUOTE

-------
  1             Upon inhalation of particulate matter such as diesel exhaust, deposition will occur
  2      throughout the respiratory tract. Because of high airflow velocities and abrupt directional
  3      changes in the nasopharyngeal and tracheobronchial regions, inertial impaction is a primary
  4      deposition mechanism (especially for particles < 2.5 urn mass median diameter [MMD]).
  5     ' Although inertial impaction is a prominent process for deposition of larger particles in the
  6      tracheobronchial region, it is of minimal significance as a determinant of regional deposition
  7      patterns for diesel exhaust particles, with an MMD < 1  |im and small aspect ratio.
  8             Because their MMD is generally < 1 urn, diesel exhaust particles are subject to deposition
  9      in the alveolar region. Based on animal data regarding  the site of origin of diesel exhaust-
10      induced tumors, particle deposition in the alveolar region may be of greatest concern relative to
11      the carcinogenic potential of diesel particulate matter and/or the adsorbed organics. However,
12      such data for humans are not available.  For such small particles, diffusion would be especially
13      prevalent in this region, whereas sedimentation would become less significant, especially for
14      particles of MMD <0.5 um.
15             Respiratory tract anatomy and ventilatory characteristics are crucial in determining
16      regional deposition patterns and are also responsible for interspecies and interindividual
17      variability in both deposition of particulate matter and inhaled dose. The variability in size and
18      branching configurations of conducting airways is an important determinant of interspecies
19      variability in deposited dose. Because of the anatomical complexity and variability in ventilatory
20      patterns, a precise categorization of airflow dynamics for a given species or for a specific portion
21      of the respiratory tract is difficult.  For more extensive discussions of deposition processes, refer
22      to reviews by Morrow (1966), Raabe (1982), U.S. Environmental Protection Agency (1982),
23      Phalen and Oldham (1983), Lippmann and Schlesinger (1984), Raabe et al. (1988), and Stdber et
24      al.(1993).
25             Exposure to whole diesel exhaust will also result in inhalation of gas-phase components
26      such as formaldehyde, acrolein, and sulfur dioxide, all of which have been demonstrated to be
27    .  sensory irritants. It is also known that these irritants affect respiratory rates (Kane and Alarie,
28      1978,1979).  The sensory irritant-induced reduction of respiratory rate is mediated through
29      stimulation of free nerve endings of the afferent trigeminal nerve (Ulrich et al., 1972),  This
30      physiologic reflex response has been shown to be a concentration-dependent response (Alarie,
31      1966,1973; Kane and Alarie, 1978). Several studies have also shown that mice appear to be
32      responsive to sensory irritants relative to alteration of respiratory patterns (Alarie,  1973; Kane
33      and Alarie, 1978, 1979). However, only very low levels of these irritants are present in diesel
34      engine exhaust; consequently, their significance in affecting delivered dose through changes in
35      respiration may be small.

   .     2/1/98        •                            4-3        DRAFT-DO NOT  CITE OR QUOTE

-------
 1      4,2,2. Species Variability in Regional Dose
 2            The variability in the anatomy of conducting airways among species results in
 3      interspecies variability of inhaled dose. Because different species breathing the same aerosol
 4      will not receive the same dose to the respiratory tract, it is generally accepted that exposure
 5      concentration is not an accurate description of respiratory tract dose (Brain and Mensah, 1983).
 6            The deposition of inhaled diesel particles in the respiratory tract of humans and
 7      mammalian species has been reviewed by Schlesinger (1985). He showed that physiological
 8      differences in the breathing mode for humans (nasal or oronasal breathers) and experimental
 9      animals (obligatory nose breathers), combined with different airway geometries, resulted in
10      significant differences in lower respiratory tract deposition for larger particles (>1 um).  In
11      particular, a much lower fraction of inhaled larger particles is deposited in the alveolar region of
12      the rat compared with humans. However, relative deposition of the much smaller diesel exhaust.
13      particles was not affected as much by the differences among species, as was demonstrated in
14      model calculations by Xu and Yu (1987). These investigators modeled the deposition efficiency
15      of inhaled diesel exhaust particles in rats, hamsters, and humans on the basis of calculations of
16      the models of Schum and Yeh (1980) and Weibel (1963). In Figure 4-1, relative deposition
17      patterns in the lower respiratory tract (trachea = generation 1; alveoli = generation 23) are very
18      similar among hamsters, rats, and humans. Variations in alveolar deposition of diesel exhaust
19      particles over one breathing cycle in these different species were predicted to be within 30% of
20      one another. Xu and Yu (1987) attributed this similarity to the fact that deposition of the
21      submicron diesel particles is dominated by diffusion rather than sedimentation or impaction.
22      Although these data assumed nose-breathing by humans, the results would not be very different
23      for mouth-breathing because of the low filtering capacity of the nose for particles in the 0.1 to
24      0.5 um range.
25            However, for dosimetric calculations and modeling, it would be of much greater
26      importance to consider the actual dose deposited per unit surface area of the respiratory tract
27      rather than the relative deposition efficiencies per lung region. Table 4-1 compares the predicted
28      deposited doses of diesel exhaust particles inhaled in 1  min for the three species, based on the
29      total lung volume, the surface area of all lung airways,  or the surface area of the epithelium of the
30      alveolar region only.  In Table 4-1, the absolute deposited dose is lower in humans than in the
31      two rodent species as a result of the greater respiratory exchange rate in rodents and smaller size
32      of the rodent lung. Such differences in the absolute deposited dose in relevant target areas are
33      important and have to be considered when extrapolating the results from diesel exhaust exposure
34      studies in animals to humans.  The differences are less  on a surface area basis than on a lung
        2/1/98      •                              4-4        DRAFT-DO NOT CITE OR QUOTE

-------
                C  10
                     •2;
                O
                (0
                C
                O
                   4A-
                o  10
                a
                o
                o
                                  Hamster

                  Fischer rat
       Human
4      8     12     16    20
   Generation Number
                                                                24
Figure 4-1. Deposition distribution patterns of inhaled diesel exhaust particles in the
airways of different species.

Source: Xu and Yu( 1987).
       Table 4-1. Predicted doses of inhaled diesel exhaust particles 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
M
(10"3 ug/mi/cm3)
3.548
3.434
0.249
M,
(10-* ng/min/cm2)
3.088
3.463
1.237
M, ' -
(10-* ng/min/cm2)
2.382
2.608
0.775
M and Mj ~ mass of particles deposited in total lung.
M2 = mass of particles deposited in the alveolar region only.
Based on the following conditions; (1) MMAD = 0.2 (im, a = 1.9,  = 0.3, and p = 1.5 g/cm3; (2) particle
concentration = 1 mg/m3; and (3) nose-breathing.
Source: Xu and Yu (1987).

2/1/98
            4-5
DRAFT-DO NOT CITE OR QUOTE

-------
 1      volume basis (Table 4-1). This is due to larger alveolar diameters in humans and concomitantly
 2      lower surface area per unit of lung volume.
 3            The alternative, perhaps more accurate physiologically, is to consider deposition rate
 4      relative to exposure concentration; the deposition rate will initiate particle redistribution
 5      processes (e.g., clearance mechanisms, phagocytosis) that transfer the particles to various
 6      subcompartments, including the alveolar macrophage pool, pulmonary interstitium, and lymph
 7      nodes. Over time, therefore, only small amounts of the original particle intake would be
 8      associated with the alveolar surface,
 9
10      4,3. RESPIRATORY TRACT CLEARANCE RATES
11      4.3.1. Tracheobronchial Clearance
12            The dynamic relationship between deposition and clearance is responsible for
13      determining lung burden at any point in time.  Clearance of highly insoluble particles from the
14      tracheobronchial region is mediated primarily by mueociliary transport and is a more rapid
15      process than those operating in alveolar regions. Mueociliary transport (often referred to as the
16     , mueociliary escalator) is accomplished by the rhythmic beating of cilia that line the respiratory
17      tract from the trachea through the terminal bronchioles. This movement propels the mucous
18      layer containing deposited particles (or particles within alveolar macrophages [AMs]) toward the
19      larynx. Clearance rate by this system is determined primarily by the flow velocity of the mucus,
20      which is greater in the proximal airways and decreases distally. These rates also exhibit
21      interspecies and individual variability. Considerable species-dependent variability in
22      tracheobronchial clearance has been reported, with dogs generally having faster clearance rates
23      than guinea pigs, rats, or rabbits (Felicetti et al., 1981). The half-time (t1/2) values for
24      tracheobronchial clearance of relatively insoluble particles are usually on the order of hours;
25      those for alveolar clearance may  be hundreds of days in humans and dogs. The clearance of
26      particulate matter from the tracheobronchial region is generally recognized as being biphasic or
27      multiphasic (Raabe, 1982). Some studies have shown that particles are cleared from large,
28      intermediate, and small  airways with t1/2 of 0.5,2.5, and 5 h, respectively. However, reports have
29      indicated that clearance from conducting airways is biphasic and that the long-term component
30      for humans may take much longer for a significant fraction of particles deposited in this region
31      and may not be complete within  24 h, as generally believed (Stahlhofen et al., 1990).
32            Although most of the particulate matter cleared from the tracheobronchial region will
33      ultimately be swallowed, the contribution of this fraction relative to carcinogenic potential is
34      unclear.  With the exception of conditions of impaired bronchial clearance, the desorptioh t1/2  for
35      particle-associated organics is generally longer than the tracheobronchial clearance times,

        2/1/98        •                            4-6       DRAFT-DO NOT CITE OR QUOTE

-------
  1      thereby making uncertain the importance of this fraction relative to carcinogenesis in the
  2      respiratory tract (Pepelko, 1987). Gerde et al. (199la) showed that for low-dose exposures,
  3      particle-associated PAHs were rapidly released, thereby suggesting retained particle-associated
  4      PAHs may be of lesser importance in tumorigenic responses than originally believed.  The
  5      relationship between the early clearance of insoluble particles (4 um aerodynamic diameter) from
  6      the tracheobronchial regions and their longer-term clearance from the alveolar region is
  7      illustrated in Figure 4-2.
  8             Cuddihy and Yeh (1986) reviewed respiratory tract clearance of particles inhaled by
  9      humans.  Depending on the type of particle (ferric oxide, Teflon discs, or albumin microspheres),
10      the technique employed, and the anatomic region (midtrachea, trachea, or main bronchi), particle
11      velocity (moved by mucociliary transport) ranged from 2.4 to 21.5 mm/min. The highest
12      velocities were recorded for midtracheal transport, and the lowest were for main bronchi.  In one
13      study, an age difference was noted for tracheal mucociliary transport velocity (5.8 mm/mm for
14      individuals less than 30 years of age and 10.1 mm/min for individuals over 55 years of age).
15             Cuddihy and Yeh (1986) described salient points to be considered when estimating
16      particle clearance velocities from tracheobronchial regions:  respiratory tract airway dimensions,
17      calculated inhaled particle deposition fractions for individual airways, and thoracic clearance
18      measurements. Predicted clearance velocities for the trachea and main bronchi were found to be
19      similar to those experimentally determined for inhaled radiolabeled particles but not for
20      intratracheally instilled particles. The velocities observed for inhalation studies were generally
21      lower than those of instillation studies.  Figure 4-3 illustrates a comparison of the short-term
22      clearance of inhaled particles by human subjects and the model predictions for this clearance.
23      However, tracheobronchial clearance via the mucociliary escalator is of limited importance for
24      long-term retention.
25             Exposure of F344 rats to whole  DPM at concentrations of 0.35,3.5, or 7,0 mg/m3 for up
26      to 24 mo did not significantly alter tracheal mucociliary clearance of 98mTc-macroaggregated
27      albumin instilled into the trachea (Wolff et al., 1987). The assessment of tracheal clearance was
28      determined by measuring the amount of material retained 1 h after instillation. The authors
29      stated that measuring retention would yield estimates of clearance efficiency comparable to
30      measuring the velocity for transport of the markers in the trachea. The results of this study were
31      in agreement with similar findings of unaltered tracheal mucociliary clearance in rats exposed to
32      DPM (0.21,1.0, or 4.4 mg/m3) for up to 4 mo (Wolff and Gray, 1980). However, the 1980 study
33      by Wolff and Gray, as well as an earlier study by Battigelli et al. (1966), showed that acute
        2/1/98                                     4-7        DRAFT-DO NOT CITE OR QUOTE

-------
                                       Tracheobronchial
                                       Deposition
                                            Alveolar Deposition
                                   20         40          60          80
                                          Hours after Inhalation
        Figure 4-2. Clearance of insoluble particles deposited in tracheobronchial and alveolar
        regions.

        Source: Chuddihy and Yeh, 1986.
 1      exposure to high concentrations of diesel exhaust soot (1.0 and 4.4 mg/m3 hi the study by Wolff
 2      and Gray [1980] and 8 to 17 mg/m3 in the study by Battigelli et al. [1966]) produced transient
 3      reductions in tracheal mucociliary clearance. Battigelli et al. (1966) also noted that the
 4      compromised tracheal clearance was not observed following cessation of exhaust exposure.
 5            The fact that tracheal clearance does not appear to be significantly impaired or is
 6      impaired only transiently following exposure to high concentrations of DPM is consistent with
 7      the absence of-pathological effects in the tracheobronchial region of the respiratory tract hi
 8  •    experimental animals. However, the apparent retention of a fraction of the deposited dose in the
 9      airways is cause for some concern regarding possible carcinogenic effects hi this region,
10      especially in light of the results from simulation studies by Gerde  et al. (1991b) that suggested
11      that release of poly cyclic aromatic hydrocarbons (PAHs) from particles may occur within
12      minutes and at the site of initial deposition. Moreover, impairment of mucociliary clearance
13      function as a result of exposure to either occupational or environmental respiratory tract toxicants
        2/1/98
4-8
DRAFT-DO NOT CITE OR QUOTE

-------
                         .
                     ,2     JV\                 Range of Three
                     *•      i  \^               Measurements
                     W      i    \  '""-~		i __ /
                     O  no—      ^-\   '   " "1      —	—	_
                     n  u.o        ^-^	      i               -—		
                     e      '            ~~   -—•	—	                    :  i
                     Q     j                              T-	—_j
                     —      i                              /
                     2  o.eH                             i
                     o      I                             ~~ Model Projection
                     H     1                               Same as Lower
                     1  o.4                                Limit of Range
                     g  0.21
                    .2    •  i
                    **
                     8     1
                                   20       40      60      80      100     120
                                         Hours after Inhalation
       Figure 4-3.  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).

 1     or to cigarette smoke will significantly enhance the retention of particles in this region.  For
 2     example, Vastag et al. (1986) demonstrated that not only smokers with clinical symptoms of
' 3     bronchitis but also symptom-free smokers have significantly reduced mucociliary clearance rates,
 4     Although impaired tracheobronchial clearance could conceivably have an impact on the effects
 5     of deposited DPM in the conducting airways, it does not appear to be relevant to the epigenetic
 6     mechanism likely in diesel exhaust-induced rat pulmonary tumors.
 7
 8     4.3.2.  Clearance From the Alveolar Region
 Q     4.3.2.1. Alveolar Clearance in Humans
10            A number of investigators have reported on the alveolar clearance kinetics of human
11     subjects.  Bohning et al. (1980) examined alveolar clearance in eight humans who had inhaled
12     <0.4 mg of 85Sr-labeled polystyrene particles (3.6 ± 1.6 nm diam.).  A double-exponential model
13     best described the clearance of the particles and provided t1/2 values of 29 ± 19 days and 298 ±
14    , 114 days for short-term and long-term phases, respectively. It was noted that of the particles
15  '   deposited in the alveolar region, 75% ± 13% were cleared via the long-term phase.  Alveolar

       2/1/98                                   4-9        DRAFT-DO NOT CITE OR QUOTE

-------
 1      retention t,/2 values of 330 and 420 days were reported for humans who had inhaled
 2      aluminosilicate particles (Bailey et al., 1982).
 3            Quantitative data on clearance rates in humans having large lung burdens of particulate
 4      matter are lacking. Bohning et al. (1982) and Cohen et al. (1979), however, did provide evidence
 5      for slower clearance in smokers, and Freedman and Robinson (1988) reported slower clearance
 6      rates in individuals who had mild pneumoconiosis. Although information on particle burden and
 7      particle overload relationships in humans is much more limited than for experimental animal
 8      models, inhibition of clearance does seem to occur. Stober et al. (1967) estimated a clearance t1/2
 9      of 4.9 years in coal miners with nil or slight silicosis,  based on postmortem lung burdens. The
10      lung burdens ranged from 2 to 50 mg/g of lung or more, well above the value for which
11      sequestration is observed in the rat. Furthermore, impaired clearance resulting from smoking or
12      exposure to other respiratory toxicants may increase the possibility of an enhanced particle
13      accumulation effect resulting from exposure to other particle sources such as diesel exhaust.
14
15      4.3.2.2. Alveolar Clearance in Animals
16            Normal alveolar clearance rates in animals have been reported by a number of
17      investigators.  Because the fat is the species for which experimentally induced lung cancer data
18      are available and for which most clearance data exist, it is the species most often used for
19      assessing human risk, and reviews of alveolar clearance studies have been generally limited to
20      this species.
21          .  Chan et al. (1981) subjected 24 male F344 rats to nose-only inhalation of DPM  (6 mg/m3)
22      labeled with 131Ba or 14C for 40 to 45 min and assessed total lung deposition, retention,  and
23      elimination. Based on radiolabel inventory, the deposition efficiency in the respiratory tract was
24      15 to 17%.  Measurement of 131Ba label in the feces during the first 4 days following exposure
25      indicated that 40% of the deposited DPM was eliminated via mucociliary clearance. Clearance
26      of the particles from the lower respiratory tract followed a two-phase elimination process
27      consisting of a rapid (ti/2 of 1 day) elimination by mucociliary transport and a slower (ti/2 of
28      62 days) macrophage-mediated alveolar clearance. This study provided data for normal alveolar
29      clearance rates of DPM not affected by prolonged exposure or particle overloading.
30             Several studies have investigated the effects of exposure concentration on the alveolar
31      clearance of DPM by laboratory animals.
32             Wolff et al. (1986, 1987) provided clearance data (ti/2) and lung burden values for F344
33      rats exposed to diesel exhaust for 7 h/day, 5 days/week for 24 mo.  Exposure concentrations of
34      0.35, 3.5, and 7.0 mg of soot/m3 were employed in this whole body-inhalation exposure
35      experiment. Intermediate (hours-days) clearance of 67Ga2O3 particles (30 min, nose-only

        2/1/98                                     4-10       DRAFT-DO NOT CITE OR QUOTE

-------
 1      inhalation) was assessed after 6,12,18, and 24 mo of exposure at all of the DPM concentrations.
 2      A two-component function described the clearance of the administered radiolabel:
                       F(f) .= A exp(-0,693 f/T,)  + B exp(-0.693  f/T2),

 3      where F(t) was the percentage retained throughout the respiratory tract, A and B were the
 4      magnitudes of the two components (component A representing the amount cleared from nasal,
 5      lung, and gastrointestinal compartments and component B representing intermediate clearance
 6      from the lung compartment), and t, and T2 were the half-times for the A and B compartments,
 7      respectively. The early retention half-times (T,), representing clearance from primary, ciliated
 8      conducting airways, were similar for rats in all exposure groups at  all time points except for
 9      those in the high exposure (7.0 mg/m3) group following 24 mo of exposure, whose clearance rate
10      was faster than that of the controls. Significantly longer B compartment retention half-times,
11      representing the early clearance from nonciliated passages such as  alveolar ducts and alveoli,
12      were noted after as few as 6 months of exposure to DPM at 7.0 mg/m3 and 18 months of
13      exposure to 3.5 mg/m3.
14            Nose-only exposures to 134Cs fused aluminosilicate particles (FAP) were used to assess
15      long-term (weeks-months) clearance. Following 24-month exposure to DPM, long-term
16      clearance of !34Cs-FAP was significantly (p<0.01) altered in the 3.5 (cumulative exposure [C x
17      T] of 11,760 mg-h/m3) and 7.0 mg/m3 (C x T = 23,520 mg-h/m3) exposure groups (ti/2 of 264 and
18      240 days, respectively) relative to the 0.35 mg/m3 and control groups (ti/2 of 81  and 79 days,
19      respectively).  Long-term clearance represents the slow component of particle removal from the
20      alveoli. The decreased clearance correlated with the greater particle burden in the lungs of the
21      3.5 and 7.0 mg/m3 exposure groups. Based on these findings, the cumulative exposure of
22      11,760 mg'h/m3 represented a particle overload condition resulting in compromised alveolar
23      clearance mechanisms.
24            Heinrich et al. (1986) exposed rats  19 h/day, 5 days/week for 2.5 years to DPM at a
25      particle concentration of about 4 mg/m3. This is equal to a C x T of 53,200 mg-h/m3. The
26      deposition in the alveolar region was estimated to equal 60 mg. The lung particle burden was
27      sufficient to result in a particle overload condition. With respect to the organic matter adsorbed
28      onto the particles, the authors estimated that over the 2.5-year period, 60 to 150 mg of particle-
29      bound organic matter had been deposited and was potentially available for biological effects.
30      This estimation was based on the analysis of the diesel exhaust used, in the experiments, values
31      for rat ventilatory functions, and estimates of deposition and clearance.
32            Accumulated burden of DPM in the lungs following an 18-mo, 7 h/day, 5 days/week
33      exposure to diesel exhaust was reported hyGriffis etal. (1983). Male and female F344 rats

        2/1/98      .                             4-11       DRAFT-DO NOT CITE OR QUOTE

-------
 1      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
 2      after exposure, and DPM was extracted from lung tissue dissolved in tetramethylammonium
 3      hydroxide. Following centrifugation and washing of the supernatant, DPM content of the tissue
 4      was quantitated using spectrophotometric techniques.  The analytical procedure was verified by
 5      comparing results to recovery studies using known amounts of DPM with lungs of unexposed
 6      rats. Long-term retention for the 0.15 and 0.94 mg/m3 groups had estimated half-times of 87 ±
 7      28 and 99 ± 8 days, respectively. The retention t]/2 for the 4.1-mg/m3 exposure group was 165 ±
 8      8 days, which was significantly (pO.OOOl) greater than those of the lower exposure groups. The
 9      18-mo exposures to 0.15 or 0.96 mg/m3 levels of DPM (C * T equivalent of 378 and 2,368
10      mg-h/m3, respectively) did not affect clearance rates, whereas the exposure to the 4.1 mg/m3
11      concentration (C x T = 10,332 mg-h/m3) resulted in significant lung DPM burdens and impaired
12      clearance.
13            In a subsequent study (Lee et al., 1983), a three-phase model was used to describe the
14      clearance of DPM (7 mg/m3 for 45 min or 2 mg/m3 for 140 min) by F344 rats (24 per group)
15      exposed by nose-only inhalation with no apparent particle overload in the lungs. The exposure
16      protocols provided comparable total doses based on a 14C radiolabel. 14CO2 resulting from
17      combustion of 14C-labeled diesel fuel was removed by a diffusion scrubber to avoid erroneous
18      assessment of 14C intake by the animals.  Retention of the radiolabeled particles was determined
19      up to 335 days after exposure and resulted in the derivation of a three-phase clearance of the
20      particles. The resulting retention t1/2 values for the three phases were 1, 6, and 80 days.  The
21      three clearance phases are taken to represent removal of tracheobronchial deposits by the
22      mucociliary escalator, removal of particles deposited in the respiratory bronchioles, and alveolar
23      clearance, respectively.  Species variability in clearance of DPM was also demonstrated by  the
24      fact that Hartley guinea pigs exhibited negligible alveolar clearance from day 10 to day 432
25      following a 45-min exposure to a DPM concentration of 7 mg/m3. Initial deposition efficiency
26      (20 ± 2%) and short-term clearance were, however, similar to those for rats.
27            Lung clearance in male F344 rats preexposed to DPM at 0.25 or 6 mg/m3 20 h/day,
28      7 days/week for periods lasting from 7 to  112 days was studied by Chan et al.  (1984). Following
29      this preexposure protocol, rats were subjected to 45-min nose-only exposure to 14C-diesel
30      exhaust and alveolar clearance of radiolabel was monitored for up to 1 year. Two models were
31      proposed:  a normal biphasic clearance model and a modified lung retention model that included
32      a slow-clearing residual component to account for sequestered aggregates of macrophages.  The
33      first model described a first-order clearance for two compartments:  R(t) = Ae'ult + Be;"2'. This
34      yielded clearance t1/2 values of 166 and 562 days for rats preexposed to 6.0 mg/m3 for 7 and 62
35 •     days, respectively. These values were significantly (p<0.05) greater than the retention t1/2 of 77 ±

        2/1/98                                    4-12        DRAFT-DO NOT CITE OR QUOTE

-------
  1      17 days for control rats.  The same retention values for rats of the 0,25 rng/m3 groups were 90 ±
  2      14 and 92 ± 15 days, respectively, for 52- and 112-day exposures and were not significantly
  3      different from controls. The two-compartment model represents overall clearance of the tracer
  4      particles, even if some of the particles were sequestered in particle-laden macrophages with
  5     ' substantially slower clearance rates. For the second model,, which excluded transport of the
  6      residual fractions in sequestered macrophage aggregates, slower clearance was observed in the
  7      group with a lung burden of 6.5 mg, and no clearance was observed in the 11.8 mg group.
  8      Clearance was shown to be dependent on the initial burden of particles, and, therefore, the
  9      clearance t1/2 would increase in higher exposure scenarios. This study emphasizes the importance
10      of particle overloading of the lung and the ramifications on clearance of particles; the significant
11      increases in half-times indicate an increasing impairment of the alveolar macrophage mobility
12      and subsequent transition into an overload condition.  Based on these data, a particle overload
13      effect was demonstrated for both the high and low exposure levels (equivalent to C x T dose of
14      840 [transitional overload] and 7,440 mg-h/m3).
15             Long-term alveolar clearance rates of particles in various laboratory animals and humans
16      have been reviewed by Pepelko (1987). Although retention t,/2 varies both among and within
17      species and is also dependent on the physicochemical properties of the inhaled particles, the
18      retention t,/2 for humans is generally much longer (>8 mo) than the average retention t1/2 of 60
19      days for rats.
20
21      4.3.2.3. Lung Burden and Pulmonary Overload Resulting in Impaired Clearance
22             Particle overload appears to be an important factor in the diesel emission-induced
23      pulmonary carcinogenicity observed in rats. Studies described hi this section provide additional
24      data showing a particle overload effect. A study by Griffis et al. (1983) demonstrated that
25      exposure  (7 h/day, 5 days/week) of rats to DPM at concentrations of 0.15,  0.94S or 4.1 mg/m3 for
26      18 mo resulted in lung burdens of 35,220, and 1,890 ug/g of lung  tissue, respectively. The
27      alveolar clearance of those rats with the highest lung burden (1,890 \ig/g of lung) was impaired,
28      as determined by a  significantly greater (pO.OOOl) retention t1/2 for DPM. This is reflected in
29      the greater lung burden/exposure concentration ratio at the highest exposure level.  Similarly, in
30      the study  by Chan et al. (1984) rats exposed for 20 h/day, 7 days/week to DPM (6 mg/m3) for
31      112 days had a total lung particle burden of 11.8 mg, with no alveolar particle clearance being
32      detected over 1 year.
33             Muhle et al. (1990) indicated that overloading of rat lungs occurred when lung particle
34      burdens reached 0.5 to 1.5 mg/g of lung tissue and that clearance mechanisms were totally
35      compromised at lung particle burdens ^ 10 mg/g for particles with a  specific density close to  1.

        2/1/98        «                            4-13       DRAFT-DO NOT CITE OR QUOTE

-------
 1            Pritchard (1989), utilizing data from a number of diesel exhaust exposure studies,
 2     examined alveolar clearance in rats as a function of cumulative exposure. The resulting analysis
 3     noted a significant increase in retention t1/2 values at exposures above 10 mg/m3-h/day and also
 4     showed that normal lung clearance mechanisms appeared to be compromised as the lung DPM
 5     burden approached 0.5 mg/g of lung.
 6            Morrow (1988) has proposed that the condition of particle overloading in the lungs is
 7     caused by a loss in the mobility of particle-engorged AMs and that such an impediment is related
 8     to the cumulative volumetric load of particles in the AM. Morrow (1988) has further estimated
 9     that the clearance function of an AM may be completely impaired when the particle burden in the
10     AM is of a volumetric size equivalent to about 60% of the normal volume of the AM. Morrow's
11     hypothesis was the initial basis for the physiology-oriented multicompartmental kinetic (POCK)
12     model derived by Stober et al. (1989) for estimating alveolar clearance and retention of
13     biologically insoluble, respirable particles. The model provides for physiology-oriented
14     multicompartmental considerations and is based on the dominant role of alveolar macrophages.
15            A revised version of this model refines the characterization of the macrophage pool by
16     including both the mobile and immobilized macrophages (Stober et al., 1994). A diagram of
17     the POCK model is shown in Figure 4-4.  For characterizing these macrophage tools, the
18     model assumes a constant maximum volume capacity of the macrophages for particle uptake
19     and a material-dependent critical macrophage load that results hi total loss of macrophage
20     mobility. Application of the model to experimental data suggested that lung overload does not
21     cause a dramatic increase in the total burden of the macrophage pool but results in a great
22     increase of the particle burden of the interstitial space, a compartment that is not available for
23     macrophage-mediated clearance.  The model predictions, however, regarding the
24     subcompartmental concentrations of particles require an experimental database at a reasonably
25     wide range of external dose rates.
26            Oberdorster and co-workers (1992) assessed the alveolar clearance of smaller (3.3 um
27     diam.) and larger (10.3 u.m diam.) polystyrene particles, the latter of which are volumetrically
28     equivalent to about 60% of the average normal volume of a rat AM, after intratracheal instillation
29     into the lungs of rats. Even though sizes of particles were found to be phagocytized by AM
30     within a day after deposition and the smaller particles were cleared at a normal rate,, only
31     minimal lung clearance of the larger particles was observed over an approximately 200-day
32     postinstillation period, thus supporting the volumetric overload hypothesis.
33            Animal studies have revealed that impairment of alveolar clearance can occur following
34     chronic exposure to DPM (Griffis et al., 1983; Wolff et al., 1987; Vostal et al., 1982; Lee et
35      al., 1983) or a variety of other diverse aerosols (Lee et al., 1986; Lee et al., 1988; Ferin and

        2/1/98     .                              4-14       DRAFT-DO NOT CITE OR QUOTE

-------
oo
j
1
1
i: i
en ' 1
® , 1
£1 	
£ i T
1 | dMP,i
1, • 1
1 Tracheobronchial
TB Tract MTB
A
•
Ms
	 	 1 	 '___'_
*r
Alveolar Surface
)jMpj AS Ms = S MSii
ii ^ ^
U|Ms,k P^MP
a
^^ Mobile Macrophages Sequestrum
^^. . s SsiMs.i
^^^^^^ imob • 'max
^^^MMP . 2 Mpj MSQ -,£bMft,

1
ui . i 0
1
H
D
O
2! •
O
/o
G
o -
H
5? • Figure 4-4. Physioiogy-or
O Source: Stober et al. (1994).
H
W
............................ I ^ 	 .............imob ........
Interstitial
?M,S
IS


Lung-Associated
LN

V V
Interstit ium
M1S
1

Lymph Nodes
iented compartmental kinetics model of retention of inhaled insoluble particles in

-------
 1      Feldstein, 1978; Muhle et al. 1990). Because high lung burdens of insoluble biochemically
 2      inert particles result in diminution of normal lung clearance kinetics or in what is now called
 3      "particle overloading," this effect appears to be more related to the mass and/or  volume of
 4      particles in the lung than to the nature of the particles per se.  It must be noted; however, that
 5      some types of particles may impair clearance at lower lung burdens (e.g., silica may impair
 6      clearance at much lower lung burdens than DPM). This could be due to toxic agents adsorbed
 7      to the particle surface, or in the case of silica, surface-associated organics.  Regardless, as
 8      pointed out by Morrow (1988), particle overloading in the lung modifies the dosimetry for
 9      particles in the lung and thereby can alter toxicologic responses.
10            Although quantitative data are limited regarding lung overload associated with impaired
11      alveolar clearance in humans, impairment of clearance mechanisms appears to occur, and at a
12      lung burden generally in the range reported to impair clearance in rats.  Stober et al. (1967), in
13      their study of coal miners,  reported lung  particle burdens of 2 to 50 mg/g lung tissue for which
14      estimated clearance t1/2 values were very  long (4.9 years). Freedman and Robinson (1988) also
15      reported slower alveolar clearance rates in coal  miners, some of whom had a mild degree of
16      pneumoconiosis. It must be noted,  however, that no lung cancer was reported for those miners
17      with apparent particle overload.
18
19      4.3.3.  Role of AMs in the  Clearance of Participate Matter
20      4.3.3.1. AM-Mediated Clearance of Particulate Matter
21            Alveolar macrophages constitute an important first-line cellular defense mechanism
22      against inhaled particles that deposit in the alveolar region of the lung.  It is well established that
23      a host of diverse materials,  including DPM, are phagocytized by the AMs shortly after deposition
24      (White and Garg, 1981; Lehnert and Morrow, 1985) and that such cell-contained particles are
25      generally rapidly sequestered from both the extracellular fluid lining in the alveolar region and
26      the potentially sensitive alveolar epithelial cells.  In addition to this role in compartmentalizing
27      particles from other lung constituents, AMs are prominently involved in mediating the clearance
28      of relatively insoluble particles from the air spaces (Lehnert and Morrow, 1985). Although the
29      details of the actual process have not been delineated, AMs with their particle burdens gain
30       access and become coupled to the mucociliary escalator and  are subsequently transported from
31      the lung via the conducting airways. Although circumstantial in nature, numerous lines of
32      evidence indicate that such AM-mediated particle clearance is normally the predominant
33      mechanism by which relatively insoluble particles are removed from the lungs  (Gibb and
34      Morrow,  1962; Perm, 1982; Harmsen et al., 1985; Lehnert and Morrow, 1985;  Powdrill et al.,
35      1989).

        2/1/98         •                           4-16       DRAFT-DO NOT  CITE OR QUOTE

-------
  1             The removal characteristics for particles deposited in the lung's alveolar region have been
  2   ,   descriptively represented by numerous investigators as a multicompartment or multicomponent
  3      process in which each component follows simple first-order kinetics (Snipes and Clem, 1981;
  4      Snipes et al., 1988; Lee et al., 1983). Although the various  compartments can be described
  5      mathematically, the actual physiologic mechanisms determining these differing clearance rates
  6      have not been well characterized.
  7             Lehnert et al. (1988, 1989) performed a study using  laboratory rats to examine particle-
  8      AM relationships over the course of alveolar clearance of low to high lung burdens of
  9      noncytotoxic mierospheres (2.13 urn diam.) to obtain information on potential AM-related
10      mechanisms that form the underlying bases for kinetic patterns of alveolar clearance as a
11      function of particle lung burdens. The intratracheally instilled lung burdens studied varied from
12      1.6 x 107 particles (about 85 ug) for the low lung burden to  2.0 x 10s particles (about 1.06 mg)
13      for the mid-dose and 6.8 x  108 particles (about 3.6 mg) for the highest lung burden. The lungs
14 .     were lavaged at various times postexposure and the numbers of spheres hi each macrophage
15      counted.  Although such experiments provide information regarding the response of the lung to
16      particulate matter, intratracheal instillation is not likely to result in the same depositional
17      characteristics as inhalation of particles would.  Therefore, it is unlikely that the response of
18      alveolar macrophages to these different depositional characteristics will be quantitatively similar.
19             The t,/2 values of both the early and later components of clearance were virtually identical
20      following deposition of the low and medium lung burdens.  For the highest lung burden,
21      significant prolongations were found in both the early, more rapid as well as the slower
22      component of alveolar clearance. The percentages of the particle burden associated with the
23      earlier and later components, however, were similar to those of the lesser lung burdens. On the
24      basis of the data, the authors concluded that translocation of AMs from alveolar spaces by way of
25      the conducting airways is fundamentally influenced by the particle burden of the cells so
26      translocated. In the case of particle overload that occurred at the highest lung burden, the
27      translocation of AMs with the heaviest cellular  burdens of particles (i.e., greater than  about 100
28      mierospheres per AM) was definitely compromised.
29             On the other hand, analysis of the disappearance of AMs with various  numbers of
30      particles indicates that they may not exclusively reflect the translocation of AM from the lung,
31      The observations are also consistent with a gradual redistribution of retained particles among the
32      lung's AMs concurrent with the removal of particle-containing AMs via the conducting airways
33      per se. Experimental support suggestive of potential processes for such particle redistribution
34      comes from a variety of investigations involving AM and other endocyte cell types (Heppleston
        2/1/98                                    4-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1     and Young, 1974; Evans et al., 1986; Aronson, 1963; Sandusky et al., 1977; Heppleston, 1961;
 2     Riley and Dean, 1978).
 3
 4     4.3.3.2.  Tramlocations of Particles to Extraalveolar Macraphage Compartment Sites
 5            Although the phagocytosis of particles by lung-free cells and the mucociliary clearance of
 6     the cells with their particulate matter burdens represent the most prominent mechanisms that
 7     govern the fate of particles deposited in the alveolar region, other mechanisms exist that can
 8     affect both the retention characteristics of relatively insoluble particles,in the lung and the lung
 9     clearance pathways for the particles. One mechanism is endocytosis of particles by Type I cells
10     (Sorokin and Brain, 1975; Adamson and Bowden, 1978,1981) that normally provide >90% of
11     the cell surface of the alveoli in the lungs of a variety of mammalian species (Crapo et al., 1983).
12     This process may be related to the size of the particles that deposit in the lungs and the numbers
13     of particles that are deposited.  Adamson and Bowden (1981) found that with Increasing loads of
14     carbon particles (0,03 urn diam.) instilled in the lungs of mice, more free particles were observed
15     in the alveoli within a few days. The relative abundance of particles endocytosed by Type I cells
16     also increased with increasing lung burdens of the particles, but instillation of large particles (1.0
17     urn) rarely resulted in their undergoing endocytosis. A 4 mg burden of 0.1 um diameter latex
18     particles is equivalent to 8 x 1Q12 particles, whereas a 4 mg burden of 1.0 (am particles is
19     composed of 8 x 109 particles. Regardless, DPM with volume median diameters between 0.05
20     and 0.3 um (Frey and Corn, 1967; Kittleson et al., 1978) would be expected to be within the size
21     range for engulfment by Type I cells should suitable encounters occur. Indeed, it has been
22     demonstrated that DPM is endocytosed by Type I cells in vivo (White and Garg, 1981).
23            Unfortunately, information on the kinetics of particle endocytosis by Type I cells relative
24     to that by AMs is scanty. Even when relatively low burdens of particulate matter are deposited
25     in the lungs, some fraction of the particles usually appears in the regional lymph nodes (Ferin
26     and Fieldstein, 1978; Lehnert,  1989). As will be discussed, endocytosis of particles by Type I
27     cells is an initial, early step hi the passage of particles to the lymph nodes.  Assuming particle
28     phagocytosis is not sufficiently rapid or perfectly efficient, increasing numbers of particles would
29  •   be expected to gain entry into the Type I epithelial cell compartment during chronic aerosol
30     exposures.  Additionally, if particles are released on a continual basis by AMs that initially
31     sequestered them after lung deposition, some fraction of the "free" particles so released could
32     also undergo passage from the alveolar space into Type I cells.
33            The endocytosis of particles by Type I cells represents only the initial stage of a process
34     that can lead to the accumulation of particles in the lung's interstitial compartment and the
35     subsequent translocation of particles to the  regional lymph nodes. As shown by Adamson and

       2/1/98                                    4-18        DRAFT-DO NOT CITE OR QUOTE

-------
  1      Bowden (1981), a vesicular transport mechanism in the Type I ceil can transfer particles from the
  2      air surface of the alveolar epithelium into the lung's interstitium, where particles may be
  3      phagoeytized by interstitial maerophages or remain in a "free" state for a poorly defined period
  4      that may be dependent on the physicochemical characteristics of the particle. The lung's
  5      interstitial compartment, accordingly, represents an anatomical site for the retention of particles
  6      in the lung. Whether or not AMs, and perhaps polymorphonuclear lymphocytes (PMNs) that
  7      have gained access to the alveolar space compartment and phagocytize particles there, also
  8      contribute to the particle translocation process into the lung's interstitium remains a controversial
  9      issue. Evidence that such migration of AMs may contribute significantly to the passage of
10      particles to the interstitial compartment and also may be involved in the subsequent translocation
11      of particles to draining lymph nodes has been obtained with the dog model (Harmsen et al.,
12      1985).
13             The fate of particles once they enter the lung's interstitial spaces remains unclear. Some
14      particles, as previously indicated, are phagoeytized by interstitial maerophages, whereas others
15      apparently remain in a free state in the interstitium for some time without being engulfed by
16      interstitial maerophages. It is unknown if interstitial maerophages subsequently enter the alveoli
17      with their engulfed burdens of particles and thereby contribute to the size of the resident AM
18      population over the course of lung clearance. Moreover, no investigations have been conducted
19      to date to assess the influence that the burden of particles with an interstitial macrophage may
20      have on the AM's ability to migrate into the  alveolar space compartment
21             At least some particles that gain entry into  the interstitial compartment can further
22      translocate to the extrapulmonary regional lymph nodes. This process apparently can involve the
23      passage of free particles as well as particle-containing cells via lymphatic channels in the lungs
24      (Harmsen et al,, 1985; Ferin and Fieldstein, 1978;  Lee et al., 1985).  It is conceivable that the
25      mobility of the interstitial maerophages could be particle-burden limited, and under conditions of
26      high cellular burdens a greater fraction of particles that accumulate in the lymph may reach these
27      sites as free particles. Whatever the process, existing evidence indicates that when lung burdens
28      of particles result in a particle-overload condition, particles accumulate both more rapidly and
29      abundantly in lymph nodes that receive lymphatic drainage from the lung (Ferin and Feldstein,
30      1978; Lee etal, 1985).
31
        2/1/98                                    4-19       DRAFT-DO NOT CITE OR QUOTE

-------
 1      4.3.3.3.  Potential Mechanisms for an AM Sequestration Compartment for Particles During
 2              Particle Overload
 3            Several factors may be involved in the particle-load-dependent retardations in the rate of
 4      particle removal from the lung and the corresponding functional appearance of an abnormally
 5      slow clearing or particle sequestration compartment.  As previously mentioned, one potential site
 6      for particle sequestration is the containment of particles in the Type I cells. Information on the
 7      retention kinetics for particles in the Type I cells is not currently available. Also, no
 8      morphometric analyses have been performed to date to estimate what fraction of a retained lung
 9      burden may be contained in the lung's Type I cell population during lung overloading.
10            Another anatomical region in the lung that may be a slow clearing site is the interstitial
11      compartment. Little is known about either the kinetics of removal of free particles or particle-
12      containing macrophages from the interstitial spaces or what fraction of a retained burden of
13      particles is contained in the lung's interstitium during particle overload.  The gradual
14      accumulation of particles in the regional lymph nodes and the appearance of particles and cells
15      with associated particles in lymphatic channels and in the peribronchial and perivascular
16      lymphoid tissue (Lee et al., 1985; White and Garg, 1981) suggest that the mobilization of
17      particles from interstitial sites via local lymphatics is a continual process.
18        .    Indeed, it is clear from histologic observations of the lungs of animals chronically
19      exposed to DPM that Type I cells, the interstitium, the lymphatic channels, and pulmonary
20      lymphoid tissues are sites that could represent subcompartments of a more generalized slow
21      clearing compartment.
22            Although these sites must be considered to be potential contributors to the increased
23      retention of particles during particle overload, a disturbance in particle-associated AM-mediated
24      clearance is undoubtedly the predominant cause, inasmuch as the AMs are the primary reservoirs
25      of deposited particles. The factors responsible for a failure of AMs to translocate from the
26      alveolar space compartment in lungs with high particulate matter burdens remain uncertain,
27      although a hypothesis concerning the process has been offered involving volumetric AM burden
28      (Morrow, 1988).
29            Other processes also may be involved in preventing particle-laden AMs from leaving the
30      alveolar compartment under conditions of particle overload in the lung.  Clusters or aggregates of
31      particle-laden AMs in the alveoli are typically found in the lungs of laboratory animals that have
32      received large lung burdens of a variety of types of particles (Lee et al.,  1985), including DPM
33      (White and Garg,  1981; McClellan et al., 1982). The aggregation of AMs may explain, in part,
34      the reduced clearance of particle-laden AM during particle overload. The definitive
35      mechanism(s) responsible for this clustering of AMs has not been elucidated to date. Whatever

        2/1/98                                   4-20      DRAFT-DO NOT CITE OR QUOTE

-------
 1      the underlying mechanism(s) for the AM agglutinating response, it is noteworthy that AMs
 2      lavaged from the lungs of diesel exhaust-exposed animals continue to demonstrate a propensity
 3      to aggregate (Strom, 1984).  This observation suggests that the surface characteristics of AMs are
 4      fundamentally altered in a manner that promotes their adherence to one another in the alveolar
 5      region and that AM aggregation may not simply be directly caused by their abundant
 6      accumulation as a result of immobilization by large particle loads.  Furthermore, even though
 7      overloaded macrophages may redistribute particle burden to other AMs, clearance may remain
 8      inhibited (Lehnert, 1988).  This may, in part, be due to attractants from, the overloaded AMs
 9      causing aggregation of those that are not carrying a particle burden.
10
11      4.3.3.4. Physiologically Based Models for Pulmonary Retention and Clearance of Insoluble
12              Particles
13             Currently available data for long-term inhalation exposures to insoluble particles (e.g.,
14      TiO2, carbon black, and DPM) show that pulmonary retention and clearance of these particles are
15      not adequately described by simple first-order kinetics and a single compartment representing the
16      alveolar macrophage particle burden. Several investigators have developed models for
17      deposition, transport, and clearance of insoluble particulate matter in the lungs. All of these
18      models identify various compartments and associated transport rates, but empirically derived
19      data are not available to validate many of the assumptions made in these models.
20             A two-compartment model was developed by Smith (1985) that includes alveolar and
21      interstitial compartments.  For uptake and clearance of particles by alveolar surface macrophages
22      and interstitial encapsulation of particles (i.e., quartz dust), available experimental data show that
23      the rate-controlling functions followed Michaelis-Menton type kinetics, while other processes
24      affecting particle transfer are assumed to be linear.  Although this model provides rate constants
25      as functions that vary depending on the conditions within the various compartments, most of the
26      described functions could not be validated with experimental data.
27             Strom et al. (1988) developed a multicompartmental model for particle retention that
28      includes the following compartments: (1) tracheobronchial tree, (2) free particulate on the
29      alveolar surface, (3) mobile phagocytic alveolar macrophages, (4)  sequestered particle-laden
30      alveolar macrophages, (5)  regional lymph nodes, and (6)'gastrointestinal tract. The model is
31      based on mass-dependent clearance (the rate coefficients reflect this relationship), which will
32      dictate  sequestration of particles and their eventual transfer to the lymph nodes.
33             Yu et al. (1989) considered a three-compartment model that includes a macrophage
34      compartment containing all of the phagocytized particles, an interstitial compartment, and a
35      lymph node compartment. Analyzing the clearance of insoluble particles from the rat lung in

        2/1/98       .                             4-21        DRAFT-DO NOT CITE OR QUOTE

-------
 1      terms of alveolar macrophage mobility kinetics indicated the reduced clearance in overload
 2      conditions to be a function of reduced macrophage mobility (due to volumetric increase in cell
 3      size from phagocytized particles) and an increased, deposition of particles in the interstitial space.
 4            The most recent version of the physiology-oriented multieompartmental kinetics model
 5      derived by Stober and co-workers (1994) (see Section 4.3.2.3., Figure 4-4) addresses alveolar
 6      clearance and retention of inhaled insoluble particles more rigorously by incorporating five -
 7      subcompartments representing alveolar particle load and the respective kinetics associated with
 8      the transfer of particle load among these compartments.  This model also emphasizes the
 9      importance of interstitial burden in the particle overload phenomenon and indicates that particle
10      overload is a function of a massive increase in particle burden of the interstitial space rather than
11      total burden of the macrophage pool. The relevance of the increased particle burden in the
12      interstitial space lies with the fact that this compartmental burden is not available for
13      macrophage-mediated clearance and, therefore, persists even after cessation of exposure.
14      Although the model predictions are tenable, experimental data are not currently available to
15      validate the proposed compartmental burdens or the transfer rates associated with these
16      compartments.  Stober and co-workers noted that the POCK model should not be generalized
17      beyond its design and cannot readily be used for interspecies extrapolation.
18
19      4.4.  BIO AVAILABILITY OF ORGANIC CONSTITUENTS PRESENT ON DIESEL
20           EXHAUST PARTICLES
21            Because it has been shown that DPM extract is not only mutagenic but also contains
22      known carcinogens, the organic fraction was  originally considered to be the primary source of
23      carcinogenicity  in animal studies. Evidence presented in more recent studies, however, indicates
24      that the insoluble carbon core of the particle may fully explain the pathogenic and carcinogenic
25      processes observed in the rat inhalation studies. (See Chapter 10 for a discussion of this issue.)
26      Although the organic constituents appear to be of limited importance in rats exposed at high
27      concentrations, their contributory potential in other species cannot be summarily discounted.
28            A key factor in determining a possible contributory effect is the bioavailability of
29      particle-adsorbed components and whether or not the bioavailable fraction represents an effective
30      dose.  An initial step, and possibly the rate-limiting step, in the bioavailability of carcinogenic
31      organics present on diesel particles is their dissociation from the particle surface (Vostal, 1983).
32      This section  will therefore focus on  studies relating to the dissociation of these organics from the
33      particles. However, the critical uncertainty in these study results is whether the organics added
34      to particles by surface adsorption will exhibit desorption characteristics similar to organics
        2/1/98                                    4-22        DRAFT-DO NOT CITE OR QUOTE

-------
  1      attached to the carbon core during pyrolytic condensation. Data are not available that provide a
  2      definitive answer to this question.
  3
  4      4.4.1. Laboratory Animal Studies
  5     '        Several studies reported on the disposition of particle-adsorbed organics following
  6      administration to various rodent species.  Most of the experiments used intratracheal instillation,
  7      although some exposed the animals by nose-only inhalation.
  8             Sun et al. (1984) compared the disposition of diesel particle-adsorbed benzo[a]pyrene
  9      (B[a]P) (0.1% by weight) and pure B[a]P following nose-only inhalation by F344 rats. Long-
10      term retention (percentage retained after 7 days) of particle-adsorbed 3H-B[a]P was
11      approximately 230-fold greater than that for pure 3H-B[a]P, Alveolar clearance of particle-
12      associated 3H was biphasic, with a long-term t1/2 of 18 days, the latter representing clearance of
13      59% of the initially deposited radiolabel.  Clearance of pure B[a]P aerosol was >99% within
14      2 h and was apparently the result of alveolar and tracheobronchial epithelial absorption into the
15      blood, rather than the result of mucociliary clearance and subsequent ingestion (Sun et al., 1982).
16      The data therefore indicate that adsorption to the carbonaceous diesel particle prolongs retention
17      of the organic components.
18             A companion study (Bond et al, 1986) examined the biological fate of l4C-l-nitropyrene
19      (I4C-NP), both in pure form and adsorbed to DPM, following 1-h nose-only inhalation by male
20      F344 rats. Concentrations of I4C-NP ranged from 0.05 to 1.1  mg/m3 of air, and DPM
21      concentrations, where used, ranged from 3.7 to 6.1 mg/m3 of air. The results indicated that long-
22      term lung retention of 14C-NP adsorbed onto DPM was 80-fold greater (t,/2 = 36 days) than that
23      for pure I4C-NP, demonstrating again that adsorption onto the diesel particles prolongs the
24      release of the PAHs.
25             Residence time is also prolonged when organics are adsorbed to other types of particles.
26      For example, Creasia et al. (1976) found that when crystalline B[a]P was instilled into the lungs
27      of mice, it was removed from the respiratory tract with a t,/2 of ~1.5 h, but when the B[a]P was
28      adsorbed to 0.5 to 1.0 um carbon particles, its t,/2 in the respiratory tract increased to  ~36 h.
29      Hence, the adsorption of B[a]P to the carbon particles increased the lung retention of the B[a]P
30      more than 20-fold.  Similar results have also been obtained with B[a]P adsorbed to other particle
31      types, including insoluble Ga2O3 (Sun et al,, 1982) and insoluble ferric oxide (Saffiotti et al.,
32      1964).  Consistent with a gradual elution of B[a]P in AMs, Creasia and co-workers (1976) found
33      that the removal of B[a]P  when bound to the carbon was faster than the lung clearance of carbon
34      particles only, which had a clearance t,/2 of ~1 days.
        2/1/98         .-                           4-23       DRAFT-DO NOT CITE OR QUOTE

-------
 1            Ball and King (1985) studied the disposition and metabolism of l4C-labeled 1-NP
 2      (>99.9% purity) coated onto DPM. A single dose of 14C-NP (380 (ag/g particle) was
 3      intratracheally administered (in 0.2-mL buffered saline) at a particle dose of 5 mg per rat.
 4      Another group of rats (number not specified) received the labeled 14C-NP in 0.5 mL of buffered
 5      saline intragastrically. Additional groups of AGUS strain rats raised conventionally or germ-free
 6      received intraperitoneal injections of 14C-NP to determine the role of gastrointestinal flora on the
 7      metabolism of 1-NP.
 8            Regardless of the route of administration, >50% of the 14C was .excreted within the first
 9      24 h; 20% to 30% of this appeared in the urine, and 40% to 60% was excreted in the feces. The
10      14C excretion pattern for the intratracheally instilled compound was nearly identical to that of the
11      orally administered compound. For animals receiving intratracheally instilled compound, 16% to
12      38% of the unexcreted dose was in the gastrointestinal tract and 5% to 8% remained in the lungs.
13      Traces of radiolabel were detected in the trachea and esophagus. Five percent to 12% of the
14      radiolabel in the lung co-purified with the protein fraction, indicating protein binding  of the 1-
15      NP-derived 14C. However, the corresponding DNA fraction contained no I4C above background
16      levels. The similar excretion kinetics and metabolic profiles for these various routes of
17      administration indicate that 1-NP becomes  bioavailable both in the lungs and the gastrointestinal
18      tract.
19            Bevan and Ruggio (1991) assessed  the bioavailability of B[a]P adsorbed to DPM from a
20      5.7-L  Oldsmobile engine. In this study, exhaust particles were supplemented with exogenous
21      3H-B[a]P to provide 2.62 (ag B[a]P/g of exhaust particles. Distribution of the radioactivity was
22      assessed at 1,  6, 24, or 72 h after intratracheal instillation of these particles into Sprague-Dawley
23      rats (1 mg of DPM suspended in 0.3 mL of 0.15  M NaCl). At 24 h after administration, 68.5%
24      of the radiolabel remained in the lungs. This is approximately a 3.5-fold greater proportion than
25      that reported by Sun et al. (1984), the difference being attributed to slower pulmonary
26      absorption, less mucociliary transport in intratracheally instilled animals, and differences in
27      administered dose.  At 3 days following administration, over 50% of the radioactivity remained
28      in the lungs, nearly 30% had been excreted into the feces, and the remainder was distributed
29      throughout the body.  Experiments using rats with cannulated bile ducts showed that
30      approximately 10% of the administered radioactivity appeared in the bile over a 10-h period and
31      that less than 5% of the radioactivity entered the feces via mucociliary transport.  The in vitro
32      elution of B[a]P into dimyristoylphosphatidylcholine (DMPC) vesicles was similar for 3H-B[a]P-
33      supplemented DPM and for native DPM (no additional B[a]P added), thereby indicating that the
34      estimation of in vivo bioavailability of B[a]P from the 3H-B[a]P-supplemented DPM was
35      reasonably accurate.                  •

        2/1/98     •                               4-24       DRAFT-DO NOT CITE OR QUOTE

-------
 1             Results of these studies using surface-bound organics showed that the retention of
 2      particle-adsorbed organics is prolonged relative to the organic chemicals alone.  The studies also
 3      indicated that some portion of the adsorbed organics become bioavailable but provided no
 4      information regarding the effective dose of these organics or whether experiments using organics
 5      adsorbed by pyrolytic condensation would exhibit similar characteristics.
 6
 7      4,4.2. Extraction of Diesel Particle-Associated Organics in Biological Fluids
 8             For mutagenicity testing or biochemical analysis, DPM is usually extracted with organic
 9      solvents such as dichloromethane. The efficiency of extraction may be much different, however,
10      from that of the fluids surrounding the particles in the in vivo state, A number of studies have
11      therefore been conducted in which attempts were made to extract DPM with serum or lung
12      lavage fluid. The efficiency of extraction was usually estimated by performing mutagenicity
13      tests on the extracts.
14             The utility of evaluating extraction by lung fluid or serum may be somewhat limited
15      because particles deposited in the alveoli are normally rapidly ingested by AMs.  However, as
16      large lung burdens of DPM are attained, such as during chronic high-concentration exposures to
17      DPM, AMs that become  heavily laden may reach their phagocytic capacity, thereby reducing
18      their phagocytic ability.  Under these conditions, an increasing fraction of deposited particles
19      could escape the phagocytic mechanism and be relatively more available over time in the
20.     extracellular lung fluid before (1) their removal from the lung by extramacrophagic clearance via
21      the tracheobronchial route, (2) their subsequent engulfment by newly recruited phagocytes, and
22      (3) their engulfment by Type I cells.  Even under such conditions, a relatively large mass of the
23      particles will be within AM phagolysosomes, where low pH and enzyme activity would be
24      expected to act on the particles and adsorbed organics.
25             Particles from a 5.7-L engine operated at idle and from a 2.1-L engine operating on a
26      cycle of varying speed and load were incubated in lavage fluid, serum, saline, albumin,
27      dipalmitoyl lecithin, or dichloromethane (Brooks et al., 1981).  The efficiency of extraction by
28      biological fluids was only 3% to 10% that of dichloromethane, based on mutagenicity testing,
29  •    and did not increase with incubation time up to 120 h. Similar findings were reported by King et
30      al. (1981). Inthisstudy,  lung lavage  fluid and lung cytosol fluid extracts of DPM were not
31      mutagenic. Serum extracts of diesel particles did exhibit some mutagenic activity, but this was
32      considerably less than that for organic solvent extracts. Furthermore, the mutagenic activity of
33      the solvent extract was significantly reduced when combined with serum or lung cytosol fluid,
34      suggesting protein binding or biotransformation of the mutagenic components.
        2/1/98                                    4-25       DRAFT-DO NOT CITE OR QUOTE

-------
 1            Siak et al, (1980) assessed the mutagenicity of material extracted from DPM by bovine
 2      serum albumin in solution, simulated lung surfactant, fetal calf serum (PCS), and physiologic
 3      saline. Only PCS was found to extract some mutagenic activity from the DPM.  These
 4      investigators concluded that the mutagens in DPM would not be readily available in vivo. This
 5     ' conclusion lacks definitive proof because extracellular lung fluid is a complex mixture of
 6      constituents that undoubtedly have a broad range of hydrophobicity (George and Hook, 1984;
 7      Wright and Clements, 1987), and it fundamentally differs from serum in terms of chemical
 8      composition (Gurley et  al, 1988).  Moreover, assessments of the ability of lavage fluids, which
 9      actually represent substantially diluted extracellular lung fluid, to extract mutagenic activity from
10      DPM clearly do not reflect the in vivo condition.
11            Creasia et al. (1976) reported that when B[a]P was adsorbed onto carbon particles larger
12      than would be expected to be easily phagocytized by AMs (15  to 30 um), the rates  of elimination
13      of the B[a]P and the particles from the lung were virtually identical. The data thus  indicate little
14      extraction from the particles not phagocytized by AM but only surrounded by epithelial lining
15      fluid.
16            In summary, because lung fluids appear to be relatively ineffective in the extraction of
17      particle-adsorbed organics and relatively few particles escape phagocytosis, free particles are
18      likely to contribute very little to the acute bioavailability of adsorbed organics.  However, during
19      a particle-overload condition, as occurred in the.chronic inhalation studies, it can be
20      hypothesized that an increased fraction of the deposited DPM will not be phagocytized and that
21      their adsorbed organics  could be released over a long period with a clearance rate that may be
22      equivalent to that of the particles themselves. In this case, the organics are bioavailable for even
23      longer periods than those from phagocytized particles, from which organics in turn are retained
24      for longer periods than nonadsorbed organics. Data are not available, however, to validate this
25      hypothesis, nor is it known that such extractions would result in effective doses. It is unknown if
26      AMs in particle-overload conditions retain the same capacity to dissolve organics or whether this
27      process is slowed as well, which would also increase the time over which carcinogenic organics
28      are released and available in the overloaded lung. Data of Creasia et al. (1976) showing
29      increased retention of organics adsorbed to large particles could be consistent with this
30      possibility because 15 um particles may still be phagocytized by AM (Snipes and Clem, 1981;
31      Oberdorsteretal., 1991).
32
        2/1/98        •                            4-26       DRAFT-DO NOT CITE OR QUOTE

-------
 1      4.4.3. Extraction of Diesel Particle-Associated Organics by Alveolar Lung Cells and Other
 2            Cell Types
 3            Another more likely mechanism by which organic carcinogens (e.g., PAHs) may be
 4      extracted from DPM in the lung is either particle dissolution or extraction of organics from the
 5      particle surface within the phagolysosomes of AMs. This mechanism presupposes that the
 6      particles are internalized by these phagocytes. Specific details about the physicochemical
 7      conditions of the  intraphagolysosomal environment, where particle dissolution in AMs
 8      presumably occurs in vivo, have not been well characterized. However, it is known that the
 9      phagolysosomes constitute an acidic (pH 4 to 5) compartment in macrophages (Nilsen et al.,
10      1988; Ohkuma and Poole, 1978). The relatively low pH in the phagolysosomes has been
11      associated with the dissolution of some types of inorganic particles (some metals) by
12      macrophages (Marafante et al., 1987; Lundborg et al., 1984). but few studies provide quantitative
13      information concerning how organic constituents of diesel particles (e.g., B[a]P) may be
14      extracted in the phagolysosomes (Bond et  al., 1983). Whatever the mechanism, the end result is
15      a prolonged exposure of the respiratory epithelium to the gradual release of carcinogenic agents.
16            Quantitative data on how readily carcinogenic organics may be extracted from DPM in
17      the human lung are not available. As shown by Creasia et al. (1976), B[a]P adsorbed onto 15 to
18      30 ^m carbon particles is removed from the mouse lung at the      rate as the particles.
19      However, B[a]P adsorbed onto 0.5 to 1.0 (am carbon particles was eliminated approximately four
20      times faster than the clearance of the carbon particles. For the rat, Sun and co-investigators
21      (1984) have reported that the t)/2 for the lung clearance of B[a]P adsorbed onto DPM over a
22      period consistent  with alveolar phase clearance was about 18 days. This latter value is similar to
23      the t,/2 for the removal of particles without B[a]P from the rat's lung during the early, more rapid
24      component of alveolar phase clearance (Snipes et al., 1988; Snipes and Clem, 1981; Ferin, 1982;
25      Lehnert et al., 1989; Lehnert, 1989). These findings may suggest that the extraction of organic
26      components from carrier particles by AMs may differ among species, although differences in the
27      lung burdens administered in the investigations mentioned here may have influenced the
28      outcomes of the studies.
29            It should be pointed out that studies designed to examine the extraction of organics by
30      AMs have generally focused on B[a]P as a representative procarcinogen associated with diesel
31      particles. Numerous other agents with carcinogenic activity are also associated with diesel
32      particles; these chemical constituents may be extracted from DPM with in vivo kinetics that
33      differ more or less from those of B[a]P.  Thus, existing dosimetry models that incorporate
34      desorption of B[a]P from DPM as a representative organic constituent (Yu and Yoon, 1988) may
        2/1/98                                    4-27       DRAFT-DO NOT CITE OR QUOTE

-------
 1      not accurately reflect the actual bioavailability of other procarcinogenic agents on DPM.  As
 2      discussed 'in the next section, however, any error in this respect is likely to be minor.
 3
 4      4.4,4. Bioavailability of Adsorbed Compounds as a Function of Particle Clearance Rates
 5            and Extraction Rates of Adsorbed Compounds
 6            The bioavailability of toxic organic compounds adsorbed to particles can be influenced
 7      by a variety of factors. Although the agent may be active while present on the particle, most
 8      particles are taken up by AMs, a cell type not generally considered to be a target site.  To reach
 9      the target site, the agent must first elute from the particle surface. Although elution can be
10      considered a necessary step, it may not always be sufficient. The agent must then diffuse out of
11      the AM into the extracellular fluid and be absorbed by a target cell (e.g., a Type I cell). In
12      analyzing phagolysosomal dissolution of various ions from particles in the lungs of Syrian
13      golden hamsters, Godleski et al. (1988) demonstrated that solubilization did not necessarily
14      result in clearance of the ions and that binding of the solubilized components to cellular and
15      extracellular structures occurred. It is reasonable to assume that phagocytized DPM particles
16      may be subject to similar processes and that these processes would be important in determining
17      the rate of bioavailability of the particle-bound constituents of DPM.  Inability of these
18      constituents to penetrate target cells or to diffuse into the bloodstream is another possible factor
19      limiting bioavailability to lung target cells. Nevertheless, until further research demonstrates
20      otherwise, it is assumed that the rate-limiting factor in the bioavailability of particle-bound
21      organics is the desorption rate from the particle surface.
22            The long-term clearance t,/2 of diesel particles from the lungs of rats in the non-
23      overloaded state was shown to range from about 2 to 3 mo (Chan et'al, 1981;  Chan et al., 1984;
24      Griffis et al., 1983; Lee et al., 1983). Clearance rate data for DPM in humans  are not available.
25      For other types of insoluble particles such as polystyrene, however, t,/2 values  are close to 1 year
26      (Bohning et al., 1982). The clearance t,/2 values for B[a]P and 1-NP from the diesel particle
27      surface, on the other hand, were reported to be only about 18 and 36 days, respectively (Sun et
28      al.,  1984; Bond et al., 1986). The lower tm values for clearance of the organics compared with
29      particles themselves indicate that most of the organics are being eluted prior to particle clearance,
30      especially in humans.
31             For humans, assuming an elution t,/2 on the order of 2 to 4 weeks, well over 90% of the
32      organics should desorb from the particles.  Even in rats, with more rapid particle clearance rates,
33      most of the organics can be expected to be eluted. Whether particle overload in the lung results
34      in a change in elution rates of the organics is not known. If lung burden of particulate matter is
35      the  proper dosimetric factor for induction of pathology or carcinogenesis, target organ dose

        2/1/98                                    4-28       DRAFT-DO NOT CITE OR QUOTE

-------
  1      would be predicted to increase more rapidly than exposure concentration under lung overload
  2      conditions,
  3             Gerde et al. (1991a,b) described models simulating the effect of particle aggregation and
  4      PAH content on the rate of PAH release in the lung.  The investigators used three models, one of
  5      which simulated a low-dose situation where only the adsorbed layer of PAH is released from the
  6      carrier particle and two of which simulated desorption of PAHs using high PAH levels (with and
  7      without an inert carrier dust). Based on the theoretical results obtained, particle retention would
  8      be of lesser importance for low-dose situations in which particle-associated PAHs would be
  9      rapidly released at the site of particle deposition and not necessarily at the site of particle
10      retention.  Frequent low-level exposure, therefore, may result in sustained exposure of target
11      cells and subsequently greater likelihood of tumor formation at the site of initial deposition. For
12      the high-dose situations, as represented by instillation experiments in animals, critical doses to
13      cells are likely to occur at the site of retention because slow release from the particles may
14      increase the dose of metabolites (and increase the risk of tumor formation) before the parent PAH
15      compound can be cleared from the lungs. Generally, the models suggested that the local
16      disposition of PAHs would be more dependent on the behavior of dissolved PAHs in the tissues
17      after their  release from the carrier particles than by interactions between the PAHs and the carrier
18      particles.  The model predictions were consistent with findings from laboratory animal studies
19      that showed longer PAH retention with higher exposures and longer retention for instillation
20      administration versus inhalation exposure.                                .
21             Studies by Gerde et al. (1993a,b,c) using beagle dogs provided additional data regarding
22      the dosimetry of inhaled PAH supportive of the previously discussed models. In the Gerde et al.
23      (1993a) study, the dogs were exposed to an aerosolized bolus of PAH crystals (phenanthrene or
24      B[a]P) in a single breath.  PAH clearance was measured by monitoring PAH levels in the
25      systemic circulation. Clearance from the alveolar region was dependent on lipophilicity of the
26      PAH;  clearance of highly lipophilic PAHs (i.e., B[a]P) was limited by  diffusion of the chemical
27      through the alveolar septa, while clearance of moderately lipophilic PAHs (i.e., phenanthrene)
28      was limited by rate of perfusion of the blood. Therefore, bronchi, with their thicker epithelia,
29      would be at greater risk than alveoli for PAH-induced toxicity at the portal of entry. In the Gerde
30      et al. (1993b) study, small volumes of saline containing either dissolved B[a]P or phenanthrene
31      or a suspension of particulate solvent green or macroaggregated albumin (MAA) were instilled
32      into the mucous lining layer of a primary bronchus or distal tracheas.  The highly lipophilic
33      B[a]P was cleared via the mucociliary escalator, some being cleared very rapidly (>90 mm/min).
34      The portion of B[a]P that penetrated the bronchial epithelium exhibited a half-time in the range
35      of 1.4 h, indicating a diffusion-limited uptake of B[a]P by the airways. Although mucociliary

        2/1/98     -                              4-29       DRAFT-DO NOT CITE OR QUOTE

-------
 1      clearance is rapid for most of the lipophilic toxicants, the long retention time for the portion that
 2      penetrates the epithelium is sufficient for substantial metabolism to occur, resulting in a potential
 3      for local toxicity. Gerde et al. (1993c) used the data from the previously described studies to
 4      validate models of alveolar clearance (Gerde et al., 1991b), mucous lining penetration (Gerde
 5      and Scholander,  1987), and bronchial wall penetration (Gerde et al., 1991b). The analysis
 6      provided a reasonable validation of the transport models for these structural regions of the lung.
 7      Specifically, alveolar clearance and mucociliary clearance are primarily via molecular diffusion,
 8      whereas clearance from bronchial walls involves diffusion, metabolism of a portion of the PAH
 9      load, and endocytosis.  Such findings suggest that the bronchial epithelium may be especially
10      vulnerable to toxicity induced by diffusion-limited lipophilic substances.
11
12      4.5. CONSIDERATIONS FOR DOSIMETRY MODELING
13             Although more than one approach is possible in the development of dosimetry models for
14      inhaled DPM, several dosimetry parameters are common to any approach. These include
15      ventilatory rates  and volumes, tracheobronchial and alveolar surface area, tracheobronchial and
16      alveolar deposition efficiency, and tracheobronchial and alveolar clearance rates. If the particle-
17      adsorbed organics are considered relevant to the carcinogenic response of the lung, elution t,/2
18      values of these chemicals should be included in the model.  The dosimetry models must consider
19      not only species  differences but also  the effects of extrapolating from high exposure
20      concentrations, which result in an inhibition of particle clearance from the lungs caused by
21      overload.  Some models include additional parameters such as transport to lung-associated lymph
22      nodes and clearance of the interstitial compartment, but experimental data are not available to
23      validate many of the assumptions used in these models.
24             An important consideration is ventilation rate.  Numerous estimates of alveolar
25      ventilation rates  are available for both humans and rats. Nevertheless, such estimates are still a
26      source of considerable potential error because they are usually resting values. Actual respiration
27      can vary widely with activity or with exposure conditions.  Human activity  levels are also highly
28      variable.  Although it may be necessary to use a single mean value in a model, it would be useful
29      to include risk estimates for individuals having much greater daily ventilation exchange rates
30      resulting from participation in endurance sports or performance of heavy labor.
31             Another variable not discussed previously is the adjustment of dose based on metabolic
32      rate. It has been EPA policy to consider that effective dose varies with metabolic rate.
33      Arguments for and against this presumption are beyond the scope of this document.  The primary
34      consideration here is the degree of adjustment. EPA has traditionally adjusted for species
35  .    differences in metabolic rate based on the 2/3 power of body weight as a surrogate for body

        2/1/98                                    4-30       DRAFT-DO NOT CITE OR QUOTE

-------
  1      surface area.  This factor is being reappraised within the Agency, and a preliminary proposal to
  2      alter this adjustment to the 3/4 power has been made.
  3             A major consideration in the development of dosimetry models is a judgment concerning
  4      which fraction of exhaust is responsible for inducing lung cancer. As is discussed in Chapter 10,
  5     • two approaches are used for quantitative assessment based on animal data. In the first, cancer is
  6      assumed to be induced by the organic constituents present on the particle surface, primarily
  7      PAHs and nitropyrenes.  As discussed herein, in an appropriate model, the effective  dose will
  8      correlate closely with the deposited dose, with only minor corrections for lung overloading. In
  9      the second approach, it is assumed that retained particle burden in the lung fully accounts for the
10      induced lung tumors.  A more specific dose parameter for this second approach may be the
11      surface area of the retained particles, as discussed in Chapter 10. In studies with fine and
12      ultraflne TiO2 particles; Oberdorster et al. (1994) reported that the ultraflne particles resulted in
13      prolonged pulmonary retention, increased translocation to and persistence in the interstitium,
14      increased Type II cell proliferation, presence of fibrotic foci, and impairment of alveolar
15      macrophage function in rats exposed for 12 weeks. Overall, there was a notable correlation
16      between particle surface area and effects. In this case, modeling must account for slowing of
17      clearance during lung overload as well as large differences in normal clearance rates between rats
18      and humans.  For this approach, if exposures are at overload levels, low-dose extrapolation may
19      result in a risk estimate considerably different from one based on target organ dose of organics.
20      A third approach could be proposed in which both exhaust components may be operating
21      simultaneously, with the PAHs initiating the carcinogenic process and the particles promoting
22      the process by inducing cell proliferation.
23
24      4.6.  SUMMARY
25             Because only very limited evidence exists for diesel exhaust-induced tumors  at non-
26      pulmonary sites or for tumor induction by the gaseous fraction alone, dosimetry considerations
27   •   were limited to either whole exhaust or particulate matter deposited in the lungs. Dosimetric
28      considerations were further limited to the alveolar region of the respiratory tract for several
29      reasons.  First of all, most deposited particulate matter is transported by mucociliary  transport
30      from the  conducting airways in less than 1 day, and exposure to diesel exhaust does not appear to
31      significantly inhibit this process.  However, as previously discussed, there may also be a long-
32      phase retention, especially if tracheobronchial clearance is impaired.  In general, the  rapid
33      clearance in the conducting airways reduces the time for extraction of organics from  the particle
34      surface (although there is evidence for some relatively rapid removal of organics), and some
        2/1/98         .                            4-31       DRAFT-DO NOT CITE OR QUOTE

-------
 1      AMs in the mucociliary escalator may contain particles desorbed of organics.  Finally, most of
 2      the pathologic and carcinogenic effects occur at or distal to the terminal bronchioles in the rat.
 3             Clearance of DPM from the alveolar region varied from about 2 mo in rats to an estimate
 4      of 1 year in humans. Under high-exposure regimes, lung overload occurred in rats, leading to
 5      slower or near cessation of clearance, thereby increasing lung burdens even further. In addition,
 6      with large lung burdens, uptake of particles by Type I cells, passage into the interstitium, and
 7      transport to lung-associated lymph nodes were increased.  Factors considered to be involved in
 8      clearance inhibition included loss of AM mobility with large particle loads and a tendency for
 9      AMs to aggregate and become immobilized.
10             Most biological fluids tested, including lung lavage fluid and serum, were relatively
11      ineffective in the extraction of organic agents adsorbed to the diesel particle surface. Particles
12      deposited in the alveolar region, however, are rapidly phagocytized by AMs, which are more
13      effective in this regard. Although actual elution t1/2 values of organics from phagocytosed
14      particles  were difficult to obtain, they were generally less than those for the particles themselves,
15      indicating that most organics are released even without inhibition of particle clearance.  The
16      gradual elution also prolonged the residence time of organics in the lungs compared with pure
17      organic agents such as B[a]P, possibly avoiding overloading of biological activation systems and
18      thus increasing their effectiveness.
19             In the development of a dosimetry model to allow both low-dose extrapolation and
20      extrapolation of DPM bioassay data from laboratory animals to humans, several parameters must
21      be accounted for. These include deposition efficiency, particle clearance rates, desorption rates
22      of organics from the particle surface (if other than a particle effect is being considered), lung
23      surface area, and particle  surface area. The respiratory rates and volumes are highly variable in
24      both laboratory animals and humans and are also  determinants of deposition efficiency. Animal
25      estimates are often based on published values collected under resting conditions.  Respiration,
26      however, may be inhibited by the irritant gases present in diesel exhaust, although less so at low
27      dilution ratios. On the other hand, respiration may be either greater or less than estimated resting
28      values, depending on whether exposures were carried out at night when the animals are likely to
29      be awake and active or during the day when they  are more likely to be asleep. Human
30      respiratory exchange rates are also quite variable, with the physically active segment of the
31      population at potentially greater risk because of higher doses resulting from higher respiration
32      rates.                                                          ,
33             Adjustment for particle clearance rate is necessary for two reasons. First of all, many of
34      the  laboratory animal  experiments were conducted under exposure regimes resulting in an
35      inhibition of clearance caused by an accompanying lung burden overload.  If lung burden of

        2/1/98     .                               4-32       DRAFT-DO NOT CITE OR QUOTE

-------
  1       particulate matter is considered to be the proper dosimetric variable, then the disproportionately

  2       large lung burdens at high levels of exposure must be adjusted for.  Second, even under low-

  3       exposure regimes, clearance is slower in humans than in rats. If the correct dosimetric variable,

  4       on the other hand, is particle-free organic matter, a smaller adjustment for variations in particle

  5       clearance rates is required because most of the organics are likely to be eluted from the particles •

  6       deposited in the alveolar region, even at normal clearance rates. Nevertheless, some adjustment

  7       is still necessary, because the organics are seldom all eluted.

  8                                                                            .

  9       4.7. REFERENCES
10
11       Adamson, IYR; Bowden, DH. (1978) Adaptive responses of the pulmonary macrophagic system to carbon. II.
12       Morphologic studies. Lab Invest 38:430-438.
13
14       Adamson, IYR; Bowden, DH. (1981) Dose response of the pulmonary macrophagic system to various particulates
15       and its relationship to transepithelial passage of free particles. Exp Lung Res 2:165-175.
16
17       Alarie, Y. (1966) Irritating properties of airborne materials to the upper respiratory tract. Arch Environ Health  .
18       13:433-449.           "                                     •                   "
19
20       Alarie, Y. (1973) Sensory irritation to the upper airways by airborne chemicals. Toxicol Exp  Pharmacol 24:279-
21   '    297,
22
23       Aronson, M. (1963) Bridge formation and cytoplasmic flow between phagocytic cells. J Exp  Med 118:1084-1102.
24
25       Bailey, MR; Fry, FA; James, AC. (1982) The long-term clearance kinetics of insoluble particles from the human
26       lung. Ann Occup Hyg 26:273-271.
27
28       Ball, LM; King, LC. (1985) Metabolism, mutagenicity, and activation of 1-nitropyrene in vivo and in vitro. Environ
29       Int 11:355-361.
30
31       Battigelli, MC; Hehnstenberg, F; Mannella, RJ; Thomas, AP. (1966) Mucociliary activation. Arch Environ Health
32       12:460-467.
33
34  •     Bevan, DR; Ruggio, DM. (1991) Bioavailability in vivo of benzo[a]pyrene adsorbed to diesel particulate. Toxicol
35       Ind Health 7:125-139.
36
37       Bohning, DE; Cohn, SH; Lee, HD; Atkins, HL. (1980) Two-phase deep-lung clearance in man. In: Sanders, CL;
38       Cross, FT; Dagle, GE; Mahaffey, JA, eds. Pulmonary toxicology of respirable particles. Technical Information
39       Center, U.S. Dept of Energy, pp. 149-161.
40  '
41       Bohning, DE; Atkins, HL; Cohn,  SH. (1982) Long-term particle clearance in man: normal and impaired. Ann
42       Occup Hyg 26:259-271.
43
44       Bond, JA; Mitchell, CE; Li, AP. (1983) Metabolism and macromolecular covalent binding of benzo(a)pyrene in
45       cultures of Fischer-344 rat lung Type II epithelial cells. Biochem Pharmacol 32:3771-3776.
46
47       Bond, JA; Sun, JD; Medinsky, MA; Jones, RK; Yeh, HC. (1986) Deposition, metabolism, and excretion of 1-
48       [MC]nitropyrene and  l-[14C]nitropyrene coated on  diesel exhaust particles as influenced by exposure concentration.
49       Toxicol Appl Pharmacol 85:102-117.


         2/1/98                                       4-33        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Brain, JD; Mensah, GA. (1983) Comparative toxicology of the respiratory tract. Am Rev Respir Dis 128:887-590
  2
  3    .   Brooks, AL; Wolff, RK; Royer, RE; Clark, CR; Sanchez, A; McClellan, RO. (1981) Deposition and biological
  4       availability of diesel particles and their associated mutagenic chemicals. Environ Int 5:263-267.
  5
  6       Chan, TL; Lee, PS; Hering, WE. (1981) Deposition and clearance of inhaled diesel exhaust particles in the
  7      • respiratory tract of Fischer rats. J Appl Toxicol 1:78-82.
  8
  9       Chan,, TL; Lee, PS; Hering, WE. (1984) Pulmonary retention of inhaled diesel particles after prolonged exposures to
10       diesel exhaust. Fundam Appl Toxicol 4:624-631.
11
12       Cohen, D; Arai, SF; Brain, JD. (1979) Smoking impairs long-term dust clearance frorn the lungs. Science 204:514-
13       517.
14
15       Crapo, JD; Young, SL; Fram, EK; Pinkerton, KE; Barry, BE; Crapo, RO. (1983) Morphometric characteristics of
16       cells in the alveolar region of mammalian lungs. Am Rev Respir Dis 128:S42-S46.
17
18       Creasia, DA; Poggenburg, JK;  Nettesheim, P, Jr. (1976) Elution of benzo[a]pyrene from carbon particles in the
19       respiratory tract of mice. J Toxicol Environ Health 1:967-975.
20
21       Cuddihy, RG; Yeh, HC. (1986) Model analysis of respiratory tract clearance of particles inhaled by people. In:
22       Muggenburg, BA; Sun, JD, eds. Annual report of the Inhalation Toxicology Research Institute. Albuquerque, NM:
23       Lovelace Biomedical and Environmental Research Institute; LMF-115, pp. 140-147.
24
25       Dahl, AR; Schlesinger, RB; Heck, HD'A; Medinsky, MA; Lucier,  G. (1991) Comparative dosimetry of inhaled
26       materials: differences among animal species and extrapolation to man. Fundam Appl Toxicol 16:1-13.
27
28       Evans, Z; Shami, SG; Martinez, LA. (1986) Enhanced proliferation of pulmonary alveolar macrophages after
29       carbon instillation in mice depleted of blood monocytes by strontium-89. Lab Invest 54:154-159.
30
31       Felicetti,  SA; Wolff, RK; Muggenburg, BA. (1981) Comparison of tracheal mucous transport in rats, guinea pigs,
32       rabbits, and dogs. J Appl Physiol 51:1612-1617 (as cited in Brain and Mensah, 1983).
33
34       Ferin, J. (1982) Pulmonary alveolar pores and alveolar macrophage-mediated particle clearance. Anat Rec 203:265-
35       272.
36
37       Ferin, J; Feldstein, ML. (1978) Pulmonary clearance and hilar lymph node content in rats after particle exposure.
38       Environ Res 16:342-352.
39
40       Freedman, AP; Robinson, SE.  (1988) Noninvasive magnetopneumographic studies of lung dust retention and
41    .   clearance in coal miners. In: Frantz, RL; Ramani, RV, eds. Respirable dust in the mineral industries; health effects,
42       characterization and control. University Park, PA: Pennsylvania State University, pp. 181-186.
43
44       Frey, JW; Corn, M. (1967) Physical and chemical characterization of particles in diesel exhaust. Am Ind Hyg Assoc
45       J 28:468-478.
46
47       George, G; Hook, GER. (1984) The pulmonary extracellular lining. Environ Health Perspect 55:227-237.
48
49       Gerde, P; Scholander, P. (1987) A hypothesis concerning asbetsos carcinogenicity: the migration of lipophilic
50       carcinogens in adsorbed lipid bilayers. Ann Occup Health 31:395-400.
51
52       Gerde, P; Medinsky, MA; Bond, JA. (1991a) Particle-associated polycyclic aromatic hydrocarbons—a reappraisal
53       of their possible role in pulmonary carcinogenesis. Toxicol Appl Pharmacol  108:1-13.
54                                                      '        .


         2/1/98          •                                4-34        DRAFT-DO NOT CITE OR  QUOTE

-------
  1      Gerde. P; Medinsky, MA; Bond, JA. (1991b) The retention of polycyclic aromatic hydrocarbons in the bronchial
  2      airways and in the alveolar region—a theoretical comparison. Toxicol Appl Pharmacol 107:239-252.
  3
  4      Gerde. P; Muggenburg, BA; Hoover, MD; Henderson, RF. (1993a) Disposition of polycyclic aromatic
  5      hydrocarbons in the respiratory tract of the beagle dog. Toxicol Appl Pharmacol 121:313-318.
  6
  7      Gerde, P; Muggenburg, BA; Sabourin, PA; Harkema, JR; Hotchkiss, JA; Hoover, MD; Henderson, RF. (1993b)
  8      Disposition of polycyclic aromatic hydrocarbons in the respiratory tract of the beagle dog. II. The conducting
  9      airways. Toxicol Appl Pharmacol 121:319-327.
10
11      Gerde, P; Muggenburg, BA; Henderson, RF.'(1993c) Disposition of polycyclic aromatic hydrocarbons in the
12      respiratory tract of the beagle dog. III. Mechanisms of the dosimetry. Toxicol Appl Pharmacol 121:328-334.
13                                                                               '
14      Gibb, FR; Morrow, PE. (1962) Alveolar clearance in dogs after inhalation of an iron 59 oxide aerosol. J Appl
15      Physiol 17:429-432.
16
•17      Godleski, JJ; Steams, RC; Katler, MR; Brain, JD. (1988) Particle dissolution in alveolar macrophages assessed by
18      electron energy loss analysis using the Zeiss CEM902 electron microscope. J Aerosol Med 1:198,
19                                                                                             .
20      Griffis, LC; Wolff, RK; Henderson, RF; Griffith, WC; Mokler, BV; McClellan, RO. (1983) Clearance of diesel soot
21      particles from rat lung after subchronic diesel exhaust exposure. Fundam Appl Toxicol 3:99-103.
22
23      Gurley, GR; Spall, WD; Valdez, JG; London, JE; Dethloff, LA; Lehnert, BE. (1988) An HPLC procedure for the
24      analysis of proteins in lung lavage fluid. Anal Biochem 172:465-478.
25
26      Harmsen, AG; Muggenburg, BA; Snipes, MB; Bice, DB. (1985) The role of macrophages in particle translocation
27      from lungs to lymph  nodes. Science 230:1277-1280.
28
29      Heinrich, U; Muhle, H; Takenaka, S; Ernst, H; Fuhst, R; Mohr, U;.Pott, F; Staber, W. (1986) Chronic effects on the
30      respiratory tract of hamsters, mice and rats after long-term inhalation of high concentrations of filtered and
31      unfiltered diesel engine emissions. J Appl Toxicol 5:383-395.
32
33      Heppleston, AG. (1961) Observations on the disposal of inhaled dust by means of the double exposure technique.
34      In; Davies, CN-, ed. Inhaled particles and vapors. New York: Pergamon; pp. 320-325.
35
36      Heppleston, AG; Young, AE. (1974) Uptake of inert particulate matter from alveolar cells: an ultrastructural study.
37     -J Fathom 1:159-164.
38
39      Iwai, K; Udagawa, T; Yamagishi, M; Yamada, H. (1986) Long-term inhalation studies of diesel exhaust on F344
40      SPF rats.  Incidence of lung cancer and lymphoma. In: Ishinishi, N; Koizumi, A; McClellan, R; Stober, W, eds.
41      Carcinogenic-and mutagenic effects of diesel engine exhaust. Amsterdam: Elsevier, pp. 471-485.
42
43      Kane, LE; Alarie, Y.  (1978) Evaluation of sensory irritation from acrolein-formaldehyde mixtures. Am Ind Hyg
44      Assoc J 39:270-274.
45
46      Kane, LE; Alarie, Y.  (1979) Interactions of sulfur dioxide and acrolein as sensor)' irritants. Toxicol Appl Pharmacol
47      48:305-315.
48
49      King, LC; Kohan, MJ; Austin, AC; Claxton, LD; Huisingh, JL. (1981) Evaluation of the release of mutagens from
50      diesel particles in the presence of physiological fluids. Environ Mutagen 3:109-121.
51
52      Kittelson, DB; Dolan, DF; Verrant, JA. (1978) Investigation of a diesel exhaust aerosol. Society of Automotive
53      Enaineers Technical  Paper No. 780109 (Transactions of SAE).
54


         2/1/98   '                                     4-35         DRAFT-DO NOT CITE OR QUOTE

-------
 1       Lee, PS; Chan, TL; Hering, WE. (1983) Long-term clearance of inhaled diesel exhaust particles In rodents, J
 2       Toxico! Environ Health 12:801-813.
 3                                    ,
 4       Lee, KP; Trochimowicz, HJ; Reinhardt, CF. (1985) Transmigration of titanium dioxide (TiO,) particles in rats after
 5       inhalation exposure. Exp Mol Pathol 42:331-343.
 6
 7       Lee, KP; Henry, NW, III; Trochimowicz, HJ; Reinhardt, CF. (1986) Pulmonary response to impaired lung clearance
 8       in rats following excessive TiO, dust deposition. Environ Res 41:144-167.
 9
10       Lee, KP; Ulrjch, CE; Geil, RG; Trochimowicz, HJ. (1988) Effects of inhaled chromium dioxide dust on rats
11       exposed for two vears. Fundam Appl Toxicol 10:125-145,
12                      ' -                                        •               "  .
13       Lehnert, BE. (1988) Distribution of particles in alveolar macrophages during lung clearance. J Aerosol Med 1:206,
14
15       Lehnert, BE. (1989) Rates of disappearance of alveolar macrophages during lung clearance as a function of
16       phagocytized paniculate burden. 1989 Annual Meeting of the American Thoracic Society; May 14-17; Cincinnati,
17       OH.
18
19       Lehnert, BE; Morrow, PE. (1985) Association of 59iron oxide with alveolar macrophages during alveolar clearance.
20       Exp Lung Res 9:1-16.
21
22       Lehnert, BE; Cline, A; London, JE. (1989) Kinetics of polymorphonuclear leukocytes and their particle burdens
23       durina the alveolar clearance of a high lung burden of particles. The Toxicologist 9:A307,
24
25       Lippmann, M; Schlesinger, KB. (1984) Interspecies comparisons of particle deposition and mucociliary clearance in
26       tracheobronchial airways. J Toxicol Environ Health 13:441-469.
27
28       Lundborg, M; Lind, B; Camner, P. (1984) Ability of rabbit alveolar macrophages to dissolve metals. Exp Lung Res
29       7:11-22.
30
31       Marafante, E; Lundborg, M; Vahter, M; Camner, P. (1987) Dissolution of two arsenic compounds by rabbit alveolar
32       macrophages in vitro. Fundam Appl Toxicol 8:382-388.
33
34   '    McClellan, RO; Brooks, AL; Cuddihy, RG; Jones, RK; Mauderly, JL; Wolff, RK. (1982) Inhalation toxicology of
35       diesel exhaust particles. In: Lewtas, J, ed. Toxicological effects of emissions from diesel engines. Amsterdam:
36      " Eisevier-North Holland, pp. 99-120.
37
38       Morrow, PE. (1966) International Commission on Radiological Protection (ICRP) task group on lung dynamics,
39       deposition and retention models for internal dosimetry of the human respiratory tract. Health Phys 12:173.
40
41       Morrow, PE. (1988) Possible mechanisms to explain dust overloading of the lungs.  Fundam Appl Toxicol 10:369-
42       384.
43           '   •        .                .
44       Muhle, H; Bellman, B; Creutzenberg, O; Heinrich, U; Keckar, M; Mermelstein, R. (1990) Dust overloading of lungs
45   '    after exposure of rats to particles of low solubility: comparative studies. J Aerosol Sci 21:374-377.
46
47       Nilsen, A; Nyberg, K; Camner, P. (1988) Intraphagosomal pH in alveolar macrophages after phagocytosis of
48       fiuorescein-labelled yeast particles. Exp Lung Res 14:197-207.
49
50       OberdOrster, G; Ferin, J; Morrow, PE. (1992) Volumetric loading of alveolar macrophages (AM): A possible basis
51       for diminished AM-mediated particle clearance. Exp Lung Res 18:87-104.
52
53    "   Oberdorster, G; Ferin, J; Lehnert, BE. (1994) Correlation between particle size, in vivo particle persistence, and
54       lung injury. Environ Health Perspect 102(Suppl. 5):173-179.


         2/1/98                                         4-36        DRAFT-DO NOT CITE OR QUOTE

-------
   1       Ohkuma, S; Poole, B, (1978) Fluorescence probe measurement of the intralysosomal pH in living cells and the
   2       perturbation of pH by various agents, Proc Natl Acad Sci USA 75:3327-3331.
   3
   4       Pepelko, WE. (1987) Feasibility of dose adjustment based on differences in long-term clearance rates of inhaled
   5       particulate matter in humans and laboratory animals. Regul Toxicol Pharmacol 7:236-252.
   6
   7       Phalen, RF; Oldham, MJ. (1983) Traeheobronehial airway structure as revealed by casting techniques. Am Rev
   8       RespirDis 128:S1-S4.
   9
 10       Powdrill, J; Buckley, C; Valdez, YE; Lehnert, BE. (1989) Airway intra-luminal macrophages: origin and role in
 11       lung clearance. The Toxicologist 9:A306.
 12
 13       Pritchard, IN. (1989) Dust overloading causes impairment of pulmonary clearance: evidence from rats and humans.
 14       Exp Pathol 37:39-42.
 15
 16       Raabe, OG. (1982) Deposition and clearance of inhaled aerosols. In: Witschi, H; Nettesheim, P, eds. Mechanisms in
 17       respiratory toxicology, Vol. 1. Boca Raton, FL: CRC Press, pp. 27-76.          •                       .
 18
 19       Raabe, OG; Al-Bayati, MA; league, SV; Rasolt, A. (1988) Regional deposition of inhaled monodisperse coarse and
 20       fine aerosol particles in small laboratory animals. In: McCallum, RI, ed. Inhaled particles IV.: Proceedings of an
 21       international  meeting. Cambridge, UK: British Occupational Hygiene Society.
 22
 23    .   Riley, PA; Dean, RT. (1978) Phagocytosis of latex particles in relation to the cell cycle in 3T3 cells. Exp Cell Biol
 24       46:367-373.
 25
 26       Saffiotti, U; Borg, SA; Grote, MI; Karp, DB. (1964) Retention rates of particulate carcinogens in lungs. Chicago
 27       Med School Quart 2.24:10-17.                •                 "             '
 28
 29       Sandusky, CB; Cowden, MW; Schwartz, SL. (1977) Effect of particle size on regurgitive exocytosis by rabbit
 30       alveolar macrophages. In: Pulmonary macrophage and epithelial cells. CONF-760927, National Technical
 31       Information Service, U.S. Dept. of Commerce, pp. 85-105.
 32
 33       Schlesinger, RB. (1985) Comparative deposition of inhaled aerosols in experimental animals and humans: a review.
 34       J Toxicol Environ Health 15:197-214.
 35
 36       Schum, GM;  Yeh, HC. (1980) Theoretical evaluation of aerosol deposition in anatomical models of mammalian
 37-      lung airways. Bull Math Biol 42:1-25.
 38
 39       Siak, JS; Chan, TL; Lee, PS. (1980) Diesel particulate extracts in bacterial test systems. In: Health effects of diesel
 40       engine emissions: proceedings of an international symposium, Vol.  1. Health Effects Research Laboratory, Office of
 41       Research and Development, U.S. Environmental Protection Agency, Cincinnati, OH, pp. 245-262.
 42
 43       Smith.TJ. (1985) Development and application of a model for estimating alveolar and interstitial dust levels. Ann
 44       Occup Hyg 29:495-516.
 45       '         •         •
 46       Snipes, MB; Clem, MF. (1981) Retention of microspheres in the rat lung after intratracheal instillations. Environ
 47       Res 24:33-41,
 48
- 49       Snipes, MB; Olson, TR; Yeh, HC. (1988) Deposition and retention patterns for 3-, 9-, and 15-mm latex
 50       microspheres inhaled by rats and guinea pigs. Exp Lung Res 14:37-50.
 51
 52       Sorokin, SP;  Brain, JD. (1975) Pathways of clearance in mouse lungs exposed to iron oxide aerosols, Anat Rec
 53       181:581-625.
 54  ,,


          2/1/98                                          4-37         DRAFT-DO NOT CITE OR QUOTE

-------
 1       Stahlhofen, W; Gebhart, J; Heyder, J. (1980) Experimental determination of the regional deposition of aerosol
 2       particles in the human respiratory tract. Am Ind Hyg Assoc J 41:385-398.
 3
 4       Stahlhofen, W; Hoebrich, R; Rudolf, G; Scheuch, G. (1990) Short-term and long-term clearance of particles from
 5       the upper human respiratory tract as a function of particle size. J Aerosol Sci 21(suppl. 1):5407-5410.
 6
 7       Stober, W; Einbrodt, HJ; Klosterkotter, W. (1967) Quantitative studies of dust retention in animal and human lungs
 8       after chronic inhalation. In: Davies, CN, ed. Inhaled particles and vapours II. Oxford: Pergamon Press, pp. 409-418.
 9
10       Stober, W; Morrow, PE; Hoover, MD. (1989) Compartmental modeling of the long-term retention of insoluble
11       particles deposited in the alveolar region of th'e lung. Fundam Appl Toxicol 13:823-842.
12
13       Stober, W; McClellan, RO; Morrow, PE. (1993) Approaches to modeling disposition of inhaled particles and  fibers
14       in the lung. In: Gardner, DE; Crapo, JD; McClellan, RO, eds. Toxicology of the lung. New York: Raven Press, pp.
15       527-601.
16
17       Stober, W; Morrow, PE; Koch, W; Morawietz, G. (1994) Alveolar clearance and retention of inhaled insoluble
18       particles in rats simulated by a model inferring macrophage particle load distributions. J Aerosol Sci 25:975-1002.
19
20       Strom,  KA. (1984) Response of pulmonary cellular defenses to the inhalation of high concentrations of diesel
21       exhaust. J Toxicol Environ Health 13:919-944.
22
23       Strom,  KA; Chan, TL; Johnson, JT. (1988) Pulmonary retention of inhaled submicron particles in rats: diesel
24       exhaust exposures and lung retention model. Ann Occup Hyg 32(Suppl. l):645-657.
25
26       Sun, JD; Wolff, RK; Kanapilly, GM. (1982) Deppsition, retention, and biological fate of inhaled benzo(a)pyrene
27       adsorbed onto ultrafine particles and as pure aerosol. Toxicol Appl Pharmacol 65:231-244.
28          .
29       Sun, JD; Wolff, RK; Kanapilly, GM; McClellan, RO. (1984) Lung retention and metabolic fate of inhaled
30       benzo(a)pyrene associated with diesel exhaust particles. Toxicol Appl Pharmacol 73:48-59.
31
32       Ulrich, C; Haddock, M; Alarie, Y. (1972) Airborne chemical irritants: role of the trigeminal nerve. Arch Environ
33       Health  24:37-42.
34
35       U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides. Research
36       Triangle Park, NC: Environmental Criteria and Assessment Office. EPA report no. EPA-600/8-82-029aF-cF.
37
38       Vastag, E; Matthys, HZ; Samboki, G; Kohler, D;  Daikeler, G. (1986). Mucociliary clearance in smokers. Eur J
39       RespirDis68:107-113.                                                           .         •
40
41       Vostal, JJ. (1983) Unavailability and biotransformation of the mutagenic component of particulate emissions
42       present in motor exhaust samples. Environ Health Perspect 47:269-281.
43
44       Vostal, JJ; Schreck, RM; Lee, PS; Chan, TL; Soderholm, SC. (1982)  Deposition and clearance of diesel particles
45       from the lung. In: Lewtas, J, ed. Toxicological effects of emissions from diesel engines. New York: Elsevier,  pp.
46       143-159.
47
48       Weibel, ER. (1963) Morphometry of the human lung. Berlin: Springer-Verlag.
49
50      White, HJ; Garg, BD. (1981) Early pulmonary response  of the rat lung to inhalation of high concentration of diesel
51       particles. J Appl Toxicol 1:104-110.
52
         2/1/98                                          4-38         DRAFT-DO NOT CITE OR QUOTE

-------
  1       Wolff, RK; Gray, RH. (1980) Tracheal clearance of particles as influenced by diesel exhaust exposure, Lovelace
  2       Inhalation Toxicology Research Institute annual report. Albuquerque, NM: Lovelace Biomedical and
  3       Environmental Research Institute; LMF-84, pp. 252-253.
  4
  5-      Wolff, RK; Henderson, RF; Snipes, MB; Sun, JD; Bond, JA; Mitchell, CE; Mauderly, JL; McClellan, RO, (1986)
  6       Lung retention of diesel soot and associated organic compounds. In: Ishinishi, N; Koizumi, A; McClellan, R;
  7      - Stober, W, eds. Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam; Elsevier, pp. 199-211.
  8
  9       Wolff, RK; Henderson, RF; Snipes, MB; Griffith, WC; Mauderly, JL; Cuddihy, RG; McClellan, RO. (1987)
10       Alterations in particle accumulation and clearance in lungs of rats chronically exposed to diesel exhaust. Fundam
11       ApplToxicol 9:154-166.
12                                                                               .
13       Wright, JR; Clements, JA. (1987) Metabolism and turnover of lung surfactant. Am Rev Respir Dis 135:426-444.
14
15       Xu, GB; Yu, CP. (1987) Deposition of diesel exhaust particles in mammalian lungs: a comparison between rodents
16       and man. Aerosol Sci Tech 7:117-123.
17
18       Yu, CP; Yoon, KJ. (1988) Determination of lung doses of diesel exhaust particulates. Presented at 5th annual
19       conference of Health Effects Institute; Colorado Springs, CO.
20
21       Yu, CP; Chen, YK; Morrow, PE. (1989) An analysis of alveolar macrophage mobility kinetics at dust overloading
22       of the lungs. Fundam ApplToxicol 13:452-459. ,
23
24
         2/1/98          •                               4-39        DRAFT-DO NOT CITE OR QUOTE

-------
                    5.  NONCANCER HEALTH EFFECTS OF DIESEL EXHAUST

 1            The objective of this chapter is to evaluate the nonearcinogenic health effects of diesel
 2      exhaust. Data pertaining to exposures to whole diesel exhaust will be presented first, followed
 3      by a comparison of the effects of filtered and unfiltered exhaust. Filtered exhaust consists of the
 4      gaseous components of the exhaust without the associated particulate matter.
 5
 6      5.1. HEALTH EFFECTS OF WHOLE EXHAUST
 7      5.1.1. Human Data
 8      5.1.1.1. Short-Term Exposures
 9            In a controlled human study, Rudell et al. (1990) exposed eight healthy, nonsmoking
10      subjects for 1 h to diluted diesel exhaust. The exposure atmosphere corresponded to about 100
11      ug/m3 of particles, 3.7 ppm NO, 1.6 ppm NO2, and 0.5 mg/m3 formaldehyde (WHO, 1994).
12      Bronchoalveolar lavage was performed prior to exposure and after exposure. Neutrophils were  .
13      significantly increased after exposure, and the phagocytosis of opsonized yeast cells in vitro by
14      alveolar macrophages was reduced after exposure. Lymphocytes were unchanged.  This study
15      provides an intriguing glimpse of the effect of low-level diesel exposure  in humans, but only one
16      exposure level was used,  the number of subjects was low, and a limited range of endpoints was
17      reported, so the data are inadequate to generalize about the human response. To date, no
18      well-controlled chamber study has been conducted using methodologies  for assessing subtle lung
19      inflammatory reactions.
20            Rudell et al. (1996) exposed volunteers to diesel exhaust for 1 h in an exposure chamber.
21      Light work on a bicycle ergometer was performed during exposure. Exposures included either
22      unaltered diesel exhaust or exhaust with particle numbers reduced 46% by a particle trap. The
23      engine used was a new Volvo model 1990, a six-cylinder direct-injection turbocharged diesel
24      with an intercooler, which was run at a steady speed of 900 rpm during the exposures.
25      Comparison of this study with others was difficult because neither exhaust dilution ratios nor
26      particle concentrations were reported. Carbon monoxide concentrations  of 27-30 ppm and NO of
27      2.6-2,7 ppm, however, suggested particle concentrations may have equaled several  mg/m3. The
28      most prominent symptoms during exposure were irritation of the eyes and nose and an
29      unpleasant smell. Both airway resistance and specific airway resistance increased significantly
30      during the exposures. Despite the 46% reduction in particle numbers by  the trap, effects on
31      symptoms and lung function were not significantly attenuated.
32            Kahn et al. (1988) reported the occurrence of 13 cases of acute overexposure to diesel
33      exhaust among Utah and  Colorado coal miners.  Twelve miners had symptoms of mucous

        2/1/98                                   5-1       DRAFT-DO NOT  CITE OR QUOTE

-------
 1      membrane irritation, headache, and lightheadedness. Eight individuals reported nausea; four
 2      reported a sensation of unreality; four reported heartburn; three reported weakness, numbness,
 3      and tingling in their extremities; three reported vomiting; two reported chest tightness; and two
 4      others reported wheezing.  Each miner lost time from work because of these symptoms, which
 5      resolved within 24 to 48 h.  No air monitoring data were presented; poor work practices were
 6      described as the predisposing conditions for overexposure.
 7            El Batawi and No weir (1966) reported that among 161 workers from two garages where
 8      diesel-powered buses were serviced and repaired, 42% complained of eye irritation, 37% of
 9      headaches, 30% of dizziness, 19% of throat irritation, and 11% of cough and phlegm. Ranges of
10      mean concentrations of diesel exhaust components in the two diesel bus garages were as follows:
11      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
12      particulate matter; the highest concentrations were obtained close to the exhaust systems of the
13      buses.
14            Eye irritation was reported by Battigelli (1965) in six subjects after 40 s of chamber
15      exposure to diluted diesel exhaust containing 4.2 ppm NO2,1 ppm SO2, 55 ppm CO, 3.2 ppm
16      total hydrocarbons, and 1 to 2 ppm total aldehydes; after 3 min and 20 s of exposure to diluted
17      diesel exhaust containing 2.8 ppm NO2,0.5 ppm SO2, 30'ppm CO,  2.5 ppm total hydrocarbons,
18      and 
-------
  1      one-half units up to 3, strong (Odor Panel of the CRC-APRAC Program Group on Composition
  2      of Diesel Exhaust, 1979; Levins, 1981).
  3             Surveys, utilizing volunteer panelists, have been taken to evaluate the general public's
  4      response to the odor of diesel exhaust. Hare and Springer (1971) and Hare et al. (1974) found
  5      that at a D rating of about 2 (TIA = 0.9, slight odor intensity), about 90% of the participants
  6      perceived the odor, and almost 60% found it objectionable. At a D rating of 3.2 (TIA = 1.2,
  7      slight to moderate odor intensity), about 95% perceived the odor, and 75% objected to it, and, at
  8      a D rating of 5 (TIA = 1.8, almost moderate), about 95% objected to it., Linnell and Scott (1962)
  9      evaluated the odor threshold for diesel exhaust in six subjects and found that a dilution factor of
10      140 to 475 was required to reduce the odor to the threshold level.
11             Linnell and Scott (1962) reported odor threshold measurement in six subjects and found
12      that the dilution factor needed to reach the threshold ranged from 140 to 475 for this small
13      sample of people. At these dilutions, the concentrations of formaldehyde ranged from 0.012 to
14      0.088 ppm.
15
16      5.1.1.1,2. Pulmonary and respiratory effects. Battigelli (1965) exposed 13 volunteers to three
17      dilutions of diesel exhaust obtained from a one-cylinder, four-cycle, 7-hp diesel engine (fuel  type
18      unspecified) and found that 15-min to 1 -h exposures had no significant effects on pulmonary
19      resistance. Pulmonary resistance was measured by plethysmography utilizing the simultaneous
20      recording of esophageal pressure and airflow determined by electrical differentiation of the
21      volume signal from a spirometer.  The units of concentration  of the constituents in the three
22      diluted exhausts were 1.3,2.8, and 4.2 ppmNO2; 0.2,0.5, and 1 ppm SO2; <20,30, and 55 ppm
23      CO; and <1.0, <1 to 2, and 1 to 2 ppm total aldehydes, respectively. Particle concentrations were
24      not reported.
25             A number of studies have evaluated changes in pulmonary function occurring over a
26      workshift in workers oceupationally exposed to diesel exhaust (specific time period not always
27      reported but assumed to be 8 h). In a study of coal miners, Reger et al. (1978) found that both
28      forced expiratory volume in  1 s (FEV,) and forced vital capacity (FVC) decreased by 0.05 L  in
29      60 diesel-exposed miners, an amount not substantially different from reductions  seen in non-
30      diesel-exposed miners (0.02  and 0.04 L, respectively). Decrements in peak expiratory flow rates
31      were similar between diesel and non-diesel exhaust-exposed miners. Miners with a history of
32      smoking had an increased number of decrements over the shift than nonsmokers did. Although
33      the monitoring data were not reported, the authors stated that  there was no relationship between
34     . the low concentrations of measured respirable dust or NO2 (personal samplers) when compared
35      with shift changes for any lung function parameter measured  for the diesel-exposed miners.  This

        2/1/98        .'•'•'                  5-3       DRAFT-DO NOT CITE OR QUOTE

-------
 1      study is limited because results were preliminary (abstract) and there was incomplete information
 2      on the control subjects.
 3            Ames et al. (1982) compared the pulmonary function of 60 coal miners exposed to diesel
 4      exhaust with that of a control group of 90 coal miners not exposed to diesel exhaust for evidence
 5      of acute respiratory effects associated with exposure to diesel exhaust. Changes over the
 6      workshift in FVC, FEVb and forced expiratory flow rate at 50% FVC (FEF50) were the indices
 7      for acute respiratory effects. The environmental concentrations of the primary pollutants were
 8      2.0 mg/m3 respirable dust (<10 urn MMAD), 0.2 ppm NO2,12 ppm CO, and 0.3 ppm
 9      formaldehyde. The investigators reported a statistically significant decline in FVC and FEV\
10      over the workshift in both the diesel-exposed and comparisons group. Current smokers had
11      greater decrements in FVC, FEV,, and FEF50 than exsmokers and nonsmokers. There was a
12      marked disparity between the ages and the time spent underground for the two study groups.
13      Diesel-exposed miners were about 15 years younger and had worked underground for 15 fewer
14     , years (4.8 versus 20.7 years) than miners not exposed to diesel exhaust. The significance of
15      these differences between the populations studied on the results is difficult to ascertain.
16            Except for the expected differences related to age, 120 underground iron ore miners
17      exposed to diesel exhaust had no workshift changes in FVC and FEV] when compared with 120
18      matched surface miners (Jorgensen and Svensson, 1970). Both groups had equal numbers (30)
19      of smokers and nonsmokers.: The frequency of bronchitis was higher among underground
20      workers, much higher among smokers than nonsmokers, and also higher among older than
21      younger workers. The authors reported that the underground miners had exposures of 0.5 to 1.5
22      ppm NO2 and between 3 and 9 mg/m3  particulate matter with 20 to 30% of the particles <5 urn
23      MMAD.  The majority of the particles were iron ore; quartz was 6 to 7% of the fraction
24      <5 am MMAD.
25            Gamble et al. (1978) measured preshift FEV, and FVC in 187 salt miners and obtained
26      peak flow forced expiratory flow rates at 25, 50, and 75% of FVC (FEF25, FEF50, or FEF75).
27      Postshift pulmonary function values were determined from total lung capacity and flows at
28      preshift percentages of FVC. The miners were exposed to mean NO2 levels of 1.5 ppm and mean
2.9      respirable particulate levels of 0.7 mg/m3.  No statistically significant changes were found
30      between changes in pulmonary function and in NO2 and respirable particles combined. Slopes of
31      the regression of NO2 and changes in FEVb FEF25, FEF50, and FEF75 were significantly different
32      from zero. The authors concluded that these small reductions in pulmonary function were
33      attributable to variations in NO2 within each of the five salt mines that contributed to the cohort.
34     .       Gamble et al. (1987a) investigated the acute effects of diesel exhaust in 232 workers in
35      four diesel bus garages using an acute  respiratory questionnaire and before and after workshift

        2/1/98                •                    5-4       DRAFT-DO NOT CITE OR QUOTE

-------
  1      spirometry. The prevalence of burning eyes, headaches, difficult or labored breathing, nausea,
  2      and wheeze experienced at work was higher in the diesel bus garage workers than in a
  3      comparison population of lead/acid battery workers who had not previously shown a statistically
  4      significant association of acute symptoms with acid exposure. Comparisons between the two
  5      groups were made without adjustment for age and smoking. There was no detectable association
  6      of exposure to NO2 (0.23 ppm ± 0.24 S.D.) or inhalable (less than 10 urn MMAD) particles (0.24
  7      mg/m3 ± 0.26 S.D.) and acute reductions in FVC, FEV,, peak flows, FEF50, and FEF75. Workers
  8      who had respiratory symptoms had slightly greater but statistically insignificant reductions in
  9      FEV, and FEF50.
10            Ulfvarson et al. (1987) evaluated workshift changes in the pulmonary function of 17 bus
11      garage workers, 25 crew members of two types of car ferries, and 37 workers on roll-on/roll-off
12      ships.  The latter group was exposed primarily to diesel exhaust; the first two groups were
13      exposed to both gasoline and diesel exhausts.  The diesel-only exposures that averaged 8 h
14      consisted of 0.13 to  1.0 mg/m3 particulate matter, 0.02 to 0.8 mg/m3 (0.016 to 0.65 ppm) NO,
15      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
16      mg/m3 (0.4 ppm) formaldehyde. The largest decrement in pulmonary function was observed
17      during a workshift following no exposure to diesel exhaust for 10 days. Forced vital capacity
18      and FEVj were significantly reduced over the workshift (0.44 L and 0.30 L^O.Ol and/xO.OOl,
19      respectively). There was no difference between smokers and nonsmokers.  Maximal
20      midexpiratory flow, closing volume expressed as the percentage of expiratory vital capacity,  and
21      alveolar plateau gradient (phase 3) were not affected. Similar but less pronounced effects on
22      FVC (- 0.16 L) were found in a second, subsequent study of stevedores (n = 24) only following 5
23      days of no exposure to diesel truck exhaust.  Pulmonary function returned to normal after 3 days
24      without occupational exposure to diesel exhaust. No exposure-related correlation was found
25      between the observed pulmonary effects and concentrations of NO, NO2, CO, or formaldehyde;
26      however, it was suggested that NO2 adsorbed onto the diesel exhaust particles may have
27      contributed to the overall dose of NO2 to the lungs. In a related study, six workers (job category
28      not defined) were placed in an exposure chamber and exposed to diluted diesel exhaust
29      containing 0.6 mg/m3 particulate matter and 3.9 mg/m3 (2.1 ppm) NO2.  The exhaust was
30      generated by a 6-cylinder, 2.38-L diesel engine, operated for 3 h and 40 min without interruption
31      at constant speed, equivalent to 60 km/h, and at about one-half full engine load. No effect on
32      pulmonary function was observed.
33
34      5.1.1.1.3. Immunological effects.  The likelihood that diesel exhaust can induce reactive airway
35      disease in humans was first reported by Wade and Newman (1993).  They identified three

        2/1/98                                    5-5       DRAFT-DO NOT CITE OR QUOTE

-------
 1      railroad workers who developed asthma following either a single exposure or a series of short-
 2      term exposures to high concentrations of DE. Asthma diagnosis was based on symptoms,
 3      pulmonary function tests, and measurement of airway hyperreactivity to methacholine or
 4      exercise.  Exposure occurred as a result of train crews riding in locomotive units trailing
 5      immediately behind the lead engine. Although the individuals had worked for the railroad for
 6      many years and presumably had been chronically exposed to lower levels of exhaust, the
 7      symptoms developed following these subacute incidents.  Unfortunately, exposure levels were
 8      not measured.                                                   .
 9             In an attempt to evaluate the potential allergenic effects of DE in humans Diaz-Sanchez
10      and associates carried out a series of clinical investigations.  In the first of these (Diaz-Sanchez et
11      al., 1994), humans were challenged by spraying either saline or 0.30 mg DPM into their nostrils.
12      This dose was considered equivalent to total exposure on  1-3 average days in Los Angeles, but
13      could occur acutely in certain nonoccupational settings such as sitting at a busy bus stop or in an
14      express tunnel. Enhanced IgE levels were noted in nasal lung lavage cells in as little as  24 h,
15      with peak production observed 4 days after DPM challenge. The effects semed to be somewhat
16      isotype-specific, because in contrast to IgE results, DPM challenge had no effect on the  levels of
17      IgG, IgA, IgM, or albumin. The selective enhancement of local IgE production was
18      demonstrated by a dramatic increase in IgE-secreting cells. Takenaka et al. (1995) reported that
19      DPM extracts enhanced IgE production from purified human B cells. Interleukin-4 plus
20      monoclonal antibody-stimulated IgE production was enhanced 20% to 360% by the addition of
21      DPM extracts over a period of 10-14 days. DPM extracts themselves did not induce IgE
22      production or synergize with interleukin-4 alone to induce IgE from purified B cells, suggesting
23      that the extracts were enhancing ongoing IgE production rather than inducing germline
24      transcription or isotype switching. The authors concluded that enhancement of IgE production in
25      the human airway resulting from the organic fraction of DPM may be an important factor in the
26      increase of allergic airway disease occurring throughout this century.
27             Although direct effects of DPM on B-cells have been demonstrated by in vitro studies, it
28      was considered likely that other cells regulating the IgE response may also be affected.  Cytokine
29      production was therefore measured in nasal lung lavage cells challenged with DPM (Diaz-
30      Sanchez et al.,  1996). Before challenge, most subjects' nasal lavage cells had detectable levels of
31      only mterferon-gamma, IL-2, and IL-13 wRNA. After challenge, the cells produced readily
32      detectable levels of wRNA for IL-2, IL-4, IL-5, IL-6, IL-10, IL-13, and interferon gamma. In
33      addition, all levels of cytokine wRNA were increased. Although the cells in the nasal lavage
34      before and after challenge do not necessarily represent the same ones either in number or type,
35      the broad increase in cytokine  production was not simply the result of an increase in T cells

        2/1/98                                     5-6       DRAFT-DO NOT CITE OR QUOTE

-------
  1      recovered in the lavage fluid. On the basis of these findings it was concluded that the increase in
  2      eytokine expression after exposure to DPM can be predicted to contribute to enhanced local IgE
  3      production and thus play a role in pollutant-induced airway diaease,
  4             In a continuation of the previous studies, the combined effects of intranasal challenge
  5      with DPM plus ragweed allergen were investigated. Subjects were selected on the basis of being
  6      allergicto ragweed. They were tested with doses of ragweed antigen sufficient to induce a
  7      detectable response, 0.3 mg DPM, or ragweed allergen plus DPM combined (Diaz-Sanchez et al.,
  8      1997).  As compared with challenge by ragweed alone, challenge with .both DPM and ragweed
  9      allergen induced markedly higher ragweed-specific IgE, but not total IgE levels.  This synergy
10      between DPM and natural allergen exposure may be a key feature in increasing allergen-induced
11      respiratory allergic  disease.
12             In the available exposure studies, considerable variability is reported in diesel exhaust
13      detection threshold. The odor scales described in some of these studies have no general use at
14      present because they are not objectively defined; however, the studies do clearly indicate
15      substantial interindividual variability in the ability to detect odor and the level at which it
16      becomes objectionable. Much of what is known about the acute  effects of diesel exhaust comes
17      from case reports that lack clear measurements of exposure concentrations.  The studies of
18      pulmonary function changes in exposed humans have looked for changes occurring over a
19      workshift or after a short-term exposure. The overall conclusion of these studies is that
20      reversible changes in pulmonary function in humans can occur in relation to diesel exhaust
21      exposure, although  it is not possible to relate these changes to specific exposure levels. Based on
22      the report by Wade and Newman (1993), reversible airflow obstruction and a syndrome
23      consistent with asthma are possible following acute, high-level exposure to diesel exhaust.
24      Recent studies by Diaz-Sanchez and co-workers have provided data indicating that DPM is a
25      likely factor in the increasing incidence of allergic hypersensitivity.  They have also shown that
26      effects are due primarily to the organic fraction and that DPM synergizes with known allergens
27      to increase their effectiveness.
28      .
29      5.1.1.2.  Long-Term Exposures
30             Several epidemiologic studies have evaluated the effects of chronic exposure to diesel
31      exhaust on occupationally exposed workers.
32            Battigelli et al. (1964) measured several indices of pulmonary function, including vital
33      capacity, FEV,, peak flow, nitrogen washout, and diffusion capacity in 210 locomotive
34      repairmen exposed to diesel exhaust in three engine houses. The average exposure of these
35      locomotive repairmen to diesel exhaust was 9.6 years. When compared with a control group

        2/1/98                                    5-7        DRAFT-DO NOT CITE OR QUOTE

-------
 1     matched for age, body size, "past extrapulmonary medical history" (no explanation given), and
 2     job status (154 railroad yard workers), no significant clinical differences were found in
 3     pulmonary function or in the prevalence of dyspnea, cough, or sputum between the diesel
 4     exhaust-exposed and nonexposed groups. Exposure to the diesel exhaust showed marked
 5     seasonal variations because the doors of the engine house were open in the summer and closed in
 6     the winter. For the exposed group, the maximum daily workplace concentrations of air
 7     pollutants measured were 1.8 ppm NO2, 1.7 ppm total aldehydes, 0.15 ppm acrolein, 4.0 ppm
 8     SO2, and 5.0 ppm total hydrocarbons.  The concentration of airborne particles was not reported.
 9            Gamble et al. (1987b) examined 283 diesel bus garage workers from four garages in two
10     cities to determine if there was excess chronic respiratory morbidity associated with exposure to
11     diesel exhaust.  Tenure of employment was used as a surrogate of exposure; mean tenure of the
12     study population was 9 years ±10 years S.D. Exposure-effect relationships within the study
13     population showed no detectable associations of symptoms with tenure. Reductions in FVC,
14     FEVb peak flow, and FEF50 (but not FEF75) were  associated with increasing tenure.  When
15     compared with a control population (716 nonexposed blue-collar workers) and after indirect
16     adjustment for age, race, and smoking, the exposed workers had a higher incidence of cough,
17     phlegm, and wheezing; however, there was no correlation between symptoms and length of
18     employment. Dyspnea showed an exposure-response trend but no apparent increase in
.19     prevalence. Mean FEV,, FVC, FEF50, and peak flow were not reduced in the total cohort
20     compared with the reference population but were  reduced in workers with 10 years or more
21     tenure.
22            Purdham et al. (1987) evaluated respiratory symptoms and pulmonary function in
23      17 stevedores employed in car ferry operations who were exposed to both diesel and gasoline
24     exhausts and in a control group of 11 on-site  office workers. Twenty-four percent of the exposed
25     group and 36% of the controls were smokers. If a particular symptom was considered to be
26     influenced by smoking, smoking status was used as a covariate in the logistic regression analysis;
27     pack-years smoked was a covariate for lung function indices. The frequency of respiratory
28     symptoms was not significantly different between the two groups; however, baseline pulmonary
29     function measurements were significantly different. The latter comparisons were measured by
30     multiple regression analysis using the actual (not percentage predicted) results and correcting for
31     age, height, and pack-years smoked. The stevedores had significantly lower FEV,, FEVt/FVC,
32     FEF50, and FEF75 (p<0.021,/?<0.023,p<0.001, andp<0.008, respectively) but not FVC. The
33     results from the stevedores were also compared with those obtained from a study of the
34     respiratory health status  of Sydney, Nova Scotia, residents.  These comparisons showed that the
35     dock workers had higher FVC, similar FEVU but lower FEV,/FVC and flow rates than the

       2/1/98                                    5-8       DRAFT-DO NOT CITE OR QUOTE

-------
  1      residents of Sydney. Based on these consistent findings, the authors concluded that the lower
  2      baseline function measurements in the stevedores provided evidence of an obstructive ventilatory
  3      defect but caution in interpretation was warranted because of the small sample size. There were
  4      no significant changes in lung function over the workshift, nor was there a difference between
  5      the two groups.  The stevedores were exposed to  significantly (p<0.04) higher concentrations of
  6      particulate matter (0.06 to 1.72 mg/m3, mean 0.50 mg/m3) than the controls (0.13 to 0.58 mg/m3,
  7      mean not reported). Exposures of stevedores to SO2, NO2, aldehydes, and PAHs were very low;
  8      occasional CO concentrations in the 20 to 100 ppm range could be detected for periods up to 1 h
  9      in areas where blockers were chaining gasoline-powered vehicles.
10            Additional epidemiological studies on the health hazards posed by exposure to diesel
11      exhaust have been conducted for mining operations. Reger et al. (1982) evaluated the respiratory
12      health status of 823 male coal miners from six diesel-equipped mines compared with 823
13      matched coal miners not exposed to diesel exhaust. The average tenure of underground work for
14      the underground miners and their controls was only about 5 years; on average, the underground
15      workers in diesel mines spent only 3 of those 5 years underground in diesel-use mines.
16  .    Underground miners exposed to diesel exhaust reported a higher incidence of symptoms of
17      cough and phlegm but proportionally fewer symptoms of moderate to severe dyspnea than their
18      matched counterparts. These differences in prevalence of symptoms were not statistically
19      significant.  The diesel-exposed underground miners, on the average, had lower FVC, FEV1;
20      FEF50, FEF75,  and FEF90 but higher peak flow and FEF25 than their matched controls.  These
21      differences, however, were not statistically significant. Health indicators for surface workers and
22      their matched controls were directionally the same as for matched underground workers. There
23      were no consistent relationships between the findings of increased respiratory symptoms,
24      decreased pulmonary function, smoking history, years of exposure, or monitored atmosphere
25      pollutants (NOX, CO, particles, and aldehydes). Mean concentrations of NOX at the six mines
26      ranged from 0 to 0.6 ppm for short-term area samples, 0.13 to 0.28 ppm for full-shift personal
27      samples, and 0.03 to 0.80 for full-shift area samples. Inhalable particles (less than 10 jam
28    .  MMAD) averaged 0.93 to 2.73 mg/m3 for personal samples and 0 to 16.1 mg/m3 for full shift
29      area samples.  Ames et al. (1984), using a portion of the miners studied by Reger, examined 280
30      diesel-exposed underground miners initially in 1977 and again in 1982. Each miner in this group
31      had at least 1 year of underground mining work history in 1977. The control group was 838
32      miners with no exposure to diesel exhaust. The miners were evaluated for the prevalence of
33      respiratory symptoms, chronic cough, phlegm, dyspnea, and changes in FVC, FEV,, and FEF50.
34      No air monitoring data were reported; exposure to diesel exhaust gases and mine dust particles
35      were described as very low. These authors found no decrements in pulmonary function or

        2/1/98   .                                 5-9       DRAFT-DO NOT CITE OR QUOTE

-------
 1      increased prevalence of respiratory symptoms attributable to exposure to diesel exhaust. In fact,
 2      the 5-year incidences of cough, phlegm, and dyspnea were greater in miners without exposure to
 3      diesel exhaust than in those exposed to diesel exhaust.
 4            Attfield (1978) studied 2,659 miners from 21 mines (8 metal, 6 potash, 5 salt, and
 5      2 trona). Diesels were employed in only 18 of the mines, but the 3 mines not using diesels were
 6      not identified. The years of diesel usage, ranging from 8 in trona mines to 16 in potash mines,
 7      were used as a surrogate for exposure to diesel exhaust. Based on a questionnaire, an increased
 8      prevalence of persistent cough was associated with exposure to aldehydes; this finding, however,
 9      was not supported by the pulmonary function data.  No adverse respiratory symptoms or
10      pulmonary function impairments were related to CO2, CO, NO2, inhalable dust, or inhalable
11      quartz.  The author failed to comment on whether the prevalence of cough was related to the high
12      incidence (70%) of smokers in the cohort.
13            Questionnaire, chest radiograph, and spirometric data were collected by Attfield et al.
14      (1982) on 630 potash miners from six potash mines. These miners were exposed for an average
15      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,
16      and total dust concentrations of 9 to 23 mg/m3. No attempt was made to measure diesel-derived
17      particles separately from other dusts. The ratio of total to inhalable  (<10 um MMAD) dust
18      ranged from 2 to 11. An increased prevalence of respiratory symptoms was related solely to
19      smoking. No association was found between symptoms and tenure of employment, dust
20      exposure, NO2, CO, or aldehydes. A higher prevalence of symptoms of cough and phlegm was
21      found, but no differences in pulmonary function (FVC and FEV\) were found in these diesel-
22      exposed potash miners when compared with the predicted values derived from a logistics model
23      based on blue-collar workers working in nondusty jobs.
24            Gamble et al. (1983) investigated respiratory morbidity in 259 miners from five salt
25      mines in terms of increased respiratory symptoms, radiographic findings, and reduced pulmonary
26      function associated with exposure to NO2, inhalable particles (<10 um MMAD), or years worked
27      underground. Two of the mines used diesel extensively; no diesels were used in one saltmine.
28      Diesels were introduced into each mine in 1956, 1957, 1963, or 1963 through 1967.  Several
29      working populations were compared with the salt miner cohort. After adjustment for age and
30      smoking, the salt miners showed no increased prevalence of cough,  phlegm, dyspnea, or airway
31      obstruction (FEV]/FVC) compared with aboveground coal miners, potash miners, or blue-collar
32      workers. The underground coal miners consistently had an elevated level of symptoms. Forced
33      expiratory volume at 1 s, FVC, FEF50, and FEF75 were uniformly lower for salt miners in relation
34      to all the comparison populations. There was, however, no association between changes in
35      pulmonary function and years worked, estimated cumulative inhalable particles, or estimated

        2/1/98        .                            5-10        DRAFT-DO NOT CITE OR QUOTE

-------
  1      NO2 exposure. The highest average exposure to partieulate matter was 1.4 mg/m3 (particle size
  2      not reported, measurement includes NaCl). Mean N02 exposure was 1,3 ppm, with a range of
  3      0.17 ppm to 2.5 ppm. In a continuation of these studies, Gamble and Jones (1983) grouped the
  4      salt miners into low-, intermediate-, and high-exposure categories based on tenure in jobs with
  5      diesel exhaust exposure. Average concentrations of inhalable particles and N02 were 0.40, 0.60,
  6      and 0.82 mg/m3 and 0.64,1.77, and 2.21 ppm for the three diesel exposure categories,
  7      respectively. A statistically significant concentration-response association was  found between
  8      the prevalence of phlegm in the salt miners and exposure to diesel exhaust (/?<0.0001) and a
  9      similar, but nonsignificant, trend for cough and dyspnea. Changes in pulmonary function
10      showed no association with diesel tenure.  In a comparison with the control group of
11      nonexposed, blue-collar workers, adjusted for age and smoking, the overall prevalence of cough
12      and phlegm (but not dyspnea) was elevated in the diesel-exposed workers. Forced expiratory
13      volumes at 1 s and FVC were within 4% of expected, which was considered to be within the
14      normal range of variation for a nonexposed population,
15            In a preliminary study of three subcohorts from bus company personnel  (clerks [lowest
16      exposure], bus drivers [intermediate exposure], and bus garage workers [highest exposure])
17      representing different levels of exposure to diesel exhaust, Edling and Axelson (1984) found a
18      fourfold higher risk ratio for cardiovascular mortality in bus garage workers, even after adjusting
19      for smoking history and allowing for at least 10 years of exposure and 15 years  or more of
20      induction-latency.  Carbon monoxide was hypothesized as the etiologic agent for the increased
21      cardiovascular disease but was not measured.  However, in a more comprehensive
22      epidemiological study, Edling et al. (1987) evaluated mortality data covering a 32-year period  for
23      a cohort of 694 bus garage employees and found no significant differences between the observed
24      and expected number of deaths from cardiovascular disease.  Information on exposure
25      components and their concentrations was not reported.
26            The absence of reported noncancerous human health effects, other than infrequently
27      occurring effects related to respiratory symptoms and pulmonary function changes, is notable.
28      Unlike studies in laboratory animals to be described later in this chapter, studies of the impact of
29      diesel exhaust on the defense mechanisms of the human lung have not been performed. No
30      direct evidence is available in humans regarding doses of diesel exhaust, gas phase, partieulate
31      phase, or total exhaust that lead to impaired particle clearance or enhanced susceptibility to
32      infection. A summary of epidemiology studies is presented in Table 5-1.
33            To date, no large-scale epidemiological study has looked for effects of chronic exposure
34      to diesel exhaust on pulmonary function. In the long-term longitudinal and cross-sectional
        2/1/98                                    5-11        DRAFT-DO NOT CITE OR QUOTE

-------
                    Table 5-1. Human studies of exposure to diesel exhaust
       Study
         Description
                   Findings
                                          Acute exposures
 Kahnetal. (1988)
 El Batawi and
 Noweir(1966)
 Battigelli(1965)
 Katzetal. (1960)
 Hare and Springer
 (1971)
 Hare etal. (1974)


 Linnell and Scott
 (1962)
 Ruddell etal, (1988)


 Rudell etal. (1996)




 Battigelli (1965)
 Wade and Newman
 (1993)

 Diaz-Sanchez et al.
 (1994)

 Takenaka et al.
 (1995)
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 diesel
exhaust in a train tunnel.
Volunteer panelists who
evaluated general public's
response to odor of diesel
exhaust.

Odor panel under highly
controlled conditions
determined odor threshold for
diesel exhaust

Volunteers exposed to about
100 ng/m3 DPM,

Volunteeers exposed to diesel
exhaust for one hour while
doing light work. Exposure
concentrations uncertain.

13 volunteers exposed to three
dilutions of diesel exhaust for
15 minto 1 h.

Three railroad workers acutely
exposed to diesel exhaust.

Volunteers challenged by a
nasal spray of 0.30 mg DPM.

Volunteers challenged by a
nasal spray of 0.30 mg DPM.
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 diesel exhaust.

Three occasions of minor eye and throat irritation;
no correlation established with concentrations of
diesel exhaust 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 diesel exhaust to an odor threshold ranged
from a factor of 140 to 475.
Neutrophils were increased and phagocytosis of
opsonized yeast cells was reduced after exposure.

Unpleasant smell along with irritation of eyes and
nose reported.  Airway resistance increased.
Reduction of particle concentrtion by trapping did
not affect results.

No significant effects on pulmonary resistance
were observed  as measured by plethysmography.


The workers developed symptoms of asthma.


Enhancement of IgE production reported due to a
dramatic increase in IgE-secreting cells.

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.	
2/1/98
                           5-12
         DRAFT-DO NOT CITE OR QUOTE

-------
              Table 5-1. Human studies of exposure to diesel exhaust (continued)
        Study
         Description
                   Findings
 Diaz-Sanchez et al,
 (1996)

 Diaz-Sanchez et al.
 (1997)
 Regeretal. (1978)
 Ames etal. (1982)
 Jdrgensen and
 Svensson(1970)
 Gamble et al. (1978)
 Gamble et al.
 (I987a).
Volunteeers 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. .
A broad increase in cytokine expression predicted
to contribute to enhanced local IgE production.

Ragweed allergen plus DPM-stiraulated 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.
                                    Studies of cross-shift changes
Five or more VC maneuvers by
each of 60 coal miners exposed
to diesel exhaust at the
beginning and end of a
workshift.

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 four 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.
FEV,, FVC, and PEFR were similar between diesel
and non-diesel-exposed miners.  Smokers had an
increased number of decrements over shift than
nonsmokers.
Significant workshift decrements occurred in
miners in both groups who smoked; no significant
differences in ventilatory function changes between
miners exposed to diesel exhaust 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
particulate 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.
2/1/98
                          5-13
         DRAFT-DO NOT CITE OR QUOTE

-------
              Table 5-1. Human studies of exposure to diesel exhaust (continued)
       Study
         Description
                   Findings
 Ulfvarson et al.
 (1987)
 Battigelli et al.
 (1964)
 Gamble et al.
 (1987b)
 Purdham et al.
 (1987)
 Regeretal. (1982)
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 diesel
exhaust, 2.1 ppm NO2, and 0.6
mg/m3 particulate matter.
Pulmonary function was affected during a
workshift exposure to diesel exhaust, 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 diesel exhaust 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 diesel exhaust.

283 male diesel bus garage
workers from four garages in
two cities were examined for
impaired pulmonary function
(FVC, FEV,, and flow rates).
Study population with a mean
tenure of 9 ± 10 years S.D. was
compared to a nonexposed
blue-collar population.

Respiratory symptoms and
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.
Differences in respiratory
symptoms and pulmonary
function were assessed in
823 coal miners from six
diesel-equipped mines.
compared to 823 matched coal
miners not exposed to diesel
exhaust.
No significant differences in VC, FEV,, peak flow,
nitrogen washout, or diffusion capacity or in the
prevalence of dyspnea, cough, or sputum were
found between the diesel exhaust-exposed and
nonexposed groups.
Analyses within the study population showed no
association of respiratory symptoms with tenure.
Reduced FEV,  and FEFj0 (but not FEF7S) 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 due to 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 diesel exhaust
thought to be responsible.
2/1/98
                           5-14
         DRAFT-DO NOT CITE OR QUOTE

-------
              Table 5-1. Human studies of exposure to diesel exhaust (continued)
        Study
         Description
                   Findings
  Ames etal. (1984)
  Attfield(1978)
 Attfield etal. (1982)
 Gamble etal. (1983)
 Gamble and Jones
 (1983)
Changes in respiratory
symptoms and function were
measured during a 5-year
period in 280 diesel-exposed
and 838 nonexposed U.S.
underground coal miners.

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 diesel exhaust.

Respiratory symptoms and
function were assessed in
630 potash miners  from six
potash mines using 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 five
salt mines by respiratory
symptoms, radiographic
findings, and spirometry.  Two
mines used diesels extensively,
two had limited use, 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.
No decrements in pulmonary function or increased
prevalence of respiratory symptoms were found
attributable to diesel exhaust.  In fact, 5-year
incidences of cough, phlegm, and dyspnea were
greater in miners without exposure to diesel
exhaust than in miners exposed to diesel exhaust.

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.
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
FEV, were found in these diesel-exposed potash
workers when compared with predicted values
from a logistic model based on blue-collar staff
working in nondusty jobs.


After adjustment for age and smoking, salt miners
showed no symptoms, increased prevalence of
cough, phlegm, dyspnea, or air obstruction
(FEVi/FVC) compared with aboveground coal
miners, potash workers, or blue-collar workers.
FEV,, 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.     	
2/1/98
                          5-15
         DRAFT-DO NOT CITE OR QUOTE

-------
             Table 5-1. Human studies of exposure to diesel exhaust (continued)
       Study
        Description
                  Findings
 Edling and AxeJson
 (1984)
 Edling etal, (1987)
Pilot study of 129 bus company
employees classified into three
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).
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 15 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.
2/1/98
                          5-16
         DRAFT-DO NOT CITE OR QUOTE

-------
  1      studies, a relationship was generally observed between work in a job with diesel exposure and
  2      respiratory symptoms (such as cough and phlegm), but there was no consistent effect on
  3      pulmonary function. The interpretation of these results is hampered by lack of measured diesel
  4      exhaust exposure levels and the short duration of exposure in these cohorts. Only active workers
  5      were included in these studies. It is possible that the relationship between work in a job with
  6      diesel exposure and respiratory symptoms was due to short-term exposure.
  7
  8      5.1.2. Laboratory Animal Studies
  9            Because  of the large number of statistical comparisons made in the laboratory animal
10      studies and to permit uniform, objective evaluations within and among studies, data will be
11      reported as significantly different (i.e.,p<0.05) unless otherwise specified. The exposure
12      regimens used and the resultant exposure conditions employed in the laboratory animal
13      inhalation studies are summarized in Appendix A.  Other than the pulmonary function studies
14      performed by Wiester et al. (1980) on guinea pigs during their exposure in inhalation chambers,
15      the pulmonary function studies performed by other investigators, although sometimes
16      unreported, were interpreted as being conducted on the following day or thereafter and not
17      immediately following exposure.
18
19      5.1.2.1. Acute Exposures
20            The acute toxicity of undiluted diesel exhaust to rabbits, guinea pigs, and mice was
21     .assessed by Pattle et al. (1957). Four engine operating conditions were used, and 4 rabbits, 10
22      guinea pigs, and 40 mice were tested under each exposure condition for 5 h (no controls were
23      used). Mortality was assessed up to 7 days after exposure. With the engine operating under light
24      load, the exhaust was highly irritating but not lethal to the test species, and only mild tracheal
25      and lung damage was observed in the exposed animals. The exhaust contained 74 mg/m3
26      particulate matter (particle size not reported), 560 ppm CO, 23 ppm NO2, and 16 ppm aldehydes.
27      Exhaust containing 5 mg/m3 particulate matter, 380 ppm CO, 43 ppm NO2, and 6.4 ppm alde-
28      hydes resulted in low mortality rates (mostly below 10%) and moderate lung damage.  Exhaust
29      containing 122 mg/m3 particulate  matter, 418 ppm CO, 51 ppm NO2, and 6.0 ppm aldehydes
30      produced high mortality rates (mostly above 50%) and severe lung damage. Exhaust containing
31      1,070 mg/m3 particulate matter, 1,700 ppm CO, 12 ppm NO2, and 154 ppm aldehydes resulted in
32      100% mortality in all three species. High CO  levels, which resulted in a carboxyhemoglobin
33      value of 60% in mice and 50% in rabbits and guinea pigs, were considered to be the main cause
34      of death in the latter case. High NO2 levels were considered to be the main cause of lung damage
        2/1/98                                   5-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1      and mortality seen in the other three tests. Aldehydes and NO2 were considered to be the main
 2      irritants in the light load test.
 3            Kobayashi et al. (1995) administered 1, 10, or 20 mg/kg DPM for 30 min to the nasal
 4      mucosa of guinea pigs.  The administration increased nasal airway resistance, augmented
 5      increased airway resistance and nasal secretion induced by a histamine aerosol, increased
 6      vascular permeability in dorsal skin, and augmented vacular permeability induced by histamine.
 7      The increases in nasal airway resistance and secretion are considered typical responses of nasal
 8      mucosa against allergic stimulation. Similar results were reported for guinea pigs exposed via
 9      inhalation for 3 h to diesel exhaust diluted to particle concentrations of either 1  or 3.2 mg/m3
10      (Kobayashi et al., 1997). These studies show that short-term exposure to diesel exhaust potently
11      induces nasal mucosal hyperrersponsiveness in guinea pigs.
12            The effects of diesel exhaust particles (DPM) and their components (extracted particles
13      and particle extracts) on the release of proinflammatory cytokines, interleukin-1, and turrior
14      necrosis factor-alpha (TNF-a) by alveolar macrophages (AM) were investigated by Yang et al.
15      (1997). Rat AM were incubated with 0, 5, 10, 20, 50, or 100 ug/106 AM/mL of DPM, methanol-
16      extracted DPM, or equivalent concentrations of DPM at 37 °C for 24 h. At high concentratrions,
17      both DPM and DPM extracts were shown to increase IL-1-like activity secreted by AM, whereas
18      extracted particles had no effect. Neither particles, particle extracts, or extracted particles
19      stimulated secretion of TNFa.  DPM inhibited lipid polysaccharide (LPS)-stimulated production
20      of IL-1 and TNF-a. In contrast, interferon-gamma stimulated production of IL-1 and TNF-a.
21            Results of this study indicate that the organic fraction of exhaust particles is responsible
22      for the effects noted. Stimulation of IL-1 but not TNF-a suggests that IL-1, but not TNF-a, may
23      play an important role in the development of DPM-induced inflammatory and immune
24      responses. The cellular mechanism involved in inhibiting increased release of IL-1 and TNF-a
25      by LPS is unknown, but may be a contributing factor to the decreased AM phagocytic activity
26      and increased susceptibility to pulmonary infection after prolonged exposure to DPM.
27            Takano et al. (1997) conducted a study designed to evaluate the effects of DPM on the
28      manifestations of allergic asthma in mice, with emphasis on antigen-induced airway
29      inflammation, the local expression of interleukin-5 (IL-5), macrophage colony stimulating factor
30      (GM-CSF), IL-2 and interferon (IFN)-gamma, and the production of antigen-specific IgE and
31      IgG.  Male ICR mice were intratracheally instilled with ovalbumin (OVA), DPM, and
32      DPM+OVA. DPM was obtained from a 4JB1 -type, light-duty 2.74 L, four-cylinder Izuzu diesel
33      engine operated at a steady speed of 1,500 rpm under a load of 10 torque (kg/m).  The OVA-
34      group mice were instilled with 1 ug OVA at 3 and 6 weeks. The mice receiving DPM alone were
35      instilled with 100 ug DPM weekly  for 6 weeks. The OVA + DPM group received the combined

        2/1/98                                   5-18       DRAFT-DO NOT CITE OR QUOTE

-------
  1      treatment in the same protocol as the OVA and the DPM groups, respectively. Additional groups
  2      were exposed for 9 weeks.  DPM aggravated ovalbumin-induced airway inflammation,
  3      characterized by infiltration of eosinophils and lymphocytes and an increase in goblet cells in the
  4      bronchial epithelium. DPM in combination with antigen markedly increased IL-5 protein levels
  5    '  in lung tissue and bronchoalveolar lavage supernatants compared with either antigen or DPM
  6      alone. The combination of DPM and antigen induced significant increases in local expression of
  7      of IL-4, GM-CSF, and IL-2, whereas expression of of IFN-gamma was not affected.  In addition,
  8      DPM exhibited adjuvant activity for the antigen-specific production of.IgG and IgE.  These
  9      results provide experimental evidence that DPM can enhance the manifestations of allergic
10      asthma and suggest that DPM is implicated in the increasing prevalence of allergic asthma in
11      recent years.
12
13      5.1.2.2. Short-Term and Subchronic Exposures
14            A number of inhalation studies have employed a regimen of 20 h/day, 7 days/week for
15      varying exposure periods up to 20 weeks to differing concentrations of airborne particulate
16      matter, vapor, and gas concentrations of diluted diesel exhaust.  Exposure regimens and
17      characterization of gas-phase components for these studies are summarized in Table 5-2.
18      Pepelko et al. (1980a) evaluated the pulmonary function of cats exposed under these conditions
19      for 28 days to 6.4 mg/m3 particulate matter.  The only significant functional change observed was
20      a decrease in maximum expiratory flow rate  at 10% vital capacity. The excised lungs of the
21      exposed cats appeared charcoal gray, with focal black spots visible on the pleural surface.
22      Pathologic changes included a predominantly peribronchial localization of black-pigmented
23      macrophages within the alveoli characteristic of focal pneumonitis or alveolitis.
24            The effects of a short-term diesel exhaust exposure on arterial blood gases, pH, blood
25      buffering, body weight changes, lung volumes, and deflation pressure-volume (PV) curves of
26      young adult rats were evaluated by Pepelko (1982a). Exposures were 20 h/day, 7 days/week for
27      8 days to a concentration of 6.4 mg/m3 particulate matter in the nonirradiated exhaust (RE) and
28    .  6.75 mg/m3 in the irradiated exhaust (IE).  In spite of the irradiation, levels of gaseous
29      compounds were not substantially different between the two groups (Table 5-2). Body weight
30      gains were significantly reduced in the reexposed rats and to an even greater degree in rats
31      exposed to IE.  Arterial blood gases and standard bicarbonate were unaffected, but arterial blood
32      pH was significantly reduced in rats exposed to IE. Residual volume and wet lung weight were
33      not affected by either exposure, but vital capacity and total lung capacity were increased
34      significantly following exposure to RE.  The shape of the deflation PV curves were nearly
35      identical for the control, RE and IE groups.

        2/1/98                                    5-19       DRAFT-DO NOT CITE OR QUOTE

-------
Table 5-2. Short-term effects of diesel exhaust on laboratory animals
1— 1
oo















1
NJ
0



O
5-
"fl
1
O
o
z
g
H
Q
M

y

d
o
H
W

Species/sex
Rat, F344, M;
Mouse, A/J, M;
Hamster, Syrian,
M
Rat, F344, M, F;
Mouse, CD-I, M,
F

Cat, Inbred, M


Rat, Sprague-
Dawley, M

Guinea Pig,
Hartley, M, F

Rat, F344,
M


Guinea Pig,
Hartley, M



Guinea Pig,
Hartley, M


Guinea Pig,
Hartley, M, F






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



20 h/day
7 days/week
8 weeks





Particles
(mg/m3)
1.5
0.19 urn MMD


0.21.
1.0
4.4

6.4


6.4
6.8'

6.8"


6.0
6.8 urn MMD


1-2 mg DPM
Intranasally



1
3.2


6.3







CxT CO NO2 SO2
(mg'h/m3) (PPm) (PPm) (PPm) Effects
2,100 to 2,730 6.9 0.49 — Increase in lung wt; increase in
thickness of alveolar walls;
minimal species difference

140 — — — No effects on lung function in
665 — — — rats (not done in mice); increase
2,926 — — — in PMNs and proteases and AM
aggregation in both species
3,584 14.6 2.1 2.1 Few effects on lung function;
focal pneumonitis or alveolitis

3,584 16.9 2.49 2.10 Decreased body wt; arterial blood
3,808 16.1" 2.76' 1.86' pH reduced; vital capacity, total
(<0.01 ppm O3)' lung capacities increased
3,808 16.7 2.9 1.9 Exposure started when animals
were 4 days old; increase in
(<0.01 ppm O3)' pulmonary flow; bradycardia
2,640 — — — Macrophage aggregation; in-
crease in PMNs; Type H cell pro-
liferation; thickened alveolar
walls
— — — — Augmented increases in nasal
airway resistance and vascular
permeability induced by a
histamine aerosol

0.5 5.9 1.4 0.13 Similar results to those reported
1.6 12.9 4.4 0.34 in the previous study using
intranasal challange

7,056 17.4 2.3 2.1 Increase in relative lung wt. AM
aggregation; hypertrophy of
goblet cells; focal hyperplasia of
alveolar epithelium





Study
Kaplan etal. (1982)



Mauderly etal. (1981)



Pepelkoetal. (1980a)


Pepelko(1982a)


Wiester etal. (1980)


White and Garg ( 1981)



Kobayashi etal. (1995)




• Kobayashi etal. (1997)



Wiester etal. (1980)








-------
00
o
                         Table 5-2. Short-term effects of diesel exhaust on laboratory animals  (continuted)
Species/sex
Mouse ICR, M




Rat, Sprague-
Dawley,
M
Exposure
period
6 weeks




24 h


Particles
(mg/m3)
100 ug DPM
intranasally



5-100 ug/106
AM/ml of DPM

CxT CO NOj SO2
(nig'h/m3) (ppni) (ppm)~ (ppnt) Effects
— — — — 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
Takanoetal.(1997)




Yang etal. (1997)


      "Irradiated exhaust.

      PMN = Polymorphonuclear leukocyte.

      AM = Alveolar macrophage.
O
o

z
o
H

n
i—i
H
W

O
?o

o
a
o
H
M

-------
 1            In related studies, Wiester et al. (1980) evaluated pulmonary function in 4-day-old guinea
 2      pigs exposed for 20 h/day, 7 days/week for 28 days to IE having a concentration of 6.3 mg/m3
 3      particulate matter. When housed in the exposure chamber, pulmonary flow resistance increased
 4      35%, and a small but significant sinus bradycardia occurred as compared with controls housed
 5      and measured in control air chambers (p<0.002). Respiratory rate, tidal volume, minute volume,
 6      and dynamic compliance were unaffected as were lead-1 electrocardiograms.
 7            A separate group of adult guinea pigs was necropsied after 56 days of exposure to IE, to
 8      diluted RE, or to clean air (Wiester et al., 1980). Exposure resulted in a significant increase in
 9      the ratio of lung weight to body weight (0.68% for controls, 0.78% for IE,  and 0.82% for RE).
10      Heart/body weight ratios were not affected by exposure. Microscopically,  there was a marked
11      accumulation of black pigment-laden alveolar macrophages (AM) throughout the lung with a
12      slight to moderate accumulation in bronchial and carinal lymph nodes. Hypertrophy of goblet
13      cells in the tracheobronchial tree was frequently observed, and focal hyperplasia of alveolar
14      lining cells was occasionally observed. No evidence of squamous metaplasia of the
15      tracheobronchial tree, emphysema, peribronchitis, or peribronchiolitis was noted.
16            White and Garg (1981) studied pathologic alterations in the lungs of rats (16 exposed and
17      8 controls) after exposure to diesel exhaust containing 6 mg/m3 particulate matter. Two rats from
18      the exposed group and one rat from the control group (filtered room air) were sacrificed after
19      each exposure interval of 6 h and 1,3,7, 14, 28, 42, and 63 days; daily exposures were for 20 h
20      and were 5.5 days/week. Evidence of AM recruitment and phagocytosis of diesel particles was
21      found at the 6-h sacrifice; after 24 h of exposure there was a focal, scattered increase in the
22      number of Type II cells. After 4 weeks of exposure, there were morphologic changes in size,
23      content,  and shape of AM, septal thickening adjacent to clusters of AMs, and an appearance of
24      inflammatory cells, primarily within the septa. At 9 weeks of exposure, focal aggregations of
25      particle-laden macrophages developed near the terminal bronchi, along with an influx of
26      accumulation of black pigment-laden alveolar macrophages (AM) throughout the lung with a
27      slight to moderate accumulation in bronchial and carinal lymph nodes. Hypertrophy of goblet
28      cells in the tracheobronchial tree was frequently observed, and focal hyperplasia of alveolar
29      lining cells was occasionally observed. No evidence of squamous metaplasia of the
30      tracheobronchial tree, emphysema, peribronchitis, or peribronchiolitis was noted.
31            White and Garg (1981) studied pathologic alterations in the lungs of rats (16 exposed and
32      8 controls) after exposure to diesel exhaust containing 6 mg/m3 particulate matter. Two rats from
33      the exposed group and one rat from the control group (filtered room air) were sacrificed after
34      each exposure interval of 6 h and 1,3,7, 14, 28, 42, and 63 days; daily exposures were for 20 h
35      and were 5.5 days/week. Evidence of AM recruitment and phagocytosis of diesel particles was

        2/1/98                                     5-22       DRAFT-DO NOT CITE OR QUOTE

-------
  1      found at the 6-h sacrifice; after 24 h of exposure there was a focal, scattered increase in the
  2      number of Type II cells. After 4 weeks of exposure, there were morphologic changes in size,
  3      content, and shape of AM, septal thickening adjacent to clusters of AMs, and an appearance of
  4      inflammatory cells, primarily within the septa.  At 9 weeks of exposure, focal aggregations of
  5      particle-laden macrophages developed near the terminal bronchi, along with an influx of
  6      polymorphonuclear leukocytes (PMNs), Type II cell proliferation, and thickening of the alveolar
  7      walls. The affected alveoli occurred in clusters that, for the most part, were located near the
  8      terminal bronchioles, but occasionally were focally located in the lung .parenchyma.
  9             Mauderlyetal. (1981) exposed rate and mice by inhalation to diluted diesel exhaust for
10      545 h over a 19-week period on a regimen of 7 h/day, 5 days/week at concentrations of 0, 0.21,
11      1.02, or 4.38 mg/rn3 particulate matter.  Indices of health effects were minimal following 19
12      weeks of exposure. There were no significant exposure-related differences in mortality or body
13      weights of the rats or mice. There also were no significant differences in respiratory function
14      (breathing patterns, dynamic lung mechanics, lung volumes, quasi-static PV relationships, forced
15      expirograms, and CO-diffusing capacity) in rats; pulmonary function was not measured in mice.
16      No effect on tracheal mucociliary or deep lung clearances were observed in the exposed groups.
17      Rats, but not mice, had elevated immune responses in lung-associated lymph nodes at the two
18      higher exposure levels.  Inflammation in the lungs of rats exposed to 4.38 mg/m3 particulate
19      matter was indicated by increases in PMNs and lung tissue proteases. Histopathologic findings
20      included AMs that contained diesel particles, an increase in Type II cells, and the presence of
21      particles in the interstitium and tracheobronchial lymph nodes.
22             Kaplan et al. (1982) evaluated the effects of subchronic exposure to diesel exhaust on
23      rats, hamsters, and mice. The  exhaust was diluted to a concentration of 1.5 mg/m3 particulate
24      matter; exposures were 20 h/day, 7 days/week.  Hamsters were exposed for 86 days, rats and
25      mice for 90 days. There were  no significant differences in mortality or growth rates between
26      exposed and control animals.   Lung weight relative to body weight of 15  rats exposed for 90 days
27      was significantly higher than the mean for a control group (15 rats).  Histological examination of
28      tissues of all three species indicated particle accumulation in the lungs and mediastinal lymph
29      nodes. Associated with the larger accumulations, there was a minimal increase in the thickness
30      of the alveolar walls, but the vast majority of the particles elicited no response.  After 6 mo of
31      recovery, considerable clearance of the diesel particles from the lungs occurred irtall three
32      species, as evaluated by gross pathology and histopathology. However, no quantitative estimate
33      of clearance was provided.
34    '         Toxic effects in animals from acute exposure to diesel exhaust appear to be primarily
35      attributable to the gaseous components (i.e., mortality from CO intoxication and lung injury

        2/1/98              «                    5-23       DRAFT-DO NOT CITE OR QUOTE

-------
 1      caused by cellular damage resulting from NO2 exposure). The results from short-term exposures
 2      indicate that rats experience no to minimal lung function impairment even at diesel exhaust
 3      levels sufficiently high to cause histological and cytological changes in the lung. In subchronic
 4      studies of durations of 4 weeks or more, frank adverse health effects are not readily apparent and
 5      when found are mild and result from exposure to concentrations of about 6 mg/m3 particulate
 6      matter and durations of exposures of 20 h/day. There is ample evidence that subchronic
 7      exposure to lower levels of diesel exhaust affects the lung, as indicated by accumulation of
 8      particles, evidence of inflammatory response, AM aggregation and accumulation near the
 9      terminal bronchioles, Type II cell proliferation, and thickening of alveolar walls adjacent to AM
10      aggregates. Little evidence exists, however, that subchronic exposure to diesel exhaust impairs
11      lung function.  Recent studies have implicated the organc fraction of DPM in the induction of
12      resoiratory allergic disease.
13
14      5.1.2.3.  Chronic Exposures
15      5.1.2.3.1. Effects on growth  and longevity. Changes in growth, body weight, absolute or
16      relative organ weights, and longevity can be measurable indicators of chronic toxic effects. Such
17      effects have been observed in some but not all of the long-term studies conducted on laboratory
18      animals exposed to diesel exhaust.  There was limited evidence for an effect on survival in the
19      published chronic animal studies; deaths occurred intermittently early in one study in female rats
20      exposed to 3.7 mg/m3 particulate matter; however, the death rate began to decrease after 15 mo,
21      and the survival rate after 30 mo was slightly higher than that of the control group (Research
22      Committee for HERP Studies, 1988).  Studies of the effects of chronic exposure to diesel exhaust
23      on survival and body weight or growth are detailed in Table 5-3.
24             Increased lung weights and lung-to-body weight ratios have been reported in rats, mice,
25      and hamsters.  These data are summarized in Table 5-4. In rats exposed for up to 36 weeks to
26      0.25 or 1.5 mg/m3 particulate matter,  lung wet weights (normalized to body weight) were
27      significantly higher in the 1.5 mg/m3  exposure group than control values after 12 weeks of
28      exposure (Misiorowski et al., 1980).  Rats and Syrian hamsters were exposed for 2 years (five
29      16-h periods per week) to diesel exhaust diluted to achieve concentrations of 0.7, 2.2, and 6.6
30      mg/m3 particulate matter (Brightwell et al., 1986). At necropsy, a significant increase in lung
31      weight was seen in both rats and hamsters exposed to diesel exhaust compared with controls.
32      This finding was more pronounced in the rats in which the increase was progressive with both
33      duration of exposure and particulate matter level. The increase was greatest at 30 mo (after the
34     • end of a 6-month observation period in the high-concentration male group where the lung weight
35      was 2.7 times the control and at 24 mo in the high-concentration female group [3.9 times

        2/1/98               •                     5-24        DRAFT-DO NOT CITE OR QUOTE

-------
CO
                      Table 5-3. Effects of chronic exposures to diesel exhaust on survival and growth of
                      laboratory animals



















Ul
1
NJ





O
2?
>
3
1
6
o
<2
o
H
O
M
O
G
O
H
M

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/M.




Rat, Wistar, F;
Mouse, NMRI,F
(7 mg/m3 only)







1

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 urn
MMD
0.25
0.75
1.5
0.19 urn MMD
6.0
12.0

8.3
0.71 urn MMD

0.35
3.5
7.0
0.25 urn MMD
4.24
0.35 urn MMD

0.7
2.2
6.6
0.1 1J
0.41d
1.08d
2.31"
3.72'
0.2-0.3 urn 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
0.5"

—
—

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









SO2
(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 ori 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 decrease
1 .06 in body weight; earlier deaths in
2.42 females exposed to 3.72 mg/mj,
4.70 stabilized by 15 mo
4.57 -
.
0.3 Reduced body weight in rats at 2. 5
1 . 1 and 6.98 mg/m3 and no effect in
3.4 mice










Study
Lewis etal. (1989)


Schreck etal. (1981)



Vinegar etal. (1981a,b)


Karagianes
etal. (1981)

Mauderly et al. (1984,
1987a)


Heinrich et al. (1986a)


Brightweli etal. (1986)


Research Committee for
HERP Studies (1988)




Heinrich etal. (1995)












-------
^o
oo
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
Exposure
period
18h/day
5 days/week
13.5 mo
(NMRI)
24 mo
(C57BL/N)
16 h/day
5 days/week
23 mo
Particles
(mg/mj)
6,98


2.44
6.33
C x T CO
(mg-h/m3) (ppm)
35,500 -NMRI 14.2
38,300 - C57


19,520 —
50,640 —
NOj SO;
(ppm) (ppm) Effects
2.3 2.8 Reduced body weight in NMRI
mice but not in C57BL/6N mice


— — Reduced survival in 6.33 mg/m*
— — after 300 days. Body weight
significantly lower at 6.33 mg/m'
Study



Nikulaetal. (1995)
       "Estimated from graphically depicted mass concentration data.
       ""Estimated from graphically presented mass concentration data for N02 (assuming 90% NO and 10% NO2).
       'Data for tests with light-duty engine; similar results with heavy-duty engine.
       'Light-duty engine.
       "Heavy-duty engine.
Is)
Os

-------
oo
                Table 5-4. Effects of chronic exposures to diesel exhaust on organ weights and
                organ-to-body-weight ratios                               *


















u»
ib
-J




O

E
H
1
i
O
o.
2; .
o
H
O
H^
w
o
70
/o
c| •
o
.H.
w

Species/sex
Rat, F344, M;
Mouse, AJi, 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/m5 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
1 04 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
0.1 9 urn MMD

«
2.0
0.23-0.36 urn
MMD
0.25
1.5
0.19 Jim MMD

2.0
0.23-0.36 urn
MMD
0.25
0.75
1.5
0.19 urn MMD
6.0
12.0

4,24
0.35 urn MMD


0.7'
2.2"
6.6





6.0'
12.0"

0,84
2.5
6.98




CXT
(mg-h/m3)
2,520-2,730



3,640


990
5,940


7,280


2,145
6,435
12,870

8,736
17,472

48,336-56,392



5,824
18,304
54,912





41,664
83,328

7,400
21,800
61.700




CO NO2 SOj
(ppm) (ppm) (ppm) Effects
— — — No effect on liver, kidney, spleen,
or heart weights


12,7 1.6 0.83 No effects on weights of lungs,
liver, heart, spleen, kidneys, and
testes
— — — , Increase in relative lung weight at
— — — L5 mg/m3 only initially seen at 12
weeks

11.5 1.5 0.81 No effects on heart weights


— — — No effects on heart mass
— • — — .
— — —

— — — Increase in lung weight and
— — — lung/body weight ratio

12.5 1.5 1.1 Increase in rat, mouse, and
hamster lung weight and dry
weights

— — — Increase in lung weight
— — — concentration related in rats; heart
32.0 — — weight/body weight ratio greatet
at 6.6 mg/m3




20.2 2.7 2.7 Decrease in lung and kidney
33.2 4.4 5.0 weights

2.6 0.3 0.3 Increased rat and mouse lung
8.3 .1.2 1.1 . weight at 7 mg/m3 from 6 mo and
21 .2 3.8 . 3.4 at 2.5 me/m3 at 22 and 24 mo





Study
Kaplan etal, (1982)



Green etal. (1983)


Misiorowski
et al. (1980)


Vallyathan
etal. (1986)

Penney etal. (1981)



Vinegar et al. (1981 a,b)


Heinrich etal. (1986a,b)
StOber(1986)


Brightwell etal. (1986)







Pepelkoetal. (1980b,
1981)
Moorman etal. (1985)
Heinrich etal. (1995)







-------
oo
                  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;
CS7BL/6N, F




Rats, F344, M


Rat


Mouse

Exposure
period
18h/day
5 days/week
13.5 mo
(NMRI)
24 mo
(C57BL/N)
16h/day
5 days/week
23 mo





Particles
(mg/ra3)
6.98





2.44
6.33

0.8
2.5
6.98 , -
6.98
4.5
CXT
(mg«h/m3)
35,500 -NMRI
•38,300 - C57




19,520
50,640






CO NO2 SOj
(ppm) (ppm) (ppm) Effects
14.2 2.3 2.8 Increased lung weight





— — — Increase in lung weight was
— — . — significant at 2 and 6 mg/m3

Increased lung weight in rats and
mice at 3.5 and 7 mg/m3



Study






Nikulaetal. (1995)


Henderson etal. (1988)




to
oo
        * I to 61 weeks of exposure.

        '62 to 124 weeks of exposure.
O
O

^
O
-i

O
I—<
H
W

O
90

/O.
a
o
H
W

-------
  1      control]). Heinrich et al. (1986a,b; see also Stober, 1986) found a significant increase in wet and
  2      dry weights of the lungs of rats and mice exposed at 4.24 mg/m3 paniculate matter for 1 year in
  3      comparison with controls. After 2 years, the difference was a factor of 2 (mice) or 3 (rats).  After
  4      the same exposure periods, the hamsters showed increases of 50 to 75%, respectively. Exposure
  5      to equivalent filtered diesel exhaust caused no significant effects in any of the species. Vinegar
  6      et al. (1980,1981a,b) exposed hamsters to two levels of diesel exhaust with resultant
  7      concentrations of about 6 and 12 mg/m3 particulate matter for 8 h/day, 7 days/week for 6 mo.
  8      Both exposures significantly increased lung weight and lung weight to.body weight ratios. The
  9      difference between lung weights of exposed and control hamsters exposed to 12 mg/m3
10      particulate matter was approximately twice that of those exposed to 6 mg/m3.
11             Heinrich et al. (1995) reported that rats exposed to 2.5 and 7 mg/m3 for  18 h/day, 5
12      days/week for 24 mo showed significantly lower body weights than control starting at day 200 in
13      the high-concentration group and at day 440 in the low-concentration group.  Body weight in the
14      low-concentration group was unaffected, as was mortality in any group. Lung weight was
15      increased in the 7 mg/m3 group starting at 3 mo and persisting throughout the study while the 2.5
16      mg/m3 group showed increased lung weight only at 22 and 24 mo of exposure.  Mice(NMRI
17      strain) exposed to 7 mg/m3 in this  study for 13.5 mo had no increase in mortality and minimal,
18      insignificant decreases in body weight.  Lung weights were dramatically affected, with increases
19      progressing throughout the study from 1.5-fold at 3 mo to 3-fold at 12 mo. Mice (NMRI and
20      C57BL/6N strains) were also exposed to 4.5 mg/m3 for 23 mo.  In NMRI mice, the body weights
21      were reported to be significantly lower than controls, but the magnitude of the change is not
22      reported so biological significance cannot be assessed.  Mortality was slightly increased, but
23      statistical significance is not reported. The C57BL/6N mice showed minimal effects on body
24      weight and mortality, which were  not reported to be statistically significant. Lung weights were
25      dramatically affected in both strains.
26             Nikula et al. (1995) exposed male and female F344 rats to diesel particle concentrations
27      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
28      effects of diesel with those of carbon black. Significantly reduced survival was observed in
29      males exposed to 6.3 mg/m3 but not in females or at the lower concentration.  Body weights were
30      decreased by exposure to 6.3 mg/m3 diesel exhaust in both male and female rats throughout the
31      exposure period. Significant increases in lung weight were first seen at 6 mo in the high-
32      exposure group and at 12 to 18 mo in the low-exposure group.
33             No evidence was found in the published literature that chronic exposure to diesel exhaust
34      affected the weight of body organs other than the lung and heart (e.g., liver, kidney, spleen, or
35      testes) (Table 5-4). Morphometric analysis of hearts from rats and guinea pigs exposed to 0.25,

        2/1/98                                   5-29       DRAFT-DO NOT CITE OR QUOTE

-------
 1      0.75, or 1.5 mg/m3 participate matter 20 h/day, 5.5 days/week for 78 weeks revealed no
 2      significant alteration in mass at any exposure level or duration of exposure (Penney et al., 1981).
 3      The analysis included relative wet weights of the right ventricle, left ventricle, combined atria,
 4      and ratio of right to left ventricle.  Vallyathan et al. (1986) found no significant differences in
 5      heart weights and the ratio of heart weight to body weight between rats exposed to 2 mg/m3
 6      particulate matter for 7 h/day, 5 days/week for 24 mo and their respective clean air chamber
 7      controls. No significant differences were found in the lungs, heart, liver, spleen, kidney, and
 8      testes of rats exposed for 52 weeks, 7 h/day, 5 days/week to diluted diesel exhaust containing 2
 9      mg/m3 particulate matter compared with their respective controls (Green et al., 1983).
10
11      5.1.2.3.2, Effects on pulmonary function. The effect of long-term exposure to diesel exhaust
12      on pulmonary function has been evaluated in laboratory studies of rats, hamsters, cats, and
13      monkeys.  These studies are summarized in Table 5-5, along with more details on the exposure
14      characteristics, in general order of increasing dose (C * T) of the diesel exhaust particulate
15      matter. The text will be presented using the same approach.
16            Lewis et al. (1989) evaluated 10 control and 10 diesel-exposed rats (2 mg/m3 particulate
17      matter, 7 h/day, 5 days/week for 52 or 104 weeks) for responses in  functional residual capacity
18      and airway resistance and conductance.  At the 104-week evaluation, the rats were also examined
19      for maximum flow volume impairments. No evidence of an impairment of-pulmonary function
20      as a result of the exposure to diesel exhaust was found in rats.  Lewis et al, (1989) exposed male
21      cynomolgus monkeys to diesel exhaust for 7 h/day, 5 days/week, for 24 mo. Groups of 15
22      monkeys were exposed to  air, diesel exhaust (2 mg/m3), coal dust, or combined coal dust and
23      diesel exhaust. Pulmonary function was evaluated prior to exposure and at 6-month intervals
24      during the  2-year exposure, including compliance and resistance, static and dynamic lung
25      volumes, distribution of ventilation, diffusing capacity, and maximum ventilatory performance.
26      There were no effects on lung volumes, diffusing capacity, or ventilation distribution, so there
27      was no evidence of restrictive disease.  There was, however, evidence of obstructive airway
28      disease as measured by low maximal flows in diesel-exposed monkeys. At 18 mo of exposure,
29      forced expiratory flow at 25% of vital capacity and forced expiratory flow normalized to FVC
30      were increased. The measurement of forced expiratory flow at 40% of total lung capacity was
31      significantly increased at 12,18, and 24 mo of exposure. The finding of an obstructive effect hi
32      monkeys contrasts with the finding of restrictive type effects in other laboratory animal species
33      (Vinegar et al., 1980,1981a; Mauderly et al., 1988; Pepelko et al.,  1980b, 1981) and suggests a
34     ' possible difference in effect between primate and small animal respiratory tracts.  In these
        2/1/98                '                    5-30       DRAFT-DO NOT CITE OR QUOTE

-------
Table 5-5. Effects of diesel exhaust on pulmonary function of laboratory animals
vo
oo



















Y*
u>
*~*



O
5
TJ
H
I
I
O

25
O

o
1— H
H
W
O
^
iQ
o
W

Species/sex
Rat, F344,
M.F

Monkey, M,
, Cynomolgus

Rat, F344.M


Rat, Wistar, F


Hamster, Chinese,
M


Rat, F344,
M, F



Hamster, Syrian, M,
F

Rat, F344, M, F;
Hamster Syrian, M,
F


Rat, Wistar, F


Cat, inbred, M




Exposure
period
7 h/day
5 days/week
104 weeks
7 h/day
5 days/week
104 weeks
20 h/day
5.5 days/week
87 weeks
7-8 h/day
5 days/week
104 weeks
8 h/day
7 days/week
26 weeks '

7 h/day
5 days/week
130 weeks


19 h/day
5 days/week
120 weeks
16 h/day
5 days/week
104 weeks


19 h/day
5 days/week
140 weeks
8 h/day
7 days/week
124 weeks


Particles
(mg/m3)
2.0
0.23-0.36 urn
MMD
2.0
0.23-0.36 urn MMD

1.5
0.19 urn MMD

3.9
0.1 urn MMD

6.0
12.0


0.35
3.5
7.0
0.23-0.26 urn
MMD
4.24
0.35 urn MMD

0.7
2.2
6.6


4.24
0.35 urn MMD

6.0"
12.0"



CxT
(mg'h/m*)
7,280


7,280


14,355


14,196-16,224


8,736
17,472


1,593
15,925
31,850


48,336


5,824
18,304
54,912


56,392


41,664
83,328



CO NO2 SO2
(ppm) (ppm) <[>pm) 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
16.5 0.34 — compliance reduced at 3.5 and 7
29.7 0.68 — mg/m3


12.5 1.5 1.1 Significant increase in airway
resistance

— — — Large number of pulmonary ftinc*
, — . — — tion changes consistent with
— — — obstructive and restrictive airway
diseases at 6.6 mg/m* (no specific
data provided)
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
33.3 4.4 5.0 lung capacity, and diffusing
capacity after 2 years; no effect on
expiratory flow


Study
Lewis etal. (1989)


Lewis etal. (1989)


Gross (1981)


Heinrich etal. (1982)


Vinegar etal. (1980,
1981a,b)


Mauderly etal. (1988)
McClellan et al. (1986)



Heinrich etal. (1986a) .


Brightwell etal. (1986)




Heinrich etal. (1986a)


Pepelkoetal. (1980b,
1981)
Moorman etal. (1985)


•1 to 61 weeks exposure.
b62 to 124 weeks of exposure. •



















-------
 1      monkeys there were no specific histopathological effects reported (see next section) although
 2      particle aggregates were reported in the distal airways, suggesting more small airway deposition.
 3            Gross (1981) exposed rats for 20 h/day, 5.5 days/week for 87 weeks to diesel exhaust
 4      containing 1.5 mg/m3 participate matter.  When the data were normalized (e.g., indices expressed
 5      in units of airflow or volume for each animal by its own forced expiratory volume), there were
 6      no apparent functionally significant changes occurring in the lungs at 38 weeks of exposure that
 7      might be attributable to the inhalation of diesel exhaust. After 87 weeks of exposure, functional
 8      residual  capacity (FRC) and its component volumes (expiratory reserve [ER] and residual
 9      volume [RV]), maximum expiratory flow (MEF) at 40% FVC, MEF at 20% FVC, and FEV0.,
10      were significantly greater in the diesel-exposed rats. An observed increase in airflow at the end
11      of the forced expiratory maneuver when a decreased airflow would be expected from the
12      increased FRC, ER, and RV data (the typical scenario of human pulmonary disease) showed
13      these data to be inconsistent with known clinically significant health effects.  Furthermore,
14      although the lung volume changes in the diesel-exposed rats could have been indicative of
15      emphysema or chronic obstructive lung disease, this interpretation was contradicted by the
16      airflow data, which suggest simultaneous lowering of the resistance of the distal airways.
17            Heinrich et al. (1982) evaluated the pulmonary function of rats exposed to a concentration
18      of 3.9 mg/m3 particulate matter for 7 to 8 h/day, 5 days/week for 2 years. When compared with a
19      control group, no significant changes in respiratory rate, minute volume, compliance, or
20      resistance occurred in the exposed group (number of rats per group was not stated).
21            Hamsters (eight or nine per group) were exposed 8 h/day, 7 days/week, for 6 mo to
22      concentrations of either about 6 mg/m3 or about 12 mg/m3 particulate matter (Vinegar et al.,
23      1980, 1981a,b).  Vital capacity, vital capacity/lung weight ratio, residual lung volume  by water
24      displacement, and CO2 diffusing  capacity decreased significantly in hamsters exposed to
25      6 mg/m3 particulate matter. Static deflation volume-pressure curves showed depressed deflation
26      volumes for diesel-exposed hamsters when volumes were corrected for body weight and even
27      greater depressed volumes when  volumes were corrected for lung weight. However, when
28      volumes were expressed as percentage of vital capacity, the diesel-exposed hamsters had higher
29      lung volumes at 0 and 5 cm H2O. In the absence of confirmatory histopathology, the authors
30      tentatively concluded that these elevated lung volumes and the significantly reduced diffusing
31      capacity in the same hamsters were indicative of possible emphysematous changes in the lung.
32      Similar lung function changes were reported in hamsters exposed at 12 mg/m3 particulate matter,
33      but detailed information was not reported. It was stated, however, that the decrease in vital
34      capacity was 176% greater in the second experiment than in the first.
        2/1/98                                    5-32       DRAFT-DO NOT CITE OR QUOTE

-------
 1             Mauderly et al. (1988; see also McClellan et al, 1986) examined the impairment of
 2      respiratory function in rats exposed for 7 h/day, 5 days/week, for 24 mo to diluted diesel exhaust
 3      with 0,35, 3.5, or 7.1 mg/m3 partieulate matter.  After 12 mo of exposure to the highest
 4      concentration of diesel exhaust, the exposed rats (n = 22) had lower total lung capacity (TLC),
 5      dynamic lung compliance (Cdyn),  FVC, and CO diffusing capacity than controls (n = 23). After
 6      24 mo of exposure to 7 mg/m3 partieulate matter, mean TLC, Cdyn, quasi-static chord compliance,
 7      and CO diffusing capacity were significantly lower than control values. Nitrogen washout and
 8      percentage of FVC expired in 0.1 s were  significantly greater than control values. There was no
 9      evidence of airflow obstruction. The functional alterations were attributed to focal fibrotic and
10      emphysematous lesions and thickened alveolar membranes observed by histological
11      examination.  Similar functional alterations and histopathologic lesions were observed in the rats
12      exposed to 3.5 mg/m3 partieulate  matter,  but such changes usually occurred later in the exposure
13      period and were generally less pronounced. There were no significant decrements in pulmonary
14      function for the 0.35 mg/m3 group at any time during the study nor were there reported
15      histopathologic changes in this group.
16             Additional studies were conducted by Heinrieh et al. (1986a,b; see also Stober, 1986) on
17      the effects of long-term exposure to diesel exhaust on the pulmonary function of hamsters and
18      rats. The exhaust was diluted to achieve  a concentration of 4.24 mg/m3 partieulate matter;
19      exposures were for 19 h/day, 5 days/week for a maximum of 120 weeks (hamsters) or  140 weeks
20      (rats).  After 1 year of exposure to the diesel exhaust, the hamsters exhibited a significant
21      increase in airway resistance and  a nonsignificant reduction in lung compliance.  For the same
22      time period, rats showed increased lung weights, a significant decrease in Cdyn, and a significant
23      increase in airway resistance.  These indices did not change during the second year of exposure.
24             Syrian hamsters and rats were exposed to 0.7,2.2, or 6.6 mg/m3 partieulate matter for
25      five 16-h periods per week for 2 years (Brightwell et al., 1986).  There were no treatment-related
26      changes in pulmonary function in the hamster.  Rats exposed to the highest concentration of
27      diesel exhaust exhibited changes in pulmonary function (data not presented) that were reported to
28      be consistent with a concentration-related obstructive and restrictive disease.
29             Pepelko et al. (1980b;  1981; see also Pepelko, 1982b) and  Moorman et al. (1985)
30      measured the lung function of adult cats chronically exposed to diesel exhaust. The cats were
31      exposed for 8 h/day and 7 days/week for  124 weeks. Exposures were at 6 mg/m3 for the first 61
32      weeks and 12 mg/m3 from weeks 62 to 124. No definitive pattern of pulmonary function
33      changes was observed following 61 weeks of exposure; however,  a classic pattern of restrictive
34      lung disease was found at 124 weeks. The-significantly reduced lung volumes (TLC, FVC, FRC,
35      and inspiratory capacity [1C]) and the significantly lower single-breath diffusing  capacity,

        2/1/98                                    5-33       DRAFT-DO NOT CITE OR QUOTE

-------
 1      coupled with normal values for dynamic ventilatory function (mechanics of breathing), indicate
 2      the presence of a lesion that restricts inspiration but does not cause airway obstruction or loss of
 3      elasticity.  This pulmonary physiological syndrome is consistent with an interstitial fibrotic
 4      response that was later verified by histopathology (Plopper et al., 1983).
 5             Pulmonary function impairment has been reported in rats, hamsters, cats, and monkeys
 6      chronically exposed to diesel exhaust. In all species but the monkey, the pulmonary function
 7      testing results have been consistent with restrictive lung disease. The monkeys demonstrated
 8      evidence of small airway obstructive responses. The disparity between the findings in monkeys
 9      and those in rats, hamsters, and cats could be in part the result of increased particle retention in
10      the smaller species resulting from (1) exposure to diesel exhaust that has higher airborne
11      concentrations of gases, vapors, and particles and/or (2) longer duration of exposure.  The nature
12      of the pulmonary impairment is also dependent on the site of deposition and routes of clearance,
13      which are determined by the anatomy and physiology of the test laboratory species and the
14      exposure regimen.  The data on pulmonary function effects raise the possibility that diesel
15      exhaust produces small airway disease in primates compared with primarily alveolar effects in
16      small animals and that similar changes might be expected in humans and monkeys.
17      Unfortunately, the available data in primates are too limited to draw clear conclusions.
18
19      5.1.2.3.3. Lung morphology, biochemistry, and lung lavage analysis. Several studies have
20      examined the morphological, histological, and histochemical changes occurring in the lungs of
21      laboratory animals chronically exposed to diesel exhaust. The histopathological effects of diesel
22      exposure in the lungs of laboratory animals are summarized in Table 5-6, ranked in order of
23      C x T. Table 5-6 also contains an expanded description of exposures.
24             Kaplan et al. (1982) performed macroscopic and microscopic examinations of the lungs
25      of rats, mice, and hamsters exposed for 20 h/day, 7 days/week for 3 mo to diesel exhaust
26      containing  1.5 mg/m3 particulate matter. Gross examination revealed diffuse and focal
27      deposition of the diesel particles that produced a grayish overall appearance of the lungs with
28      scattered, denser black areas.  There was clearance of particles via the lymphatics to regional
29      lymph nodes. Microscopic examination revealed no anatomic changes in the upper respiratory
30      tract; the mucociliary border was normal in appearance. Most of the particles were in
31      macrophages, but some were free as small aggregates on alveolar and bronchiolar surfaces. The
32      particle-laden macrophages were often in masses near the entrances of the lymphatic drainage
33      and respiratory ducts.  Associated with these masses was a minimal increase in the thickness of
34      the alveolar walls; however, the vast majority of the particles elicited no response. After 6 mo
        2/1/98                                    5-34       DRAFT-DO NOT CITE OR QUOTE

-------
Table 5-6. Histopathological effects of diesel exhaust in the lungs of laboratory animals
oo





s












1
Ul




D
§>
TJ
H
1
D
.O
2J
O
H
n



T1
r"
(O
G
O
H
M

Species/sex
Rat, F344, M;
Mouse, A/J, M;
Hamster, Syrian,
M
Monkey,
Cynomoigus, 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
(mg/m3)
1.5
0,19 urn MDD


2.0
0.23-0.36 urn
MDD
2.0
• 0.23-0.36 urn
MPD
6.0





6.0
12.0


3.9
O.I urn MDD

8.3
0.71 Jim MDD




4.9



0.35
3.5
7.1
0.23 urn MDD






CxT CO NO2 SOj
(mg'h/m3) (ppm) (('["«) (ppm) Effects
2,520-2,730 — — — Inflammatory changes; increase
in lung weight; increase in
thickness of alveolar walls

7,280 11.5 1.5 0.8 AM aggregation; no fibrosis,
inflammation, or emphysema

3,640 11.5 1.5 0.8 Multifocal histiocytosis;
inflammatory changes; Type II
cell proliferation; fibrosis
13,104 — — — 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
6,240 . — — — Inflammatory changes; AM
12,480 — — — accumulation; thickened alveolar
lining; Type II cell hyperplasia;
edema; increase in collagen
16,380-18,720 18.5 ' 1.2 3.1 Inflammatory changes, 60%
adenomatous cell proliferation

21,663 . 50.0 4.0-6.0 — Inflammatory changes; AM
aggregation; alveolar cell
hypertrophy; interstitial fibrosis,
emphysema (diagnostic method-
ology not described)

28,538 7.0 1.8 13.1 Type II cell proliferation;
inflammatory changes; bronchial
hyperplasia; fibrosis

1,592 2,9 0.05 — Alveolar and bronchiolar
15,925 16.5 0.34 — epithelial metaplasia in rats at 3. 5
31,850 29.7 0.68 — and 7.0 mg/m3; fibrosis at
7.0 mg/m3 in rats and mice;
inflammatory changes
'





Study
Kaplan etal. (1982)



Lewis et al. (1989)


Bhatnagar etal. (1980)
Pepelko(l982a)

Bhatnagar etai. (1980)
Pepelko(1982a)




Pepelko (1982b)



Heinrich etal. (1982)


Karagianes etal. (1981)





Iwaietal. (1986)



Mauderly et al.
(1987a)
Henderson etal. (1988)








-------
                Table 5-6.  Histopathological effects of diesel exhaust in the lungs of laboratory animals (continued)
00
w
OS
•*^f

$
O
O

2

s
n
HH
H
W
CJ

.3
m
Species/sex
Rat^ Wistar, F;
Mouse, NMRI, F
(7 mg/m3 only)

Mouse, NMRI, F;
C57BL/6N, F




Mouse

Rat, M, F,
F344/Jcl.



Hamster, Syrian,
M,F



Mouse, NMRI, F


Rat, Wistar, F



Guinea Pig,
Hartley, M




r
Exposure
period
18 h/day
5 days/week
24 mo

18 h/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


19 h/day
5 days/week
120 weeks
19 h/day
5 days/week
140 weeks

20 h/day
5.5
days/week
104 weeks



Particles
(mg/m3)
0.8
2.5
6.98
.
6.98





4.5

0.11"
0.411
1.08-
2-.31"
3.72"
4.24




4.24


4.24



0.25
0.75
1.5
6.0



Cxj
(mg-h/m3)
7,400
21,800
61,700

35,500 -NMRI
38,300 - C57






1,373
5,117
13,478
28,829
46,336
48,336




48,336


56,392



2,860
8,580
17,160
68,640



CO
(ppm)
2.6
8.3
21.2

14.2







1.23
2.12
3.96
7.10
12.9
12.5




12.5


. 12.5



__
—
—
—



NO2
(ppm)
0.3
1.2
3.8

2.3







' -0.08
0.26
0.70
1.41
3.00
1.5




1.5


1.5



—
—
—
—



SO2
(ppm)
0.3
1.1
3.4

2.8







0.38
1.06
2.42
4.70
4.57
1.1




1.1


1.1



—
—
—
—



Effects
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 11
cell hyperplasia and lung tumors
seen at >0.4 mg/m3; shortening
and loss of cilia in trachea and
bronchi
Inflammatory changes; thickened
alveolar septa;
bronchioloalveolar hyperplasia;
emphysema (diagnostic
methodology not described)
Inflammatory changes; bronchio-
loalveolar hyperplasia; alveolar
lipoproteinosis; fibrosis
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
Study
Heinrichetal. (1995)











Research Committee
for HERP Studies
(1988)


Heinrich et al. (1986a)




Heinrichetal. (1986a)


Heinrich et al. (1986a)



Barnhartet al. (1981,
1982)
Vostaletal. (1981)





-------
                  Table 5-6.  Histopathological effects of diesel exhaust in the lungs of laboratory animals (continued)
tx
OJ
o

z
o
H
O
>— I
H
M

O
90
Exposure
Species/sex period
Cat, inbred, M - 8 h/day
7 days/week
124 weeks
Rat, F344, M 16 h/day
5 days/week
23 mo

Particles
(mg/mj)
6.0°
12.0d
2.44
6.33

C*T
(mg-h/m1)
41,664
83,328
19,520
50,640

CO
(ppm)
20.2
33.2
—

NO2 SOj .
(ppm) (ppm) Effects
2.7 2.1 Inflammatory changes; AM
4.4 5.0 aggregation; bronchiolar
epithelial metaplasia; Type II cell
hyperplasia; peribronchiolar
fibrosis
— — AM hyperplasia, epithelial
— — hyperplasia, inflammation, septal
fibrosis, bronehoalveolar
metaplasia
Study
Plopper.etal. (1983)
Hyde etal. (1985)
Nikulaetal. (1995)

         "Light-duty engine.

         'Heavy-duty engine.

         M to 61 weeks exposure.

         d62 to 124 weeks of exposure,


         AM = Alveolar macrophage.

         PMN = Polymorphonuclear leukocyte.
o
H
W

-------
  1      of recovery, the lungs of all three species contained considerably less pigment, as assessed by
  2      gross pathological and histopathological examinations,
  3             Lewis et al. (1989; see also Green et al, 1983) performed serial histological examinations
  4      of rat lung tissue exposed to diesel exhaust containing 2 mg/m3 particulate matter for 7 h/day, 7
  5      days/week for 2 years. Accumulations of black-pigmented AMs were seen in the alveolar ducts
  6      adjacent to terminal bronchioles as early as 3  mo of exposure, and particles were seen within the
  7      interstitium of the alveolar ducts.  These macular lesions increased in size up to 12 mo of
  8      exposure.  Collagen or reticulum fibers were seen only rarely in association with deposited
  9      particles; the vast majority of lesions showed no evidence of fibrosis.  There was no evidence of
10      focal emphysema with the macules.  Multifocal histiocytosis (24% of exposed rats) was observed
11      only after 24 mo of exposure. These lesions were most commonly observed subpleurally and
12      were composed of collections of degenerating macrophages and amorphous granular material
13      within alveoli, together with fibrosis and chronic inflammatory cells in the interstitium.
14      Epithelial lining cells adjacent to collections of pigmented macrophages showed a marked Type
15      II cell hyperplasia; degenerative changes were not observed in Type I cells.  Histological
16      examination of lung tissue from monkeys exposed for 24 mo in the same regimen as used for rats
17      revealed aggregates of black particles, principally in the distal airways of the lung.  Particles
18      were present within the cytoplasm of macrophages in the alveolar spaces as well as the
19      interstitium. Fibrosis, focal emphysema, or inflammation was not observed.  No specific
20      histopathological lesions were reported for the monkey.
21             Nikula et al. (1997) reevaluated the lung tissue from this study. They concluded that
22      there were no significant differences in the amount of retained particulate matter between
23      monkeys and rats exposed under the same conditions. The rats, however, retained a greater
24      portion of the particulate matter in lumens of the alveolar ducts and alveoli than did the monkeys.
25      Conversely, monkeys retained a greater portion of the particulate material in the interstitum than
26      did rats. Aggregations of particle-laden macrophages in the alveoli were rare, and there were few
27      signs of particle-associated inflammation in the monkeys. Minimal histopathologic lesions were
28      detected in the interstitium.  Although the lungs of the monkeys showed a marginal and
29      significantly lesser inflammatory response than rats exposed to the same exposure regime, the
30      results should be interpreted with caution because 2 years is near the normal lifetime for rats, but
31      less than 10% of the normal lifespan of Cynomolgus monkeys,
32             Histopathological effects of diesel exhaust on the lungs of rats have been investigated by
,33      the Health Effects Research Program on Diesel Exhaust (HERP) in Japan.  Both light-duty (LD)
34      and heavy-duty (HD) diesel engines were used. The exhaust was diluted to achieve nominal
35      concentrations of 0.1 (LD only), 0.4 (LD and HD), 1 (LD and HD), 2 (LD and HD), and 4 (HD

        2/1/98                                   5-38      DRAFT-DO NOT CITE OR QUOTE

-------
  1      only) mg/m3 paniculate matter. Rats were exposed for 16 h/day, 6 days/week for 30 mo. No
  2      histopathological changes were observed in the lungs of rats exposed to 0.4 mg/m3 particulate
  3      matter or less. At concentrations above 0,4 mg/m3 particulate matter, severe morphological
  4      changes were observed. These changes consisted of shortened and absent cilia in the tracheal
  5      and bronchial epithelium, marked hyperplasia of the bronchiolar epithelium, and swelling of the
  6      Type II cellular epithelium. These lesions appeared to increase in severity with increases in
  7      exhaust concentration and duration of exposure. There was no difference in the degree of
  8      changes in pulmonary pathology at the same level of concentrations between the LD and the HD
  9      series.
10             Histological examination of the respiratory tract of hamsters revealed significantly higher
11      numbers of hamsters exhibiting definite proliferative changes in the lungs in the group exposed
12      to diesel exhaust than were observed in the group exposed to particle-free diesel exhaust or clean
13      air (Heinrich et al., 1982).  Sixty percent of these changes were described as adenomatous
14      proliferations.  Exposures were for 7 to 8 h/day, 5 days/week for 104 weeks to diesel exhaust
15      diluted to achieve a concentration of 3.9 mg/m3 particulate matter.
16            Heinrich et al. (1995) reported increased incidence and  severity of bronchioloalveolar
17      hyperplasia in rats exposed to 0.8, 2.5, and 7 mg/m3. The lesion in the lowest concentration
18      group was described as very slight to moderate.  Slight to moderate interstitial fibrosis also
19      increased in incidence and severity in all exposed groups, but incidences were not reported.  This
20      chronic study also exposed NMRI mice to  7 mg/m3 for 13.5 mo and both NMRI and C56BL/6N
21      mice to 4.5 mg/m3 for 24 mo. Noncancer histological endpoints are not discussed in any detail in
22      the report, which is focused on the carcinogenicity on diesel as compared with titanium dioxide
23      and carbon black.
24            Iwai et al, (1986) performed serial histopathology on the lungs of rats at 1, 3,6,12, and
25      24 mo of exposure to diesel exhaust. Exposures were for 8 h/day, 7 days/week for 24 mo; the
26      exposure atmosphere contained 4.9 mg/m3  particulate matter. At 1 and 3 mo of exposure, there
27      were minimal histological changes in the lungs of the exposed rats. After 6 mo of exposure,
28      there were particle-laden macrophages distributed irregularly throughout the lung and a
29      proliferation of Type II cells with adenomatous metaplasia in areas where the macrophages had
30      accumulated,  After 1 year of exposure, foci of heterotrophic hyperplasia of ciliated or
31      nonciliated bronchiolar epithelium on the adjacent alveolar walls were more common, the
32      quantity of deposited particulate matter increased, and the number of degenerative AMs and
33      proliferative lesions of Type II or bronchiolar epithelial cells increased.  After 2 years of
34      exposure, there was a fibrous thickening of the alveolar walls, mast cell infiltration with
35      epithelial hyperplasia in areas where the macrophages had accumulated, and neoplasms.

        2/1/98                                     5-39        DRAFT-DO NOT CITE OR QUOTE

-------
 1             Heinrich et al. (1986a; see also Stober, 1986) performed histopathologic examinations of
 2      the respiratory tract of hamsters, mice, and rats exposed to diesel exhaust that had 4 mg/m3
 3      participate matter. Exposures were for 19 h/day, 5 days/week; the maximum exposure period
 4      was 120 weeks for hamsters and mice and 140 weeks for rats. Histological examination revealed
 5      different levels of response among the three species. In hamsters, the exhaust produced
 6      thickened alveolar septa, bronchioloalveolar hyperplasia, and what were termed ernphysernatous
 7      lesions (diagnostic methodology not described).  In mice, bronchioloalveolar hyperplasia
 8      occurred in 64% of the mice exposed to the exhaust and in 5% of the controls. Multifocal
 9      alveolar lipoproteinosis occurred in 71% and multifocal interstitial fibrosis occurred in 43% of
10      the mice exposed to exhaust but in only 4% of the controls.  In exposed rats, there were severe
11      inflammatory changes in the lungs, as well as thickened septa, foci of macrophages, and
12      hyperplastic and metaplastic lesions.
13             Nikula et al. (1995) reported in detail the nonneoplastic effects in male and female F344
14      rats exposed to 2.4 or 6.3 mg/m3 of diesel exhaust particles.  At 3 mo in the low-concentration
15      group, enlarged particle-containing macrophages were found with minimal aggregation.  With
16      higher concentration and longer duration of exposure, the number and size of macrophages and
17      aggregates increased. Alveolar epithelial hyperplasia was found starting at 3 mo and in all rats at
18      6 mo. These lesions progressed to chronic active inflammation, alveolar proteinosis, and septal
19      fibrosis at 12 mo.  Other lesions observed late in the study included bronchiolar-alveolar
20      metaplasia, squamous metaplasia, and squamous cysts. This study reports in detail the
21      progression of lesions in diesel exhaust exposure and finds relatively little difference between the
22      lesions caused by diesel exhaust exposure and exposure to similar levels of carbon black
23      particles.
24             The effects of diesel exhaust on the lungs of 18-week-old rats exposed to 8.3 ±2.0 mg/m3
25      particulate matter were investigated by Karagianes et al. (1981). Exposures were for 6 h/day, 5
26      days/week, for 4, 8, 16, or 20 mo.  Histological examinations of lung tissue noted focal
27      aggregation of particle-laden AMs, alveolar histiocytosis, interstitial fibrosis, and alveolar
28      emphysema (diagnostic methodology not described). Lesion severity was related to length of
29      exposure. No significant differences were noted in lesion severity among the diesel exhaust, the
30      diesel exhaust plus coal dust (5.8 ± 3.5 mg/m3), or the high-concentration (14.9 ± 6.2 mg/m3)
31      coal dust exposure groups following 20. mo of exposure.
32             Histological changes in the lungs of guinea pigs exposed to diluted diesel exhaust
33      containing either 0.25, 0.75, 1.5, or 6.0 mg/m3 particulate matter were reported by Barnhart et al.
34     • (1981, 1982).  Exposures at 0.75 and 1.5 mg/m3 for 2 weeks to  6 mo resulted in an uptake of
35      exhaust particles by three alveolar cell types (AMs, Type I cells, and interstitial macrophages)

        2/1/98               '                      5-40       DRAFT-DO NOT CITE OR QUOTE

-------
  1      and also by granulocytic leukocytes (eosinophils). The alveolar-capillary membrane increased in
  2      thickness as a result of an increase in the absolute tissue volume of interstitium and Type II cells.
  3      In a continuation of these studies, guinea pigs were exposed to diesel exhaust (up to 6.0 mg/m3
  4      particulate matter) for 2 years (Barnhart et al,, 1982).  A minimal tissue response occurred at the
  5      concentration of 0,25 mg/m3 After 9 mo of exposure, there was a significant increase, about
  Q      30%, in Type I and II cells, endothelial cells, and interstitial cells over concurrent age-matched
  7.      controls; by 24 mo only macrophages and Type II cells were significantly increased. As in the
  8      earlier study, ultrastructural evaluation showed that Type I cells, AMs,. and eosinophils
  9      phagocytized the diesel particles. Exposure to 0.75 mg/m3 for 6 mo resulted in fibrosis in
10      regions of macrophage clusters and in focal Type II cell proliferation.  No additional information
11      was provided regarding the fibrotic changes with increasing concentration or duration of
12      exposure.  With increasing concentration/duration of diesel exhaust exposure, Type II cell
13      clusters occurred in some alveoli. Intraalveolar debris was particularly prominent after exposures
14      at 1.5 and 6.0 mg/m3 and consisted of secretory products from Type II cells.
15            In studies conducted on hamsters, Pepelko (1982b) found that the lungs of hamsters
16      exposed for 8 h/day, 7 days/week for 6 mo to 6 or 12 mg/m3 particulate matter were
17      characterized by large numbers of black AM in the alveolar spaces, thickening of the alveolar
18      epithelium, hyperplasia of Type II cells, and edema,
19            Lungs from rats and'mice exposed to 0.35, 3.5, or 7.1 mg/m3 (0.23 to 0.26 um mass
20      median diameter [MMD]) for 7 h/day and 5 days/week showed pathologic lesions (Mauderly et
21      al., 1987a; Henderson et al., 1988). After 1 year of exposure at 7.1 mg/m3, the lungs of the rats
22      exhibited focal areas of fibrosis; fibrosis increased with increasing duration of exposure and was
23      observable in the 3.5-mg/m3 group of rats at 18 mo. The severity of inflammatory responses and
24      fibrosis was directly related to the exposure level. In the 0.35 mg/m3 group of rats, there was no
25      inflammation or fibrosis.  Although the mouse lungs contained high lung burdens of diesel
26      particles per gram of lung weight at each equivalent exposure concentration, there was
27      substantially less inflammatory reaction and fibrosis than was the case in rats.  Fibrosis was
28      observed only in the lungs of mice exposed at 7 mg/m3 and consisted of fine fibrillar thickening
29      of occasional alveolar septa.
30            Histological examinations were performed on the lungs of cats initially exposed to
31      6 mg/m3 particulate matter for 61 weeks and subsequently increased to 12 mg/m3 for Weeks 62
32      to 124 of exposure. Plopper et al. (1983; see also Hyde et al., 1985) concluded  from the results
33      of this study that exposure to diesel exhaust produced changes in both epithelial and interstitial
34    '  tissue compartments and that the focus of these lesions in the peripheral lung was the centriacinar
35      region where the alveolar ducts join the terminal conducting airways. This conclusion was based

        2/1/98               '                     5-41        DRAFT-DO NOT CITE  OR QUOTE

-------
 1      on the following evidence. The epithelium of the terminal and respiratory bronchioles in
 2      exposed cats consisted of three cell types (ciliated, basal, and Clara cells) compared with only
 3   _   one type (Clara cells) in the controls. The proximal acinar region showed evidence of
 4      peribronchial fibrosis and bronchiolar epithelial metaplasia.  Type II cell hyperplasia was present
 5      in the proximal interalveolar septa. The more distal alveolar ducts and the majority of the rest of
 6      the parenchyma were unchanged from controls. Peribronchial fibrosis was greater at the end of 6
 7      mo in clean air following exposure, whereas the bronchiolar epithelial metaplasia was most
 8      severe at the end of exposure.  Following an additional 6 mo in clean air, the bronchiolar
 9      epithelium more closely resembled the control epithelial cell population.
10             Wallace et al. (1987) used transmission electron microscopy (TEM) to determine the
11      effect of diesel exhaust on the intravascular and interstitial cellular populations of the lungs of
12      exposed rats and guinea pigs.  Exposed animals and matched controls were exposed to 0.25,
13      0.75, 1.5, or 6.0 mg/m3 particulate matter for 2, 6, or 10 weeks or 18 mo.  The results inferred the
14      following: (1) exposure to 6.0 mg/m3 for 2 weeks was insufficient to elicit any cellular response,
15      (2) both species demonstrated an adaptive multicellular response to diesel exhaust, (3) increased
16      numbers of fibroblasts were found in the interstitium from week 6 of exposure through month
17      18, and (4) there was no significant difference in either cell type or number in alveolar
18      capillaries, but there was a significant increase at 18 mo in the mononuclear population in the
19      interstitium of both species.
20             Additional means for assessing the adverse effects of diesel exhaust on the lung are to
21      examine biochemical and cytological changes in bronchoalveolar lavage fluid (BALF) and in
22      lung tissue. Fedan et al. (1985) performed studies to determine whether chronic exposure of rats
23      affected the pharmacologic characteristics of the rat's airway smooth muscle.  Concentration-
24      response relationships for tension changes induced with acetylcholine, 5-hydroxytryptamine,
25      potassium chloride, and isoproterenol were assessed in vitro on isolated preparations of airway
26      smooth muscle (trachealis). Chronic exposure to diesel exhaust significantly increased the
27      maximal contractile responses to acetylcholine compared with control values; exposure did not
28      alter the sensitivity (EC50 values) of the muscles to the agonists.  Exposures were to diesel
29      exhaust containing 2 mg/m3 particulate matter  for 7 h/day, 5 days/week for 2 years.
30             Biochemical studies of BALF obtained from hamsters and rats revealed that exposures to
31      diesel exhaust  caused significant increases in lactic dehydrogenase, alkaline phosphatase,
32      glucose-6-phosphate dehydrogenase (G6P-DH), total protein, collagen, and protease (pH 5.1)
33      after approximately 1 year and 2 years of exposure (Heinrich et al., 1986a).  These responses
34      were generally much greater in rats than in-hamsters.  Exposures were to diesel exhaust
        2/1/98                                    5-42       DRAFT-DO NOT CITE OR QUOTE

-------
 1      containing 4.24 mg/m3 participate matter for 19 h/day, 5 days/week for 120 (hamsters) to 140  ~
 2      (rats) weeks.
 3             Protein, p-glucuronidase activity, and acid phosphatase activity were significantly
 4      elevated in BALF obtained from rats exposed to diesel exhaust containing 0,75 or 1.5 mg/m3
 5      participate matter for 12 mo (Strom, 1984). Exposure for 6 mo resulted in significant increases
 6      in acid phosphatase activity at 0.75 mg/m3 and  in protein, p-glucuronidase, and acid phosphatase
 7      activity at the 1.5 mg/m3 concentration.  Exposure at 0.25 mg/m3 particulate matter did not affect
 8      the three indices measured at either time period. The exposures were for 20 h/day, 5.5 days/week
 9      for 52 weeks.
10             Additional biochemical studies (Misiorowski et al.,  1980) were conducted on laboratory
11      animals exposed under the same conditions and at the same site as reported on by Strom (1984).
12      In most cases, exposures at 0.25 mg/m3 did not cause any significant changes.  The DNA content
13      in lung tissue and the rate  of collagen synthesis were significantly increased at 1.5 mg/m3
14      particulate matter after 6 mo. Collagen deposition was not affected.  Total lung collagen content
15      increased in proportion to  the increase in lung weight. The  activity of prolyl hydroxylase was
16      significantly increased at 12 weeks at 0.25 and  1.5 mg/m3; it then decreased with age. Lysal
17      oxidase activity did not change. After 9 mo of exposure, there were significant increases in lung
18      phospholipids in rats and guinea pigs exposed to 0.75 mg/m3 and in lung cholesterol in rats and
19      guinea pigs exposed to 1.5 mg/m3.  Pulmonary  prostaglandin dehydrogenase activity was
20      stimulated by an exposure at 0,25 mg/m3 but was not affected by exposure at 1.5 mg/m3
21      (Chaudhari et al., 1980,1981). Exposures for 12  or 24 weeks resulted in a concentration-
22      dependent lowering of this enzyme activity.  Exposure of male rats and guinea pigs at 0.75
23      mg/m3 for 12 weeks did not cause any changes  in glutathione levels of the lung, heart, or liver.
24      Rats exposed for 2 mo at 6 mg/m3 showed a significant depletion of hepatic glutathione, whereas
25      the lung showed an increase of glutathione (Chaudhari and Dutta, 1982). Schneider and Felt
26      (1981) reported that similar exposures did not substantially  change adenylate cyclase and
27      guanylate cyclase activities in lung or liver tissue of exposed rats and guinea pigs.
28             Bhatnagar et al. (1980; see also Pepelko, 1982a) evaluated changes in the biochemistry  of
29      lung connective tissue of diesel-exposed rats and mice. The mice were exposed for 8 h/day and 7
30      days/week for up to 9 mo to exhaust containing 6 mg/m3 particulate matter. Total lung protein
31      content was measured as was labeled proline and labeled leucine. Le.ucine incorporation is an
32      index of total protein synthesis, although collagen is very low in leucine. Proline incorporation
33      reflects collagen synthesis. Amino acid incorporation was measured in vivo in the rat and in
34      short-term organ culture in mice. Both rats and mice showed a large increase in total protein (41
35      to 47% in rats), while leucine incorporation declined and proline incorporation was unchanged.

        2/1/98                                    5-43       DRAFT-DO NOT CITE OR QUOTE

-------
. 1      These data are consistent with an overall depression of protein synthesis in diesel-exposed
 2      animals and also with a relative increase in collagen synthesis compared to other proteins. The
 3      increase in collagen synthesis suggested proliferation of connective tissue and possible fibrosis
 4      (Pepelko, 1982a).
 5             A number of reports (McClellan et al., 1986; Mauderly et al., 1987a, 1990a; Henderson et
 6      al., 1988) have addressed biochemical and cytological changes in lung tissue and BALF of
 7      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
 8      mg/m3 paniculate matter.  At the lowest exposure level (0.35 mg/m3), no biochemical or
 9      cytological changes occurred in the BALF or in lung tissue in either Fischer 344 rats  or CD-I
10      mice. Henderson et al. (1988) provide considerable time-course information on inflammatory
11      events taking place throughout a chronic exposure.  A chronic inflammatory response was seen at
12      the two higher exposure levels in both species, as evidenced by increases in inflammatory cells
13      (macrophages and neutrophils), cytoplasmic and lysosomal enzymes (lactate dehydrogenase,
14      glutathione reductase, and p-glucuronidase), and protein (hydroxyproline) in BALF.  Analysis of
15      lung tissue indicated similar changes in enzyme levels as well as an increase in  total lung
16      collagen content. After 18 mo of exposure, lung tissue glutathione was depleted in a
17      concentration-dependent fashion in rats but was slightly increased in mice. Lavage fluid levels
18      of glutathione and glutathione reductase activity increased in a concentration-dependent manner
19      and were higher in mice than in rats.  Rats exposed for 24 mo to diesel exhaust  (3.5 mg/m3
20      particulate 'matter) had a fivefold increase in the bronchoconstrictive prostaglandin PGF2a and a
21      twofold increase in the inflammatory leukotriene LTB4. In similarly exposed mice, there was a
22      twofold increase in both parameters.  These investigators concluded that the release of larger
23      amounts of such mediators of inflammation from the alveolar phagocytic cells of rats accounted
24      for the greater fibrogenic response seen in that species.
25             Biochemical analysis of lung tissue from cats exposed for 124 weeks and held in clean air
26      for an additional 26 weeks indicated increases of lung collagen; this finding was confirmed by an
27      observed increase in total lung wet weight and in connective tissue fibers estimated
28      morphometrically in these cats (Hyde et al., 1985).  Exposures were for 7 h/day, 5 days/week at 6
29      mg/m3 particulate matter for 61 weeks and at 12 mg/m3 for weeks 62 to 124.
30             Heinrich et al. (1995) reported on bronchoalveolar lavage in animals exposed for 24 mo
31      and found exposure-related increases in lactate dehydrogenase, P-glucuronidase, protein, and
32      hydroxyproline in groups exposed to 2.5 or 7 mg/m3, although detailed data are not presented.
33      Lavage analyses were not carried out in concurrent studies in mice.
34             Further effects of exposure to diesel exhaust on pulmonary cytology and lung
35      biochemistry may be found in Section 5.1.2.3.

        2/1/98                                    5-44       DRAFT-DO NOT CITE  OR QUOTE

-------
 1             The pathogenic sequence following the inhalation of diesel exhaust as determined
 2      histopathologically and biochemically begins with the phagocytosis of diesel particles by AMs.
 3      These activated macrophages release chemotactic factors that attract neutrophils and additional
 4      AMs. As the lung burden of diesel particles increases, there is an aggregation of particle-laden
 5      AMs in alveoli adjacent to terminal bronchioles, increases in the number of Type II cells lining
 6      particle-laden alveoli, and the presence of particles within alveolar and peribronchial interstitial
 7      tissues and associated lymph nodes. The neutrophils and macrophages release mediators of
 8      inflammation and oxygen radicals that deplete a biochemical defense mechanism of the lung
 9      (i.e., glutathione). As will be described later in more detail, other defense mechanisms are
10      affected, particularly the decreased viability of AMs, which leads to decreased phagocytic
11      activity and death of the macrophage.  The latter series of events may result in the presence of
12      pulmonary inflammatory, fibrotic, or emphysematous lesions. The data suggest that there may
13      be a threshold of exposure to diesel exhaust below which adverse structural and biochemical
14      effects may not occur in the lung; however, differences in the anatomy and pathological
15      responses of laboratory animals coupled with their lifespans compared with humans make a
16      determination of human levels of exposure to diesel exhaust without resultant pulmonary injury a
17      difficult and challenging endeavor.
is
19      5.1.2.3.4. Effects on pulmonary defense mechanisms.  The respiratory system has a number of
20      defense mechanisms that negate or compensate for the effects produced by the injurious
21      substances that repeatedly insult the upper respiratory tract, the tracheobroncbial airways, and the
22      alveoli. The effects of exposure on the pulmonary defense mechanisms of laboratory animals as
23      well as more details on exposure atmosphere are summarized in Table 5-7 and ranked by
24      cumulative exposure (C x T).
25             Several studies have been conducted investigating the. effect of inhaled diesel exhaust on
26      the deposition and fate of inert tracer particles or diesel particles themselves. Lung clearance of
27      deposited particles occurs in two distinct phases:  a rapid phase (hours to days) from the
28      tracheobronchial region via the mucociliary escalator and a much slower phase (weeks to mo)
29      from the nonciliated pulmonary region via primarily but not solely AMs. Battigelli et al. (1966)
30      reported impaired tracheal mucociliary clearance in vitro in excised trachea from rats exposed for
31      single or repeated exposures of 4 to 6 hours at two dilutions of diesel exhaust that resulted in
32      exposures of approximately 8 and 17 mg/m3 diesel particles. The exposure to 17 mg/m3 resulted
33      in decreased clearance after a single exposure as well as  after a cumulative exposure of 34 or 100
34      hours.  Clearance was reduced to a lesser extent and in fewer tracheas from animals exposed to 8
35      mg/m3 for a cumulative exposure of 40 hours.  Lewis etal. (1989) found  no difference in the

        2/1/98                                    5-45       DRAFT-DO  NOT CITE OR QUOTE

-------
oo
            Table 5-7.  Effects of exposure to diesel exhaust on the pulmonary
            defense mechanisms of laboratory animals
•t.
o\
Species/sex
Exposure
period
Particles
(mg/ra3)
CxT
(mg-h/iii1)
CO
(Ppm)
NO,
(ppm)
SOZ
(ppm)
Effects
Study
                                                        ALVEOLAR MACROPHAGE STATUS

Guinea Pig,        20 h/day            0.25                  220        2.9        —        —      No significant changes in absolute       Chen et. al.
Hartley           5.5                 1.5                 1,320        7.5        —        —      numbers of AMs                     (1980)
                 days/week      0.19 urn MOD
                 8 weeks

Rat, F344, M       7 h/day              2.0                 7,280        11.5        1.5        0.81     Little effect on viability, cell number,    Castranova et al.
                 5 days/week     0.23-0.36 |im         -                                .             oxygen consumption, membrane        (1985)
                 104 weeks         MOD   '                                                      integrity, lyzomal enzyme activity, or
                                                                                                 protein content of AMs; decreased ceil
                                                                                                 volume and ruffling of cell membrane
                                                                                                 and depressed luminescence of AM

Rat, F344, M       20 h/day            0.25'           715-8,580   .2.9        —        . —      AM eel! counts proportional to         Strom (1984)
                 5.5                0.75*                            4.8        —        —      concentration of DPM at 0.75 and       Vostal et al.
                 days/week           1.5b                             7.5        —        —      1.5 mg/m3; AM increased in lungs in     (1982)
                 26,48, or       0,19umMDD                                                      response to rate of DPM mass entering
                 52 weeks                                                                        lung rather than total DPM burden in
                                                                                   ,              lung; increased PMNs were propor-
                                                                                                 tional to inhaled concentrations and/or
                                                                                                 duration of exposure; PMNs affiliated
                                                                                                 with clusters  of aggregated AM rather
                                                                                                 than DPM
Rat F344/Crl,
M,F
Mouse, CD, M, F






Rat, Wistar, F


Rat, F344/Crl, M


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
0.35
3.5
7.0
0.25 urn MOD





0.8
2.5
7.1
3.49


1,274"
12,740"
25,480'=






7,400
21,800
61,700
12,704


2.9
16.5
29.7






2.6
8.3
21.2
9.8


0.05
0.34
0.68






0,3
. 1.1
3.4
1.2


— Significant increases of AM in rats and
— mice exposed to 7.0 mg/m3 DPM for
— 24 and 18 mo, respectively, bufnot at
concentrations of 3.5 or 0.35 mg/m'
DPM for the same exposure durations;
PMNs increased in a dose-dependent
fashion in both rats and mice exposed
to 3,5 or 7.0 mg/m5 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

Henderson et al.
(1988)







Heinrich et al.
(1995)

Mauderly et al.
(1990a)


-------
OO
                    Table 5-7.  Effects of exposure to diesel exhaust on the pulmonary
                    defense mechanisms of laboratory animals (continued)
           Species/sex
 Exposure
  period
  Particles
  (mg/m3)
(mg-h/ra3)
 CO
(ppm)
 NO2
(ppm)
 S02
(ppm)
                                                                                                                Effects
Study
         Rat, M, F
                                                                       CLEARANCE
7h/day
5 days/week
12 weeks
   ,0.2
    1.0
    4.5
0.25um MOD
      84
     420
    1,890
                            Evidence of apparent speeding of
                            tracheal clearance at the 4.5 mg/m3
                            level after 1 week of "Tc
                            macroaggregated-albumin and reduced
                            clearance of tracer aerosol in each of
                            the three exposure levels at 12 weeks;
                            indication of a lower percentage of
                            ciliated cells at the 1.0 and 4.5 mg/m3
                            levels
                                                   Wolff and Gray
                                                   (1980)
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


18h/day
5 days/week
24 mo
7h/day
5 days/week
6 mo

7 h/day
5 days/week
18 weeks


7 h/day
5 days/week
26-104
weeks



4-6 h/day
7 days/week
0.1 to 14.3
weeks
0.8
2.5
7.1
3.55



0.15
0.94
4.1
<0.5 urn MOD

2.0
0.23-0.36 urn
MOD




0.9
8.0
17.0

7,400 2.6 0.3
21,800 8.3 1.2
61,700 21.2 3.8
3,321 7.9 9.5



94.5 — —
592 — —
2,583 — —


1,820-7,280 11.5 1.5






2.5-10,210 — 5.0
— 2.7
— 8.0

0.3 Significant increase in clearance half-
1.1 time of inhaled labeled aerosols in all
3.4 groups at 3-18 mo.
Clearance of 2 um, aluminosilicate
particles. Half-time significantly
increased in adult, not different in
developing rats
— Lung burdens of DPM were
— concentration-related; clearance half-
— time of DPM almost double in
4.1 mg/m3 group compared to 0.15
mg/m3 group
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
0.2 Impairment of tracheal mucociliary
0.6 clearance in a concentration-response
1 .0 manner

Heinrich et al.
(1995)

Mauderly et al.
(1987b)


Griffts et al.
(1983)



Lewis et al.
(1989)





Battigelli et al.
(1966)



-------
            Table 5-7.  Effects of exposure to diesel exhaust on the pulmonary
            defense mechanisms of laboratory animals (continued)
t/1

oo
Exposure
Species/sex period
Rat, F344, 7 h/day
M, f 5 days/week
130 weeks







Rat, F344/Crl, M 7 h/day
5 days/week
24 mo


Particles
(mg/m3)
0.35
3.5
7.1
0.25 urn MOD






3.49




CxT CO
(mg-h/ni1) (ppm)
1,593 . 2.9
15,925 16.5
31,850 29.7

,





12,704 9.8




NO2 SO,
(ppm) (ppm) Effects
0.1 — No changes in trachea! mucociliary
0.3 — clearance after 6, 12, 18, 24, or 30 mo
0.7 — of exposure; 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
1.2 — Doubling of long-term clearance half-
time for clearance of 1 ,0 urn alumino-
silicate particles. Less effect on
clearance in animals with
experimentally induced emphysema
Study
Wolff etai.
(1987)








Mauderly et a).
(1990a)



                                                         MICROBIAL-INDUCED MORTALITY
Mice, CD-1, F
Mice CD-I, F
Mice, CR/GD-1,
F
                            7 h/day
                            5 days/week
                            4, 12, or •
                            26 weeks
                            8 h/day
                            7 days/week
                            2 h up to
                            46 weeks
    2.0
0;23-0.36 urn
   MOD
280-1,820         11.5        1.5        0.1
                                   5.3 to 7.9
                    11-20,350
                  19
                  to
                  22
1.8
to
3.6
0.9
to
2.8
 No change in mortality in mice          Hatch et al.
 exposed intratracheally to 100 fig of     (1985)
 DPM prior to exposure to aerosolized
 Streptococcus sp.

 Mortality similar at each exposure       Hahon et al.
 duration when challenged with          (1985)
 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

 Enhanced susceptibility to lethal         Campbell et ai.
 effects of S. pyogenes infections at all     (1980, 1981)
 exposure 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

-------
to .
                  Table 5-7.  Effects of exposure to diesel exhaust on the pulmonary
•*>                 defense mechanisms of laboratory animals (continued)


   "Chronic exposure lasted 52 weeks.
   'Chronic exposure lasted 48 weeks.
   "Calculated for 104-week exposure.
   DPM = Diesel paniculate matter.
   AM = Alveolar macrophage.
   PMN = Polymorphonuclear leukocyte.

-------
 1      clearance of 59Fe3O4 particles (1.5 um MMAD, og 1.8) 1 day after dosing control and diesel
 2      exhaust-exposed rats (2 mg/m3, 7 h/day, 5 days/week for 8 weeks).
 3             Wolff et al. (1987) and Wolff and Gray (1980) studied the effects of both subchronic and
 4      chronic diesel exhaust exposure on the tracheal clearance of particles. Tracheal clearance
 5      assessments were made by measuring the retention of radiolabeled technetium macroaggregated-
 6      albumin remaining 1 h after instillation in the distal trachea of rats.  In the subchronic studies,
 7      rats were exposed to 4.5, 1.0, or 0.2 mg/m3 particulate matter on a 7 h/day, 5 days/week schedule
 8      for up to  12 weeks. After 1 week there was an apparent speeding of tracheal clearance at the 4.5
 9      mg/m3 exposure level (p=0.10), which returned toward baseline after 6 weeks and was slightly
10      below the baseline rate at 12 weeks. In the 1.0 mg/m3 group, there was a progressive significant
11      reduction in the clearance rate at 6 and 12 weeks of exposure.  There was a trend toward reduced
12      clearance in the 0.2 mg/m3 group. Scanning electron micrographs indicated minimal changes in
13      ciliary morphology; however, there was an indication of a lower percentage of ciliated cells at
14      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
15      mg/m3 for 7 h/day, 5 days/week for 30 mo. There were no significant differences in tracheal
16      clearance rates between the control group and  any of the exposure groups after 6, 12, 18, 24, or
17      30 mo of exposure. The preexposure measurements for all groups, however, were significantly
18      lower than those during the exposure period, suggesting a possible age effect.  The preexposure
19      value for the 3.5-mg/m3 group was also significantly lower than the  control group.
20             There is a substantial body of evidence for an impairment of particle clearance from the
21      bronchioloalveolar region of rats following exposure to diesel exhaust.  Griffis et al. (1983)
22      exposed rats 7 h/day, 5 days/week for 18 weeks to diesel exhaust at  0.15, 0.94, or 4.1 mg/m3
23      particulate matter.  Lung burdens of the 0.15, 0.94, and 4.1 mg/m3 levels were 35, 220, and 1,890
24      ug/g lung, respectively, 1 day after the 18-week exposure.  The clearance half-tune  of the DPM
25      was significantly greater, almost double, for the 4.1 mg/m3 exposure group than for those of the
26      lower exposure groups, 165 ± 8 days versus 99 ± 8 days (0.94 mg/m3) and 87 ± 28 days (0.15
27      mg/m3), respectively.
28             Chan et al. (1981) showed a dose-related slowing of 14C-diesel particle clearance in rats
29      preexposed to diesel exhaust at 0.25 or 6 mg/m3 particulate matter for 20 h/day, 7 days/week for
30      7 to 112 days.  Clearance was inhibited in the  6 mg/m3 group when compared by length of
31      exposure or compared with the 0.25 mg/m3 or control rats at the  same time periods.
32             Heinrich et al. (1982) evaluated lung clearance in rats exposed for approximately 18 mo
33      at 3.9 mg/m3 DPM for 7 to 8 h/day, 5 days/week. Following exposure to 59Fe2O3-aerosol, the rats
34     ' were returned to the diesel exhaust exposure and the radioactivity was measured over the
        2/1/98               '                    5-50       DRAFT-DO NOT CITE OR QUOTE

-------
  1      thoracic area at subsequent times. The biological half-life of the iron oxide deposited in the rats'
  2      lungs was nearly twice that of controls.
  3             Heinrich also used labeled iron oxide aerosols to study clearance in rats exposed to 0.8,
  4      2,5, or 7 mg/m3 diesel DPM for 24 mo (Heinrich et al., 1995).  Clearance measurements were
  5      carried out at 3,12, and 18 mo of exposure.  Half-times of clearance were increased in a
  6      concentration- and duration-related way in all exposed groups, with a range of a 50% increase in
  7      the 0.8  mg/m3 group at 3 mo to an 11-fold increase in the 7 mg/m3 group at 19 mo. The
  8      differential cell counts in these animals were stated to have shown clear effects in the 2,5 and 7
  9      mg/m3 groups, but specific information about the changes is not reported.
10             Wolff et al. (1987) investigated alterations in DPM clearance from the lungs of rats
11      chronically exposed to diesel exhaust at 0,0.35, 3.5, or 7.0 mg/m3 DPM for 7 h/day, 5 days/week
12      for up to 24 mo. Progressive increases in lung burdens were observed over time in the 3.5 and
13      7.0 mg/m3 exposure groups.  Levels of DPM in terms of milligrams per lung  were 0.60,11.5, and
14      20,5 after 24 mo of exposure at the 0.35,3.5, or 7.0 mg/m3 exposure levels, respectively. There
15      were significant increases in 16-day clearance half-times of inhaled radiolabeled particles of
16      67Ga2O3 (0.1 um MMD) as early as 6 mo at the 7.0 mg/m3 level and 18 mo at the 3.5 mg/m3
17      level; no significant changes were seen at the 0.35 mg/m3 level. Rats inhaled fused
18      aluminosilicate particles (2 um MMAD) labeled with 134Cs after 24 mo of diesel exhaust
19      exposure; long-term clearance half-times were 79, 81,264, and 240 days for the 0, 0.35, 3.5, and.
20      7.0 mg/m3 groups, respectively. Differences were significant between the control and the 3.5 and
21      7.0 mg/m3 groups (p< 0.01).
22             Mauderly et al. (1987b) compared the effects of diesel exhaust in the developing lung to
23      the adult lung by exposing groups of male F344 rats to 3.5 mg/m3 for 7 h/day, 5 days/week for 6
24      mo. One group (adult) was exposed between 6 and 12 mo of age, and the other was exposed
25      beginning in utero and until 6 mo of age. Clearance of an inhaled monodisperse 2 um
26      aluminosilicate particle was measured after exposure for 6 mo. The clearance half-time of the
27      slow phase was found to be doubled in adult rats compared with age-matched controls and was
28      not significantly affected in developing rat lungs.
29             Mauderly et al. compared the effects  of diesel exhaust in normal lungs with rats in which
30      emphysema had been induced experimentally by instillation of elastase 6 weeks before diesel
31      exhaust exposures.  The rats were exposed to 3.5 mg/m3 DPM for 7 h/day, 5 days/week for 24
32      mo. Measurements included histopathology, clearance, pulmonary function,  lung lavage, and
33      immune response.  In the rats that were not pretreated with elastase, there was a significant
34     • reduction in the number of macrophages recovered by pulmonary lavage in contrast to the
35      increases in macrophages reported by Strom (1984) and Henderson et al. (1988).  The half-time

        2/1/98                •                    5-51       DRAFT-DO NOT CITE OR QUOTE

-------
 1      of the slow phase of clearance of inhaled, 1 um, monodisperse particles was doubled in the
 2      exposure animals without elastase pretreatment.  The elastase pretreatment did not affect
 3      clearance in unexposed animals but significantly reduced the effect of diesel.  The clearance half-
 4      time was significantly less in elastase-pretreated, diesel-exposed animals than in diesel-exposed
 5      normal animals. Many other effects measured in this study were also less affected by diesel
 6      exposure in elastase-treated animals.  Measurements of lung burden of DPM showed that
 7      elastase-pretreated animals accumulated less than half as much DPM mass as normal animals
 8      exposed at the same time, suggesting that the difference in effect could be explained by
 9      differences in dose to the lung.
10             Lewis et al.  (1989) conducted lung burden and 59Fe3O4 tracer studies in rats exposed for
11      12 and  24 mo to 2 mg/m3 DPM (7 h/day, 5 days/week). The slope of the Fe3O4 clearance curve
12      was significantly steeper than that of the controls, indicating a more rapid alveolar clearance of
13      the deposited 59Fe3O4.  After 120 days from the inhalation of the tracer particle, 19% and 8% of
14      the initially deposited 59Fe3O4 were present in the lungs of control and diesel exhaust-exposed
15      rats, respectively. The lung burden of DPM, however, increased significantly between 12 and 24
16      mo of exposure (0.52 to 0.97% lung dry weight), indicating a later dose-dependent inhibition of
17      clearance.
18             Alveolar macrophages, because of their phagocytic and digestive capabilities, are one of
19      the prime defense mechanisms of the alveolar region of the lung against inhaled particles. Thus,
20      characterization of the effects of diesel exhaust on various properties of AMs provides
21      information on the integrity or compromise of a key pulmonary defense mechanism. The
22      physiological viability of AM from diesel-exposed rats was assessed after 2 years of exposure by
23      Castranova et al. (1985). The 7 h/day, 5 days/week exposure at 2 mg/m3 DPM had little effect on
24      the following:  viability, cell number, oxygen consumption, membrane integrity, lysosomal
25      enzyme activity, or protein content of the AM. A slight decrease in cell volume, a decrease in
26      chemiluminescence indicative of a decreased secretion of reactive oxygen species, and a decrease
27      in ruffling of the cell membrane were observed. These findings  could be reflective of an overall
28      reduction in phagocytic activity.
29             Exposure to diesel exhaust has been reported both to increase the number of recoverable
30      AMs from the lung (Strom, 1984; Vostal et al., 1982; Henderson et al., 1988)  or to produce no
31      change in numbers  (Chen et al., 1980; Castranova et al.,  1985).  Strom (1984) found that in rats
32      exposed to 0.25 mg/m3 DPM for 20 h/day, 5.5 days/week for 6 mb or 1 year, as well as in the
33      controls, BAL cells consisted entirely of AMs, with no differences in the cell counts in the lavage
34      fluid. At the higher concentrations, 0.75 or 1.5 mg/m3, the count of AM increased proportionally
35      with the exposure concentration; the results were identical for AMs at both 6 and 11 or 12 mo of

        2/1/98                                     5-52       DRAFT-DO NOT CITE OR QUOTE

-------
  1     exposure. The increase in AM counts was much larger after exposure to 1.5 mg/m3 DPM for 6
  2     mo than after exposure to 0.75 mg/m3 for 1 year, although the total mass (calculated as C * T) of
  3  .   deposited particulate burden was the same. These data suggested to the authors that the number
  4     of lavaged AM was proportional to the mass influx of particles rather than to the actual DPM
  5     burden in the lung. These results further implied that there may be a threshold for the rate of
  6     mass influx of DPM into the lungs of rats above which there was an increased recruitment of
  7     AMs. Henderson et al. (1988) reported similar findings of significant increases of AMs in rats
  8     and mice exposed to 7.1 mg/m3 DPM for 18 and 24 mo, respectively, for 7 h/day, 5 days/week,
  9     but not at concentrations of 3.5 or 0.35 mg/m3 for the same exposure durations. Chen et al.
 10     (1980), using an exposure regimen of 0.25 and 1.5 mg/m3 DPM for 2 mo and 20 h/day and 5.5
 11     days/week, found no significant changes in absolute numbers of AMs from guinea pig BALF nor
 12     did Castranova et al. (1985) in rat BALF following exposure to 2 mg/m3 DPM for 7 h/day, 5
 13     days/week for 2 years.
 14            A similar inflammatory response was noted by Henderson etal, (1988) and Strom (1984),
 15     as evidenced by an increased number of PMNs present in BALF from rodents exposed to diesel
 16     exhaust.  Henderson et al. (1988) found these changes in rats and mice exposed to 7.1  and 3.5
 17     mg/m3 DPM for 7 h/day, 5 days/week.  Significant increases in BALF PMNs were observed  in
 18     mice at 6 mo of exposure and thereafter at the 7.1 and 3.5 mg/m3 exposure levels, but in rats only
 19     the 7.1 mg/m3 exposure level showed an increase in BALF PMNs at 6 mo of exposure and
 20     thereafter. Significant increases in BALF PMNs occurred in rats at 12,18, and 24 mo of
 21     exposure to 3.5 mg/m3 DPM. Although increases in PMNs were usually greater in mice in terms
. 22     of absolute numbers, the PMN response in terms of increase relative to controls was only about
 23     one-third that of rats.  Strom (1984) reported that the increased numbers of PMNs in BALF were
 24     proportional to the inhaled concentrations and/or duration of exposure.  The PMNs also appeared
 25     to be affiliated with clusters of aggregated AMs rather than to the diesel particles per se.
 26     Proliferation of Type II cells likewise occurred in response to the formed aggregates of AMs
 27     (White and Garg, 1981).
 28            The integrity of pulmonary defense mechanisms can also be ascertained by assessing  if
 29     exposure to diesel exhaust affects the colonization and clearance of pathogens and alters the
 30     challenged animals' response to respiratory tract infections. Campbell et al. (1980,1981)
 31     exposed mice to diesel exhaust followed by infectious challenge with Salmonella typhimurium,
 32     Streptococcuspyogenes, or A/PR8-3 influenza virus and measured microbial-induced mortality.
 33     Exposures to the diesel exhaust were to 6 mg/m3 DPM for 8 h/day, 7 days/week for up to 321
 34     days. Exposure to the diesel exhaust resulted in enhanced susceptibility to the lethal effects of S.
 35     pyogenes infection at all exposure durations (2 h, 6 h; 8,15,16,  307, and 321 days). Tests with

        2/1/98                                   5-53       DRAFT-DO NOT CITE OR QUOTE

-------
 1      S. typhimurium were inconclusive because of the high mortality rates in the controls. The mice
 2      exposed to diesel exhaust did not exhibit an enhanced mortality when challenged with the
 3      influenza virus.  Hatch et al. (1985) found no changes in the susceptibility of mice to Group C
 4      Streptococcus sp. infection following intratracheal injection of 100 ug of DPM suspended in
 5      unbuffered saline.
 6            Hahon et al. (1985) assessed virus-induced mortality, virus multiplication with
 7      concomitant interferon (IFN) levels (lungs and sera), antibody response, and lung histopathology
 8      in mice exposed to diesel exhaust prior to infectious challenge with Ao./PR/8/34 influenza virus.
 9      Weanling mice were exposed to the diesel exhaust containing 2 mg/m3 DPM for 7 h/day, 5
10      days/week.  In mice exposed for 1, 3, and 6 mo, mortality was similar between the exposed and
11      control mice.  In mice exposed for 3 and 6 mo, however, there were significant increases in the
12      percentage of mice having lung consolidation, higher virus growth, depressed interferon levels,
13      and a fourfold reduction in hemagglutinin antibody levels; these effects were not seen after the
14      1-mo exposure.
15            The effects of diesel exhaust on the pulmonary defense mechanisms are determined by
16      three critical factors related to exposure: the concentrations of the pollutants, the exposure
17      duration, and the exposure pattern. Higher doses of diesel exhaust as determined by an increase
18      in one or more of these three variables have been reported to increase the numbers of AMs,
19      PMNs, and Type II cells in the lung, whereas lower doses fail to produce such changes.  The
20      single most significant contributor to the impairment of the pulmonary defense mechanisms
21      appears to be an excessive accumulation of DPM, particularly as particle-laden aggregates of
22      AMs.  Such an accumulation would result from an increase in deposition and/or a reduction in
23      clearance. The deposition of particles does not appear to change significantly following
24      exposure to equivalent diesel exhaust doses over time. Because of the significant nonlinearity in
25      particle accumulation between low and high doses of diesel exhaust exposure, coupled with no
26      evidence of increased particle deposition, an impairment in one or more of the mechanisms of
27      pulmonary defense appears to be responsible for the DPM accumulation and subsequent
28      pathological sequelae.  The time of onset of pulmonary clearance impairment was dependent
29      both on the magnitude and on the duration of exposures. For example, for rats exposed for
30      7 h/day, 5 days/week for 104 weeks, the concentration needed to induce pulmonary clearance
31      impairment appears to  lie between 0.35  and 2.0 mg/m3 DPM.
32
33      5.1.2.3.5. Effects on the immune system. The effects of diesel exhaust on the immune  system
34      of guinea pigs were investigated by Dziedzic (1981). Exposures were to 1.5 mg/m3 DPM for 20
35      h/day, 5.5 days/week for up to 8 weeks. There was no effect of diesel exposure when compared

        2/1/98                                    5-54       DRAFT-DO NOT CITE OR QUOTE

-------
  1      with matched controls for the number of B and T lymphocytes and null cells isolated from the
  2      tracheobronchial lymph nodes, spleen, and blood. Cell viability as measured by trypan blue
  3      exclusion was comparable between the exposed and control groups. The results of this study and
  4      others on the effects of exposure to diesel exhaust on the immune system are summarized in
  5      Table 5-8,
  6            Mentnech et al. (1984) examined the effect of diesel exhaust on the immune system of
  7      rats. Exposures were to 2 mg/m3 DPM for 7 h/day, 5 days/week for up to 2 years. Rats exposed
  8      for 12 and 24 mo were tested for immunocompetency by determining antibody-producing cells
  9      in the spleen 4 days after  immunization with sheep erythroeytes. The proliferative response of
10      splenic T-lymphocytes to the mitogens concanavalin A and phytohemagglutinin was assessed in
11      rats exposed for 24 mo. There were no significant differences between the exposed and control
12      animals. Results obtained from these two assays indicate that neither humoral immunity
13      (assessed by enumerating antibody-producing cells) nor cellular immunity (assessed by the
14      lymphocyte blast transformation assay) were markedly affected by the exposures.
15            Bice et al. (1985) evaluated whether or not exposure to diesel exhaust would alter
16      antibody immune responses induced after lung immunization of rats and mice.  Exposures were
17      to 0.35, 3.5, or 7.1 mg/m3 DPM for 7 h/day, 5 days/week for 24 mo. Chamber controls and
18      exposed animals were immunized by intratracheal instillation of sheep red blood cells (SRBC)
19      after 6,12,18, or 24 mo of exposure. No suppression in the immune response occurred in either
20      species. After 12,18, and 24 mo of exposure, the total number of anti-SRBC IgM antibody
21      forming cells (AFCs) was elevated in rats, but not in mice, exposed to 3.5 or 7.1 mg/m3 DPM;
22      after 6 mo of exposure, only the 7.1 mg/m3 level was found to have caused this response in rats.
23      The number of AFC per 106 lymphoid cells in lung-associated lymph nodes and the levels of
24      specific IgM, IgG, or IgA in rat sera were not significantly altered. The investigators concluded
25      that the increased cellularity and the presence of DPM in the lung-associated lymph nodes had
26      only a minimal effect on the immune and antigen filtration function of these tissues, Takafuji et
27      al. (1987) evaluated the IgE antibody response of mice inoculated intranasally at intervals of 3
28      weeks with varying doses of a suspension of DPM in ovalbumin.
29             Antiovalbumin IgE antibody titers, assayed by passive cutaneous anaphylaxis, were
30      enhanced by doses as low as 1 ug of particles compared with immunization with ovalbumin
31      alone.
32            The inhalation of diesel exhaust appeared to have only minimal effects on the immune
33      status of rats and guinea pigs. Conversely, intranasally delivered doses as low as 1 ug of DPM
34      exerted an adjuvant activity for IgE antibody production in mice. Further studies of the effects of


        2/1/98                                   5-55        DRAFT-DO NOT CITE OR QUOTE

-------
                     Table 5-8. Effects of exposures to diesel exhaust on the immune system of laboratory animals
VO
OQ
Ln

ON
Species/sex
Mouse, BDF1,
F

Guinea Pig,
Hartley, M



Rat, F344, M




Rat, F344;
Mouse, CD-I







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
1 04 weeks






Particles
(tng/mj)
—


' 1.5
0.1 9 urn MOD



2.0
0.23-0.36 jim
MOD


0.35
3.5
•7.1
0,25 urn MOD





C x T CO 'NO, SO2
(mg-h/rn*) (ppm) (ppn) (PPm) Effects
— — — — Intranasally delivered doses of DPM as
low as 1 u.g exerted an adjuvant
activity for IgE antibody production
660 or 7,280 7.5 — — No alterations in numbers of B, T, and
null lymphocytes or cell viability
among lymphocytes isolated from
tracheobronchial lymph nodes, spleen,
or blood
3,640 or 7,280 1 1 .5 1 .5 0.8 Neither humoral immunity (assessed .
by enumerating antibody-producing
cells) nor cellular immunity (assessed
by the lymphocyte blast transformation
assay) were markedly affected
1,274 2.9 0.05 -T— Total number of anti-sheep red blood
12,740 16.5 0.34 — ceil IgM AFC in the lung-associated
25,480 29.7 0.68 — 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
Study
Takafuji et al.
(1987)

Dziedzic(1981)




Mentnech et al.
(1984)



Bice etal. (1985)








        DPM = Diesel paniculate matter.
        AFC = Antibody-forming cells.

-------
  1      diesel exhaust on the immune system are needed to clarify the impact of such variables as route
  2      of exposure, species, dose, and atopy,
  3
  4      5.1.2.3.6. Effects on the liver.  Meiss et al. (1981) examined alterations in the hepatic
  5      parenchyma of hamsters by using thin-section and freeze-fracture histological techniques.
  6      Exposures to diesel exhaust were for 7 to 8 h/day, 5 days/week, for 5 mo at about 4 or 11 mg/m3
  7      DPM. The livers of the hamsters exposed to both concentrations of diesel exhaust exhibited
  8      moderate dilatation of the sinusoids, with activation of the Kupffer cells and slight changes in the
  9      cell nuclei.  Fatty deposits were observed in the sinusoids, and small fat droplets were
10      occasionally observed in the peripheral hepatocytes.  Mitochondria often had a loss of eristae and
11      exhibited a pleomorphic character. Giant microbodies were seen in the hepatocytes, which were
12      moderately enlarged, and gap junctions between hepatocytes exhibited a wide range in structural
13      diversity. The results of this study and others on the effect of exposure of diesel exhaust on the
14      liver of laboratory animals are summarized in Table 5-9.
15            Green et al. (1983) and Plopper et al. (1983) reported no changes in liver weights of rats
16      exposed to 2 mg/m3 DPM for 7 h/day, 5 days/week for 52 weeks or of cats exposed to 6 to 12
17      mg/m3, 8 h/day, 7  days/week for 124 weeks.
18            The use of light and electron microscopy revealed that long-term inhalation of varying
19      high concentrations of diesel exhaust caused numerous alterations to the hepatic parenchyma of
20      guinea pigs. A less sensitive index of liver toxicity, increased liver weight, failed to denote an
21      effect of diesel exhaust on the liver of the rat and cat following long-term exposure to diesel
22      exhaust. These results are too limited to understand potential impacts on the liver.
23
24      5.1.2.3.7.  Blood and cardiovascular systems. Several studies have evaluated the effects of
25      diesel exhaust exposure on hematological and cardiovascular parameters of laboratory animals.
26      These studies are summarized in Table 5-10.  Standard hematological indices of toxicological
27      effects on red and  white blood cells failed to denote dramatic and consistent responses.
28    .  Erythrocyte (RBC) counts were reported as being unaffected in cats (Pepelko and Peirano, 1983),
29      rats and  monkeys (Lewis et al., 1989), guinea pigs and rats (Penney et al., 1981), and rats
30      (Karagianes et al., 1981); lowered in rats (Heinrich et al., 1982); and elevated in rats (Research
31      Committee for HERP Studies, 1988; Brightwell et al., 1986). Mean corpuscular volume was
32      significantly increased in monkeys, 69 versus 64 (Lewis et al., 1989), and hamsters (Heinrich et
33      al., 1982) and lowered in rats (Research Committee for HERP Studies, 1988). The only other
34      parameters of erythrocyte status and related events were lowered mean corpuscular hemoglobin
        2/1/98   .                                 5-57       DRAFT-DO NOT CITE OR QUOTE

-------
Is)
                            Table 5-9.  Effects of exposure to diesel exhaust on the liver of laboratory animals
Ocr
Cft


-------
oo
                  Table 5-10. Effects of exposure to diesel exhaust on the hematological and
                  cardiovascular systems of laboratory animals ,















1
M3



o
>
H
1
1
O
O
25 .
O -
H
G

M

50
r w
O
O
M

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









Exposure
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
5.5 •
days/week
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







Particles
(mg/m3)
2.
0.23-0.36' |im
MOD
2
0.23-0.36 urn
MOD

6.3'
6.8"

3.9
0.1 urn MOD

0.25
0.75
1.5
0,1 9 umMDD
8.3
0.71 urn MOD

0.11"
0.41"
1.08C
. 2.31'
3.72" •
0.1 umMDD

0.7
' 2.2
6.6







CXT
(mg-h/in3)
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







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







NO2
(ppm)
1.5


1.5



2.3" .
2.9

1.2


o.ii
0.27
0.49

4-6


0.08
0.26
0.70
1.41
3.00


—
—
—







SO2
(ppm) Effects
0.8 Increased MCV


0.8 Increase in banded neutrophils; no
effect on heart or pulmonary arteries


2.1 No effect on heart mass or ECG; small
1 .9 decrease in heart rate (IE only)

3,1 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


0.38 At higher concentrations, RBC, Hb,
1.06 Hct slightly elevated; MCV and mean
2.42 corpuscular hemoglobin and
4.70 concentration were lowered
4.57
*

— 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



Study
Lewis et al.
(1989)

Lewis et al.
(1989)
Vallyathan et al.
(1986)
Wiester et al.
(1980)

Heinrichet al.
(1982)

Penney et al.
(1981)


Karagianes et al.
(1981)

Research
Committee for
HERP Studies
(1988)



Brightwell et al.
(1986)









-------

Ui
6\
o
o
H
O
)«H
H
W
o
H
W
                   Table 5-10. Effects of exposure to diesel exhaust on the hematological and cardiovascular
                   systems of laboratory animals (continued)

Species/sex
Cat, Inbred, M


Exposure
period
8 h/day
7 days/week
124 weeks
Particles
(mg/m3)
6.0'
12,0r

CXT
(mg'h/mj)
41,664
83,328

CO
(ppm)
20.2
33,3

NO2
(ppm)
2.7
4.4

SOj
(ppra)
2.1
5.0


Effects
Increases in banded neutrophils;
significant at 12 mo, but not 24 mo


Study
Pepelko and
Peirano (1983)

         "Nonirradiated diesel exhaust.
         'Irradiated diesel exhaust.
         "Light-duty engine.
         •"Heavy-duty engine.
         °l to 61 weeks of exposure,
         '62 to 124 weeks of exposure.

         Key:  MCV = Mean corpuscular volume.

-------
  1      and mean corpuscular hemoglobin concentration in rats (Research Committee for HERP Studies,
  2      1988), a 3 to 5% increase in earboxyhemoglobin saturation in rats (Karagianes et al., 1981), and a
  3      suggestion of an increase in prothrombin time (Brightwell et al., 1986). The biological
  4      significance of these findings regarding adverse health effects is deemed to be inconsequential,
  5             Three investigators (Pepelko and Peirano, 1983; Lewis et al., 1989; Brightwell et al.,
  6      1986) reported an increase in the percentage of banded neutrophils in cats and rats.  This effect
  7      was not observed in monkeys (Lewis et al., 1989). The health implications of an increase in
  8      abnormal maturation of circulating neutrophils are uncertain but indicate a toxic response of
  9      leukocytes following exposures to diesel exhaust. Leukocyte counts were reported to be reduced
10      in hamsters (Heinrich et al., 1982); increased in rats (Brightwell et al., 1986); and unaffected in
11      cats, rats, and monkeys (Pepelko and Peirano, 1983; Research Committee for HERP Studies,
12      1988; Lewis et al., 1989). These inconsistent findings indicate that the leukocyte counts are more
13      indicative of the clinical status of the laboratory animals than any direct effect of exposure to
14      diesel exhaust.
15             An important consequence of particle retention in the lungs  of exposed subjects can be
16      the development of pulmonary hypertension and cor pulmonale. Such pathology usually arises
17      from pulmonary fibrosis  or emphysema obliterating the pulmonary vascular bed or by chronic
18      anoxia. No significant changes in heart mass were found in guinea pigs or rats exposed to diesel .
19      exhaust (Wiester et al., 1980; Penney et al,, 1981; Lewis et al., 1989). Rats exposed to diesel
20      exhaust showed a greater increase in the medial wall thickness of pulmonary arteries of differing
21      diameters and right ventricular wall thickness; these increases, however, did not  achieve
22      statistically significant levels (Vallyathan et al., 1986).  Brightwell et al. (1986) reported
23      increased heart/body weight and right ventricular/heart weight ratios and decreased left
24      ventricular contractility in rats exposed to 6.6 mg/m3 DPM for 16 h/day, 5 days/week for 104
25      weeks.      .
26
27      5.1.2.3.8. Serum chemistry.  A number of investigators have studied the effects of exposure to
28      diesel exhaust on serum biochemistry and no consistent effects have been found. Such studies
29      are summarized in Table  5-11.
30             The biological significance of changes in serum chemistry in female but not male rats
31      exposed at 2 mg/m3 DPM for 7 h/day, 5 days/week for 104 weeks (Lewis et al., 1989) is difficult
,32      to interpret. Not only were the effects noted in one sex (females) only, but the serum enzymes,
33      lactate dehydrogenase (LDH), serum glutamic-oxaloacetic transaminase (SGOT), and serum
34      glutamic-pyruvic transaminase (SGPT), were elevated in the control group, a circumstance
35      contrary to denoting organ damage in the exposed female rats. The elevations of liver-related

        2/1/98               '                    5-61       DRAFT-DO NOT CITE OR QUOTE

-------
 oo
to
o
1
o
H
n
H
o
g
H
M
                                Table 5-11.  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






Exposure
period
7h/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





Particles
(mg/m3)
2.0
0.23
0.36 urn MOD
3.9
0.1 urn MOD

0.11"
0.41s
1.08'
2.31"
3.72b
0. 19-0.28 urn
MOD
0.7
2.2
6.6





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





CO
(ppm)
11.5


18.5


1.23
2.12
3.96
7.10
12.9


_^_
—
32.0





NO2
(ppm)
1.5


1.2


0.08
0.26
3.96
7.10
3.00


—
—
—





SO2
(ppm)
0.8


3.1


0.38
1.06
2.42
4.70
4.57


—
—
—





Effects
Decreased phosphate, LDH, SGOT, and
SGPT; increased sodium in females but not
males
After 29 weeks, increases in SCOT, LDH,
alkaline phosphatase, gamma-glutamyl
transferase, and BUN
Lower cholinesterase activity in males in both
the light- and heavy-duty series and elevated
gamma globulin and electrolyte levels in males
and females in both series



Rats, 6.6 mg/m3, reduction in blood glucose,
blood proteins, triglycerides, and cholesterol;
increase in BUN, alkaline phosphate alamine,
and aspartate aminotransferases (SGPT and
SGOT); hamsters, 6.6 mg/m3, decrease in
potassium, LDH, aspartate aminotransferase;
increase in albumin and gamma-glutamyl
transferase
Study
Lewis etal. (1989)


Heinrich et al.
(1982)

Research
Committee for
HERP Studies
(1988)



Brightwell etal.
(1986)






            Cat inbred, M
                  8 h/day
                  7 days/week
                  124 weeks
 6.0C
12.0'
41,664      20.2     2.7       2.1     BUN unaltered; SGOT and SGPT unaffected;    Pepelko and
83,328      33.3     4.4       5.0     LHD increase after 1 year of exposure         Peirano (1983)
Tight-duty engine.
'Heavy-duty engine.
cl to 61 weeks of exposure.
d62 to 124 weeks of exposure.

Key:   LDH  = Lactate dehydrogenase.
      SGOT = Serum glutamic-oxaloacetic transaminase.
      BUN  = Blood urea nitrogen.
      SGPT = Serum glutamk-pyruvic transaminase

-------
  1      serum enzymes in the control versus the exposed female rats appear to be a random event among
  2      these aged subjects.  The incidence of age-related disease, such as mononuclear cell leukemia,
  3      can markedly affect such enzyme levels, seriously compromising the usefulness of a comparison
  4      to historical controls. The serum sodium values of 144 versus 148 mmol/L in control and
  5      exposed rats, respectively, although statistically different, would have no biological import.
  6             The increased serum enzyme activities, alkaline phosphatase, SGOT, SGPT, gamma-
  7      glutamyl transpeptidase, and decreased cholinesterase activity suggest an impaired liver;
  8      however, such an impairment was not established histopathologically (Heinrich et al, 1982;
  9      Research Committee for HERP Studies, 1988; Brightwell et al., 1986). The increased urea
10      nitrogen, electrolyte levels, and gamma globulin concentration and reduction in total blood
11      proteins are indicative of impaired kidney function.  Again there was no histopathological confir-
12      mation of impaired kidneys in these studies,
13             Clinical chemistry studies suggest impairment of both liver and kidney functions in rats
14      and hamsters chronically exposed to high concentrations of diesel exhaust.  The absence of
15      histopathological confirmation, the appearance of such effects near the end of the lifespan of the
16      laboratory animal, and the failure to find such biochemical  changes in cats exposed to a higher
17      dose, however, tend to discredit the probability of hepatic and renal hazards to humans exposed
18      at atmospheric levels of diesel exhaust.
19
20      5.1.2.3.9. Effects on microsomal enzymes.  Several studies have examined the effects of diesel
21      exhaust exposure on microsomal enzymes associated with the metabolism and possible
22      activation of xenobiotics, especially polynuclear aromatic hydrocarbons. These studies are
23      summarized in Table 5-12.  Lee et al. (1980) measured the  activities of aryl hydrocarbon
24      hydroxylase (AHH) and epoxide hydrase (EH) in liver, lung, testis, and prostate gland of adult
25      male rats exposed to 6.32 mg/m3 DPM 20 h/day for 42 days*  Maximal significant AHH activities
26      (pmol/min/mg microsomal protein) occurred at different times during the exposure period, and
27      differences between controls and exposed rats, respectively, were as follows: prostate
28      0.29 versus 1.31, lung 3.67 versus 5.11, and liver 113.9 versus 164.0.  There was no difference in
29      AHH activity in the testis between exposed and control rats. Epoxide hydrase activity was not
30      significantly different from control values for any of the organs tested.
31             Pepelko and Peirano  (1983) found no statistical differences in liver microsomal
32      cytochrome P448-450 levels and liver microsomal AHH between control and diesel-exposed
33      mice either at 6 and 8 mo of exposure. Small differences were noted in the lung microsomal
34      AHH activities, but these were believed to be artifactual differences, due to  increases in
        2/1/98                                    5-63       DRAFT-DO NOT CITE OR QUOTE

-------
00
                            Table 5-12.  Effects of chronic exposures to diesel exhaust
                            on microsomal enzymes of laboratory animals
t
. Species/sex
Exposure
period
Particles
(mg/m3)
CxJ
(mg-h/ni3)
CO
(ppm)
N02
(ppm)
SO2
(ppm)
Effects
Study
           Rat, F344, M
Intratracheal administration of DPM
extract required doses greater than 6
mg/mj before the lung AHH was barely
doubled; liver AHH activity was
unchanged
Chen (1986)
Mouse, CD-I, F






Rat, Sprague-
Dawley, M





Rat, F344, M









Rat, F344, F



Rat, F344, M






7h/day
5 days/week
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
12,26, or
104 weeks
20 h/day
5.5
days/week
8-53 weeks



2.0 280 11.5 1.5
0.2-0.36 nm
MOD




6.3 882-6,174 17.4 2.3






0.75 330-6,435 4.8 • —
1.5 7.5 —
0.19^mMDD

0.75 330-6,435 . 4.8 —
1.5 7.5 —
0.19 urn MOD



2.0 840-7,280 11.5 1.5
0.23-0.36 urn
MOD

0.25 220-8,745 2.9 —
1.5 7.5 —
0.1 9 urn MOD




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
2.1 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-
450 was unchanged in lungs and liver
following inhalation or ip administration

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/m] DPM impaired
lung microsomal metabolism of B[a]P
Rabovsky et al.
(1986)





Lee etal. (1980)






Chen and Vostal
(1981)








Rabovsky et al.
(1984)


Navarro et al.
(1981)






-------
oo
Os
o

5Z!
o
H

o
i_H
H
W


s

o
C!
O
H
W
                          Table 5-12. Effects of chronic exposures to diesel exhaust on microsomal enzymes

                          of laboratory animals (continued)
. Species/sex
Mouse, A/J, M

Exposure
period
8 days/week
7 days/week
26 or
35 weeks
Particles
(mg/m3)
6.0

CxT
(ittg-h/m3)
17.4

CO
(ppm)
17.4

NO2
(ppm)
2.3

SO2
(ppm)
2.1

Effects
No differences in lung and liver AHH
activities and liver P-448, P-450 levels

Study
Pepelko and
Peirano(1983)

         AHH = Aryl hydrocarbon hydroclase.

         B[a]P = Benzo[a]pyrene.

-------
 1      nonmicrosomal lung protein present in the microsomal preparations. Exposures to 6 mg/m3
 2      DPM were for 8 h/day, 7 days/week.
 3             Rabovsky et al. (1984) investigated the effect of chronic exposure to diesel exhaust on
 4      microsomal cytochrome P450-associated benzo[a]pyrene hydroxylase and 7-ethoxycoumarin
 5      deethylase activities in rat lung and liver. Male rats were exposed for 7 h/day, 5 days/week for
 6      104 weeks to 2 mg/m3 DPM. The exposure had no effect on B[a]P hydroxylase or 7-
 7      ethoxycoumarin deethylase activities in lung or liver. In related studies, Rabovsky et al. (1986)
 8      examined the effects of diesel exhaust on vitally induced enzyme activity and interferon
 9      production in female mice. The mice were exposed for 7 h/day, 5 days/week for 1 month to
10      diesel exhaust diluted to achieve a concentration of 2 mg/m3 DPM.  After the exposure, the mice
11      were inoculated  intranasally with influenza virus.  Changes in serum levels of interferon and liver
12      microsomal activities of 7-ethoxycoumarin, ethylmorphine demethylase, and nicotinamide
13      adenine dinucleotide phosphate (NADPH)-dependent cytochrome c reductase were measured.  In
14      the absence of viral inoculation, exposure to diesel exhaust had no significant effects on the
15      activity levels of the two liver microsomal monooxygenases and NADPH-dependent cytochrome
16      c reductase. Exposure to diesel exhaust produced smaller increases in ethylmorphine
17      demethylase activity on days 2 to 4 postvirus infection and also abolished the day 4 postinfection
18      increase in NADPH-dependent cytochrome c reductase when compared  with nonexposed mice.
19      These data suggested to the authors that the relationship that exists between metabolic
20      detoxification and resistance to infection in unexposed mice was altered during a short-term
21      exposure to diesel exhaust.
22             Chen and Vostal (1981) measured the activity of AHH and the content of cytochrome
23      P450 in the lungs and livers of rats exposed by inhalation or intraperitoneal  (i.p.) injection of a
24      dichloromethane extract of DPM.  In the inhalation exposures, the exhaust was diluted to achieve
25      concentrations of 0.75 or 1.5 mg/m3 DPM, and the exposure regimen was 20 h/day, 5.5
26      days/week for up to 9 mo. The concentration of total hydrocarbons and  particle-phase
27      hydrocarbons was not reported. Parenteral administration involved repeated i.p. injections at
28      several dose levels for 4 days.  Inhalation exposure had no significant effect on liver microsomal
29      AHH activity; however, lung AHH activity was slightly reduced after 6  mo  exposure to
30      1.5 mg/m3.  An i.p. dose of DPM extract, estimated to be equivalent to the inhalation exposure,
31      had no  effect on AHH activity in liver or lungs. No changes were observed in cytochrome P450
32      contents in lungs or liver following inhalation  exposure or i.p. treatment. Direct intratracheal
33      administration of a dichloromethane DPM extract required doses greater than 6 mg/kg body
34      weight before the activity of induced AHH in the lung was barely doubled; liver AHH activity
35      remained unchanged (Chen, 1986).

        2/1/98   .                                 5-66      DRAFT-DO NOT CITE OR QUOTE

-------
  1             In related studies, Navarre et al. (1981) evaluated the effect of exposure to diesel exhaust
  2      on rat hepatic and pulmonary microsomal enzyme activities. The same exposure regimen was
  3      employed (20 h/day, 5.5 days/week, for up to 1 year), and the exhaust was diluted to achieve
  4      concentrations of 0.25 and 1.5 mg/m3 DPM (a few studies were also conducted at 0.75 mg/m3).
  5      After 8 weeks of exposure, there was no evidence for the induction of cytochrome P450,
  6      cytochrome P448, or NADPH-dependent cytochrome c reductase in rat liver microsomes. One
  7      year of exposure had little, if any, effect on the hepatic metabolism of B[a]P.  However, 1 year of
  8      exposure to 0.25 and 1.5 mg/m3 significantly impaired the ability of lung microsomes to
  9      metabolize B[a]P (0.15 and 0.02 nmole/30 min/mg protein, respectively, versus 0.32 nmole/30
10      min/mg protein for the controls).
11             There are conflicting results regarding the induction of microsomal AHH activities in the
12      lungs and liver of rodents exposed to diesel exhaust. One study reported induced AHH activity
13      in the lungs, liver, and prostate of rats exposed to diesel exhaust containing 6.32 mg/m3 DPM for
14      20 h/day for 42 days; however, no induction of AHH was observed in the lungs of rats and mice
15      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
16      periods up to 2 years.  Exposure to diesel exhaust has not been shown to produce adverse effects
17      on microsomal cytochrome P450 in the lungs or liver of rats or mice. The weight of evidence
18      suggests that the absence of enzyme induction in the rodent lung exposed to diesel exhaust is
19      caused either by the unavailability of the adsorbed hydrocarbons or by their presence in
20      insufficient quantities for enzyme induction.
21
22      5.1.2.3.10. Effects on behavior and neurophysiology. Studies on the effects of exposure to
23      diesel exhaust on the behavior and neurophysiology of laboratory animals are summarized in
24      Table 5-13. Laurie et al. (1978) and Laurie et al. (1980) examined behavioral alterations in adult
25      and neonatal rats exposed to diesel exhaust. Exposure for 20 h/day, 7 days/week, for 6 weeks to
26      exhaust containing 6 mg/m3 DPM produced a significant reduction in adult spontaneous loco-
27      motor activity (SLA) and in neonatal pivoting (Laurie et al., 1978). In a follow-up study, Laurie
28      et al. (1980) found that shorter exposure (8 h/day) to 6 mg/m3 DPM also resulted in a reduction
29      of SLA in adult rats. Laurie et al. (1980) conducted additional behavioral tests on adult rats
30      exposed during their neonatal period. For two of three exposure situations (20 h/day for 17 days
31      postparturition, or 8 h/day for the first 28 or 42 days postparturition), significantly lower SLA
32      was observed in the majority of the tests conducted on the adults after week 5 of measurement.
33      When compared with control rats, adult 15-month-old rats that had been exposed as neonates (20
34      h/day for 17 days) also exhibited a significantly slower rate of acquisition of a bar-pressing task
35      to obtain food. The investigators noted that the evidence was insufficient to determine whether

        2/1/98   .                                 5-67       DRAFT-DO NOT CITE OR QUOTE

-------
                     Table 5-13. 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
period (mg/m3)
8 h/day 6
7 days/week
1-4 weeks
20 h/day 6
7 days week
6 weeks
8 or 20 h/day 6
7 days/week
3, 4, 6, or
16 weeks
C*T CO NO2 SO,
(mg'h/ni3) . (ppm) (ppm) (ppm) Effects
336-1,344 19 2.5 1.8 Somatosensory and visual evoked poten-
tials revealed longer pulse latencies in
pups exposed neonatally
5,040 19 2.5 1.8 Reduction in adult SLA and in neonatal
pivoting
1,008-13,440 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 etal.
(1980)
SLA = Spontaneous locomotor activity. • . -
OS
00



>M*




£
O
O

Z
O
H

O
»..(
H
W
O
H
W

-------
  1      the differences were the result of a learning deficit or due to some other cause (e.g., motivational
  2   ,   or arousal differences).
  3             These data are difficult to interpret in terms of health hazards to humans under ambient
  4      environmental conditions because of the high concentration of diesel exhaust to which the
  5      laboratory rats were exposed.  Additionally, there are no further concentration-response studies to
  6      assess at what exposure levels these observed results persist or abate.  A permanent alteration in
  7      both learning ability and activity resulting from exposures early in life is a health hazard whose
  8      significance to humans should be pursued further.
  9             Neurophysiological effects from exposure to diesel exhaust were investigated in rats by
10      Laurie and Boyes (1980, 1981).  Rats were exposed to diluted diesel exhaust containing 6 mg/m3
11      DPM for 8 h/day, 7 days/week from birth up until 28 days of age.  Somatosensory evoked
12      potential, as elicited by a 1 mA electrical pulse to the tibial nerve in the left hind limb, and visual
13      evoked potential, as elicited by a flash of light, were the end points tested. An increased pulse
14      latency was reported for the rats exposed to diesel exhaust, and this was thought to be caused by
15      a reduction in the degree of nerve myelinization. There was no neuropathological examination,
16      however, to confirm this supposition.                                                   .
17             Based on the data presented, it is not possible to specify the particular neurological
18      impairment(s) induced by the exposure to diesel exhaust.  Again, these results occurred following
19      exposure to a high level of diesel exhaust and no additional concentration-response studies were
20      performed.
21
22      5.1.2.3.11. Effects on reproduction and development.  Studies of the effects of exposure to
23      diesel exhaust on reproduction and development are summarized in Table 5-14.  Twenty rats
24      were exposed 8 h/day on days 6 through 15 of gestation to diluted diesel exhaust containing 6
25      mg/m3 DPM (Werchowski et al., 1980a,b; Pepelko and Peirano, 1983). There were no signs of
26      maternal toxicity or decreased fertility. No skeletal or visceral teratogenic effects were observed
27      in 20-day-old fetuses (Werchowski et al.,  1980a). In a second study, 42 rabbits were exposed to
28      6 mg/m3 DPM for 8 h/day, on gestation days 6 through 18. No adverse effects on body weight
29      gain or fertility were seen in the does exposed to diesel  exhaust. No visceral or skeletal
30      developmental abnormalities were observed in the fetuses (Werchowski et al., 1980b).
31             Pepelko and Peirano (1983) evaluated the potential for diesel exhaust to affect
32      reproductive performance in mice exposed from 100 days prior to exposure throughout maturity
33      of the F2 generation. The mice were exposed for 8 h/day, 7 days/week to 12 mg/m3 DPM. In
34      general, treatment-related effects were minimal.  Some  differences in organ and body weights
        2/1/98                 .                   5-69       DRAFT-DO NOT CITE OR QUOTE

-------
00'
Table 5-14. Effects of chronic exposures to diesel exhaust on reproduction and
development in laboratory animals


'
o
O
3 •
g '
1'
9
w
o
/o
c;
o
H
W
Exposure Particles
Species/sex period (mg/m3)
Mouse, [C57BL]/ 5 days 50, 100, or
6XC3H]F,,M 200mg/kg
in corn oil; i.p.
injection
Rat, Sprague- 8 h/day 6
Dawley, F 7 days/week
1.7 weeks
Rabbit, New 8 h/day 6
Zealand Albino, F 7 days/week
1.9 weeks
Monkey, 7 h/day . 2
Cynomolgus, M 5 days/week
104 weeks
Mouse, A/Strong, 8 h/day 6
M 7 days/week
31or
38 weeks
Mouse, CD-I, 8 h/day 12
M, F 7 days/week
6 to 28
weeks





C x T CO NO2 SOt
(mg'h/nf) (ppm) (ppm) (ppm) Effects
— — ' — — Dose-related increase in sperm
abnormalities; decrease in sperm number
at highest dose; testicular weights
unaffected
571 , 20 2.7 2.1 No signs of maternal toxicity or
decreased fertility; no skeletal or visceral
teratogenic effects in 20-day-old fetuses
638 20 2.7 2,1 No adverse effects on maternal weight
gain or fertility; no skeletal or visceral '
teratogenic effects in the fetuses
7,280 11.5 1.5 0.8 No effects on sperm motility, velocity,
density, morphology, or incidence of
abnormalities
10,416-12,768 20 2.7 2.1 No effect on sperm morphology; high
rate of spontaneous sperm abnormalities
may have masked small effects
4,032-18,816 33 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 anthracosts
•




Study
Quinto and De
Marinis (1984)
Werchowski et
a!. (1980a)
Pepelko and
Peirano (1983)
Werchowski et
al. (1980a)
Pepelko and
Peirano (1983)
Lewis et al,
(1989)
Pereiraetal.
(1981)
Pepelko and
Peirano (1983)






-------
  1      were noted, but overall fertility and survival rates were not altered by exposure to diesel exhaust,
  2      The only consistent change, an increase in lung weights, was accompanied by a gross
  3      pathological diagnosis of anthracosis.  These data denoted that exposure to diesel exhaust at a
  4      concentration of 12 mg/m3 did not affect reproduction.  See Section 5.3, which reports a lack of
  5      effects of exposure to diesel exhaust on rat lung development (Mauderly et al., 1987b).
  6            Several studies have evaluated the effect of exposure to diesel exhaust on sperm.  Lewis
  7      et al. (1989) found no adverse sperm effects (sperm motility, velocity, densities, morphology, or
  8      incidence of abnormal sperm) in monkeys exposed for 7 h/day, 5 days/(week, for 104 weeks to 2
  Q      mg/m3 DPM. In another study in which A/Strong mice were exposed to diesel exhaust
10      containing 6 mg/mj DPM for 8 h/day for 31  or 38 weeks, no significant differences were
11      observed in sperm morphology between exposed and control mice (Pereira et al., 1981).  It was
12      noted, however, that there was a high rate of spontaneous sperm abnormalities in this strain of
13      mice, and this may have masked any small positive effect. Quinto and De Marinis (1984)
14      reported a statistically significant and dose-related increase in sperm abnormalities in mice
15      injected intraperitoneally for 5 days with 50,100, or 200 mg/kg of DPM suspended in corn oil.
16            A significant decrease in sperm number was seen at the highest dose, but testicular weight
17      was unaffected by the treatment.
18            No teratogenic, embryotoxic, fetotoxic, or female reproductive effects were observed in
19.      mice, rats,  or rabbits at exposure levels up to 12 mg/m3 DPM. Effects on sperm morphology and
20      number were reported in hamsters and mice exposed to high doses of DPM; however, no adverse
21      effects were observed in sperm obtained from monkeys exposed at 2 mg/m3. for 7 h/day,
22      5 days/week for 104 weeks. Concentrations of 12 mg/m3 DPM did not affect male rat
23      reproductive fertility in the F0 and F, generation breeders.  Thus, exposure to diesel exhaust
24      would not appear to be a reproductive or developmental hazard.
25
26      5.2. COMPARISON OF HEALTH EFFECTS OF FILTERED AND UNFILTERED
27   -       DIESEL EXHAUST
28            In four chronic toxicity studies of diesel exhaust, the  experimental protocol included
29      exposing test animals to exhaust containing no particles. Comparisons were then made between
30      the effects  caused by whole, unfiltered exhaust and those caused by the gaseous components of
31      the exhaust. Concentrations of components of the exposure atmospheres in these four studies are
32      given in Table 5-15.
33            Heinrich et al. (1982) compared the toxic effects of whole and filtered diesel exhaust on
34      hamsters and rats.  Exposures were for 7 to 8 h/day and 5 days/week.  Rats exposed for 24 mo to
35      either whole or filtered exhaust exhibited no significant changes in respiratory frequency,

        2/1/98               •                   5-71        DRAFT-DO NOT CITE OR QUOTE

-------
oo
-u
                                        Table 5-15.  Composition of exposure atmospheres in studies
                                        comparing unfiltered and filtered diesel exhaust8
Species/sex
Rat Wistar, F;
Hamster, Syrian


Rat, F344, F




Rat, F344, M, F;
Hamster, Syrian,
M, F



Rat, Wistar, F;
Hamster, Syrian,
F;
Mouse NMRI, F





Mouse, NMRI, F,
C57BL/6N, F




Exposure1*
period
7h/day
5 days/week
104 weeks

8h/day
7 days/week
104 weeks


16h/day
5 days/week
104 weeks



19 h/day
5 days/week
120 to
140 weeks





18 h/day
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
FJ —
£ 	






UF 4.5
F 0.01
C 0.01



C * T CO NO2
(mg-h/m3) (ppra) (ppra)
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



SO,
(ppm) Effects
3.1 No effect on pulmonary function or heart rate in
2.8 rats; increases in 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 hemoglobin; 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
1.02 hamsters; increased 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 etal.
(1982)


Iwai etal. (1986)




Brightwell et al.
(1986)




Heinrich etal.
(1986a)







Heinrich etal.
(1995)




          'Mean 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.

-------
  1      respirator}' minute volume, compliance or resistance as measured by a whole-body
  2      plethysmography, or heart rate. In the hamsters, histological changes (adenomatous
  3      proliferations) were seen in the lungs of animals exposed to either whole or filtered exhaust;
  4      however, in all groups exposed to the whole exhaust the number of hamsters exhibiting such
  5      lesions was significantly higher than for the corresponding groups exposed to filtered exhaust or
  6      clean air.  Severity of the lesions was, however, not reported.
  7             In  a second study, Heinrich et al. (1986a, see also Stober, 1986) compared the toxic
  8      effects of whole and filtered diesel exhaust on hamsters, rats, and mice.  The test animals (96 per
  9      test group) were exposed for 19 h/day, 5 days/week for 120 (hamsters and mice) or 140 (rats)
10      weeks. Body weights of hamsters were unaffected by either exposure. Body weights of rats and
11      mice were reduced by the whole exhaust but not by the filtered exhaust. Exposure-related higher
12      mortality rates occurred in mice after 2 years of exposure to whole exhaust. After 1 year of
13      exposure to the whole exhaust, hamsters exhibited increased lung weights, a significant increase
14      in airway resistance, and a nonsignificant reduction in lung compliance. For the same time
15      period, rats exhibited increased lung weights, a significant decrease in dynamic lung compliance,
16      and a significant increase in airway resistance. Test animals exposed to filtered exhaust did not
17      exhibit such effects. Histopathological examination indicated that different levels of response
18      occurred in the three species. In hamsters, filtered exhaust caused no significant
19      histopathological effects in the lung; whole exhaust caused thickened alveolar septa,
20      bronchioloalveolar hyperplasia, and emphysematous lesions. In mice, whole exhaust, but not
21      filtered exhaust, caused multifocal bronchioloalveolar hyperplasia, multifocal alveolar
22      lipoproteinosis, and multifocal interstitial fibrosis. In rats, there were no significant
23      morphological changes in the lungs following exposure to filtered exhaust. In rats exposed to
24      whole exhaust, there were severe inflammatory changes in the lungs, thickened alveolar septa,
25      foci of macrophages, crystals of cholesterol, and hyperplastic and metaplastic lesions.
26      Biochemical studies of lung lavage fluids of hamsters and mice indicated that exposure to filtered
27      exhaust caused fewer changes than did exposure to whole exhaust. The latter produced
28      significant increases in lactate dehydrogenase, alkaline phosphatase, glucose-6-phosphate
29      dehydrogenase, total protein, protease (pH 5.1), and collagen.  The filtered exhaust had a slight
30      but nonsignificant effect on G6P-DH, total protein, and collagen.  Similarly, cytological studies
31      showed that while the filtered exhaust had no effect on differential cell counts, the whole exhaust
32      resulted in an increase in leukocytes (161 ± 43.3/uL versus 55.7 ± 12.8/uL in the controls), a
33      decrease in AM (30.0 ± 12,5 versus 51.3 ± 12.5/uL in the controls), and an increase in
34      granulocytes (125 ± 39.7 versus 1.23 ± 1.14/^L in the controls). All  values presented for this
35      study are the mean with its standard deviation. The differences were significant for each cell

        2/1/98                                     5-73       DRAFT-DO NOT CITE OR QUOTE

-------
 1      type. There was also a small increase in lymphocytes (5.81 ± 4.72 versus 3.01 ± 1.23/uL in the
 2      controls).
 3             Iwai et al. (1986) exposed rats (24 per group) to whole or filtered diesel exhaust 8 h/day,
 4      7 days/week for 24 mo. The whole exhaust was diluted to achieve a concentration of 4.9 ±1.6
 5      mg/m3  DPM. Body weights in the whole exhaust group began to decrease after 6 mo and in both
 6      exposed groups began to decrease after 18 mo, when compared with controls.  Lung-to-body
 7      weight ratios of the rats exposed to the whole exhaust showed a significant increase (p<0.01)
 8      after 12 mo in comparison with control values.  Spleen-to-body weight ratios of both exposed
 9      groups were higher than control values after 24 mo. After 6 mo of exposure to whole exhaust,
10      DPM accumulated in AM, and Type II cell hyperplasia was observed. After 2 years of exposure,
11      the alveolar walls had become fibrotic with mast cell infiltration and epithelial hyperplasia. In
12      rats exposed to filtered exhaust, after 2 years there were only minimal histologic changes in the  .
13      lungs, with slight hyperplasia and stratification of bronchiolar epithelium and infiltration of
14      atypical lymphocytic cells in the spleen.
15             Brightwell et al. (1986) evaluated the toxic effects of whole and filtered diesel exhaust on
16      rats and hamsters. Three exhaust dilutions were tested, producing concentrations of 0.7, 2.2, and
17      6.6 mg/m3 DPM. The test animals (144 rats and 312 hamsters per exposure group) were exposed
18      for five 16-h periods per week for 2 years.  The four exposure types were gasoline, gasoline
19      catalyst, diesel, and filtered diesel. The results presented were limited to statistically significant
20      differences between exhaust-exposed and control animals. The inference from the discussion
21      section of the paper was that there was a minimum of toxicity in the animals exposed to filtered
22      diesel exhaust:  "It is clear from the results presented that statistically significant differences
23      between exhaust-exposed and control animals are almost exclusively limited to animals exposed
24      to either gasoline or unfiltered diesel exhaust."  Additional results are described in Section 5.4.
25             Heinrich et al. (1995) exposed female NMRI and C57BL/6N mice to a diesel exhaust
26      dilution that resulted in a DPM concentration of 4.5 mg/m3 and to the same dilution after filtering
27      to remove the particles. This study is focused on the carcinogenic effects of diesel particle
28      exposure, and inadequate information was presented to compare noncancer effects in filtered
29      versus  unfiltered exhaust.
30   "          A comparison of the toxic responses in laboratory animals exposed to whole exhaust or
31      filtered exhaust containing no particles demonstrates across studies that when the exhaust is
32      sufficiently diluted to'limit the concentrations of gaseous irritants (NO2 and SO2), irritant vapors
33      (aldehydes), CO, or other systemic toxicants, the diesel particles are the prime etiologic agents of
34      noncancer health effects, although additivity or synergism with the gases cannot be ruled out.
35      These toxic responses are both functional and pathological and represent cascading sequelae of
36      lung pathology based on concentration and species. The diesel particles plus gas exposures
        2/1/98                                     5-74        DRAFT-DO NOT CITE OR QUOTE

-------
  1      produced biochemical and cytological changes in the lung that are much more prominent than
  2      those evoked by the gas phase alone. Such marked differences between whole and filtered diesel
  3      exhaust are also evident from general toxicological indices, such as decreases in body weight and
  4      increases in lung weights, pulmonary function measurements, and pulmonary histopathology
  5      (e.g., proliferative changes in Type II cells and respiratory bronchiolar epithelium, fibrosis).
  6      Hamsters, under equivalent exposure regimens, have lower levels of retained DPM in their lungs
  7      than rats and mice do and, consequently, less pulmonary function impairment and pulmonary
  8      pathology. These differences may result from lower DPM inspiration and deposition during
  9      exposure, greater DPM clearance, or lung tissue less susceptible to the cytotoxicity of deposited
10      DPM.              .
11
12      5.3. INTERACTIVE EFFECTS OF DIESEL EXHAUST'
13             A multitude of factors may influence the susceptibility to exposure to diesel exhaust as
14      well as the resulting response. Some of these have already been discussed in detail (e.g., the
15      composition of diesel exhaust and concentration-response data); others will be addressed in this
16      section (e.g., the interaction of diesel exhaust with factors particular to the exposed individual
17      and the interaction of diesel exhaust components with other airborne contaminants).
18             Mauderly et al. (1990a) compared the susceptibility of normal rats and rats with
19      preexisting laboratory-induced pulmonary emphysema exposed for 7 h/day, 5 days/week for 24
20      mo to diesel exhaust  containing 3.5 mg/m3 DPM or to clean air (controls). Emphysema was
21      induced in one-half of the rats by intratracheal instillation of elastase 6 weeks before exhaust
22      exposure. Measurements included lung burdens of diesel particles, respiratory function,
23      bronchoalveolar lavage, clearance of radiolabeled particles, pulmonary immune responses, lung
24      collagen, excised lung weight and volume, histopathology, and mean linear intercept of terminal
25      air spaces. None of the data for the 63 parameters measured suggest that rats with
26      emphysematous lungs were more susceptible than rats with normal lungs to the effects of diesel
27      exhaust exposure. In fact, each of the 14 emphysema-exhaust interactions detected by statistical
28      analysis of variance indicated that emphysema acted to reduce the effects of diesel exhaust
29      exposure. Diesel particulate matter accumulated much less rapidly in the lungs of
30      emphysematous rats than in those of normal rats. The mean lung burdens of DPM in the
31      emphysematous rats were 39, 36, and 37% of the lung burdens of normal rats at 12, 18, and 24
32      mo, respectively. No significant interactions were observed among lung morphometric
33      parameters. Emphysema prevented the exhaust-induced increase for three respiratory indices of
34      expiratory flow rate at low lung volumes, reduced the exhaust-induced increase in nine lavage
35      fluid indicators of lung damage, prevented the expression of an exhaust-induced increase in lung
36      collagen, and reduced the exhaust-induced delay in DPM clearance.
        2/1/98                                    5-75      DRAFT-DO NOT CITE OR QUOTE

-------
 1             Mauderly et al. (1987b) evaluated the relative susceptibility of developing and adult rat
 2     . lungs to damage by exposure to diesel exhaust. Rats (48 per test group) were exposed to diesel
 3      exhaust containing 3.5 mg/m3 DPM and about 0.8 ppm NO2. Exposures were for 7 h/day,
 4      5 days/week through gestation to the age of 6 mo, or from the age of 6 to 12 mo. Comparative
 5      studies were conducted on respiratory function, immune response, lung clearance, airway fluid
 6      enzymes, protein and cytology, lung tissue collagen, and proteinases in both age groups. After
 7      the 6-month exposure, adult rats, compared with controls, exhibited (1) more focal aggregates of
 8      particle-containing AMs in the alveolar ducts near the terminal bronchioles, (2) a sixfold increase
 9      in the neutrophils (as a percentage of total leukocytes) in the airway fluids, (3) a significantly
10      higher number of total lymphoid cells in the pulmonary lymph nodes, (4) delayed clearance of
11      diesel particles and radiolabeled particles  (t1/2 = 90 days versus 47 days for controls), and (5)
12      increased lung weights. These effects were not seen in the young rats. On a weight for weight
13      (milligrams of DPM per gram of lung) basis, DPM accumulation in the lungs was similar in
14      young and adult rats immediately after the exposure. During the 6-month postexposure period,
15      DPM clearance was much more rapid  in the neonatal rats, approximately 2.5-fold. During
16      postexposure, diesel particle-laden macrophages became aggregated in the neonatal rats, but
17      these aggregations were located primarily in a subpleural position.  The authors concluded that
18      exposure to diesel exhaust, using pulmonary function, structural (qualitative or quantitative)
19      biochemistry as the indices, did not  affect the developing rat lung more severely than the adult rat
20      lung.
                                                                             r
21             As a result of the increasing  trend of using diesel-powered equipment in coal mining
22      operations and the concern for adverse health effects in coal miners exposed to both coal dust  or
23      coal mine dust and diesel  exhaust, Lewis et al. (1989) and Karagianes et al. (1981) investigated
24      the interaction of coal dust and diesel exhaust. Lewis et al. (1989) exposed rats, mice, and
25      cynomolgus monkeys to (1) filtered ambient air, (2) 2 mg/m3 DPM, (3) 2 mg/m3 respirable coal
26      dust, and (4) 1 mg/m3 of both DPM  and respirable coal dust. Gaseous and vapor concentrations
27      were identical in both diesel exhaust exposures.  Exposures were for 7 h/day, 5 days/week for  up
28      to 24 mo.  Synergistic effects between diesel exhaust and coal dust were not demonstrated;
29      additive toxic effects were the predominant effects noted.
30             Karagianes et al. (1981) exposed rats (24 per group) to diesel exhaust containing
31      8.3 mg/m3 of DPM alone  or in combination with about 6 mg/m3 of coal  dust. No synergistic
32      effects were found between diesel exhaust and coal dust; additive effects in terms of visual dust
33      burdens in necropsied lungs were related to dose (i.e., length of exposure and airborne paniculate
34      concentrations).
35             The health effects  of airborne contaminants from sources other than diesel engines may
36      be altered in the presence  of DPM by their adsorption onto the diesel particles.  When adsorbed
        2/1/98                                     5-76        DRAFT-DO NOT CITE OR QUOTE

-------
  1      onto diesel particles, the gases and vapors can be transported and deposited deeper into the lungs,
  2      and because they are more concentrated on the particle surface, the resultant cytotoxic effects or
  3      physiological responses may be enhanced. Nitrogen dioxide adsorbed onto carbon particles
  4      caused pulmonary parenchymal lesions in mice, whereas NO2 alone produced edema and
  5      inflammation but no lesions (Boren, 1964).  Exposure to formaldehyde and acrolein adsorbed
  6      onto carbon particles (1 to 4 um) resulted in the  recruitment of PMNs to tracheal and
  7      intrapulmonary epithelial tissues but not when the aldehydes were tested alone (Kilbum and
  8      McKenzie, 1978).
  9            There is no direct evidence that diesel exhaust interacts with other substances in an
10      exposure environment or the physiological status of the exposed subject other than impaired
11      resistance to respiratory tract infections. Although there is experimental evidence that gases and
12      vapors can be adsorbed onto carbonaceous particles, enhancing the toxicity of these particles
13      when deposited in the lung, there is no evidence for an increased health risk from such
14      interactions with DPM under ambient urban atmospheric conditions. Likewise, there is no
15      experimental evidence in laboratory animals that the youth or preexisting emphysema of an
16      exposed individual enhances the risk of exposure to diesel exhaust.
17
18      5.4. COMPARATIVE RESPONSIVENESS AMONG SPECIES TO THE PULMONARY
19          EFFECTS OF DIESEL EXHAUST
20            There is some evidence indicating that species may differ in pulmonary responses to DE.
21      Mauderly et al. (1993) compared the pulmonary histopathology of rats and mice after 18 mo of
22      exposure to DE.  There was less aggregation of macrophages. Diffuse septal thickening was
23      noted in the mice, but there were few inflammatory cells, no focal fibrosis, little epithelial
24      hyperplasia, and no epithelial metaplasia,  as was observed in rats. Heinrich et al. (1986a)
25      reported that wet lung weight of hamsters increased only 1.8-fold following chronic exposure to
26      DE, compared with an increase of 3,4-fold in rats.  Smaller increases in neutrophils, lactic acid
27      dehydrpgenase, collagen, and protein supported the conclusion of a lesser inflammatory response
28      in Syrian hamsters.  The histopathologic changes in the lungs of Chinese hamsters after 6 mo
29      exposure to DE,  on the other hand, was similar to that of rats (Pepelko and Peirano, 1983).
30      Guinea pigs respond to  chronic DE exposure  with a well-defined epithelial proliferation, but it is
31      based on an eosinophilic response in contrast to the neutrophil-based responses in other species.
32      Epithelial hyperplasia and metaplasia were quite striking in the terminal and respiratory
33      bronchioles of cats exposed for 27 mo to DE  (Plopper et al., 1983). This study is of particular
34      interest because the terminal airways of cats are more similar to those of humans than rodent
35      species are.  It should be noted, however, that exposure concentrations were very high (12
36      mg/m3) for most of the period.  Lewis et d. (1989) exposed rats and Cynamolgous monkeys 8
        2/1/98                                    5-77       DRAFT-DO NOT CITE OR QUOTE

-------
 1      hours per day, 5 days per week for 2 years to DE at a particle concentration of 2 mg/m3.
 2      Unfortunately, this exposure rate was sufficiently low that few effects were noted in either
 3      species other than focal accumulations of particles, primarily in the alveolar macrophages,
 4      interstitium, and lymphoid tissue. To achieve chronic exposure conditions in monkeys would
 5      also require a considerably longer exposure period. It is apparent that species do vary in their
 6      pulmonary responses to DE exposure, despite the difficulty in making direct comparisons
 7      because of differences in exposure regimes, lifespans, and pulmonary anatomy. Most species do
 8      respond, however, suggesting that humans are likely to be susceptible to induction of pulmonary
 9      pathology during chronic exposure to DE.
10
11      5.5.  DOSE-RATE AND PARTICULATE CAUSATIVE ISSUES
12             The purpose of animal toxicological experimentation is to identify the hazards and dose-
13      response effects posed by a chemical substance or complex mixture and to extrapolate these
14      effects to humans for subsequent health assessments. The cardinal principle in  such a process is
15      that the intensity and character of the toxic action are a function of the dose of the toxic agent(s)
16      that reaches the  critical site of action. The considerable body of evidence reviewed clearly
17      denotes that major noncancerous health hazards may be presented to the lung following the
18      inhalation of diesel exhaust. Based on pulmonary function and histopathological and
19      histochemical effects, a determination can be made concerning what dose/exposure rates of
20      diesel exhaust (expressed in terms of the DPM concentration) result in an injury to the lung and
21      which appear to elicit no effect.  The inhalation of poorly soluble particles, such as those found in
22      diesel exhaust, increases the pulmonary particulate burden.  When the dosing rate exceeds the
23      ability of the pulmonary defense mechanisms to achieve a steady-state lung burden of particles,
24      there is a slowing of clearance and the progressive retention of particles in the lung that can
25      ultimately approach a complete  cessation of lung clearance (Morrow, 1988). This phenomenon
26      has practical significance both for the interpretation of experimental inhalation data and for the
27      prevention of disease in humans exposed to airborne particles.
28             The data for exposure intensities that cause adverse pulmonary effects demonstrate that
29      they are less than the exposure intensities reported to be necessary to  induce lung tumors. Using
30      the most widely studied laboratory animal species and the one reported  to be the most sensitive
31      to tumor induction, the laboratory rat, the no-adverse-effect exposure intensity for adverse
32      pulmonary effects was 56 mg-h-m"3/week (Brightwell et al., 1986). The lowest-observed-effect
33      level for adverse pulmonary effects (noncancer) in rats was 70 mg-h-m"3/week (Lewis et al.,
34      1989), and for pulmonary tumors, 122.5 mg-h-m"3/week (Mauderly et al., 1987a).  The results
35      clearly show that noncancerous  pulmonary effects are produced at lower exposure intensities
36      than are pulmonary tumors.  Such data support the position that inflammatory and proliferative
        2/1/98   .                                 5-78       DRAFT-DO NOT CITE OR QUOTE

-------
  1      changes in the lung may play a key role in the etiology of pulmonary tumors in exposed rats
  2      (Mauderly et al., 1990b).
  3             Adults who have a preexisting condition that may predispose their lungs to increased
  4      particle retention (e.g., smoking or high particulate burdens from nondiesel sources),
  5      inflammation (e.g., repeated respiratory infections), epithelial proliferation (e.g., chronic
  6      bronchitis), and fibrosis (e.g., silica exposure) and infants and children, due to their developing
  7      pulmonary and immunologic systems, may have a greater susceptibility to the toxic actions of
  8      diesel exhaust. It should be noted that both the developing lung and a model of a preexisting
  9      disease state have been studied with regard to their effect on the lungs' response to diesel exhaust
10      (Mauderly et al., 1990a, 1987b). Mauderly et al. (1987b) showed that diesel did not affect the
11      developing lung more severely than the adult rat lung, and in fact, that clearance was faster in the
12      younger lung.  Mauderly et al. (1990a) compared the pulmonary response to inhalation of diesel
13      exhaust in rats with elastase-induced emphysema with normal rats.  They found that respiratory
14      tract effects were not more severe in emphysematous rats and that the lung burden of particles
15      was less in the compromised rat. These studies provide limited evidence that some factors that
16      are often considered to result in a wider distribution of sensitivity among members of the
17      population may not have this effect with diesel exposure. However, these studies have no
18      counterpart in human studies and extrapolation to humans remains uncertain.
19             There is also the issue of whether the noncancerous health effects related to exposure to
20      diesel exhaust are caused by the carbonaceous core of the particle or substances adsorbed onto
21      the core, or both.
22             Current understanding suggests that much of the toxicity resulting from the inhalation  of
23      diesel exhaust relates to the carbonaceous core of the particles. Several studies on inhaled
24      aerosols demonstrate that lung reactions characterized by an appearance of particle-laden AMs
25      and their infiltration into the alveolar ducts, adjoining alveoli and tracheobronchial lymph nodes,
26      hyperplasia of Type II cells, and the impairment of pulmonary clearance mechanisms are not
27      limited to exposure to diesel particles.  Such responses have also been observed in rats following
28      the inhalation of coal dust (Lewis et al., 1989; Karagianes et al., 1986), titanium dioxide
29      (Heinrich et al., 1995; Lee et al., 1985), carbon black (Nikula et al., 1995; Heinrich et al., 1995),
30      titanium tetrachloride hydrolysis products (Lee et al., 1986), quartz (Klosterkotter and Buneman,
31      1961), volcanic ash (Wehner et al., 1986), amosite (Bolton et al., 1983), and manmade mineral
32      fibers (Lee et al., 1988) among others.  In more recent studies, animals have been exposed to
33      carbon black that is similar to the carbon core of the diesel exhaust particle. Nikula et al. (1995)
34      exposed rats for 24 mo to carbon black or diesel exhaust at target exposure concentrations of 2.5
35      and 6 mg/rn3 (exposure rates of 200 or 520 mg-h-m"3/week).  Both concentrations induced AM

        2/1/98                                   5-79       DRAFT-DO NOT CITE OR QUOTE

-------
 1      accumulation, epithelial proliferation, inflammation, and fibrosis. They observed essentially no
 2      difference in potency of nonneoplastic or in tumor responses based on a regression analysis.
 3            Dungworth et al. (1994) reported moderate to severe inflammation characterized by
 4      multifocal bronchoalveolar hyperplasia, alveolar histiocytosis, and focal segmental fibrosis in
 5      rats exposed to carbon black for up to 20 mo at exposure rates of 510 to 540 mg-h-m'Vweek. The
 6      observed lung pathology reflects notable dose-response relationships and usually evolves in a
 7      similar manner. With increasing dose, there is an increased accumulation and aggregation of
 8      particle-laden AMs,  Type II cell hyperplasia, a foamy (degenerative) macrophage response,
 9      alveolar proteinosis, alveolar bronchiolization, cholesterol granulomas, and often squamous cell
10      carcinomas and bronchioalveolar adenomas derived from metaplastic squamous cells in the areas
11      of alveolar bronchiolization.
12            Heinrich et al. (1995) compared effects of diesel exposure in rats and mice with exposure
13      to titanium dioxide or carbon black. Exposures to TiO2 and carbon black were adjusted during
14      the exposure to result in a similar lung burden for the three types of particles.  At similar lung
15      burdens in the rat, DPM, TiO2, and carbon black had nearly identical effects on lung weights and
16      on the incidence of lesions, both noncancer and cancer. Also, a similar effect on clearance of a
17      labeled test aerosol was measured for the different particles. A comparison of the effect of DPM,
18      TiO2, and carbon black exposures in mice also showed a similar effect on lung weight, but
19      noncancer effects were not reported and no significant increase in tumors was observed.
20            These experiments provide strong support for the idea that diesel exhaust toxicity results
21      from a mechanism that is  analogous to that of other relatively inert particles in the lung. This
22      qualitative similarity exists along with some apparent quantitative differences in the potency of
23      various particles for producing effects on the lung or on particle clearance.
24            Particle size, volume, surface area, and composition may be the critical elements in the
25      overload phenomenon following exposure to particles, which could explain those quantitative
26      differences.  The overloaded AMs secrete a variety of cytokines, oxidants, and proteolytic
27      enzymes that are responsible for inducing particle aggregation and damaging adjacent epithelial
28      tissue (Oberdorster and Yu, 1991).  For a more detailed discussion of mechanism, see Chapter
29      10.
30            The principal noncancefous health hazard to humans posed by exposure to diesel exhaust
31      is a structural or functional injury to the lung based on the laboratory animal data.  Such effects
32      are demonstrable at  dose rates or cumulative doses of DPM lower than those reported to be
33      necessary to induce  lung tumors. Current knowledge indicates that the carbonaceous core of
34      diesel particles is the major causative factor in the injury to the lung and that other factors such as
35      the cytotoxicity of adsorbed substances on the particles also may play a role. The lung injury
36      appears  to be mediated through effects on pulmonary AMs. Because noncancerous pulmonary
        2/1/98                                    5-80        DRAFT-DO NOT CITE OR QUOTE

-------
  1      effects occur at lower doses than tumor induction does in the rat, and because these effects may
  2      be cofactors in the etiology of diesel exhaust-induced tumors, noncancerous pulmonary effects
  3      must be considered in the total evaluation of diesel exhaust, notably the particulate component.
  4
  5      5.6. SUMMARY AND DISCUSSION
  6      5.6.1. Effects of Diesel Exhaust on Humans
  7            The most readily identified acute noncancer health effect of diesel exhaust on humans is
  8      its ability to elicit subjective complaints of eye, throat, and bronchial irritation and
  9      neurophysiological symptoms such as headache, lightheadedness, nausea, vomiting, and
10      numbness and tingling of the extremities. Studies of the perception and offensiveness of the odor
11      of diesel exhaust and a human volunteer study in an exposure chamber have demonstrated that .
12      the time of onset of the human subj ective symptoms is inversely related to increasing
13      concentrations of diesel exhaust and the severity is directly related to increasing concentrations of
14      diesel exhaust. In one study in which a diesel engine was operated under varying load
15      conditions, a dilution factor of 140 to 475 was needed to reduce the exhaust level to an odor-
16      detection threshold level.
17            A public health issue is whether short-term exposure to diesel exhaust might result in an
18      acute decrement in ventilatory function and whether the frequent repetition of such acute
19      respiratory effects could result in chronic lung function impairment. One convenient means of
20      studying acute decrements in ventilatory function is to monitor differences in pulmonary function
21      in occupationally exposed workers at the beginning and end of a workshift In studies of
22      underground miners, bus garage workers, dock workers,  and locomotive repairmen, increases in
23      respiratory symptoms (cough, phlegm, and dyspnea) and decreases in lung function (FVC, FEVt,
24      PEFR, and FEF25_75) over the course of a workshift were  generally found to be minimal and not
25      statistically significant.  In a study of acute respiratory responses in diesel bus garage workers,
26      there was an increased reporting of cough, labored breathing, chest tightness, and wheezing, but
27      no reductions in pulmonary function were associated with exposure to diesel  exhaust.
28      Pulmonary function was affected in stevedores over a workshift exposure to diesel exhaust but
29      normalized after a few days without exposure to diesel exhaust fumes. In a third study, there was
30      a trend toward greater ventilatory function changes during a workshift among coal miners, but
31      the decrements were similar in miners exposed and not exposed to diesel exhaust.
32            Smokers appeared to demonstrate larger workshift respiratory function decrements and
33      increased incidents of respiratory symptoms. Acute sensory and respiratory symptoms were
34      earlier and more sensitive indicators of potential health risks from diesel exposure than were
35      decrements in pulmonary function. Studies on the acute  health effects of exposure to diesel
36      exhaust in humans, experimental and epidemiologic, have failed to demonstrate a consistent
        2/1/98               '                    5-81       DRAFT-DO NOT CITE OR QUOTE

-------
 1      pattern of adverse effects on respiratory morbidity; the majority of studies offer, at best,
 2      equivocal evidence for an exposure-response relationship. The environmental contaminants have
 3      frequently been below permissible workplace exposure limits; in those few cases where health
 4      effects have been reported, the authors have failed to identify conclusively the individual or
 5      collective causative agents in the diesel exhaust.
 6             Chronic effects of diesel exhaust exposure have been evaluated in epidemiologic studies
 7      of occupationally exposed workers (metal and nonmetal miners, railroad yard workers,
 8      stevedores, and bus garage mechanics). Most of the epidemiologic data indicate an absence of an
 9      excess risk of chronic respiratory disease associated with exposure to diesel exhaust.  In a few
10      studies, a higher prevalence of respiratory symptoms, primarily cough, phlegm, or chronic
11      bronchitis, was observed among the exposed. These increased symptoms, however, were usually
12      not accompanied by significant changes in pulmonary function.  Reductions in FEVj  and FVC
13      and, to a lesser extent, FEF50 and FEF75, also have been reported. Two studies detected
14      statistically significant decrements in baseline pulmonary function consistent with obstructive
15      airway disease. One study of stevedores had a limited sample size of 17 exposed and 11
16      controls.  The second study in coal miners showed that both underground and surface workers at
17      diesel-use mines had somewhat lower pulmonary performance than their matched controls.  The
18      proportion of workers in or at diesel-use mines, however, showed equivalent evidence of
19      obstructive airway disease and for this reason the authors of the  second paper felt that factors
20      other than diesel exposure might have been responsible. A doubling of minor restrictive airway
21      disease was also observed in workers in or at diesel-use mines. These two studies, coupled with
22      other reported nonsignificant trends in respiratory flow-volume measurements, suggest that
23      exposure to diesel exhaust may impair pulmonary function among occupational populations.
24      Epidemiologic studies of the effects of diesel exhaust on organ systems other than the pulmonary
25      system are scant. Whereas a preliminary study of the association of cardiovascular mortality and
26      exposure to diesel exhaust found a fourfold higher risk ratio, a more comprehensive
27      epidemiologic study by the same investigators found no significant difference between the
28      observed and expected number of deaths caused by cardiovascular disease.
29             Caution is warranted in the interpretation of results from the epidemiologic studies that
30      have addressed noncarcinogenic health effects from exposure to diesel exhaust. These
31      investigations suffer from myriad methodological problems,  including (1) incomplete
32      information on the  extent of exposure to diesel exhaust, necessitating in some studies estimations
33      of exposures from job titles and resultant misclassification; (2) the presence of confounding
34      variables such as smoking or occupational exposures to other toxic substances (e.g., mine dusts);
35      and (3) the short duration and low intensity of exposure.  These limitations restrict drawing

        2/1/98               •                    5-82        DRAFT-DO NOT CITE OR QUOTE

-------
  1      definitive conclusions as to the cause of any noneareinogenic diesel exhaust effect, observed or
  2      reported.
  3
  4      5.6.2. Effects of Diesel Exhaust on Laboratory Animals
  5             Laboratory animal studies of the toxic effects of diesel exhaust have involved acute, sub-
  6      chronic, and chronic exposure regimens. In acute exposure studies, toxic effects appear to have
  7      been associated primarily with high concentrations of carbon monoxide, nitrogen dioxide, and
  8      aliphatic aldehydes.  In short- and long-term studies, toxic effects have, been associated with
  9      exposure to the complex exhaust mixture. Effects of diesel exhaust in various animal species are
10      summarized in Tables 5-2 to 5-14.  In short-term studies, health effects are not readily apparent,
11      and when found, are mild and result from concentrations of about 6 mg/m3 DPM and durations of
1.2      exposure approximating 20 h/day.  There is ample evidence, however, that short-term exposures
13      at lower levels of diesel exhaust affect the lung, as indicated by an accumulation of DPM,
14      evidence of inflammatory response, AM aggregation and accumulation near the terminal
15      bronchioles, Type II cell proliferation, and the thickening of alveolar walls adjacent to AM
16      aggregation. Little evidence exists, however, from short-term  studies that exposure to diesel
17      exhaust impairs lung function. Chronic exposures cause lung pathology that results in altered
18      pulmonary function and increased DPM retention in the lung.  Exposures to diesel exhaust have
19      also been associated with increased susceptibility to respiratory tract infection, neurological or
20      behavioral changes, an increase in banded neutrophils, and morphological alterations in the liver.
21
22      5.6.2.1.  Effects on Survival and Growth
23            The data presented in Table 5-3  show limited effects on survival in mice and rats and
24      some evidence of reduced body weight in rats following chronic exposures to concentrations of
25      1.5 mg/m3 DPM or higher and exposure durations of 16 to 20 h/day, 5 days/week for 104 to 130
26      weeks. Increased lung weights and lung to body weight ratios in rats, mice, and hamsters; an
27      increased heart to body weight ratio in rats; and decreased lung and kidney weights in cats have
28      been reported following chronic exposure to diesel exhaust.  No evidence was found of an effect
29      of diesel exhaust on other body organs (Table 5-4). The lowest-observed-effect level in rats
30      approximated 1 to 2 mg/m3 DPM for 7 h/day, 5 days/week for 104 weeks.
31
32      5.6.2.2.  Effects on Pulmonary Function
33            Pulmonary function impairment has been reported in rats, hamsters, cats, and monkeys
34      exposed to diesel exhaust and included lung mechanical properties (compliance and resistance),
35      diffusing capacity, lung volumes, and ventilatory performance (Table 5-5).  The effects generally
36      appeared only after prolonged exposures. The lowest exposure levels (expressed in terms of
        2/1/98                                    5-83       DRAFT-DO NOT CITE OR QUOTE

-------
 1      DPM concentrations) that resulted in impairment of pulmonary function occurred at 2 mg/m3 in
 2      cynomolgus monkeys (the only level tested), 1.5 and 3.5 mg/m3 in rats, 4.24 and 6 mg/m3 in
 3  .    hamsters, and 11.7 mg/m3 in cats. Exposures in monkeys, cats, and rats (3.5 mg/m3) were for 7
 4      to 8 h/day, 5 days/week for 104 to 130 weeks. While this duration is considered to constitute a
 5      lifetime study in rodents, it is a small part of the lifetime of a monkey or cat, so these effect levels
 6      are essentially comparing chronic rodent studies with a subchronic monkey study. Exposures in
 7      hamsters and rats (1.5 mg/m3) varied in hours per day (8 to 20) and weeks of exposure (26 to
 8      130). In all species but the monkey, the testing results were consistent.with restrictive lung
 9      disease; alteration in expiratory flow rates indicated that 1.5 mg/m3 DPM was a LOAEL for a
10      chronic exposure (Gross, 1981).  Monkeys demonstrated evidence of obstructive airway disease.
11      The nature of the pulmonary impairment is dependent on the dose of toxicants delivered to and
12      retained in the lung, the site of deposition and effective clearance or repair, and the anatomy and
13      physiology of the affected species; these variables appear to be factors in the disparity of the
14  .    airway disease in monkey versus the other species tested.
15
16      5.6.2.3. Hlstopathological and Histochemical Effects
17             Histological studies have demonstrated that chronic exposure to diesel exhaust can result
18      in effects on respiratory tract tissue (Table 5-6).  Typical findings include alveolar histiocytosis,
19      AM aggregation, tissue inflammation, increase in PMNs, hyperplasia of bronchiolar and alveolar
20      Type II cells, thickened alveolar septa, edema, fibrosis, and emphysema.  Lesions in the trachea
21      and bronchi were observed in some studies. Associated with these histopathological findings
22      were various histochemical changes in the lung, including increases in lung DNA, total  protein,
23      alkaline and acid phosphatase, glucose-6-phosphate dehydrogenase; increased synthesis of
24      collagen; and release of inflammatory mediators such as leukotriene LTB and prostaglandin
25      PGF2a. Although the overall laboratory evidence is that prolonged exposure to DPM results in
26      histopathological and histochemical changes in the lungs of exposed animals, some studies have
27      also demonstrated that there may be a threshold of exposure to DPM below which pathologic
28      changes do not occur. These no-observed-adverse-effect levels for histopathological effects were
29      reported to be 2 mg/m3 for cynomolgus monkeys (the only concentration tested), 0.11 to 0.35
30      mg/m3  for rats, and 0.25 mg/m3 DPM for guinea pigs exposed for 7 to 20 h/day, 5 to 5.5
31      days/week for 104 to 13 0 weeks.
32
33      5.6.2.4. Effects on Defense Mechanisms
34             The pathological effects of DPM appear to be strongly dependent on the relative rates of
35      pulmonary deposition and clearance (Table 5-7). Clearance of particles from the alveolar region
36      of the lungs is a multiphasic process involving phagocytosis by AMs. Chronic exposure to DPM
        2/1/98                                    5-84        DRAFT-DO NOT CITE OR QUOTE

-------
  1      concentrations of about 1 mg/m3 or above, under varying exposure durations, causes pulmonary
  2      clearance to be reduced with concomitant focal aggregations of particle-laden AMs, particularly
  3      in the peribronchiolar and alveolar regions, as well as in the hilar and mediastinal lymph nodes.
  4      The exposure concentration at which focal aggregates of particle-laden AMs occur may vary
  5      from species to species, depending on rate of uptake and pulmonary deposition, pulmonary
  6      clearance rates, the relative size of the AM population per unit of lung tissue, the rate of
  7      recruitment of AMs and leukocytes, and the relative efficiencies for removal of particles by the
  8      mucociliary and lymphatic transport system. The principal mechanism of reduced particle
  9      clearance appears to be an effect on pulmonary AMs. Impairment of particle clearance seems to
10      be nonspecific and applies primarily to dusts that are persistently retained in the lungs. Lung dust
11      levels of approximately 0.1 to 1 mg/g lung tissue appear to produce this effect in the Fischer 344
12      rat (HEI, 1995). Morrow (1988) suggested that the inability of particle-laden AMs to translocate
13      to the mucociliary escalator is correlated to an average composite particle volume per AM in the
14      lung.  When this particle volume exceeds approximately 60 um3 per AM in the Fischer 344 rat,
15      impairment of clearance appears to be initiated.  When the particulate volume exceeds
16      approximately 600 um3 per cell, evidence suggests that AM-mediated particulate clearance
17      virtually ceases and agglomerated particle-laden macrophages remain in the alveolar region and
18      increasingly nonphagocytized dust particles translocate to the pulmonary interstitiurn.  Data for
19      other laboratory animal species  and humans are, unfortunately,  limited.
20             Several laboratory animal studies have indicated that exposure to DPM can reduce an
21      animal's resistance to respiratory infections. This effect, which can occur even after only 2 or 6 h
22      of exposure to diesel exhaust containing 5 to 8 mg/m3 DPM, does not appear to be caused by
23      direct impairment of the lymphoid or splenic immune systems;  however, in one study of
24      influenza virus infection, interferon levels and hemagglutinin antibody levels were adversely
25      affected in the exposed mice. Studies on the effects of exposure to diesel exhaust or DPM on the
26      immune system of laboratory animals have produced equivocal results (Table 5-8).
27                                            .
28      5.6.2.5. Neurological and Behavioral Effects
29             Behavioral effects have been observed in rats exposed to diesel exhaust from birth to 28
30      days of age (Table 5-13). Exposure caused a decreased level of spontaneous locomotor activity
31      and a detrimental effect on learning in adulthood. In agreement with the behavioral changes was
32      physiological evidence for delayed neuronal maturation. Exposures were to 6 mg/m3 DPM for 8
33      h/day, 7 days/week from birth to about 7,14,21, or 28 days of age.
34             .                                   .
        2/1/98                                    5-85       DRAFT-DO NOT CITE OR QUOTE

-------
 1      5.6.2.6.  Other Noncancerous Effects
 2            Essentially no effects (based on the weight of evidence of a number of studies) were
 3      noted for reproductive and teratogenic effects in mice, rats, rabbits, and monkeys; clinical
 4      chemistry and hematology in the rat, cat, hamster, and monkeys; and enzyme induction in the rat
 5      and mouse (Tables 5-10 through 5-12 and 5-14).
 6
 7      5.6.3. Comparison of Filtered and Unfiltered Diesel Exhaust
 8            The comparison of the toxic  responses in laboratory animals exposed to whole diesel
 9      exhaust or filtered exhaust containing no particles demonstrates across laboratories that diesel
10      particles are the principal etiologic agent of noncancerous health effects in laboratory animals
11      exposed to diesel exhaust (Table 5-15). Whether the particles act additively or synergistically
12      with the gases cannot be determined from the designs of the studies.  Under equivalent exposure
13      regimens, hamsters have lower levels of retained DPM in their lungs than rats and mice do and
14      consequently less pulmonary function impairment and pulmonary pathology.  These differences
15      may result from a lower intake rate of DPM, lower deposition rate and/or more rapid clearance
16      rate, or lung tissue that is less susceptible to the cytotoxicity of DPM. Observations of a
17      decreased respiration  in hamsters when exposed by inhalation favor lower intake and deposition
18      rates.
19
20      5.6.4. Interactive Effects of Diesel Exhaust
21           . There is no direct evidence that diesel exhaust interacts with other substances in an
22      exposure environment, other than an impaired resistance to respiratory tract infections. Young
23      animals were not more susceptible.  In several ways, animals with laboratory-induced
24      emphysema were more resistant. There is experimental evidence that both inorganic and organic
25      compounds can be adsorbed onto carbonaceous particles. When such substances become
26      affiliated with particles, these substances can be carried deeper into the lungs where they might
27      have a more direct and potent effect on epithelial cells or on AM ingesting the particles. Few
28      specific studies to test interactive effects of diesel exhaust with atmospheric contaminants, other
29      than coal dust, have been conducted. Coal dust and DPM had an additive effect only.
30
31      5.6.5. Summary
32            The principal noncancerous health hazard to humans from exposure to diesel exhaust is a
33      structural, functional, or biochemical injury to the lung. Current knowledge indicates that the
34      carbonaceous core of the diesel particle is the prime causative agent of lung injury.  The lung
35      injury appears to be mediated by a progressive impairment of AMs.  Because noncancerous
36      pulmonary effects occur at lower doses than those inducing tumors in rats and appear to be
        2/1/98                                    5-86       DRAFT-DO NOT CITE OR QUOTE

-------
  1      cofactors in the etiology of diesel exhaust-induced tumors in rats, noncancerous pulmonary

  2      effects are relevant factors in the development of risk assessments,

  3

  4      5.7. REFERENCES  •                   .
  5
  6      Ames, RG; Attfield, MD; Hankinson, JL; Hearl, FJ; Reger, RB. (1982) Acute respiratory effects of exposure to
  7      diesel emissions in coal miners. Am Rev Respir Dis 125:39-42.
  8
  9      Ames, RG; Reger, RB; Hall, DS, (1984) Chronic respiratory effects of exposure to diesel emissions in coal mines.
10      Arch Environ Health 39:389-394.
11
12      Attfield, MD. (1978) The effect of exposure to silica and diesel exhaust in underground metal and nonmetal miners.
13      In: Kelley, WD, ed. Industrial hygiene for mining and tunneling: proceedings of an American Conference of
14      Governmental Industrial Hygienists topical symposium; November 1978; Denver, CO; pp. 129-135.
15
16      Attfield, MD; Trabant, GD; Wheeler, RW. (1982) Exposure to diesel fumes and dust at six potash mines. Ann Occup
17      Hyg 26:817-831.
18
19      Barnes, DG; Daston, GP; Evans, JS; Jarabek, AM; Kavlock, RJ; Kimmel, CA; Park, C; Spitzer, HL. (1995)
20      Benchmark Dose Workshop: criteria for use of a benchmark dose to estimate a reference dose. Reg Toxicol
21      Pharmacol 21:296-306.
22
23      Barnhart, MI; Chen, S-T; Salley, SO; Puro, H. (1981) infrastructure and morphometry of the alveolar lung of guinea
24      pigs chronically exposed to diesel engine exhaust: six month's experience. J Appl Toxicol 1:88-103.
25
26      Barnhart, MI; Salley, SO; Chen, S-T; Puro, H. (1982) Morphometric ultrastructural analysis of alveolar lungs of
27      guinea pigs chronically exposed by inhalation to diesel exhaust (DE). In: Lewtas, J, ed. Toxicological effects of
28      emissions from diesel engines: proceedings of the Environmental Protection Agency diesel emissions symposium;
29      October 1981; Raleigh, NC. 183-200. (Developments in toxicology and environmental science: v. 10). New York:
30      Elsevier Biomedical; pp.
31
32      Battigelli, MC. (1965) Effects of diesel exhaust Arch Environ Health 10:165-167.
33
34      Battigelli, MC; Mannella, RJ; Hatch, TF, (1964) Environmental and clinical investigation of workmen exposed to
35      diesel exhaust in railroad engine houses. IndMedSurg 33:121-124,
36
37      Battigelli, MC; Hengstenberg, F; Mannela, RJ; Thomas, AP. (1966) Mucociliary activity. Arch Environ Health
38      12:460-466.
39
40      Bhatnagar, RS. (1980) Collagen and prolyl hydroxylase levels in lungs in dogs exposed to automobile exhaust and
41      other noxious gas mixtures. In: Stara, JF; Dungworth, DL; Orthoefer, JG; Tyler, WS, eds. Long-term effects of air
42      pollutants in canine species. Cincinnati, OH: U.S. Environmental Protection Agency, Environmental Criteria and
43      Assessment Office; report no. EPA-600/8-80-014; pp. 71-87.
44
45      Bhatnagar, RS; Hussain, MZ; Sorensen, KR; Von Dohlen, FM; Danner, RM; McMillan, L; Lee, SD, (1980)
46      Biochemical alterations in lung connective tissue in rats and mice exposed to diesel emissions. In: Pepelko, WE;
47      Danner, RM; Clarke, NA, eds. Health effects of diesel engine emissions: proceedings of an international symposium,
48      v. 1; December 1979; Cincinnati, OH. Cincinnati,  OH: U.S. Environmental Protection Agency, Health Effects
49      Research Laboratory; pp. 557-570; EPA report no. EPA-600/9-80-057a. Available from: NTIS, Springfield, VA;
50      PB81-173809.
51
52      Bice, DE; Mauderly, JL; Jones, RK; McClellan, RO. (1985) Effects of inhaled diesel exhaust on immune responses
53      after lung immunization. Fundam Appl Toxicol  5:1075-1086.


         2/1/98    .                                     5-87        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Bolton, RE; Vincent, JH; Jones, AD; Addison, J; Beckett, ST. (1983) An overload hypothesis for pulmonary
  2       clearance of UICC amosite fibres inhaled by rats. Br J Ind Med 40:264-272.
  3
  4       Boren, HG. (1964) Carbon as a carrier mechanism  for irritant gases. Arch Environ Health 8:119-124.
  5
  6       Brightwell, J; Fouillet, X; Cassano-Zoppi, A-L; Gatz, R; Duchosal, F. (1986) Neoplastic and functional changes in
  7       rodents after chronic inhalation of engine exhaust emissions. In: Ishinishi, N; Koizumi, A; McClellan, RO; Stober,
  8       W, eds. Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the international satellite
  9       symposium on toxicological effects of emissions from diesel engines; July; Tsukuba Science City, Japan.
10       (Developments in toxicology and environmental science: v. 13). Amsterdam: Elsevier Science Publishers B. V.; pp.
11       471-485.
12
13       Campbell, KI; George, EL; Washington, IS, Jr. (1980) Enhanced susceptibility to infection in mice after exposure to
14       dilute exhaust from light duty diesel engines. In: Pepelko, WE; Danner, RM; Clarke, NA, eds. Health effects of
15       diesel engine emissions: proceedings of an international symposium, v. 2; December 1979; Cincinnati, OH.
16       Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp. 772-785; EPA
17       report no. EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.
18
19       Campbell, KI; George, EL; Washington, IS, Jr. (1981) Enhanced susceptibility to infection in mice after exposure to
20       dilute exhaust from light duty diesel engines. Environ Int 5:377-382.
21
22       Castleman, WL; Dungworth, DL; Schwartz, LW; Tyler, WS. (1980) Acute respiratory bronchiolitis: an
23       ultrastructural and autoradiographic study of epithelial cell injury and renewal in rhesus monkeys exposed to ozone.
24       Am JPathol 98:811-840.
25
26       Castranova, V; Bowman, L; Reasor, MJ; Lewis, T; Tucker, J; Miles, PR. (1985) The response of rat alveolar
27       macrophages to chronic inhalation of coal dust and/or diesel exhaust. Environ Res 36:405-419.
28
29       Chan, TL; Lee, PS; Hering, WE. (1981) Deposition and clearance of inhaled diesel exhaust particles in the
30       respiratory tract of Fischer rats. J Appl Toxicol 1:77-82.
31
32       Chaudhari, A; Dutta, S.  (1982) Alterations in tissue glutathione and angiotensin converting enzyme due to inhalation
33       of diesel engine exhaust. J Toxicol Environ Health 9:327-337.
34
35       Chaudhari, A; Fairer, RG; Dutta, S. (1980) Effect of exposure to diesel exhaust on pulmonary prostaglandin
36       dehydrogenase (PGDH) activity. In: Pepelko, WE; Danner, RM; Clarke, NA, eds. Health effects of diesel engine
37       emissions: proceedings of an international symposium, v. 1; December 1979; Cincinnati, OH. Cincinnati, OH: U.S.
38       Environmental Protection Agency, Health Effects Research Laboratory; pp. 532-537; EPA report no.
39       EPA-600/9-80-057a. Available from: NTIS, Springfield, VA; PB81-173809.
40
41       Chaudhari, A; Farrer, RG; Dutta, S. (1981) Effect of exposure to diesel exhaust on pulmonary prostaglandin
42       dehydrogenase (PGDH) activity. J Appl Toxicol 1:132-134.
43
44      . Chen, KC. (1986) Induction of aryl hydrocarbon hydroxylase in rat tissue following intratracheal instillation of diesel
45       paniculate extract and benzo[a]pyrene. J Appl Toxicol 6:259-262.
46
47       Chen, KC; Vostal, JJ. (1981) Aryl hydrocarbon hydroxylase activity induced by injected diesel paniculate extract vs
48       inhalation of diluted diesel exhaust. J Appl Toxicol 1:127-131.
49
50       Chen, S; Weller,  MA; Barnhart, MI. (1980) Effects of diesel engine exhaust on pulmonary alveolar macrophages.
5.1       Scanning Electron Microsc 3:327-338.
52
5.3       Cuddihy, RG; Griffith, WC; Clark, CR; McClellan, RO. (1981) Potential health and environmental effects of light
54       duty diesel vehicles II. Albuquerque, NM: Lovelace Biomedical and Environmental Research Institute, Inhalation
55       Toxicology Research Institute; report no. LMF-89.


         2/1/98   .                                       5-88        DRAFT-DO NOT CITE OR QUOTE

-------
  1      Diaz-Sanchez, D; Dotson, AR; Takenaka, H; Saxon, A..(1994) Diesel exhaust particles induce local IgE production
  2      in vivo and alter the pattern of IgE messenger RNA isoforms. JClin Invest 94:1417-1425.
  3
  4      Diaz-Sanchez, D; Tsien, A, Casillas, A; Dotson, AR; Saxon, A, (1996) Enhanced nasal cytokine production in
  5      human beings after in vivo challenge with diesel particles. J. Allergy Clin Immunol 98:114-123.
  6
  7      Diaz-Sanchez, D; Tsien, A; Fleming, J; Saxon, A, (1997) Combined exhaust paniculate and ragweed allergen
  8      challenge markedly enhances human in vivo nasal ragweed-specific IgE and skews cytokine production to a T helper
  9      cell 2-type pattern. J Immunol 158:2406-2413.
 10                                                .
 11      Dungworth, DL; Mohr, U; Heinrich, U; Ernst, H; Kittel, B. (1994) Pathologic effects of inhaled particles in rat
 12      lungs: associations between inflammatory and neoplastic processes. In: Mohr, U; Dungworth, DL; Mauderly, JL;
 13      Oberdorster, G, eds. Toxic and carcinogenic effects of solid particles in the respiratory tract: [proceedings of the 4th
 14      international inhalation symposium]; March 1993; Hannover, Germany. Washington, DC: International Life
 15      Sciences Institute Press; pp. 75-98.
 16
 17      Dziedzic, D. (1981) Differential counts of B and T lymphocytes in the lymph nodes, circulating blood and spleen
 18      after inhalation of high concentrations of diesel exhaust. J Appl Toxicol 1:111-115.
 19
 20      Edling, C; Axelson, O. (1984) Risk factors of coronary heart disease among personnel in a bus company. Int Arch
 21      Occup Environ Health 54:181-183.
 22
 23      Edling, C; Anjou, C-G; Axelson, O; Kling, H. (1987) Mortality among personnel exposed to diesel exhaust. Int Arch
 24      Occup Environ Health 59:559-565.
 25
 26      El Batawi, MA; Noweir, MH. (1966) Health problems resulting from prolonged exposure to air pollution in diesel
 27      bus garages, Ind Health 4:1-10.
 28             '
 29      Eskelson, CD; Strom,  KA; Vostal, JJ; Misiorowski, RL. (1981) Lipids in the lung and lung gavage fluid of animals
 30      exposed to diesel particulates. Toxicologist 1:74-75. .
 31
 32      Fedan, JSj.Frazier, DG; Moorman, WJ; Attfield, MD; Franczak, MS; Kosten, CJ; Cahill, JF; Lewis, TR; Green,
 33      FHY. (1985) Effects of a two-year inhalation exposure of rats to coal dust and/or diesel exhaust on tension responses
 34      of isolated airway smooth muscle. Am Rev RespirDis  131:651-655.
 35
 36      Forrest, L; Lee, WB; Smalley, WM. (1980) Assessment of environmental impacts of light-duty vehicle dieselization
 37      [final report], Washington, DC: U.S. Department of Transportation, National Highway Traffic Safety
 38      Administration; report no. DOT-TSC-NHTSA-80-5. Available from: NTIS, Springfield, VA; PB80-220759.
 39                                                                                            .       .
40      Gamble, JF; Jones, WG. (1983) Respiratory effects of diesel exhaust in salt miners. Am Rev Respir Dis
41      128:389-394.   '                                                              .
42
43      Gamble, J; Jones, W; Hudak, J; Merchant, J. (1978) Acute changes in pulmonary function in salt miners. Presented
44      at: ACGIH conference on mining and tunneling; November; Denver, CO. Cincinnati, OH: American Conference of
45      Governmental Industrial Hygienists; pp. 119-128.
46
47      Gamble, J; Jones, W; Hudak, J. (1983) An epidemiological study of salt miners in diesel and nondiesel mines. Am J
48      Ind Med 4:435-458.
49
 50      Gamble, J; Jones, W; Minshall, S. (1987a) Epidemiological-environmental study of diesel bus garage workers: acute
51      effects of NO2 and respirable particulate on the respiratory system. Environ Res 42:201-214.
52
53      Gamble, J; Jones, W; Minshall, S. (1987b) Epidemiological-environmental study of diesel bus garage workers:
54      chronic effects of diesel exhaust on the respiratory system. Environ Res 44:6-17.
55


         2/1/98                                          5-89        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Garg, BD. (1983) Histologic quantitation of macrophage aggregates in the lungs of diesel exhaust exposed rats. Acta
  2       Stereol. 2(suppl. l):235-238.
  3
  4       Gillespie, JR. (1980) Review of the cardiovascular and pulmonary function studies on beagles exposed for 68
  5       months to auto exhaust and other air pollutants. In: Stara, JF; Dungworth, DL; Orthoefer, JG; Tyler, WS, eds.
  6       Long-term effects of air pollutants: in canine species. Washington, DC: U.S. Environmental Protection Agency,
  7       Environmental Criteria and Assessment Office; pp. 115-153; EPA report no. EPA-600/8-80-014. Available from:
  8       NTIS, Springfield, VA; PBS 1-144875.
  9
10       Green, FHY; Boyd, RL; Danner-Rabovsky, J; Fisher, MJ; Moorman, WJ; Ong, T-M; Tucker, J; Vallyathan, V;
11       Whong, W-Z; Zoldak, J; Lewis, T. (1983) Inhalation studies of diesel exhaust and coal dust in rats. Scand J Work
12       Environ Health 9:181-188.
13
14       Griffis, LC; Wolff, RK; Henderson, RF; Griffith, WC; Mokler, BV; McClellan, RO. (1983) Clearance of diesel soot
15       particles from rat lung after a subchronic diesel exhaust exposure. Fundam Appl Toxicol 3:99-103.
16
17       Gross, KB. (1981) Pulmonary function testing of animals chronically exposed to diluted diesel exhaust. J Appl
18       Toxicol 1:116-123.
19
20       Hahon, N; Booth, JA; Green, F; Lewis, TR. (1985) Influenza virus infection in mice after exposure to coal dust and
21       diesel engine emissions. Environ Res 37:44-60.
22
23       Hare, CT; Springer, KJ. (1971) Public response to diesel engine exhaust odors [final report]. San Antonio, TX:
24       Southwest Research Institute; report no. AR-804. Available from: NTIS, Springfield, VA; PB-204012.
25
26       Hare, CT; Springer, KJ; Somers, JH;  Huls, TA. (1974) Public opinion of diesel odor.
27
28       Hatch.'GE; Boykin, E; Graham, JA; Lewtas, J; Pott, F; Loud, K; Mumford, JL. (1985) Inhalable particles and
29       pulmonary host defense: in vivo and  in vitro effects of ambient air and combustion particles. Environ Res 36:67-80.
30                                                     .
31       Heinrich, U; Peters, L; Funcke, W; Pott, F; Mohr, U; Stober, W. (1982) Investigation of toxic and carcinogenic
32       effects of diesel exhaust in long-term inhalation exposure of rodents. In: Lewtas, J, ed. Toxicological effects of
33       emissions from diesel engines: proceedings of the Environmental  Protection Agency diesel emissions symposium;
34       October 1981; Raleigh, NC.(Developments in toxicology and environmental science: v. 10). New York: Elsevier
35       Biomedical; pp. 225-242.
36
37       Heinrich, U; Pott, F; Mohr, U; Stober, W. (1985) Experimental methods for the detection of the carcinogenicity
38       and/or cocarcinogenicity of inhaled polycyclic-aromatic-hydrocarbon-containing emissions. Carcinogenesis
39       (London) 8:131-146.
40                                            .
41       Heinrich, U; Muhle, H; Takenaka, S; Ernst, H; Fuhst, R; Mohr, U; Pott, F; Stober, W. (1986a) Chronic effects on the
42       respiratory tract of hamsters, mice, and rats after long-term inhalation of high concentrations of filtered and
43       unfiltered diesel engine emissions. J Appl Toxicol 6:383-395.
44
45       Heinrich, U; Pott, Fj.Rittinghausen, S. (1986b) Comparison of chronic inhalation effects in rodents after long-term
46       exposure to either coal oven flue gas mixed with pyrolized pitch or diesel engine exhaust. In: Ishinishi, N; Koizumi,
47       A; McClellan, RO; Stober,  W, eds. Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the
48       international satellite syposium on toxicological effects of emissions from diesel engines; July; Tsukuba Science
49       City, Japan.  (Developments in toxicology and environmental science: v.  13). Amsterdam: Elsevier Science
50       Publishers B.V.; pp. 441-457.
51
52       Heinrich, U; Fuhst, R; Rittinghausen, S; Creutzenberg, O; Bellman, B; Koch, W; Levsen, K. (1995) Chronic
53       inhalation exposure of Wistar rats  and two different strains of mouse to diesel engine exhaust, carbon black, and
54       titanium dioxide. Inhalation Toxicol 7:533-556.
55


         2/1/98                                          5-90         DRAFT-DO NOT CITE OR QUOTE

-------
  1       Henderson, RF; Sun, JD; Jones, RK; Mauderly, JL; McClellan, RO. (1983) Biochemical and cytological response in
  2      airways of rodents exposed in life span studies to diluted diesel exhaust. Am Rev Respir Dis 127:164.
  3
  4     ' Henderson, RF; Pickrell, JA; Jones, RK; Sun, JD; Benson, JM; Mauderly, JL; McClellan, RO. (1988) Response of
  5      rodents to inhaled diluted diesel exhaust: biochemical and cytological changes in bronchoalveolar lavage fluid and in
  6      lung tissue. Fundam Appl Toxicol 11:546-567.
  7
  8      Hinners, RG; Burkart, JK; Malanchuk, M. (1980) Facilities for diesel exhaust studies. In: Pepelko, WE; Danner,
  9      RM; Clarke, NA, eds. Health effects of diesel engine emissions: proceedings of an international symposium;
10      December 1979. Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp.
11       681-697;  EPA report no. EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PBS 1-173817.
12
13      Hyde, D;  Orthoefer, J; Dungworth, D; Tyler, W; Carter, R; Lum, H. (1978) Morphometfic and morphologic
14      evaluation of pulmonary lesions in beagle dogs chronically exposed to high ambient levels of air pollutants. Lab
15      Invest 38:455-469.
16
17      Hyde, DM; Plopper, CG; Weir, AJ; Murnane, RD; Warren, DL; Last, JA; Pepelko, WE. (1985) Peribronchiolar
18      fibrosis in lungs of cats chronically exposed to diesel exhaust. Lab Invest 52:195-206.
19
20      Ingalls, 'MR (1985) Improved mobile source exposure estimation. Ann Arbor, MI: U.S. Environmental Protection
21       Agency, Emission Control Technology Division; EPA report no. EPA-460/3-85-002. Available from: NTIS,
22    '  Springfield, VA; PB85-248441.
23
24      Iwai, K; Udagawa, T; Yamagishi, M; Yamada, H. (1986) Long-term inhalation studies of diesel exhaust on F344
25      SPF rats.  Incidence of lung cancer and lymphoma. In: Ishinishi, N; Koizumi, A; McClellan, RO; Stober, W, eds.
26      Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the international satellite symposium on
27      lexicological effects of emissions from diesel engines; July; Tsukuba Science City, Japan.(Developments in
28      toxicology and environmental science: v. 13). Amsterdam: Elsevier Science Publishers B. V.; pp. 349-360.
29
30      JOrgensen, H; Svensson, A. (1970) Studies on pulmonary function and respiratory tract symptoms of workers in an
31       iron ore mine where diesel trucks are used underground. J Occup Med 12:348-354.
32
33      Kahn, G;  Orris, P; Weeks, J. (1988) Acute overexposure to diesel exhaust: report of 13 cases. Am J Ind Med
34      13:405-406.
35
36      Kane, LE; Alarie, Y. (1978) Evaluation of sensory irritation from acrolein-formaldehyde mixtures. Am Ind Hyg
37      Assoc J 39:270-274.
38
39      Kane, LE; Alarie, Y. (1979) Interactions of sulfur dioxide and acrolein as sensory irritants. Toxicol Appl Pharmacol
40      48:305-315.
41
42      Kaplan, HL; MacKenzie, WF; Springer, KJ; Schreck, RM; Vostal, JJ. (1982) A subchronic study of the effects of
43      exposure  of three species of rodents to diesel exhaust. In: Lewtas, J, ed. Toxicological effects of emissions from
44      diesel engines: proceedings of the Environmental Protection Agency diesel emission symposium; October 1981;
45      Raleigh, NC; (Developments in toxicology and environmental science: v. 10). New York: Elsevier Biomedical; pp.
46      161-182.
47
48      Karagianes, MT; Palmer, RF; Busch, RH. (1981) Effects of inhaled diesel emissions and coal dust in rats. Am Ind
49      Hyg Assoc J 42:382-391.
50
51       Katz, M; Rennie,  RP; Jegier, Z. (1960) Air pollution and associated health hazards from diesel locomotive traffic in
52      a railroad tunnel. Occup Health Rev 11:2-15.
53
54      Kilburn, KH; McKenzie, WN. (1978) Leukocyte recruitment to airways by aldehyde-carbon combinations that
55      mimic cigarette smoke. Lab Invest 38:134-142.


         2/1/98                   '                      5-91        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Klosterkotter, W; Biinemann, G. (1961) Animal experiments on the elimination of inhaled dust. In: Davies, CN., ed.
  2       Inhaled particles and vapours: proceedings of an international symposium; March-April 1960; Oxford, United
  3       Kingdom. New York, NY: Pergamon Press; pp. 327-341.
  4
  5       Kobayashi, T; Ito, T. (1995) Diesel exhaust particulates induce nasal mucosal hyperresponsiveness to inhaled
  6       histamine aerosol. Fundam Appl Toxicol 27:195-202.
  7
  8       Kobayashi, T; Ikeue, T; Ito, T; Ikeda, A; Murakami, M; Kato, A; Maejima, K; Nakajima, T; Suzuki, T. (1997)
  9       Short-term exposure to diesel exhaust induces nasal mucosal hyperresponsiveness to histamine in guinea pigs.
10       Fundam Appl Toxicol 38:166-172.
11
12       Laurie, RD; Boyes,  WK. (1980) Neurophysiological alterations due to diesel exhaust exposure during the neonatal
13       life of the rat. In:  Pepelko, WE; Danner, RM; Clarke, NA, eds. Health effects of diesel engine emissions:
14       proceedings of an international  symposium, v. 2; December 1979; Cincinnati, OH. Cincinnati, OH: U.S.
15       Environmental Protection Agency, Health Effects Research Laboratory; pp. 713-727; EPA report no.
16       EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.
17
18       Laurie, RD; Boyes,  WK. (1981) Neurophysiological alterations due to diesel exhaust exposure during the neonatal
19       life of the rat. Environ Int 5:363-368.
20
21       Laurie, RD; Lewkowski, JP; Cooper, GP; Hastings, L. (1978) Effects of diesel exhaust on behavior of the rat.
22       Presented at: 71st annual meeting of the Air Pollution Control Asscociation; June; Houston,  TX. Pittsburgh, PA: Air
23       Pollution Control Association.
24
25       Laurie, RD; Boyes,  WK; Wessendarp, T. (1980) Behavioral alterations due to diesel exhaust exposure. In: Pepelko,
26       WE; Danner, RM; Clarke, NA,  eds.  Health effects of diesel engine emissions: proceedings of an international
27       symposium, v. 2; December 1979; Cincinnati, OH. Cincinnati, OH: U.S. Environmental Protection Agency, Health
28       Effects Research Laboratory; pp. 698-712; EPA report no. EPA-600/9-80-057b. Available from: NTIS, Springfield,
29       VA;PB81-173817.
30
31       Lee, IP; Suzuki, K;  Lee, SD; Dixon, RL. (1980) Aryl hydrocarbon hydroxylase induction in rat lung, liver, and male
32       reproductive organs following inhalation exposure to diesel emission. Toxicol Appl Pharmacol 52:181-184.
33
34       Lee, KP; Trochimowicz, HJ; Reinhardt, CF. (1985) Pulmonary response of rats exposed to titanium dioxide (TiO2)
35       by inhalation for two years. Toxicol Appl Pharmacol 79:179-192.
36
37       Lee, KP; Kelly, DP; Schneider, PW; Trochimowicz, HJ. (1986) Inhalation toxicity study on  rats exposed to titanium
38       tetrachloride atmospheric hydrolysis products for two years. Toxicol Appl Pharmacol 83:30-45.
39            .
40       Lee, KP; Kelly, DP; O'Neal, FO; Stadler, JC; Kennedy, GL, Jr. (1988)  Lung response to ultrafine Kevlar aramid
41       synthetic fibrils following 2-year inhalation exposure in rats. Fundam Appl Toxicol  11:1-20.
42
43       Lee, SD; Campbell, KI; Laurie, D; Hinners, RG; Malanchuk, M; Moore, W; Bhatnagar, RJ;  Lee, I. (1978)
44       Toxicological assessment of diesel emissions. Presented at: 71st annual meeting of the Air Pollution Control
45       Association; June; Houston, TX. Pittsburgh, PA: Air Pollution Control  Association;  paper no. 78-33.4.
46
47       Levins, PL. (1981) Review of diesel odor and toxic vapor emissions. Washington, DC: U.S.  Department of
48       Transportation, National Highway Traffic Safety Administration; report no. DOT-TSC-NHTSA-81-9.
49
50       Lewis, TR; Moorman, WJ. (1980) Pulmonary and cardiovascular physiology studies during  exposure. In: Stara, JF;
51       Dungworth, DL; Orthoefer, JG; Tyler, WS, eds. Long-term effects of air pollutants:  in canine species. Cincinnati,
52       OH: U.S. Environmental Protection Agency, Office of Health and Environmental Assessment, Environmental
53       Criteria and Assessment Office; pp.  97-113; EPA report no. EPA-600/8-80-014. Available from: NTIS, Springfield,
54      ' VA; PBS 1-144875.
55


         2/1/98                 '                       5-92        DRAFT-DO NOT CITE OR QUOTE

-------
 ' 1       Lewis, TR; Campbell, KI; Vaughan, TR, Jr. (1969) Effects on canine pulmonary function via induced N02
  2      impairment, paniculate interaction, and subsequent SOX. Arch Environ Health 18:596-601,
  3
  4      Lewis, TR; Green, FHY; Moorman, WJ; Burg, JR; Lynch, DW. (1989) A chronic inhalation toxicity study of diesel
  5      engine emissions and coal dust, alone and combined. J Am Coll Toxicol 8:345-375,
  6
  7      Linnell, RH; Scott, WE. (1962) Diesel exhaust composition and odor studies. J Air Pollut Control Assoc
  8      12:510-515,545.
  9
10      Mauderly, JL; Benson, JM; Bice, DE; Henderson, RF; Jones, RK; McClellan, RO; Mokler, BV; Pickrell, JA;
11       Redman, HC; Wolff, RK. (1981) Observations on rodents exposed for 19 weeks to diluted diesel exhaust  In:
12      Inhalation Toxicology Research Institute annual report 1980-1981. Albuquerque, NM: Lovelace Biomedical and
13      Environmental Research Institute; report no. LMF-91; pp. 305-311.
14
15      Mauderly, JL; Benson, JM; Bice, DE; Carpenter, RL; Evans, MJ; Henderson, RF; Jones, RK; McClellan, RO;
16      Pickrell, JA; Redman, HC; Shami, SG; Wolff, RK. (1983) Life-span study of rodents inhaling diesel exhaust: results
17      through 30 months. In: Inhalation Toxicology Research Institute annual report 1982-1983. Albuquerque, NM:
18      Lovelace Biomedical and Environmental Research Institute; report no. LMF-107.
19
20      Mauderly, JL; Benson, JM; Bice, DE; Carpenter, RL; Evans, MJ; Henderson, RF; Jones, RK; McClellan, RO;
21       Pickrell, JA; Shami, SG; Wolff, RK. (1984) Life-span study of rodents inhaling diesel exhaust: effect on body weight
22      and survival. Albuquerque, NM: Lovelace Biomedical and Environmental Research Institute; report no. LMF-113;
23      pp. 287-291.
24
25      Mauderly, JL; Jones, RK; Griffith, WC; Henderson, RF; McClellan,  RO. (1987a) Diesel exhaust is a pulmonary
26      carcinogen in rats exposed chronically by inhalation. Fundam Appl Toxicol 9:208-221.
27
28      Mauderly, JL; Bice, DE; Carpenter, RL; Gillett, NA; Henderson, RF; Pickrell, JA; Wolff, RK. (1987b) Effects of
29      inhaled nitrogen dioxide and diesel exhaust on developing lung. Cambridge, MA: Health Effects Institute; research
30      report no. 8.
31
32      Mauderly, JL; Gillett, NA; Henderson, RF; Jones, RK; McClellan, RO. (1988) Relationships of lung structural and
33      functional changes to accumulation of diesel exhaust particles. In: Dodgson, J; McCallum, RI; Bailey, MR; Fisher,
34      DR, eds. Inhaled particles VI: proceedings of an international symposium and workshop on lung dosimetry;
35      September 1985; Cambridge, United Kingdom. Ann Occup Hyg 32(suppl. l):659-669.
36
37      Mauderly, JL; Bice, DE; Cheng, YS; Gillett, NA; Griffith, WC; Henderson, RF; Pickrell, JA; Wolff, RK. (1990a)
38      Influence of preexisting pulmonary emphysema on susceptibility of rats to diesel exhaust. Am Rev Respir Dis
39      141:1333-1341.
40
41       Mauderly, JL; Cheng, YS; Snipes, MB. (1990b) Particle overload in  toxicology studies: friend or foe and how can
42      we know? In: Proceedings of the particle lung interaction symposium; May; Rochester, NY.
43
44      McClellan, RO; Bice, DE; Cuddihy, RG; Gillett, NA; Henderson, RF; Jones, RK; Mauderly, JL; Pickrell, JA; Shami,
45      SG; Wolff, RK. (1986) Health effects of diesel exhaust. In: Lee, SD; Schneider, T; Grant, LD; Verkerk,  PJ, eds.
46      Aerosols: research, risk assessment and control strategies: proceedings of the second U.S.-Dutch international
47      symposium; May 1985; Williamsburg, VA. Chelsea, MI: Lewis Publishers, Inc.; pp. 597-615.
48
49      Meiss, R; Robenek, H; Schubert, M; Themann, H; Heinrich, U. (1981) Ultrastructural alterations in the livers of
50      golden hamsters following experimental chronic inhalation of diluted diesel exhaust emission. Int Arch Occup
51       Environ Health 48:147-157.
52
53      Mentnech, MS; Lewis, DM; Olenchock, SA; Mull, JC; Roller, WA; Lewis, TR. (1984) Effects of coal dust and
54      diesel exhaust on immune competence in rats. J Toxicol Environ Health 13:31-41.
55


         2/1/98                                         5-93        DRAFT-DO NOT CITE OR  QUOTE

-------
  1       Misiorowski, RL; Strom, KA; Vostal, JJ; Chvapil, M. (1980) Lung biochemistry of rats chronically exposed to diesel
  2       participates. In: Pepelko, WE; Danner, RM; Clarke, NA, eds. Health effects of diesel engine emissions: proceedings
  3       of an international symposium; December 1979. Cincinnati, OH: U.S. Environmental Protection Agency, Health
  4       Effects Research Laboratory; pp. 465-480; EPA report no. EPA-600/9-80-057a. Available from: NTIS, Springfield,
  5   '    VA;PB81-173809.
  6
  7       Misiorowski, RL; Strom, KS; Vostal, JJ; Tillema, L; Chvapil, M. (1981) Collagen parameters in the lung of rats
  8       chronically exposed to diesel particulates. Toxicologist 1:75.
  9
10       Moorman, WJ; Clark, JC; Pepelko, WE; Mattox, J. (1985) Pulmonary Sanction responses in cats following long-term
11       exposure to diesel exhaust. JAppl Toxicol 5:301-305.
12
13       Morrow, PE. (1988) Possible mechanisms to explain dust overloading of the lungs. Fundam Appl Toxicol
14       10:369-384.
15
16       Navarro, C; Charboneau, J; McCauley, R. (1981) The effect of in vivo exposure to diesel exhaust on rat hepatic and
17       pulmonary microsomal activities. J Appl Toxicol 1:124-126.
18
19       Nikula, KJ; Snipes, MB; Barr, EB; Griffith, WC; Henderson, RF; Mauderly, JL. (1994) Influence of particle-
20       associated organic compounds on the carcinogenicity of diesel exhaust. In: Mohr, U; Dungworth, DL; Mauderly, JL;
21       Oberdorster, G, eds. Toxic and carcinogenic effects of solid particles in the respiratory tract [proceedings of the 4th
22       international inhalation symposium]; March 1993; Hannover, Germany. Washington, DC: International Life
23       Sciences Institute Press; pp. 565-568.
24
25       Nikula, KJ; Snipes, MB; Barr, EB; Griffith, WC; Henderson, RF; Mauderly, JL. (1995) Comparative pulmonary
26       toxicities and carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl
27       Toxicol 25:80-94.
28
29       Nikula, KJ; Avila, KJ; Griffith, WC; Mauderly JL. (1997) Lung  tissue responses and sites of particle retention differ
30       between rats and Cynomolgus monkeys exposed chronically to diesel exhaust and coal dust. Fundam Appl Toxicol
31       37:37-53.           '                                                                                  ,
32
33       Oberdorster, G; Yu, CP. (1991) The carcinogenic potential of inhaled diesel exhaust: a particle effect? J Aerosol Sci
34       21(suppl. 1):S397-S401.
35
36       Odor Panel of the CRC-APRAC Program Group on Composition of Diesel Exhaust. (1979) Development and
37       evaluation of a method for measurement of diesel exhaust odor using a diesel odor analysis system (DOAS). New
38       York, NY: Coordinating Research Council, Inc., Air Pollution Research Advisory Committee; CRC-APRAC project
39       no. CAPI-1-64.                                                 '
40
41       Orthoefer, JG; Stara, JF; Yang, YY; Campbell, KI. (1980) Clinical summary of beagle study. Physical examinations,
42       ocular examinations, hematology examination (post-exposure) and blood chemistries. In: Stara, JF; Dungworth, DL;
43       Orthoefer, JG; Tyler, WS, eds. Long-term effects of air pollutants:  in canine species. Cincinnati, OH: U.S.
44       Environmental Protection Agency, Office of Health and Environmental Criteria, Environmental Criteria and
45       Assessment Office; pp. 55-70; EPA report no. EPA-600/8-80-014. Available from: NTIS, Springfield,  VA;
46       PBS 1-144875.
47
48       Pattle, RE; Stretch, H; Burgess, F; Sinclair, K; Edginton, JAG. (1957) The toxicity of fumes from a diesel engine
49       under four different running conditions. Br J Ind Med 14:47-55.
50
51       Penney, DG; Baylerian, MS; Fanning, KE; Thill, JE; Yedavally, S; Fanning, CM. (1981) A study of heart and blood
52       of rodents inhaling diesel engine exhaust particulates. Environ Res 26:453-462.
53
54       Pepelko, WE. (1982a) Effects of 28 days exposure to diesel engine emissions in rats. Environ Res 27:16-23.
55


         2/1/98                                         5-94         DRAFT-DO NOT CITE OR QUOTE

-------
  1      Pepelko, WE. (1982b) EPA studies on the toxicological effects of inhaled diesel engine emissions. In: Lewtas, J, ed.
  2      Toxicological effects of emissions from diesel engines: proceedings of the Environmental Protection Agency diesel
  3      emissions symposium; October 1981; Raleigh, NC. (Developments in toxicology and environmental science: v. 10).
  4      New York: Elsevier Biomedical; pp.  121-142.
  5
  6      Pepelko, WE; Peirano, WB. (1983) Health effects of exposure to diesel engine emissions: a summary of animal
  7      studies conducted by the U.'S, Environmental Protection Agency's Health Effects Research Laboratories at
  8      Cincinnati, Ohio. J Am Coll Toxicol 2:253-306.
  9
10      Pepelko, WE; Mattox, JK; Yang, YY; Moore, W, Jr. (1980a) Pulmonary function and pathology in cats exposed 28
11      days to diesel exhaust. J Environ Pathol Toxicol 4:449-45 8.
12
13      Pepelko, WE; Mattox, J; Moorman, WJ; Clark, JC. (1980b) Pulmonary function evaluation of cats after one year of
14      exposure to diesel exhaust. In: Pepelko, WE; Danner, RM; Clarke, NA, eds. Health effects of diesel engine
15      emissions: proceedings of an international symposium, v. 2; December 1979; Cincinnati, OH. Cincinnati, OH: U.S.
16      Environmental Protection Agency, Health Effeqts Research Laboratory; pp. 757-765; EPA report no.
17      EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.
18
19      Pepelko, WE; Mattox, J; Moorman, WJ; Clark, JC. (1981) Pulmonary function evaluation of cats after one year of
20      exposure to diesel exhaust. Environ Int 5:373-376.
21
22      Pereira, MA; Sabharwal, PS; Gordon, L; Wyrobek, AJ. (1981) The effect of diesel exhaust on sperm-shape
23      abnormalities in mice. Environ Int 5:459-460.
24
25      Plopper, CG; Hyde, DM; Weir, AJ. (1983) Centriacinar alterations hi lungs of cats chronically exposed to diesel
26      exhaust. Lab Invest 49:391-399.
27
28      Purdham, JT; Holness, DL; Pilger, CW. (1987) Environmental and medical assessment of stevedores employed in
29      ferry operations. ApplIndHyg 2:133-139.
30
31      Quinto,  I; De Marinis, E. (1984) Sperm abnormalities hi mice exposed to  diesel particulate. Mutat Res 130:242.
32
33      Rabovsky, J; Petersen, MR; Lewis, TR; Marion, KJ; Groseclose, RD. (1984) Chronic inhalation of diesel exhaust
34      and coal dust: effect of age and exposure on selected enzyme activities associated with microsomal cytochrome
35      P-45G in rat lung and liver. J Toxicol Environ Health 14:655-666.
36
37      Rabovsky, J; Judy, DJ; Rodak, DJ; Petersen, M. (1986) Influenza virus-induced alterations of cytochrome P-450
38      enzyme activities following exposure of mice to coal and diesel particulates. Environ Res 40:136-144.
39
40      Reger, R. (1979) Ventilatory function changes over a work shift for coal miners exposed to diesel emissions. In:
41      Bridbord, K; French, J, eds. Toxicological and carcinogenic health hazards in the workplace: proceedings of the first
42      annual NIOSH scientific symposium; 1978; Cincinnati, OH. Park Forest South, IL: Pathotox Publishers, Inc.; pp.
43      346-347.
44
45      Reger, R; Hancock, J; Hankinson, J; Hearl, F; Merchant, J. (1982) Coal miners exposed to diesel exhaust emissions.
46      Ann OccupHyg 26:799-815.
47
48      Research Committee for HERP Studies. (1988) Diesel exhaust and health risks: results of the HERP studies.
49      Tsukuba, Ibaraki, Japan: Japan Automobile Research Institute, Inc.  .
50
51      Rudell, B; Sandstrom, T; Stjernberb, N; Kolmodin-Hedman, B. (1990) Controlled diesel exhaust exposure in an
52      exposure chamber: pulmonary effects investigated with bronchoalveolar lavage. J Aerosol Science 21:411-414.
53              -                                        •
54      Rudell, B; Ledin, M-C; Hammerstrom, U; Stjeraberg, N; Lundback, B; Sandstrom, T. (1996) Effects on symptoms
55      and lung function in humans  experimentally exposed to diesel exhaust. Occup Environ Med 53:658-662


         2/1/98                                         5-95         DRAFT-DO NOT CITE OR QUOTE

-------
  1       Schneider, DR; Felt, BT. (1981) Effect of diesel participate exposure on adenylate and guanylate cyclase of rat and
  2       guinea pig liver and lung. J Appl Toxicol 1:135-139.
  3
  4       Schreck, RM; Soderholm,  SC; Chan, TL; Hering, WE; D'Arcy, JB; Smiler, KL. (1980) Introduction to and
  5       experimental conditions in the GMR chronic inhalation studies of diesel exhaust. In: Pepelko, WE; Danner, RM;
  6       Clarke, NA, eds. Health effects of diesel engine emissions: proceedings of an international symposium; December
  7       1979. Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp. 573-591;
  8       EPA report no. EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.
  9
10       Schreck, RM; Soderholm,  SC; Chan, TL; Smiler, KL; D'Arcy, JB. (1981) Experimental conditions in GMR chronic
11       inhalation studies of diesel exhaust. J Appl Toxicol 1:67-76.
12                                                                                .
13       Stephens, RJ; Freeman, G; Evans, MJ. (1972) Early response of lungs to low levels of nitrogen dioxide: light and
14       electron microscopy. Arch Environ Health 24:160-179.
15
16       Stober, W. (1986) Experimental induction of tumors in hamsters, mice and rats after long-term inhalation of filtered
17       and unfiltered diesel engine exhaust. In: Ishinishi, N; Koizumi, A; McClellan, RO; Stober, W, eds. Carcinogenic and
18       mutagenic effects of diesel engine exhaust: proceedings of the international satellite symposium on toxicological
19       effects of emissions from diesel engines; July; Tsukuba Science City, Japan. Amsterdam, The Netherlands: Elsevier
20       Science Publishers B. V.; pp. 421-439. (Developments in toxicology and environmental science: v. 13).
21
22       Stokinger, HE. (1977) Routes of entry and modes of action.  In: Key, MM; Henschel, AF; Butler, J; Ligo, RN;
23       Tabershaw, IR; Ede, L, eds. Occupational diseases: a guide to their recognition. Washington, DC: U.S. Department
24       of Health, Education, and Welfare, National Institute for Occupational Safety and Health; pp. 11-42; DHEW
25       (NIOSH) publication no. 77-181.
26
27       Strom, KA. (1984) Response of pulmonary cellular defenses to the inhalation of high concentrations of diesel
28       exhaust. J Toxicol Environ Health 13:919-944.
29
30       Takafuji, S; Suzuki, S; Koizumi, K;Tadokoro, K; Miyamoto, T; Dcemori, R;  Muranaka, M. (1987) Diesel-exhaust
31       particulates inoculated by the intranasal route have an adjuvant activity for IgE production in mice. J Allergy Clin
32       Immunol 79:639-645.
33
34       Takano, H; Yoshikawa, T; Ichinose, T; Miyabara, Y; Imaoka, K; Sagai, M. (1997) Diesel exhaust particles enhance
35       antigen-induced airway inflammation and local cytokine expression in mice. Am J Respir Crit Care Med 156:36-42.
36
37       Takenaka, H; Zhang, K; Diaz-Sanchez, D; Tsien, A; Saxon,  A. (1995) Enhanced human IgE production results from
38       exposure to the aromatic hydrocarbons from diesel exhaust:  Direct effects on B-cell IgE production. J Allergy Clin
39       Immunol 95:103-115.                           .
40
41       Turk, A. (1967) Selection and training of judges for sensory evaluation of the intensity and character of diesel
42       exhaust odors. Cincinnati,  OH: U.S. Department of Health, Education, and Welfare, National Center for Air
43       Pollution Control; Public Health Service publication no. 999-AP-32. Available from: NTIS, Springfield, VA;
44       PB-174707.
45
46       U.S. Environmental Protection Agency (1983) Diesel paniculate study. Ann Arbor, MI: Office of Mobile Sources.
47
48       Ulfvarson, U; Alexandersson, R; Aringer, L; Svensson, E; Hedenstiema, G;  Hogstedt, C; Holmberg, B; Rosen, G;
49       Sorsa, M. (1987) Effects of exposure to vehicle exhaust on health. Scand J Work Environ Health 13:505-512.
50
51       Vallyathan, V; Virmani, R; Rochlani, S;  Green, FHY; Lewis, T. (1986) Effect of diesel emissions and coal dust
52       inhalation on heart and pulmonary arteries of rats. J Toxicol Environ Health 19:33-41.
53
54       Vinegar, A; Carson, AI; Pepelko, WE. (1980) Pulmonary function changes in Chinese hamsters exposed six months
55       to diesel exhaust. In: Pepelko, WE; Danner, RM; Clarke, NA, eds. Health effects of diesel engine emissions:


         2/1/98                                         5-96        DRAFT-DO NOT CITE OR  QUOTE

-------
  1       proceedings of an international symposium, v, 2; December, 1979; Cincinnati, OH. Cincinnati, OH: U.S.
  2      Environmental Protection Agency, Health Effects Research Laboratory; pp. 749-756; EPA report no.
  3      EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.
  4
  5      Vinegar, A; Carson, A; Pepelko, WE; Orthoefer, JO. (1981a) Effect of six months of exposure to two levels of diesel
  6      exhaust on pulmonary function of Chinese hamsters. Fed Proc 40:593.
  7
  8      Vinegar, A; Carson, A; Pepelko, WE. (1981 b) Pulmonary function changes in Chinese hamsters exposed six months
  9      to diesel exhaust. Environ Int 5:369-371.
10
11       Vostal, JJ; Chan, TL; Garg, BD; Lee, PS; Strom, KA. (1981) Lymphatic transport of inhaled diesel particles in the
12      lungs of rats and guinea pigs exposed to diluted diesel exhaust. Environ Int 5:339-34,7.
13
14      Vostal, JJ; White, HJ; Strom, KA; Siak, J-S; Chen, K-C; Dziedzic, D. (1982) Response of the pulmonary defense
15      system to diesel paniculate exposure. In: Lewtas, J, ed. Toxicological effects of emissions from diesel engines:
16      proceedings of the Environmental Protection Agency diesel emissions symposium; October 1981; Raleigh, NC. New
17      York, NY: Elsevier Biomedical; pp. 201-221. (Developments in toxicology and environmental science: v. 10).
18
19      Wade, JF, III; Newman, L.S. (1993) Diesel asthma. Reactive airway disease following overexposure to locomotive
20      exhaust. J Occup Med 35:149-154.
21
22      Wallace, MA; Salley, SO; Barnhart, MI. (1987) Analysis of the effects of inhaled diesel exhaust on the alveolar
23      intravascular and interstitial cellular components of rodent lungs. Scanning Microsc  1:1387-1395.
24
25      Waller, RE; Commins, BT; Lawther, PJ. (1961) Air pollution in road tunnels. Br J Ind Med 18:250-259.
26
27      Wehner, AP; Dagle, GE; Clark, ML; Buschbom, RL. (1986) Lung changes in rats following inhalation exposure to
28      volcanic ash for two years. Environ Res 40:499-517.
29
30 .     Weller, MA; Chen, S-T; Barnhart, MI. (1981) Acid phosphatase in alveolar macrophages exposed in vivo to diesel
31       engine exhaust. Micron 12:89-90.
32
33      Werchowski, KM; Chaffee, V W; Briggs, GB. (1980a) Teratologic effects of long-term exposure to diesel exhaust
34      emissions (rats). Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory; EPA
35      report no. EPA-600/1-80-010. Available from: NTIS, Springfield, VA; PB80-159965.
36
37      Werchowski, KM; Henne, SP; Briggs, GB. (1980b) Teratologic effects of long-term exposure to diesel exhaust
38      emissions (rabbits). Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory;
39      EPA report no. EPA-600/1-80-011. Available from: NTIS, Springfield, VA; PB80-168529.
40
41       White, HJ; Garg, BD. (1981) Early pulmonary response of the rat lung to inhalation of high concentration of diesel
42      particles: JApplToxicol 1:104-110.
43                        '    '              .
44      -Wiester, MJ; Iltis, R; Moore, W. (1980) Altered function and histology in guinea pigs after inhalation of diesel
45      exhaust. Environ Res 22:285-297.
46
47      Wolff, RK; Gray, RL. (1980) Tracheal clearance of particles. In: Diel, JH; Bice, DE; Martinez, BS, eds. Inhalation
48      Toxicology Research Institute annual report: 1979-1980. Albuquerque, NM: Lovelace Biomedical and
49      Environmental Research Institute; p. 252; report no. LMF-84.
50
51       Wolff, RK; Henderson, RF; Snipes, MB; Griffith, WC; Mauderly, JL; Cuddihy, RG; McClellan, RO. (1987)
52      Alterations hi particle accumulation and clearance in lungs of rats chronically exposed to diesel exhaust. Fundam
53      ApplToxicol9:154-166.
54                 ,
         2/1/98    .                                     5-97        DRAFT-DO NOT CITE OR QUOTE

-------
1      Wright, ES. (1986) Effects of short-term exposure to diesel exhaust on lung cell proliferation and phospholipid
2      metabolism. Exp Lung Res 10:39-55.
3
4      Yang, HM; Ma, JYC; Castranova, V; Ma, JKH. (1997) Effects of diesel exhaust particles on the release of
5      interleukin-1 and tumor necrosis factor-alpha from rat alveolar macrophages. Exp Lung Res 23:269-284.
6
        2/1/98   -                                   5-98        DRAFT-DO NOT CITE OR QUOTE

-------
             6.  QUALITATIVE AND QUANTITATIVE ASSESSMENT OF NONCANCER
                    HEALTH EFFECTS—DERIVATION OF THE INHALATION
                                 REFERENCE CONCENTRATION

  1      6.1. INTRODUCTION
  2            Noncancer endpoints have been studied in detail in controlled laboratory animal studies
  3      of diesel exhaust, and the progression of events from initial particle deposition through chronic
  4      structural and functional alterations has been described. Some of these effects are seen early in
  5      the course of a lifetime exposure and progress throughout the lifetime of the animal in the
  6      absence of a tumor response. These findings raise the possibility of noncancer respiratory
  7      disease as a human health hazard of long-term exposure to diesel exhaust. This chapter presents
  8      a qualitative and quantitative assessment of the  toxicological data on noncancer endpoints for
  9      diesel emissions.
10            The quantitative assessment of noncancer health effects from exposure to diesel exhaust
11      emissions involves the development of an inhalation reference concentration (RfC). An RfC is
12      defined as an estimate (with uncertainty spanning perhaps an order of magnitude) of a continuous
13      inhalation exposure to the human population (including sensitive subgroups) that is likely to be
14      without appreciable risks of deleterious noneaneer effects during a lifetime. The RfC approach is
15      based on the assumption that a threshold exists for the human population below which no effect
16      will occur. The RfC is an estimate of a likely subthreshold concentration. To derive the RfC, the
17      database on toxicological effects is reviewed and the most relevant and sensitive endpoints for
18      human risk assessment are identified. The lowest-observed-adverse-effect level (LOAEL, the
19      lowest concentration producing an adverse effect) or the no-observed-adverse-effect level
20      (NOAEL, the highest concentration that did not produce any adverse effect) is used as the basis
21      for deriving the RfC.  The NOAEL (or LOAEL) for the database is selected after the human
22      equivalent concentration is calculated for the exposure regimens used in the experimental
23      studies. The NOAEL is considered to be an operational estimate of a subthreshold exposure.
24      The human equivalent concentration of the NOAEL is then divided by the uncertainty factors to
25      account for any uncertainties or data gaps necessary to extrapolate from the experimental
26      conditions to a no-adverse-effect level in a chronically and continuously exposed sensitive
27      human. Once consensus on the RfD derivation  has been reached following both external and
28      Agency peer review, the RfC is said to be verified and is made public through EPA's Integrated
29      Risk Information System (IRIS).
30            The benchmark dose/concentration (BMC) approach may also be used to derive the RfC,
31      as has been done for carbon disulfide, chlorodifluoromethane, and several other chemicals (U.S.

        2/1/98                                  6-1        DRAFT-DO NOT CITE OR QUOTE

-------
 1      EPA, 1995a). The BMC approach, applies a dose-response model to the data from key studies
 2      and then uses the dose-response relationship to interpolate an exposure concentration that is
 3      predicted to result in a predefined response level, which is termed the benchmark response
 4      (BMR), such as a 10% incidence of a lesion or a 10% change in a continuous response variable
 5      (e.g., lung weight).  The lower confidence limit on the concentration predicted to result in the
 6      BMR is the BMC, which is used as the basis for deriving the RfC. Methods for performing the
 7      BMC approach, as well as scientific consensus and Agency, policy regarding the implementation
 8      of the BMC approach in risk assessment, are under development (U.S. EPA, 1995b; Barnes et
 9      al., 1995). Benchmark analyses like the  one contained in this section for diesel exhaust may
10      serve as the basis for deriving risk assessment values, as in the cases noted above, or as a point of
11      comparison for values derived from the NOAEL/LOAEL approach.
12             The study or studies  identifying the LOAEL, NOAEL, and/or BMC selected as the basis
13      for deriving the RfC are termed the principal study or studies. The principal studies are those
14      that identify the threshold region of the concentration-response curve and are representative of
15      the entire database in this regard.  Other  studies that are pertinent to identifying the threshold for
16      the effect are termed supporting studies.  Supporting studies may provide additional evidence
17      identifying the concentration-response relationship, the relative sensitivity of various effects or
18      species, or the occurrence of other noncancer endpoints, such as reproductive or developmental
19      toxicity.  Principal and supporting studies used in deriving the RfC for diesel engine emissions
20      are discussed in Sections 6.4 and 6.5, respectively, and the derivation of the RfC is discussed in
21      Section 6.6.
22
23      6.2. DETERMINATION OF CRITICAL TARGET SITE
24             The noncarcinogenic effects of inhalation of diesel exhaust have been studied in many
25      chronic and subchronic experiments in several laboratory animal species (Chapter 5). The
26      pathogenic sequence following the inhalation of diesel exhaust as determined histopathologically
27      and biochemically begins with the phagocytosis of diesel particles by alveolar macrophages
28      (AMs). These activated AMs release chemotactic  factors that attract neutrophils and additional
29      AMs. As the lung burden of diesel particulate matter (DPM) increases, there are aggregations of
30      particle-laden AMs in alveoli adjacent to terminal bronchioles, increases in the number of Type  II
31      cells lining particle-laden alveoli, and the presence of particles within alveolar and peribronchial
32      interstitial tissues and associated lymph nodes. The neutrophils and AMs release mediators of
33      inflammation and oxygen radicals, and particle-laden macrophages are functionally altered,
34      resulting  in decreased viability and impaired phagocytosis and clearance of particles. The latter
35      series of events may result in pulmonary inflammatory, fibrotic, or emphysematous lesions.

        2/1/98                                     6-2        DRAFT-DO NOT CITE OR QUOTE

-------
  1      Studies showing these effects are described in Chapter 5.  Epidemiologic studies of people
  2      exposed in various occupations in which diesel engines are used provide suggestive evidence for
  3      a respiratory effect. Although detailed information describing the pathogenesis of respiratory
  4      effects in humans is lacking, the effects in human studies lend qualitative support to the findings
  5      in controlled animal studies.
  6            The weight of evidence from the available toxicological data on diesel exhaust indicates
  7      with high confidence that inhalation of diesel exhaust can be a respiratory hazard, based on
  8      findings in multiple controlled laboratory animal studies in several species with suggestive
  9      evidence from human occupational studies. The endpoints of concern include biochemical,
10      histological, and functional changes in the pulmonary and tracheobronchial regions. There is
11      also some evidence for effects on respiratory system-related immune function. Although there is
12      some suggestive evidence of liver and kidney changes in animals exposed to diesel exhaust, as
13      well as some indication of neurotoxic effects at high concentrations, these data are inadequate to
14      indicate that a hazard exists for these endpoints. Studies of other endpoints, including
15      reproductive and developmental toxicity, in controlled animal exposures have shown no evidence
16      of potential hazard.
17
18      6.3.  APPROACH FOR DERIVATION OF THE INHALATION REFERENCE
19-           CONCENTRATION - •
20            Twelve long-term (>1 year) laboratory animal inhalation studies of diesel engine
21      emissions have been conducted. The focus of these studies has been on the respiratory tract
22      effects in the pulmonary region. Effects in the upper respiratory tract and in other organs were
23      not found consistently in chronic animal exposures.  The research programs on the toxicology of
24      diesel emissions at the Inhalation Toxicology Research Institute (ITRI) and the Japanese Health
25      Effects Research Program (HERP) consisted of large-scale chronic exposures, with exposed
26      animals being designated for the study of various endpoints and at various time points (Ishinishi
27      et al.,  1986, 1988; Mauderly et al., 1987a,b, 1988; Henderson et al., 1988; Wolff et al, 1987).
28      Each research program is represented by multiple published accounts of results. These programs
29      were selected as the principal basis for deriving the RfC because each contains studies that
30      identify an LOAEL and an NOAEL for respiratory effects after chronic exposure (see Section
31      5.2) as well as pulmonary histopathology.
32            Diesel particulate matter is composed of an insoluble carbon core with a surface coating
33      of relatively soluble organic constituents. Because macrophage accumulation, epithelial
34      histopathology, and reduced clearance have been observed in rodents exposed to high
35      concentrations of chemically inert particles (Morrow, 1992), it appears possible that the toxicity

        2/1/98                '                    6-3        DRAFT-DO NOT CITE OR QUOTE

-------
 1      of DPM results from the carbon core rather than from the associated organics. However, the
 2      organic component of diesel particles, consisting of a large number of polycyclic aromatic
 3      hydrocarbons and heterocyclic compounds and their derivatives (Chapter 2), may also play a role
 4      in the pulmonary toxicity of DPM. It is not possible to separate the carbon core from the
 5      adsorbed organics to compare the toxicity. Therefore, the whole particle was used as the
 6      dosimeter. See Chapters 5 and 10 for further details.
 7            The use of a specific retention or physiologically based pharmacokinetic model is
 8      considered the optimum method for RfC derivation, and default approaches are described for
 9      chemicals without applicable models. A model developed by Yu and Yoon (1990) that accounts
10      for species differences in deposition efficiency, normal and particle overload lung clearance
11      rates, respiratory exchange rates, and particle transport to lung-associated lymph nodes was
12      selected for development of the RfC.  Because the dependence of mechanical alveolar clearance
13      on particle lung burden in humans is not known, it was assumed in development of the model for
14      humans that the particle overload phenomenon occurs in humans and in rats at equivalent lung
15      burdens expressed as mass per unit surface area (Yu and Yoon, 1990).  This assumption allows
16      for the development of a diesel particle-specific human retention model and therefore allows
17      extrapolation from the rat studies to human exposures. The model has not been extended to other
18      species at this time because data describing the dependence of the particle overload phenomenon
19      on lung particle burden for species other than the rat are not available.  See Chapter 4 for further
20      discussion of the model.
21            The input data required to run the dosimetric model include the particle size
22      characterization expressed as mass median aerodynamic diameter (MMAD) and the geometric
23      standard deviation (og). In the principal and supporting  studies used for the RfC derivation,
24      these parameters are measured using different methods and are reported in different levels  of
25      detail. Simulation data presented by Yu and.Xu (1986) show that across a range of MMAD and
26      og inclusive  of the values reported in these studies, the pulmonary deposition fraction differs by
27      no more than 20%. The minimal effect of even a large distribution of particle size on deposition
28      probably results because the particles are still mostly in the submicron range and deposition is
29      influenced primarily by diffusion. However, it has also been shown that the particle
30      characteristics in a diesel exhaust exposure study depend very much on the procedures used to
31      generate the  chamber atmosphere. Especially important are the volume and temperature of the
32      dilution gas,  because of the rapid coagulation of particles. The differences reported in particle
33      sizes and distributions in various studies likely reflected real differences in the exposure
34      chambers as  well as different analytical methods. Because the particle  diameter and size
35      distribution were not reported in the two lowest exposure concentrations in the  HERP studies, it

        2/1/98                •                     6-4       DRAFT-DO NOT CITE OR QUOTE

-------
  1      was decided to use a default particle size of MMAD = 0.2 um and og = 2,3 for modeling of lung
  2      burden.  For consistency, the lung burdens for the other studies were also calculated using the
  3      default particle size assumption. The difference in the human equivalent concentration using the
  4      default particle size compared with the actual reported particle size is no more than 4% in the
  5      HERP study and 19% in the ITPJ study.
  6
  7      6.4.  THE PRINCIPAL STUDIES FOR INHALATION REFERENCE
  8           CONCENTRATION DERIVATION
  9            The experimental protocol and results for the principal studies are discussed in Chapter 5
10      and Appendix A and are briefly reviewed here. In studies conducted at ITRI, rats and mice were
11      exposed to target DPM concentrations of 0, 0.35, 3.5, or 7 mg/m3 for 7 h/day, 5 days/week for up
12      to 30 mo (rats) or 24 mo (mice) (Mauderly et al., 1988). A total of 364 to 367 rats per exposure
13      level were exposed and used for various studies examining different endpoints such as
14      earcinogenicity, respiratory tract histopathology and morphometric analysis, particle clearance,
15      lung burden of DPM, pulmonary function testing, lung biochemistry, lung lavage biochemistry
16      and cytology, immune function, and lung cell labeling index.  Subsets of animals were examined
17      at 6, 12,  18, and 24 mo of exposure and surviving rats were examined at 30  mo.  Diesel
18      emissions from a 5.7-L engine operated on a Federal Test Procedure urban driving cycle were
19      diluted and fed into the exposure chambers. Particle concentrations were measured daily using a
20      filter sample, and weekly grab samples were taken to measure gaseous components including
21      carbon monoxide, carbon dioxide, nitrogen oxides, ammonia, and hydrocarbons. The actual
22      DPM concentrations for the low-, medium-, and high-exposure levels were 0.353,3.47, and 7.08
23      mg/m3, respectively. Mass median diameters (geometric standard deviations) determined using
24      an impactor/parallel flow diffusion battery were 0.262 (4.2), 0.249 (4.5), and 0.234 (4.4) for the
25      low-, medium-, and high-exposure groups, respectively.
26            Lung wet weight to dry weight ratio was increased significantly in the two highest
27      exposure groups. Qualitative descriptions of the histological results in the respiratory tract are
28      found in Mauderly et al. (1987a, 1988), Henderson et al. (1988), and McClellan et al. (1986).
29      Aggregates of particle-laden AMs were seen after 6 mo in rats exposed to 7 mg/m3 DPM target
30      concentrations, and after 1 year of exposure histological changes were seen, including, focal areas
31      of epithelial metaplasia. Fibrosis and metaplasia increased with increasing duration of exposure
32      and were observable in the 3.5 and 7 mg/m3 groups of rats at 24 mo.  Changes in the epithelium
33      included extension of bronchiolar cell types into the alveoli. Focal thickening of the alveolar
34      septa was also observed.  Histological effects were seen in areas near aggregations of particle-
35      laden AMs. The severity of inflammatory responses and fibrosis was directly related to the

        2/1/98                                    6-5       DRAFT-DO NOT CITE OR QUOTE

-------
 1      exposure level. In the 0,35 mg/m3 group of rats, there was no inflammation or fibrosis.
 2      Although the mouse lungs contained higher lung burdens of DPM per gram of lung weight at
 3      each equivalent exposure concentration, there was substantially less inflammatory reaction and
 4      fibrosis than was the case in rats. Fibrosis was observed only in the lungs of mice exposed at 7
 5      mg/m3 DPM and consisted of fine fibrillar thickening of occasional alveolar septa.
 6            Groups of 16 rats and mice (8/sex) were subjected to bronchoalveolar lavage after 6,  12,
 7      18, and 24 (rats only) mo of exposure (Henderson et al., 1988). Lung wet weights were increased
 8      at 7 mg/m3 in mice and rats at all time points and in mice at 3.5 mg/m3 at all time points after 6
 9      mo. An increase in lavagable neutrophils, indicating an inflammatory response in the lung, was
10      seen at 3.5 and 7 mg/m3 in rats and mice at most time points. An increase in protein content of
11      the bronchoalveolar lavage fluid was observed in rats exposed to 3.5 or 7 mg/m3 at 12 and 18
12      mo but not at 24 mo. Increased protein content was also seen in mice at the two higher
13      concentrations at all time points. Increases in lavage fluid content of lactate dehydrogenase,
14      glutathione reductase, p-glucuronidase, glutathione, and hydroxyproline were observed in rats
15      and mice exposed to 3.5 or 7 mg/m3 at various time points. At the lowest exposure level, no
16      biochemical or cytological changes occurred in the lavage fluid or in lung tissue in either Fischer
17      344 rats or CD-I mice.
18            Mauderly et al. (1988; see also McClellan et al., 1986) examined the impairment of
19      respiratory function in rats exposed according to the protocol described above. After 24 mo of
20      exposure to 7 mg/m3 DPM, mean TLC, Cdyn, quasi-static chord compliance, and CO diffusing
21      capacity were significantly lower than control values, and nitrogen washout and percentage of
22      forced vital capacity expired in 0.1 s were significantly greater than control values.  There was no
23      evidence of airflow obstruction.  Similar functional alterations were observed in the rats exposed
24      to 3.5 mg/m3 DPM, but such changes usually occurred later in the exposure period and were
25      generally less pronounced. There were no significant decrements in pulmonary function for the
26      0.35 mg/m3 group  at any time during the study.
27            Wolff et al. (1987) investigated alterations in particle clearance from the lungs of rats in
28      the ITRI study. Progressive increases in lung burdens were observed over time in the 3.5 and 7.0
29      mg/m3 exposure groups. There were significant increases in  16-day clearance half-times of
30      inhaled radiolabeled particles of gallium oxide (0.1 um MMAD) as early as 6 mo at the 7.0
31      mg/m3 level and 18 mo at the 3.5 mg/m3 level; no significant changes were seen at the 0.35
32      mg/m3 level. Rats that inhaled fused aluminosilicate particles (2 |im MMAD) radiolabeled with
33      cesium after 24 mo of diesel exhaust exposure showed increased clearance half-times in the 3.5
34      and 7.0 mg/m3 groups.
        2/1/98                                     6-6        DRAFT-DO NOT CITE OR QUOTE

-------
  1             In the HERP studies, histopathological effects of diesel exhaust on the lungs of rats were
  2      investigated,.(Ishinishi et ai, 1986,1988).  In this study, both light-duty (LD, 1.8-L) and heavy-
  3      duty (HD, 11 -L) diesel engines were operated under constant velocity and load conditions. The
  4      exhaust was diluted to achieve target concentrations of 0.1 (LD only), 0.4 (LD and HD), 1 (LD
  5      and HD), 2 (LD and HD), and 4 (HD only) mg/m3 DPM. Particle concentrations were
  6      determined by filter samples. Actual concentrations were 0.11,0.41,1.18, and 2.32 mg/m3 for
  7      the light-duty engine and 0.46,0.96,1.84, and 3.72 mg/m3 for the heavy-duty engine. Fischer
  8      344 rats (120 males and 95 females per exposure level for each engine type) were exposed for 16
  9      h/day, 6 days/week for 30 mo. Particle size distributions were determined using an Andersen
10      cascade impactor and an electrical aerosol analyzer. At the 24-mo sampling, the MMAD and
11      distribution (og) were 0.22 (2.93) and 0.19 (2.71) for the light-duty engine groups at 2.32 and
12      1.18 mg/m3, respectively, and 0.27 (3.18) and 0.22 (2.93) for the heavy-duty engine groups at
13      3.72 and 1.84 mg/m3, respectively (Ishinishi et al, 1988). The number and timing of the samples
14      are not clear from the published reports, nor is it clear which method was used for the results
15      reported above. Particle size data were not reported for the other exposure groups.  Hematology,
16      clinical chemistry, urinalysis, and light and electron microscopic examinations were performed.
17      The body weight  of females exposed to 4 mg/m3 DPM was 15% to 20% less than that of controls
18      throughout the study.  No histopathological changes were observed in the lungs of rats exposed
19      to 0.4 mg/m3 DPM or less. At concentrations above 0.4 mg/m3 DPM, accumulation of particle-
20      laden AMs was observed. In areas of AM accumulation, there was bronchiolization of the
21      alveolar ducts, with bronchiolar  epithelium replacing alveolar epithelium.  Proliferation of
22      brochiolar epithelium and Type II cells was observed. In these areas, edematous thickening and
23      fibrosis of the alveolar septum were seen.  Fibrosis of the alveolar septum developed into small
24      fibrotic lesions. These lesions are collectively referred to as hyperplastic lesions by the authors
25      and their incidence is reported. From a total of 123 to 125  animals examined (approximately
26      equal numbers of males and females), hyperplastic lesions were reported in 4,4, 6,12, and 87
27      animals in the light-duty engine  groups exposed to 0, 0.11,0.41,1.18, and 2.32 mg/m3 DPM,
28      respectively, and  in 1, 3, 7,14, and 25 animals in the heavy-duty engine groups exposed to 0,
29      0.46, 0.96, 1.84, and 3.72 mg/m3 DPM, respectively. Statistical analysis of these results was not
30      reported, but there was no difference in the severity ascribed to changes in pulmonary pathology
31      at similar exposure concentrations between the LD and the HD series.
32            The ITRI and HERP studies are complementary for identifying the critical effect and its
33      LOAEL and NOAEL.  The ITRI study provides results on many different endpoints reflecting
34      pulmonary toxicity, and the effect levels are the same, but the LOAEL and NOAEL are different
35      by a factor of 10.  In the HERP study, the concentrations differ by a factor of 2-4, but only

        2/1/98                                    6-7        DRAFT-DO NOT CITE OR QUOTE

-------
 1     histopathology is reported. Taken together, these two studies (including several published
 2     reports for the ITRI study) provide good definition of the low-concentration effects of diesel
 3     emissions.
 4            The HERP study identifies LOAELs for rats exposed chronically at 1.18 and 0.96 mg/m3
 5     (actual exposure) for the LD and HD series, respectively, and NOAELs at 0.41  and 0.46 mg/m3
 6     (actual) for the LD and HD series. The ITRI studies identify a NOAEL for biochemical,
 7     histological, and functional changes in the pulmonary region at 0.35 mg/m3 (LOAEL = 3.5
 8     mg/m3).  The human equivalent concentrations (HECs) for the principal studies were obtained
 9     using the deposition and retention model of Yu and Yoon (1990), as discussed previously. The
10     HEC calculation is based on the assumption that the estimate for the human exposure scenario
11     (a 70-year continuous exposure) should result in an equivalent dose metric, expressed as mass of
12     diesel particle carbon core per unit of pulmonary region surface area, to that associated with no
13     effect at the end of the 2-year rat study.  To obtain the HEC, the lung burden in the rat study is
14     calculated using the exposure regimen (concentration, number of hours per day, and days per
15     week) and values for rat tidal volume, functional residual capacity, and breathing frequency.  A
16     continuous human exposure resulting in the same final lung burden is calculated and is the HEC.
17     The HEC values corresponding to the animals' exposure levels in the principal studies are shown
18     in Table  6-1, along with a designation of the concentrations as AEL (adverse-effects level) or
19     NOAEL; the LOAELs (HEC) are 0.30, 0.36, and 0.36 mg/m3. These values, along with the
20     LOAELs from other studies (discussed below), show strong support for an experimental
21     threshold in rats in the range of 0.15 to 0.3 mg/m3 DPM.  The highest NO AEL  (HEC), which is
22     below all LOAELs (HEC), is 0.155 mg/m3 DPM from the HERP heavy-duty diesel study. This
23     NOAEL  (HEC) is selected as the basis for the RfC calculation.
24
25     6.5. SUPPORTING STUDIES FOR INHALATION REFERENCE CONCENTRATION
26          DERIVATION
27            Chronic inhalation studies using male F344 rats and male Hartley guinea pigs were
28     carried out at the General  Motors (GM) Research Laboratories (Barnhart et al., 1981, 1982).
29     Exposures to target concentrations of 0.25, 0.75, and 1.5 mg/m3 DPM were generated 20 h/day,
30     5.5 days/week for up to 2  years. Exposures at 0.75 and 1.5 mg/m3 for 2 weeks to 6 mo were
31     reported  by Barnhart et al. (1981, 1982).  The focus of these studies is on electron micrographic
32     morphometry, and very little descriptive light microscopic histology is reported. These data
33     show that no appreciable changes in morphometric parameters occurred after a 2-year exposure
34     to 0.25.mg/m3, while exposure to 0.75 or 1.5 mg/m3 DPM resulted in increased thickness of
35     alveolar  septa and increased number of various types of alveolar cells. Increased numbers of

       2/1/98                                    6-8        DRAFT-DO NOT CITE OR QUOTE

-------
          Table 6-1. Human equivalent continuous concentrations from the principal studies
Study
HERP-Light Duty



HERP-Heavy Duty



ITRI


Exposure
concentration
(mg/m3)
0.11
0.41.
1.18
2.32
0.46
0.96
1.84
' 3.72
0.353
3.47
7.08
AEL/NOAEL*
NOAEL
NOAEL
AEL
AEL
NOAEL
AEL
AEL
AEL
NOAEL
AEL
AEL
HEC"
(mg/m3)
0.038
0.139
0.359
0.571
0.155
0.303
0.493
0.911
0.042
0.360
0.582
       "AEL: adverse-effects level; NOAEL: no-observed-adverse-effect level.
       bHEC: human equivalent concentration.
 1     PMNs and monocytes were lavaged from rats exposed to 0.75 or 1.5 mg/m3, and biochemical
 2     changes occurred in lung tissue at these concentrations (Misiorowski et al, 1980; Eskelson et al.,
 3     1981; Strom, 1984). These studies demonstrate an LOAEL of 0.796 mg/m3 DPM and a NOAEL
 4     of 0.258 mg/m3 DPM for male guinea pigs in a chronic study for respiratory endpoints, including
 5     light and electron microscopy, lavage cytology, and lung tissue biochemistry.
 6            A 15-mo inhalation study was performed by Southwest Research Institute for General
 7     _Motors (Kaplan et al., 1983). Male F344 rats, Syrian golden hamsters, and A/J mice were
 8     exposed to diluted diesel exhaust at target concentrations of 0.25,0.75, and  1.5 mg/m3 for 20
 9     h/day and 7 days/week. Focal accumulation of particle-laden AMs was associated with minimal
10     to mild fibrosis of the alveolar wall. Based on accumulation of particle-laden macrophages, this
11     study identifies an LOAEL at 0.735 mg/m3 and an NOAEL at 0.242 mg/m3.
12            In a study performed by NIOSH (Lewis et al., 1986,1989; Green et al., 1983), male and
13     female F344 rats and male Cynomolgus monkeys were exposed to target levels of 2 mg/m3 diesel
14     particles. Accumulations of black-pigmented alveolar macrophages were seen in the alveolar

       2/1/98                                   6-9        DRAFT-DO NOT CITE OR QUOTE

-------
 1      ducts of rats adjacent to terminal bronchioles, and epithelial lining cells adjacent to collections of
 2      pigmented macrophages showed marked Type II cell hyperplasia. No evidence of impaired
 3      pulmonary function as a result of the exposure to diesel exhaust was found in rats.  Histological
 4      examination of lung tissue from monkeys exposed for 24 mo in the same regimen used for rats
 5      revealed aggregates of black particles, principally in the distal airways of the lung. Fibrosis,
 6      focal emphysema, or inflammation was not observed. The monkeys exposed to diesel exhaust
 7      demonstrated small airway obstructive disease.  This study demonstrates an LOAEL for rats and
 8      monkeys at a diesel particle concentration of 2 mg/m3. Although the data suggest that the
 9      pulmonary function effect in primates more closely resembles that in humans, this study had only
10      one exposed group, making evaluation of dose response impossible. Thus, it was not considered
11      sufficient to eliminate consideration of the strong rodent database.
12            Heinrich et al. (1986; see also Stober, 1986) exposed male and female Syrian golden
13      hamsters, female NMRI mice, and female Wistar rats to diesel engine emissions with a
14      4.2 mg/m3 particulate concentration. Lung weights were increased by a factor of 2 or 3 in rats
15      and mice after 2 years of exposure, and in hamsters the lung weights were increased by 50% to
16      70%. 'Although histological examination revealed different levels of response among the three
17      species, histological effects were seen in all species and effects on pulmonary function were
18      observed in rats  and hamsters. This study demonstrates an LOAEL of 4.2 mg/m3 in rats for
19      respiratory system effects.
20            The effects of diesel exhaust on the lungs of 18-week-old male Wistar rats exposed to 8.3
21      ± 2.0 mg/m3 particulate matter were investigated by Karagianes et al. (1981). Histological
22      examinations of lung tissue noted focal aggregation of particle-laden alveolar macrophages,
23      alveolar histiocytosis, interstitial fibrosis, and alveolar emphysema.  Lesion severity was related
24      to length of exposure. No exposure-related effects were seen in the nose, larynx, or trachea.
25      This study demonstrates an LOAEL of 8.3 mg/m3 DPM for respiratory effects after chronic
26      exposure of rats to diesel emissions.
27            Lung function was studied in adult cats chronically exposed to diesel exhaust
28      concentrations of'6.34 mg/m3 for the first 61 weeks and 6.7 mg/m3 from weeks 62 to 124. No
29     ' definitive pattern of pulmonary function changes was observed following 61 weeks of exposure;
30      however, a classic pattern of restrictive lung disease was found at 124 weeks (Pepelko et al.,
31      1980),
32            Heinrich et al. (1995) exposed Wistar rats to diesel exhaust at DPM concentrations of 0.8,
33      2.5, and 7 mg/m3, 18 h/day, 5 days/week for 24 mo. Body weights were significantly decreased
34      in the two higher exposure groups. Brpnchoalveolar hyperplasia and interstitial fibrosis of
        2/1/98                                    6-10       DRAFT-DO NOT CITE OR QUOTE

-------
 1      increasing incidence and severity at greater concentrations were seen in all exposure groups.
 2      This study demonstrates an LOAEL of 0.8 mg/m3.
 3            Nikula et al. (1995) exposed Fischer 344 rats to diesel exhaust at DPM concentrations of
 4      2.4 and 6.3 mg/m3 16 h/day, 5 days/week for 23 mo. Survival was decreased in the high-
 5 •     exposure males, while body weights were reduced in both males and females in the high-
 6      exposure group. Pulmonary hyperplasia, inflammation, and fibrosis were seen in a high
 7      percentage of rats in both exposure groups. The high exposure concentrations precluded use of
 8      this study for the development of an RfC.
 9            Werchowski et al. (1980a) reported a developmental study in rabbits exposed on days 6
10      through  18 of gestation to a l-in-10 dilution of diesel exhaust (DPM concentration -12 mg/m3).
11      Exposure to diesel emissions had no effect on maternal toxicity or on the developing fetuses. In
12      a companion study (Werchowski et al., 1980b), 20 SD rats were exposed for 8 h/day during days
13      5 to 16 to a target concentration of 12 mg/m3 of DPM.  Fetuses were examined for external,
14      internal, and skeletal malformations, and the number of live and dead fetuses, resorptions,
15      implants, corpora lutea,  fetal weight, litter weight, sex ratio, and maternal toxicity were recorded.
16      No conclusive evidence of developmental effects was observed in this study.
17            In an EPA-sponsored reproductive study summarized by Pepelko and Peraino (1983),
18      CD-I mice were exposed to a target concentration of 12 mg/m3 DPM for 8 h/day and
19      7 days/week.  The F0 and F, animals were exposed for 100 days prior to breeding, and
20      100 mating pairs were randomly assigned to four exposure groups of 25 each. Viability counts
21      and pup  weights were recorded at 4, 7, and 14 days after birth and at weaning. No treatment-
22      related effects on body weight in F0 mice or in F, animals through weaning or in mating animals
23      through  gestation were found. No treatment-related effects on gestation length, percent fertile,
24      litter size, or pup survival were observed. The only organ weight difference was an increase in
25      lung weight in exposed F0 and F, mice (lung weight and lung weight/body weight) and in F2
26      males (lung weight/body weight). Based on this study, an NOAEL for reproductive effects in
27      rats is identified at 12 mg/m3 DPM.
28            The reproductive and developmental studies described in Chapter 5 show that the effects
29      in the respiratory system are the most sensitive effects that result from diesel exhaust exposures.
30      These studies add to the confidence that a variety of noncancer effects have been studied and are
31      required for a designation of high confidence in the database and the RfC.
32            Several epidemiologic studies have evaluated the effects of chronic exposure to diesel
33      exhaust on occupationally exposed workers. The human studies, taken together, are suggestive
34      but inconclusive of an effect on pulmonary function, as described in Chapter 5. The studies are
35      not directly useful for deriving the RfC because of inadequate ability to directly relate  the

        2/1/98                                ,6-11        DRAFT-DO NOT CITE OR QUOTE

-------
 1      observed effects to known concentrations of DPM. The studies are confounded by coexposures
 2      to other particles or by a lack of measurement of particle exposure.
 3
 4      6.5.1.  Respiratory Tract Effects in Species Other Than the Rat
 5            In several of the chronic inhalation studies described in Chapter 5, one or more species
 6      other than the rat were also exposed and examined for toxic effects. These should provide a
 7      basis for comparison of the effects in rats with the effects in other species.  In the study
 8      performed at ITRI (Henderson et al., 1988; Mauderly et al., 1988), male and female CD-I mice
 9      were exposed similarly to the rats. The LOAEL and NOAEL in rats and mice from this study
1.0      would be the same, with the NOAEL for respiratory tract effects being 0.35 mg/m3 DPM
11      (duration-adjusted NOAEL is 0.074 mg/m3), although some differences in the severity of the
12      effect were apparent.
13            In the study conducted by the GM Biomedical Science Department (Barnhart et al., 1981,
14      1982; Strom, 1984; Gross, 1981), male Hartley guinea pigs as well as F344 rats were chronically
15      exposed to 0.258,0.796, and 1.53 mg/m3 DPM. The evidence from this study leads to the
16      conclusion that the LOAEL and NOAEL for rats and guinea pigs are the same, although
17      important differences in the endpoints were reported in the two species. The NOAEL is 0.258
18      mg/m3 (duration adjusted NOAEL is 0.17 mg/m3).
19            Kaplan et al. (1982) reported a subchronic study in F344 rats, A/J mice, and Syrian
20      golden hamsters exposed to 1.5 mg/m3 DPM. The histological observations, including AM
21      accumulation and associated thickening  of the alveolar wall, were described together, with no
22      distinction between species, suggesting that the observed effects were similar in the species
23      examined. Kaplan et al. (1983) reported a 15-mo study in which F344 rats, A/J mice, and Syrian
24      golden hamsters were exposed to 0.25, 0.75, or 1.5 mg/m3 DPM. No exposure-related lesions
25      were found in tissues other than the respiratory tract.  Based on particle-laden AM accumulation,
26      this study identifies an LOAEL at 0.735 mg/m3 and an NOAEL at 0.242 mg/m3. The
27      descriptions provided suggest that the pulmonary effects were similar across the three species
28      examined, but this conclusion is compromised by the lack of detailed reporting and the
29      possibility of intercurrent infection in rats and poor animal health (as evidenced by poor growth)
30      in hamsters.  The duration adjusted NOAEL is 0.202 mg/m3.
31            Lewis et al. (1986,1989) exposed rats and monkeys to 2 mg/m3 DPM for 2 years and
32      reported pulmonary function and histopathology.  Pulmonary function was affected in both
33      species, although with a different pattern of response, as discussed in Chapter 5. Significant
34      differences were observed in the histopathological response. In monkeys, slight particle
35      accumulation was observed, but no fibrosis, focal emphysema, or inflammation was present. Rat

        2/1/98                                   6-12        DRAFT-DO NOT CITE OR QUOTE

-------
  1      lungs in this experiment showed AM accumulation, multifocal histiocytosis, and associated
  2      fibrosis and inflammatory cells in the interstitium.
  3             Heinrich et al. (1986) exposed Wistar rats, Syrian golden hamsters, and NMRI mice
  4      chronically to 4 mg/m3 DPM. Lung weight was increased two-fold in mice, 1.5-fold in hamsters,
  5      and three-fold in rats. The activity of enzymes recovered in bronchoalveolar lavage was
  6      increased to roughly the same extent in rats, mice, and hamsters. Hamsters showed thickened
  7      alveolar septa and slight epithelial hyperplasia, with no AM accumulation.  Mice also showed
  8      epithelial hyperplasia and interstitial fibrosis. Rat lungs had severe inflammatory changes,
  9      thickened alveolar septa, hyperplasia, and metaplasia.  This study presents the clearest indication
10      of a possibly greater severity in rats compared with other rodent species for noncancer effects. It
11      also suggests that the effect in rats may be qualitatively different, with AM accumulation playing
12      a greater role in pathogenesis in rats than in other rodent species.
13             Heinrich et al. (1995) also compared effects of chronic diesel exposure on rats and two
14      strains of mice exposed to fairly high concentrations of diesel particles. Similar lung burdens
15      were reported in rats and mice on the basis of particle mass per unit lung wet weight.  Lung
16      weight was increased to about the same extent in rats and mice. However, the study is focused
17      on cancer effects, and insufficient information is provided to make a detailed comparison of
18      noncancer histopathology in rats and mice.
19.            Several of the studies described above and in Chapter 7 suggest a significant difference in
20      the carcinogenic response of rats and other experimental animal species. It is less clear whether
21      such a difference holds for noncancer effects at lower exposure levels. The studies described
22      above show similar effect levels for different species for effects that occur earlier or at lower
23      exposure concentration, including accumulation of particles, bronchoalveolar lavage
24      measurements, lung weight, and minor epithelial thickening and hyperplasia. At higher diesel
25      concentrations there are clear differences between rats and the  other species tested, especially in
26      the progression to more severe histologically observed endpoints, such as hyperplasia,
27      metaplasia, and inflammatory response.  Thus the NOAEL for chronic effects of diesel does not
28      appear to be substantially different among species, although there is some suggestion in the
29      literature of a more sensitive as well as qualitatively different response in rats. This comparison
30      is weakened by the fact that the published reports often give less emphasis to noncancer
31      responses and because the effects in rats and other species are not always measured or reported in
32      the same way.  The pathogenesis of diesel exhaust effects has not been studied as thoroughly in
33      any other species as it has in the rat. For example, no specific  measurement of particle clearance
34      from the lung has been reported in any species other than the rat. Within the resolving power of
35      the available studies,  it is concluded that there is limited evidence for a difference in the NOAEL

        2/1/98                •                    6-13       DRAFT-DO NOT CITE OR QUOTE

-------
 1     for noncancer effects across species, but the evidence is not adequate to quantitatively define the
 2     difference, especially at low exposure concentrations.  Hence there is no clearly more appropriate
 3     species on which the RfC derivation for noncancer effects should be based.
 4            Mice were included in the ITRI, Kaplan et al. (1982), and Heinrich et al. (1986, 1995)
 5     studies. The Heinrich studies used a single exposure to high concentrations and are supportive of
 6     the other results in mice but are not appropriate to define an LOAEL for mice.  The Kaplan study
 7     defines an LOAEL and NOAEL of 0.735 and 0.242 mg/m3 DPM, respectively. The duration-
 8     adjusted LOAEL and NOAEL are 0.613 and 0.202 mg/m3, respectively. The ITRI study defined
 9     the adjusted LOAEL and NOAEL at 0.723 and 0.074 mg/m3, respectively. Because the dose
10     spacing is so wide in the ITRI study, the Kaplan study is more appropriate for defining an
11     NOAEL. Likewise, the Kaplan et al. study is the only multiple-dose study in hamsters, and it
12     defines the same LOAEL and NOAEL  for hamsters as for mice.  The GM study is the only
13     chronic study in guinea pigs, and it defines the LOAEL and NOAEL for this species at 0.796 and
14     0.258 mg/m3, respectively. The adjusted LOAEL and NOAEL for guinea pigs from the GM
15     study are 0.52 and 0.17 mg/m3, respectively. The effects levels for mice, hamsters, and guinea
16     pigs are similar to the duration-adjusted LOAEL and NOAEL for rats, which are 0.723 mg/m3
17     (from ITRI study) and 0.26 mg/m3 (from Ishinishi et al., 1988), respectively. If the RfC were to
18     be derived based on the duration-adjusted NOAEL, the rat data would be preferred because of the
19     more complete database of chronic rat studies and the more complete presentation of the
20     noncancer endpoints in the rat studies.
21            The method for deriving inhalation RfCs (U.S. EPA, 1994) includes dosimetric
22     adjustments of animal exposure to arrive at a human equivalent concentration.  The default
23     calculation of an HEC for a particle exposure uses the ratio of animal to human regional
24     deposited dose (RDDR) to a specific region of the respiratory tract. The methods also allow
25     replacement of the default approach when a better model is available. The derivation of the RfC
26     in this case makes use of the Yu and Yoon (1990) model to calculate the HEC from the rat
27     studies. Since the Yu and Yoon model has been developed only for the rat-to-human
28     extrapolation, the chosen approach assumes that dosimetric differences between rats and other
29     small animal species would not result in a substantially lower HEC.  The LOAEL (HEC) and
30     NOAEL (HEC) from the rat studies based on the Yu and Yoon model are 0.36 and 0.155 mg/m3,
31     respectively.
32
33     6.6.  DERIVATION OF THE INHALATION REFERENCE CONCENTRATION
34            Studies of chronic exposures to diesel emissions performed at ITRI and HERP were
35     selected as the basis of the RfC because they identify both an NOAEL and an LOAEL for rats

       2/1/98               •                   6-14      DRAFT-DO NOT CITE OR QUOTE

-------
 1      exposed chronically.  The only other study identifying both an NOAEL and an LOAEL was the
 2      GM study, which was not used because information characterizing the pulmonary lesions in rats
 3      was limited. The availability of the dosimetric model for rats and not for other species, along
 4      with the apparent comparability between the rat and other rodent species in response, resulted in
 5      choosing the rat as the basis for developing the RfC.  Although the data from the monkey in the
 6      Lewis et al. (1989) study suggest that the pulmonary function effect in primates more closely
 7      resembles that in humans, this study had only one exposed group, making evaluation of dose
 8      response impossible.  Thus, this was not considered to be a strong enough basis to eliminate
 9      consideration of the strong rodent database. The pulmonary effects, including histological
10      lesions, biochemical changes, pulmonary function impairment, and impaired particle clearance,
11      were determined to be the critical  noncancer effect.  Sufficient documentation from other studies
12      showed that there is no effect in the extrathoracic region of the respiratory system or in other
13      organs at the lowest levels that produce pulmonary effects in chronic exposures. In addition,
14    .  adequate information is available  from the EPA studies showing no effect on development in two
15      species or on reproduction in a two-generation reproductive study.
16            Because the RfC is based on an NOAEL from a chronic animal study, uncertainty exists
17      in the extrapolation from animals  to humans and for extrapolation to sensitive members of the
18      population (inter- and intraspecies extrapolation). A default factor of 1.0 is normally applied for
19      each area of uncertainty (i.e., a total uncertainty factor of 100) when a chronic animal NOAEL is
20      available. Since a dosimetry model specifically for diesel particles is available, the use of this
21      model is considered to reduce the  uncertainty in extrapolating between animals and humans,
22      compared to a case in which no chemical or species-specific data on dosimetry are available.
23      The default uncertainty factor of 10 includes aspects of pharmacokinetics and
24      pharmacodynamics.  An uncertainty factor of 1 rather than 10 was adopted for interspecies
25      extrapolation and was used for the diesel RfC. The uncertainty factor for interspecies
26      extrapolation is normally reduced when kinetic data are available, as in this case, which reduces
27      the uncertainty in extrapolating from animals to humans by accounting for the kinetic differences
28      with data. A further reduction was considered appropriate, as recommended by peer reviewers.,
29      on the basis that substantial information suggests that the rat may be a very sensitive species,
30      compared with humans, to the effects of inhalation of diesel particles. There is some evidence,
31      as discussed above, that rats may be more sensitive than other rodent species. There is also
32      evidence from the Lewis et al. (1989) study that rats may be more sensitive than monkeys.
33      Humans would be expected to be more similar to monkeys because their respiratory tract
34      structure is more similar and because of their closer phylogenetic relationship.  In the comparison
35      of rat effects with those on other rodents and monkeys, there is also limited evidence that the

        2/1/98                                    6-15        DRAFT-DO NOT CITE OR QUOTE

-------
 1      responses are qualitatively different, with a much greater role for macrophage dysfunction and
 2      accumulation and associated epithelial effects in rats. These lines of evidence, which are fairly
 3      limited individually, lead to a reasonably strong argument for a reduction in the uncertainty factor
 4      when considered together. A total uncertainty factor of 10 results for intraspecies extrapolation.
 5             With the NOAEL (HEC) of 0.155 mg/m3 DPM from the HERP study, an RfC of 16
 6      (j,g/m3 was calculated. The RfC also includes confidence statements associated with the principal
 7      study, the database, and the resulting RfC.  The studies used as the basis for the RfC were well-
 8      conducted chronic studies with adequate numbers of animals, and the LOAELs and NOAELs
 9      were consistent across studies, thereby resulting in high confidence. The database contains
10      several chronic studies, including multiple species, that support the LOAEL observed in the
11      principal studies. There are also developmental and reproductive studies, resulting in a high-
12      confidence data base.  Because of the high confidence in the studies and database, the RfC has
13      high confidence.
14
15      6.6.1.  Application of the Benchmark Dose Approach to Derivation of the RfC
16             An alternative to deriving the RfC based on the NOAEL identified in the animal studies
17      is application of the benchmark dose/concentration approach.  The BMC was described by
18      Crump (1984) and recently discussed by EPA (1995b).  The BMC approach involves fitting a
19      dose-response function to dose and effect information from a single study and using the
20      dose-response curve to predict the dose that will result in a level of response that is defined a
21      priori as the benchmark response.  For example, a 10% increase in incidence of epithelial
22      hyperplasia might be defined as the benchmark response, and a dose-response curve relating
23      inhaled DPM to hyperplasia  in rats exposed chronically to diesel exhaust would be used to
24      estimate the exposure concentration resulting in a 10% increase.  The lower confidence limit of
25      that concentration is the BMC, and it is used as the representative value for the dose-response
26      assessment.
27             For diesel exhaust there are several  chronic exposure studies in animals that could be
28      used to estimate a benchmark concentration. Software for performing benchmark dose
29      calculations is available commercially, and benchmark dose programs also can be developed
30      using standard statistical software.  Different approaches are used for modeling dichotomous
31      versus continuous data.  Dichotomous models estimate the probability of the effect being
32      modeled at a given dose, and continuous models estimate the magnitude of the response at a
33      given dose. This fundamentally different output causes difficulty in cases such as the database
34      for diesel exhaust that include both types of data, because comparisons of the two types of results
35      must be made to select the most appropriate BMC for the dose-response assessment. Crump

        2/1/98                                    6-16       DRAFT-DO NOT CITE OR QUOTE

-------
  1      (1995) has developed an approach to modeling continuous data that models dose against the
  2      distribution of the continuous variable and estimates the probability of an abnormal response at a
  3      given dose using a preselected magnitude of response to define an abnormal response. This
  4      approach has the distinct advantage of expressing the results for dichotomous and continuous
  5      data in the same terms, but it has not been evaluated extensively and is not readily available,
  6            A first step in carrying out a BMC analysis is selecting studies and data sets that are
  7      appropriate to model. Minimum data criteria have not been clearly established.  As provisional
  8      criteria, the analyses that follow will require that complete information on the response of interest
  9      should be available and that at least two exposure levels with responses that differ from those of
10      the controls are needed. Based on this criterion, studies with a single exposure concentration
11      (including Lewis et al.5 1989; Hemrich et al., 1986; Iwai et al., 1986; Karagianes et al., 1981;
12      Pepelko et al., 1980; and the mouse data from Heinrich et al.,  1995) are not amenable to the
13      BMC approach. In addition, the rat data from Heinrich et al. (1995) and Creutzenberg et al.
14      (1990) are not amenable to BMC analysis because the information on noncancer histopathology
15      is not reported in detail and the information on lung clearance rates is reported as group means
16      with no standard deviation.
17            To perform a dose-response analysis for a continuous variable, either a measure of
18      variability or the individual animal measurements are needed. The study reported by Kaplan et
19      al. (1983) is also not considered useful for BMC analysis.  The principal result of interest in the
20      Kaplan study is reported as the incidence of pneumoconiosis in rats, mice, and hamsters. The
21      term pneumoconiosis appears to be applied to the presence of DPM in the lung and in AMs as
22      well as AM accumulation and any secondary effects in the epithelium. Epithelial effects in the
23      area near particle-laden AMs are discussed in the text of the Kaplan study but are not listed as a
24      separate entity in the tables of effects.  Thus the effect termed pneumoconiosis appears to include
25      both the expected dose-related increase in particles in the lung and any adverse effects occurring
26      secondary to particle deposition, and therefore does not represent a clearly adverse effect. The
27      existence of some studies for which the BMC approach is feasible and some for which it is not
28      feasible introduces a potentially serious difficulty in deciding on the most appropriate
29      dose-response value to use in deriving the RfC. In the case of diesel exhaust, howevers this
30      concern is reduced because the studies identified as the most appropriate for RfC derivation
31      based on NOAEL/LOAEL levels are also amenable to BMC analysis.
32            The studies for which BMC analyses were performed were the ITRI, General Motors,
33      Ishinishi (1986), and Nikulaetal. (1995) studies.  These studies contained a variety of endpoints
34      that could be modeled using the BMC approach. A total of 41 data sets were selected from these
35      studies for BMC determination. Because a variety of models are  available, and each model can

        2/1/98                                   6-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1      be varied (for example, by including a background term or a threshold term), the number of
 2      model runs for a database as complete as that for diesel exhaust can become unmanageable. In
 3      addition to this large number of endpoints and models, several dose or exposure terms could be
 4      used (e.g., for rats one could use the exposure concentration, the exposure concentration x time,
 5      the duration-average exposure concentration, the lung burden  estimated by a deposition model, or
 6      the human equivalent concentration based on a deposition and clearance model). Clearly, the
 7      number of possible model runs expands geometrically, and some decisions must be made to limit
 8      and focus the extent of the analyses.
 9            In these analyses, the dose metric used for rat data was the human equivalent continuous
10      exposure concentration based on the Yu and Yoon (1990) deposition/clearance model. For mice,
11      the duration-averaged exposure concentration was used.  The  41 data sets were modeled using
12      the polynomial model dichotomous or continuous data (Howe, 1990a, 1990b), based on extra
13      risk, and including or excluding a threshold term. The inclusion of a threshold term improved
14      model fit substantially only in cases with low-dose groups showing no response over controls. In
15      most cases a background term was also estimated, and in some analyses with dichotomous data it
16      was omitted. As expected, the inclusion of a background term improved model fit for
17      dichotomous data only when there was a nonzero response in  the controls. Based on the results
18      using the polynomial model, 12 data sets were selected for analysis with the Weibull model
19      (Howe, 1990c, 1990d) to determine the sensitivity of the BMC estimate to the choice of model.
20      The availability of different dose metrics and different models for different endpoints introduces
21      additional difficulty in determining the most appropriate BMC for deriving the RfC. This
22      difficulty is lessened by the fact that the rat data for BMC analysis are the most extensive, they
23      include several studies for which both LOAEL and NOAEL are identified, and they are also the
24      data for which the deposition and clearance model is available. As discussed previously, the data
25      from other species do not suggest large differences in species  sensitivity at low concentrations.
26            Perhaps the most critical decision in the BMC approach is the level of response defined
27      as the benchmark response. As discussed by EPA (1995b), there is an emerging consensus that a
28      BMR  of 0.05 or 0.1 is probably appropriate for dichotomous data for most endpoints. There is
29      no clear consensus about the appropriate choice of BMR or on how to select a BMR for a
30      continuous effect. This dilemma is the main reason for the appeal of models such as that
31      presented by Crump (1995) or by Gaylor and Slikker (1990),  which derive a BMC in terms of a
32      probability statement. However, both of those models require that some magnitude of response
33      be selected to delineate between "responders" and "nonresponders," so the issue remains as to
34      how one can consistently define a response level for the myriad endpoints found in the
35      toxicological literature.  One approach is to work backward from the way the BMC will be used

        2/1/98                                   6-18       DRAFT-DO NOT CITE OR QUOTE

-------
 1      to derive the RfC. The discussion by EPA (1995b) and the precedents for the use of BMC in
 2      derivation of RfCs that are now on IRIS suggest that the BMC will be used like an NOAEL has
 3      been used in the past. It follows that the BMR should be set at a level that would not be
 4      considered adverse for the effect in question, or for an effect that is very mildly adverse such that
 5      the use of the lower confidence limit on dose results in a BMC that is in the nonadverse range
 6      under the conditions of the experiment. This provides general guidance for selecting BMRs for
 7      continuous endpoints, although the issue of maintaining consistency between endpoints remains
 8      an extremely important one because of the interrelatedness of different lexicological endpoints.
 9      The benchmark responses used in the BMC analyses for diesel exhaust are shown in Table 6-2.
10             It should be noted that the lack of clear policy on selecting the BMC and the lack of
11      specific guidance for comparing different respiratory tract effects in the form of target
12      organ-specific guidelines, along with the other issues raised above, make BMC analysis a
13      possible but not automatic alternative to the derivation of the RfC as presented. The BMC
14      analysis for diesel is presented here for comparison with the RfC based on the LOAEL/NOAEL
15      from the rat studies as presented above.
16             In the ITRI study, in particular, many endpoints were amenable to benchmark modeling.
11      Henderson et al. (1988) present data on lung weight and lavage biochemistry and cytology from
18      rats and mice sacrificed at various time points during a chronic study. Wolff et al. (1987) present
19      data on particle clearance half-times in rats. The results shown in Table 6-3 are based on the
20      best-fitting exponential polynomial model using the BMR from Table 6-2 applied to the ITRI
21      data (Henderson etal., 1988; Wolff etal., 1987).
              Table 6-2. Definition of benchmark response levels for endpoints
              important in the diesel exhaust BMC analysis

        Endpoint	;	Benchmark definition        Extra risk
        Lung weight                                 10% increase                    0.10
        Bronchoalveolar lavage-number ofmaerophages 50% increase                    0.50
        Bronchoalveolar lavage-number of neutrophils   200% increase (3 x control)       2.0
        Bronchoalveolar lavage-protein                100% increase (2 x control)       1.0
        Bronchoalveolar lavage-enzymes               100% increase (2 * control)       1.0
        Clearance half-time                           20% increase                    0.20
        Body burden of a particle after 200 days         20% increase                    0.20
        Incidence of hyperplasia                       10% incidence                   0.10
        Alveolar-capillary thickness	20% increase	0.20
        2/1/98         .                           6-19       DRAFT-DO NOT CITE OR QUOTE

-------
      Table 6-3.  Results of benchmark concentration analyses using the
      polynomial model and data from the ITRI study
Data set, model
Rats
Male rat lung weight @ 24 mo
Polynomial, no threshold
Female rat lung weight @ 24 mo
Polynomial, no threshold
Rat BAL macrophages @ 24 mo, M and F combined
Polynomial, no threshold
Rat BAL neutrophils @ 24 mo, M and F combined
Polynomial, threshold
Rat BAL protein @ 1 8 mo, M and F combined
Polynomial, threshold
Rat BAL LDH @ 1 8 mo, M and F combined
Polynomial, no threshold
Rat BAL P-glucuronidase @ 24 mo, M and F combined
Polynomial, threshold
Rat Ga2O3 clearance half-time @ 6 mo, M and F combined
Polynomial, no threshold
Rat Ga2O3 clearance half-time @ 12 mo, M and F combined
Polynomial, no threshold
Rat Ga2O3 clearance half-time @ 24 mo, M and F combined
Polynomial, no threshold
Rat Cs-FAP clearance half-time @ 24 mo, M and F combined
Polynomial, no threshold
Rat Cs-FAP clearance, % of initial body burden after 200 days
Polynomial, no threshold
Mice
Male mouse lung weight @ 24 mo
Polynomial, no threshold
Female mouse lung weight @ 24 mo
Polynomial, no threshold
Mouse BAL macrophage @ 24 mo, M and F combined
Polynomial, no threshold
Mouse BAL neutrophils @ 24 mo, M and F combined
Polynomial, no threshold
Mouse BAL protein @ 1 8 mo, M and F combined
Polynomial, no threshold
Mouse BAL LDH @ 1 8 mo, M and F combined
Polynomial, no threshold
Mouse BAL P-glucuronidase @ 24 mo, M and F combined
Polynomial, threshold
Benchmark
response

10
10
50
200
100
100
100
20
20
20
20
20

10
10
10
200
100
100
100
MLE

0.10
0.061
0.315
0.0942
0.309
0.221
0.172
0.262
0.168
0.074
0.048
0.044

0.13
0.16
BMC

0.08
0.05
0.27
0.06
0.18
0.15
0.05
0.15
0.10
0.04 .
0.04
0.03

0.11
0.13
Not able to
converge
0.640
0.353
0.455
0.112
0.47
0.27
0.36
0.11
Source: Henderson et al. (1988) and Wolff et al. (1987).
2/1/98
6-20
DRAFT-DO NOT CITE OR QUOTE

-------
  1            The BMC values are the lower 95% confidence limit on the exposure concentration
  2      predicted to result in the BMR response level.. All of the responses modeled above were data
  3      presented as continuous variables, and the model was used to estimate the exposure predicted to
  4      result in a predefined response magnitude, the BMR, Without any other adjustment for the
  5      severity of the effect, it is implicit in the BMC approach for continuous data that BMRs are
  6      assumed to represent effects of equivalent severity.  Clearly, such a comparison is very subjective
  7      and cannot be made precisely. Nevertheless, the judgment of the appropriate BMR is
  8      fundamental to the application of the BMC approach.
  9            The BMCs from the ITRI rat data were calculated using the human equivalent
10      concentration of the rat exposures based on the Yu and Yoon (1990) dosimetfy model. The
11      NOAEL from the ITRI study was 0,35 mg/m3 (duration-adjusted NOAEL is 0.074 mg/m3) and
12      the human equivalent NOAEL is 0.042 mg/m3.  This NOAEL (HEC) is lower than the NOAEL
13      from the Ishinishi et al. study, which was  used to derive the RfC because the NOAEL and
14      LOAEL from the ITRI study differed by a factor of 10. The rat NOAEL used to derive the RfC
15      was a NOAEL (HEC) of 0.155 mg/m3.  The BMC values based on rat lung weight were below
16      the NOAEL from Ishinishi et al. (1986), suggesting that the lung weight is a sensitive effect. The
17      sensitivity of the lung weight is consistent with other chronic rat studies as well as those in mice
18      (Heinrich et al.,  1995). The bronchoalveolar lavage  indicators showed large variability in BMC
19      value, with macrophages, protein, and LDH having BMCs above the range of the rat NOAEL.
20      Lavagable neutrophils and p-glucuronidase, however, were quite sensitive.  Measures of particle
21      clearance in rat lungs were also sensitive,  resulting in a range of BMC values between 0.033 and
22      0.146 mg/m3. It was noted that there are no BMC estimates based on histopathological effects,
23      even though these effects tend to be sensitive indicators of diesel exposure, because these effects
24      were not reported in an adequately quantitative manner in the various publications describing the
25      ITRI study.
26            The BMC levels based on mouse data from the ITRI study also showed a wide range,
27      with lung weight and lavage P-glucuronidase being the most sensitive. The mouse BMCs were
28      calculated using the duration-averaged exposure concentration as the dose term, so the
29      appropriate NOAEL for comparison is the duration-adjusted NOAEL from the ITRI study, which
30      was 0.074 mg/m3. The duration-averaged LOAEL for  the same study is 0.723 mg/m3. All mouse
31      BMC values from the ITRI study fall between the NOAEL (HEC) and the LOAEL (HEC).
32            The basis for the RfC was the NOAEL (HEC) in the Ishinishi study, which was 0.155
33      mg/m3.  This NOAEL (HEC) was selected because it was lower than the valid LOAELs in rats
34      from all studies considered and higher than the other NOAEL (HEC) values. The only data in
       2/1/98                                  6-21       DRAFT-DO NOT CITE OR QUOTE

-------
 1      the Ishinishi et al. (1988) study that were amenable to BMC analysis were the incidence of
 2      hyperplastic lesions in male, female, and combined. These results are shown in Table 6-4.
 3             The BMC for the most sensitive effect is very similar to the NOAEL(HEC).  Most of the
 4      BMC values fall between the LOAEL (HEC) and the NOAEL (HEC) from the same data set.
 5      Two of the BMCs for the heavy-duty diesel experiment exceed the LOAEL (HEC). This could
 6      result from the large number of animals and close dose spacing, which allows identification of a
 7      lower LOAEL (HEC). These data sets were characterized by very low incidence of lesions at the
 8      lowest exposure level, and the assignment of the LOAEL (HEC) was difficult because of the low
 9      incidence and the lack of detailed description of the extent or severity of the response. The BMC
10      procedure makes determination of the effect level more objective for these data.
11             Two other studies contained information that was amenable to the BMC approach: the
12      chronic study done at GM  and the Nikula et al. (1995) study. Results of these analyses are
13      shown in Table 6-5.
               Table 6-4.  Results of benchmark concentration analyses using the
               polynomial model and data from the HERP study
         Data set, model
            Benchmark
              response    MLE"      BMC*
HERP light-duty diesel, hyperplastic lesions, male and female
combined
 Polynomial, threshold, background

HERP light-duty diesel, hyperplastic lesions, male rats
 Polynomial, no threshold, background

HERP light-duty diesel, hyperplastic lesions, female rats
 Polynomial, no threshold

HERP heavy-duty diesel, hyperplastic lesions, male and female
combined
 Polynomial, no threshold

HERP heavy-duty diesel, hyperplastic lesions, male rats
 Polynomial, no threshold

HERP heavy-duty diesel, hyperplastic lesions, female rats
 Polynomial, no threshold	
                                                                   10
                10
                10
                10 .
                                                                   10
                                                                   10
                         0.39
                                                                            0.35
                                                                            0.24
                                                                            0.49
                          0.67
                          0.43
                        0.34
                        0.32
                        0.19
                        0.38
                         0.46
                         0.30
        "ing/m3.
        Source: Ishinishi et al. (1988).
        2/1/98
6-22
DRAFT-DO NOT CITE OR QUOTE

-------
      Table 6-5. Results of benchmark concentration analyses using the
      polynomial model and data from the GM and Nikula studies
Benchmark
Data set, model response
GM study — continuous variables
Male rat alveolar-capillary thickness @ 6 mo 20
Polynomial, threshold
Male rat BAL PMNs @ 48 weeks . 200
Polynomial, no threshold
Male rat BAL macrophages @ 48 weeks 50
Polynomial, no threshold
Nikula study — continuous variables
Male RT Lu wt. @ 23 mo 10
Polynomial, no threshold
Female RT Lu wt. @ 23 mo • .10
Polynomial, no threshold
Male RT Lu wt. @ 1 8 mo 10
Polynomial, no threshold
Female RT Lu wt. @ 18 mo ' 10
Polynomial, no threshold
Nikula study — dichotomous variables
Male RT chronic inflammation @ > 1 8 mo 10
Polynomial, no threshold
Female RT chronic inflammation @ > 1 8 mo 10
Polynomial, no threshold
Female RT alveolar proteinosis @ > 1 8 mo 10
Polynomial, no threshold
Male RT bronchoalveolar metaplasia @ > 1 8 mo 10
Polynomial, no threshold
Female RT bronchoalveolar metaplasia @ > 1 8 mo 10
Polynomial, no threshold
Male RT focal fibrosis and epithelial hyperplasia @ > 1 8 mo 10
Polynomial, no threshold
Female RT focal fibrosis and epithelial hyperplasia @ >1 8 mo ' 10
Polynomial, no threshold
MLEa BMC8

0.069 0.06
0.087 0.08
0.244 0.12

0.088 0,06

0.039 0.03

. 0.170 0.13

0.122 • 0.04


0.232 0.14

0.095 0.08
0.122 0.03
0.195 0.12
0,026 0.02
0.721 0.46
0.336 0.20
"mg/rn3.
Source: Barnhartetal. (1981, 1982) and Nikula (1995).
2/1/98
6-23
DRAFT-DO NOT CITE OR QUOTE

-------
 1         Overall, the benchmark concentrations from several experiments and a variety of endpoints
 2      support the NOAEL and LOAEL identified from the rat database for chronic diesel exposure
 3      studies.  It is common practice to apply several models to the key data sets in a benchmark
 4      concentration analysis to determine whether the BMC is significantly model dependent, and if so,
 5      to select the most appropriate model. One reason for using a BMR that is close to the observable
 6      range is that the various models used for dose-response analysis tend to diverge more as they are
 7      extrapolated to lower doses and at the BMR such extrapolation should not be necessary, so
 8      substantial model dependence of the result is less likely.  Several of the data sets that resulted in
 9      lower BMC estimates were modeled using the Weibull model, and the results are shown in Table
10      6-6.
11         These results indicate that there is relatively little difference among models for a variety of
12      studies and endpoints.  The BMC analyses shown in Table 6-6 were selected because they
13      resulted in the lowest BMC levels of the studies and endpoints evaluated. It has been suggested
14      that the lowest BMC level should be used to derive the RfC, analogous to the use of the lowest
15      LOAEL to derive the RfC.  Selection of the most appropriate BMC to derive the RfC is
16      complicated by the fact that there may be many BMCs from a given experiment that identifies a
17      single LOAEL and NOAEL.  The ITRI study, for example, identifies an LOAEL of 3.5 mg/m3
18      and an NOAEL of 0.35 mg/m3 (exposure concentrations), but because of the large number of
19      endpoints studied, many BMC values are available.  On the other hand, a study such as Ishinishi
20      et al. (1986) has only one noncancer endpoint that can be used for BMC analysis, and the
21      Heinrich et al. (1995) study had no noncancer endpoints that were presented in a sufficiently
22      quantitative manner for BMC analysis. If these studies are not  considered in an analysis because
23      BMC models cannot be applied, substantial uncertainty would be introduced because significant
24      information is being ignored. In addition, several caveats must be considered relating to the
25      studies for which BMCs were calculated. The endpoints of chronic inflammation and alveolar
26      proteinosis in the female rats in the Nikula study were based on BMC models fit to two exposure
27      groups, and the incidences in the lowest exposed group were 49% and 83%, respectively.  Thus,
28      the BMC is estimated at a concentration well below the lowest  data point, and this extrapolation
29      may be cause for concern. This concern applies equally to the lung weight data in the Nikula
30      study, which show a greater than twofold increase in the low-concentration females. The data on
31      pulmonary particle clearance from the ITRI study were reported as an estimate of the clearance
32      half-time with no measure of variability. To run the BMC model, a coefficient of variability of
33      20% was assumed. Thus the data themselves do not meet the minimum requirements for
34      application of the BMC approach as discussed above. These data sets result in the lowest BMC
35      levels in the diesel database.  For several other endpoints, such as alveolar-capillary thickness

        2/1/98        .                             6-24        DRAFT-DO NOT CITE OR QUOTE

-------
      Table 6-6.  Comparison of benchmark concentration calculated using the
      polynomial and Weibull models for selected endpoints
Data set, model
Dichotomous variables
HERP light-duty diesel, hyperplastic lesions, female rats
Polynomial, no threshold
Weibull, no threshold
HERP heavy-duty diesel, hyperplastic lesions, female rats
Polynomial, no threshold
Weibull, no threshold
Nikula study, female RT chronic inflammation @ >18 mo
Polynomial, no threshold
Weibull, no threshold
Nikula study, female RT alveolar proteinosis @ >18 mo
Polynomial, no threshold
Weibull, no threshold
Continuous variables
ITRI rat Ga2O3 clearance half-time @ 24 mo, M and F combined
Polynomial, no threshold
Weibull, threshold
ITRI rat Cs-FAP clearance half-time @ 24 mo, M and F combined
Polynomial, no threshold
Weibull, no threshold
ITRI male rat lung weight @ 24 mo
Polynomial, no threshold
Weibull, no threshold
ITRI female rat lung weight @ 24 mo
Polynomial, no threshold
Weibull, no threshold
ITRI rat BAL neutrophils @ 24 mo, M and F combined
Polynomial, threshold
Weibull, no threshold
ITRI rat BAL p-glucuronidase @ 24 mo, M and F combined
Polynomial, threshold
Weibull, threshold
G.M. study, male rat alveolar-capillary thickness @ 6 mo
Polynomial, threshold
Weibull, no threshold
G.M. study, male rat BAL PMNs @ 48 weeks
Polynomial, no threshold
Weibull, no threshold
Benchmark
response


10
10

. 10
10

10
10

10
10


20
20

20
20

10
10

10
10

200
200

100
100

20
20

200
200
MLEa


0.247
0.350

0.425
0.273

0.095
0.102

0.122
0.175


0.074
0.176

0.048
0.050

0.103
0.103

0.061
0.068

0.094
0.094

0.172
0.176

0.069
0.069

0.087
0.087
BMCa


0.19
0.29

0.30
0.17

0.08
0.09

0.03
0.03


0.04
0.07

0.03
0.03

0.08
0.08

0.05
0.05

0.06
0.06

0.05
0.07

0.06
0.0-6

0.08
0.09
aing/m3.
2/1/98                               6-25      DRAFT-DO NOT CITE OR QUOTE

-------
 1      and lavage p-glucuronidase level, it is not clear that the BMR selected (20% and 100% increase,
 2      respectively) is the appropriate level in the context of applying the BMC to derive the RfC.
 3      Despite these caveats, it is clear that the BMC values for the more sensitive endpoints tend
 4      toward concentrations that are at or below the NOAEL identified by the combined rat database.
 5      This may well result in part because the ITRI study, which is the best documented study, has a
 6      tenfold difference between the LOAEL and NOAEL, and the Ishinishi study from which the rat
 7      NOAEL was obtained uses much closer dose spacing (a factor of 2.5 between LOAEL and
 8      NOAEL) and reports in much less detail on the noncancer effects. Thus the lack of detailed
 9      investigation in the Ishinishi study might allow a higher NOAEL to be identified at a level only
10      about one-half of the LOAELs from the ITRI, GM, Lewis et al. (1989),  and Heinrich (1995)
11      studies.  In other words, the BMC analyses suggest that a lower rat NOAEL might have been
12      identified if different dose spacing and more detailed investigation of noncancer endpoints had
13      been used in the existing studies.
14            Several limitations have been mentioned in the preceding text regarding the use of the
15      BMC analysis for deriving an RfC. The principal limitations are the following:
16            •   Some key studies in rats have inadequate quantitative data for BMC.
17            •   Some endpoints are amenable to BMC and others are not.
18           . •   The policy for selecting BMC from many endpoints is not yet clear.
19            •   It is not clear how to compare dichotomous and continuous  BMCs.
20            •   There is a lack of precedent or guidance for selecting BMR  levels.
21            •   A deposition model is available only for rats (it is not clear how to compare BMCs
22                based on deposition/retention models with BMC based on default duration-adjusted
23                concentrations).
24      Because of the issues and questions raised  by these aspects of the BMC approach, the BMC will
25      not be used to derive the RfC at this time.
26
27      6.7. SUMMARY
28            A large number of studies of chronic DPM inhalation in laboratory animals are available.
29      These studies characterize the respiratory effects and the concentration-response relationship of
30      those effects in detail.  Many.epidemiologic studies of occupationally exposed humans also are
31      available. The epidemiologic studies provide qualitative evidence that supports the identification
32      of a hazard to the respiratory system from laboratory animal studies. The human studies are of
33      limited value quantitatively because of their inadequate exposure characterization and
34      confounding by concurrent exposure to other pollutants. The  laboratory animal studies are used
35      to derive an RfC. The chronic studies from ITRI and HERP were selected as the principal

        2/1/98                                     6-26       DRAFT-DO NOT CITE OR QUOTE

-------
  1      studies for RfC development. The deposition and retention model discussed in Chapter 4 and

  2      Appendix B was used to calculate human equivalent concentrations and identified a NOAEL

  3      (HEC) of 0.155,mg/m3 from the HERP studies and a LOAEL(HEC) of 0.36 mg/m3 from the ITRI

  4      studies.  An uncertainty factor of 10 was applied to account for sensitive members of the

  5      population, resulting in an RfC of 16 |ig/m3.  The RfC is considered to have high confidence

  6      attributable to high confidence in the study and database.

  7

  8      6.8.  REFERENCES
  9
10      Barnes, DO; Daston, GP; Evans, JS; Jarabek, AM; Kavlock, RJ; Kimmel, CA; Park, C; Spitzer, HL. (1995)
11      Benchmark dose workshop: criteria for use of a benchmark dose to estimate a reference dose. Reg Toxicol
12      Pharmacol 21:296-306.
13
14      Barnhart,  MI; Chen, S-T; Salley, SO; Puro, H, (1981) Ultrastracture and morphometry of the alveolar lung of
15      guinea pigs chronically exposed to diesel engine exhaust: six months' experience.  J Appl Toxicol 1:88-103.
16
17      Barnhart,  MI; Salley, SO; Chen, S-T; Puro, H. (1982) Morphometric ultrastructural analysis of alveolar lungs of
18      guinea pigs chronically exposed by inhalation to diesel exhaust (DE). In: Lewtas,  J, ed. Toxicological effects of
19      emissions  from diesel engines: proceedings of the Environmental Protection Agency diesel emissions symposium;
20      October 1981; Raleigh, NC. (Developments in toxicology and environmental science: v.  10.) New York: Elsevier
21      Biomedical; pp. 183-200.                                        ,
22
23      Creutzenberg, O; Bellmann, B; Heinrieh, U; Fuhst, R; Koch, W; MuMe,  H. (1990) Clearance and retention of
24      inhaled diesel exhaust particles, carbon black, and titanium dioxide in rats  at lung overload conditions. J Aerosol
25      Sci 21(Suppl):S455-S458.
26                                                                                          .         '
27      Crump, KS. (1984) A new method for determining allowable daily intakes, Fundam Appl Toxicol 4:854-871.
28
29      Cramp, KS. (1995) Calculation of benchmark doses from continuous data. Risk Anal 15:79-89.
30
31      Eskelson,  CD; Strom, KA; Vostal, JJ; Misiorowski, RL. (1981) Lipids in die lung and lung gavage fluid of
32      animals exposed to diesel particulates. Toxicologist 1:74-75.
33
34      Gaylor, DW;  Slikker, W, Jr. (1990) Risk assessment for neurotoxic effects. Neurotoxicology 11:211-218.
35
36      Green, FHY;  Boyd, RL; Danner-Rabovsky, J;  Fisher, MJ; Moorman, WJ; Ong, T-M; Tucker, J; Vallyathan, V;
37      Whong, W-Z; Zoldak, J; Lewis, T. (1983) Inhalation studies of diesel exhaust and coal dust in rats. Scand J
38      Work Environ Health 9:181-188.
39
40      Gross, KB. (1981) Pulmonary function testing of animals chronically exposed to diluted diesel exhaust. J Appl
41      Toxicol 1:116-123.
42
43      Heinrieh,  U; Muhle, H; Takenaka, S; Ernst, H; Fuhst, R; Mohr, U; Pott, F; Stober, W, (1986) Chronic effects
44      on the respiratory tract of hamsters, mice, and rats after long-term inhalation of high concentrations of filtered
45      and unfiltered diesel engine emissions. J Appl Toxicol 6:383-395.
46
47      Heinrieh,  U; Fuhst, R; Rittinghausen, S; Creutzenberg, O; Bellmann, B; Koch, W; Levsen, K. (1995) Chronic
48      inhalation  exposure of Wistar rats and two strains of mice to diesel engine  exhaust, carbon black, and titanium
49      dioxide. Inhal Toxicol 7:553-556.
50


         2/1/98                                       6-27        DRAFT-DO NOT CITE OR QUOTE

-------
 1       Henderson, RF; Pickrell, JA; Jones, RK; Sun, JD; Benson, JM; Mauderly, JL; McClellan, RO. (1988) Response
 2       of rodents to  inhaled diluted diesel exhaust: biochemical and cytological changes in bronchoalveolar lavage fluid
 3       and in lung tissue. Fundam Appl Toxicol 11:546-567.
 4
 5       Howe, RB. (1990a) THRESH: a computer program to compute a reference dose from quantal animal toxicity data
 6       using the benchmark dose method. Clement International Corporation, K.S. Crump Division, Ruston, LA.
 7
 8       Howe, RB. (1990b) THC: a computer program to compute a reference dose from continuous animal toxicity data
 9       using the benchmark dose method. Clement International Corporation, K.S. Crump Division, Ruston, LA.
10
11       Howe, RB. (1990c) THRESHW:  a computer program to compute reference doses from quantal animal toxicity
12       data using the benchmark dose method. Clement International Corporation, K.S. Crump Division, Ruston, LA.
13
14       Howe, RB. (1990d) THWC: a computer program to compute a reference dose from continuous animal toxicity
15       data using the benchmark dose method. Clement International Corporation, K.S. Crump Division, Ruston, LA.
16
17       Ishinishi, N;  Kuwabara, N; Nagase, S; Suzuki, T; Ishiwata, S; Kohno, T.  (1986)  Long-term inhalation studies on
18       effects of exhaust from heavy and light duty diesel engines on F344 rats. In: Ishinishi, N; Koizumi, A;
19       McClellan, RO; Sto'ber, W, eds. Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the
20       international  satellite symposium on lexicological effects of emissions from diesel  engines; July; Tsukuba Science
21       City, Japan. (Developments in toxicology and environmental science: v.  13.) Amsterdam: Elsevier Science
22       Publishers  BV; pp. 329-348
23
24       Ishinishi, N;  Kuwabara, N; Takaki, Y; Nagase,  S; Suzuki, T; Nakajima, T; Maejima, K; Kato, A; Nakamura,
25       M.  (1988)  Long-term inhalation experiments on  diesel exhaust. In: Diesel exhaust and health risks: results of the
26       HERP studies.  Tsukuba, Ibaraki,  Japan: Japan Automobile Research Institute, Inc., Research Committee for
27       HERP Studies; pp. 11-84.
28
29       Iwai, K; Udagawa, T; Yamagishi; M; Yamada,  H. (1986) Long-term inhalation studies of diesel exhaust on F344
30       SPF rats. Incidence of lung cancer and lymphoma. In: Ishinishi, N; Koizumi, A; McClellan, RO; Stober, W, eds.
31       Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the international satellite symposium
32       on lexicological effects of emissions from diesel engines; July; Tsukuba Science City, Japan. (Developments in
33       toxicology and environmenlal science: v. 13.) Amslerdam: Elsevier Science Publishers BV;  pp. 349-360
34
35       Kaplan,  HL;  MacKenzie, WF; Springer, KJ; Schreck, RM; Vostal, JJ. (1982) A subchronic study of the effects
36       of exposure of Ihree species of rodenls lo diesel exhaust. In: Lewtas, J, ed. Toxicological effects of emissions
37       from diesel engines: proceedings of the Environmental Protection Agency diesel symposium; October 1981;
38       Raleigh, NC. New York: Elsevier Biomedical; pp. 161-182.
39
40       Kaplan,  HL;  Springer,  KJ; MacKenzie, WF. (1983) Studies of potential  health effects of long-term exposure to
41       diesel exhaust emissions. San Antonio, TX: Southwest Research Institute; SwRI project no. 01-0750-103.
42
43       Karagianes, MT; Palmer, RF; Busch, RH. (1981)  Effects of inhaled diesel emissions and coal dust in rats. Am
44       IndHygAssocJ 42:382-391.
45
46       Lewis, TR; Green, FHY; Moorman, WJ; Burg, JAR; Lynch, DW. (1986) A chronic inhalation toxicity study of
47       diesel engine emissions and coal dust, alone and combined. In: Ishinishi, N; Koizumi, A; McClellan, RO; Stober,
48       W, eds.  Carcinogenic and mutagenic effects of diesel engine exhaust: proceedings of the international satellite
49       symposium on lexicological effects of emissions from diesel engines; July; Tsukuba Science City, Japan.
50       (Developments in toxicology and  environmenlal  science: v. 13.) Amslerdam: Elsevier Science Publishers BV; pp.
51       361-380
52                      •
53     '  Lewis, TR; Green, FHY; Moorman, WJ; Burg, JR; Lynch, DW. (1989) A chronic inhalation toxicity study of
54       diesel engine emissions and coal dust, alone and combined. J Am Coll Toxicol 8:345-375.


         2/1/98                  .                        6-28         DRAFT-DO NOT CITE OR QUOTE

-------
 .  1       Mauderly, JL; Jones, RK; Griffith, WC; Henderson, RF; McClellan, RO. (1987a) Diesel exhaust is a pulmonary
   2      carcinogen in rats exposed chronically by inhalation. Fundam Appl Toxicol 9:208-221.
   3
   4      Mauderly, JL; Bice, DE; Carpenter, RL; Gillett, NA; Henderson, RF; Pickrell, JA; Wolff, RK. (1987b) Effects
   5      of inhaled nitrogen dioxide and diesel exhaust on developing lung. Cambridge, MA:  Health Effects Institute;
   6      research report no. 8.                                   •                      '
   7
   8      Mauderly, JL; Gillett, NA; Henderson,  RF; Jones, RK; McClellan, RO, (1988) Relationships of lung structural
   9      and functional changes to accumulation of diesel exhaust particles. In: Dodgson, J; McCallum, RI; Bailey, MR;
 10      Fisher, DR, eds. Inhaled particles VI: proceedings of an international symposium and workshop on lung
 11       dosimetry; September 1985; Cambridge, United Kingdom. Ann Occup Hyg 32(suppl. l):659-669.
 12
 13      McClellan, RO; Bice, DE; Cuddihy, RG; Gillett,  NA; Henderson, RF; Jones, RK;'Mauderly, JL; Pickrell, JA;
 14      Shami, SG; Wolff, RK.  (1986) Health effects of diesel exhaust. In: Lee, SD; Schneider, T; Grant, LD; Verkerk,
 15      PJ, eds. Aerosols: research,  risk assessment and control strategies: proceedings of the second U.S.-Dutch
 16      international symposium; May 1985; Williamsburg, VA. Chelsea, MI: Lewis Publishers, Inc.; pp. 597-615.
 17
 18      Misiorowski,  RL; Strom, KA; Vostal, JJ; Chvapil, M. (1980) Lung biochemistry of rats chronically exposed to
 19      diesel particulates. In: Pepelko, WE; Danner, RM; Clarke, NA, eds. Health effects of diesel engine emissions:
 20      proceedings of an international symposium; December 1979. Cincinnati, OH: U.S. Environmental Protection
 21       Agency, Health  Effects Research Laboratory; pp. 465-480; EPA report no. EPA-600/9-80-057a. Available from:
 22      NTIS, Springfield, VA;  PBS 1-173809.
 23
 24      Morrow,  PE.  (1992) Dust overloading of the lungs: update and appraisal. Toxicol Appl Pharmacol 113:1-12.
 25
 26      Nikula, KJ; Snipes, MB; Barr, EB; Griffith, WC; Henderson, RP; Mauderly, JL.  (1995) Comparative pulmonary
 27      toxicities  and carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl
.28      Toxicol 25:80-94.                                                       .
 29
 30 .      Pepelko,  WE; Peirano, WB. (1983) Health effects of exposure to diesel engine emissions:  a summary of animal
 31       studies conducted by the U.S. Environmental Protection Agency's Health Effects Research Laboratories at
 32      Cincinnati, Ohio. J Am Coll Toxicol 2:253-306.
 33
 34      Pepelko,  WE; Mattox, J; Moorman, WJ; Clark, JC. (1980) Pulmonary function evaluation of cats after one year
 35      of exposure to diesel exhaust. In: Pepelko, WE; Danner, RM; Clarke,  NA., eds. Health effects of diesel engine
 36      emissions: proceedings of an international symposium, v. 2; December 1979; Cincinnati, OH. Cincinnati, OH:
 37      U.S. Environmental Protection Agency,  Health Effects Research Laboratory; pp. 757-765; EPA report no.
 38      EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.
 39
 40      Stober, W. (1986) Experimental induction-of tumors in hamsters, mice and rats after long-term inhalation of
 41       filtered and unfiltered diesel engine exhaust. ID: Ishinishi, N.; Koizumi, A.; McClellan, R. O.; Stober, W., eds.
 42      Carcinogenic and mutagenie effects of diesel engine exhaust: proceedings of the international satellite symposium
 43      on lexicological  effects of emissions from diesel engines; July; Tsukuba Science City, Japan. (Developments in
 44     • toxicology and environmental science: v. 13.) Amsterdam: Elsevier Science Publishers B.  V.; pp. 421-439.
 45
 46      Strom, KA. (1984) Response of pulmonary cellular defenses to the inhalation of high concentrations of diesel
 47      exhaust. J Toxicol Environ Health 13:919-944.
 48
 49      U.S. Environmental Protection Agency.  (1994) Methods for derivation of inhalation reference concentrations and
 50      application of inhalation dosimetry. Research Triangle Park, NC: Office of Research and Development, National
 51       Center for Environmental Assessment; EPA report no. EPA/600/8-90/066F,
 52
 53      U.S. Environmental Protection Agency.  (1995a) Integrated Risk Information System (IRIS). Online. Cincinnati,
 54      OH:  National  Center for Environmental  Assessment.


          2/1/98                                          6-29         DRAFT-DO NOT CITE OR QUOTE

-------
  1       U.S. Environmental Protection Agency. (1995b) The use of the benchmark dose approach in health risk
  2       assessment. Washington, DC: Office of Research and Development, Risk Assessment Forum.
  3
  4       Werchowski, KM; Chaffee, VW; Briggs, GB. (1980a) Teratologic effects of long-term exposure to diesel exhaust
  5   '    emissions (rats). Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory;
  6       EPA report no. EPA-600/1-80-010. Available from: NTIS, Springfield, VA; PB80-159965.
  7
  8       Werchowski, KM; Henne, SP; Briggs, GB. (1980b) Teratologic effects of long-term exposure to diesel exhaust
  9       emissions (rabbits). Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory;
10       EPA report no. EPA-600/1-80-011. Available from: NTIS, Springfield, VA; PB80-168529.
11
12       Wolff, RK; Henderson, RF; Snipes, MB; Griffith, WC; Mauderly, JL; Cuddihy, RG; McClellan, RO. (1987)
13       Alterations in particle accumulation and clearance in lungs of rats chronically exposed to diesel exhaust. Fundam
14       ApplToxicol9:154-166.
15
16       Yu, CP; Xu, GB. (1986) Predictive models for deposition of diesel exhaust particulates in human and rat lungs.
17       Aerosol Sci Technol 5:337-347.
18
19       Yu, CP; Yoon, KJ. (1990) Retention modeling of diesel exhaust particles in rats and humans. Amherst, NY: State
20       University of New York at Buffalo (Health Effects Institute research report no. 40).
         2/1/98                                         6-30        DRAFT-DO NOT CITE OR QUOTE

-------
 1         7.  CARCINOGENICITY OF DIESEL EMISSIONS IN LABORATORY ANIMALS
 2
 3      7.1. INTRODUCTION
 4            The particulate phase of diesel exhaust is composed of aggregates of carbon particles; the
 5      primary particle diameter ranges from 10 to 80 run, and aggregates of these primary particles
 6      have mass median diameters averaging 0.2 to 0.3 [am (Vuk et al., 1976; Carpenter and Johnson,
 7    ,  1979), although some may approach 1.0 um. A variety of organic compounds, including
 8      polycyclic aromatic hydrocarbons (PAHs), are adsorbed to this carbon.core (see Tables 2-8 and
 9      2-10) and comprise  5% to 65% of the total particle mass (Cuddihy et al., 1984). Some of these
10      organic compounds, such as benzo[a]pyrene (B[a]P), dinitropyrenes, and 1 -nitropyrene, have
11      received special attention regarding their carcinogenic and mutagenic potential. These organics
12      may be strongly or weakly bound to the carbon core and represent varying amounts of the total
13      particle mass. Qualitative and quantitative relationships for these organics depend on such
14      variables as fuel composition, engine design, and engine operating conditions.  Although less
15      emphasis has been placed on the gaseous phase, potential carcinogens such as formaldehyde,
16      acetaldehyde, benzene, as well as lower molecular weight PAHs, may also be present in this
17      fraction.
18            The respirability of these particles and their associated organics provides a basis for
19      health hazard concerns, and the reported mutagenicity (Huisingh et al., 1978) and skin papilloma
20      induction (Kotin et al., 1955) of solvent extracts of diesel particulate matter (DPM) suggests  a
21      potential for eareinogenieity. Zamora et al. (1983) provided evidence that DPM extracts
22      contained components that acted as weak tumor promoters in vitro.  Recently, emphasis has been
23      directed toward assessing the carcinogenic potential of whole and filtered diesel exhaust using
24      whole-animal studies and understanding the mechanisms and implications of deposition,
25      retention, and clearance of the particulate phase of diesel exhaust.
26            This chapter summarizes studies that assess the carcinogenic potential of diesel exhaust in
27      laboratory  animals.  Experimental protocols for the inhalation studies usually consisted of
28      exposure (usually chronic) to diluted exhaust in whole-body exposure chambers using rats, mice,
29      and hamsters as model species. Some of these studies used both filtered (free of particulate
30      matter) diesel exhaust and unfiltered (whole) diesel exhaust to differentiate gaseous-phase effects
31      from effects induced by DPM and its adsorbed components. Inhalation exposure to DPM alone,
32      however, was not reported.  Particulate matter concentrations in the diesel exhaust used in these
33      studies ranged from 0.1 to 12 mg/m3.  Clean air (usually filtered) was used in the control
34      exposures. Studies providing both positive, negative, or inconclusive findings have been
35      reported. In this chapter, any indication of statistical significance implies that p< 0:05 was

        271/98                                    7-1        DRAFT-DO NOT CITE OR QUOTE

-------
  1     reported in the reviewed publications. The experimental protocols and exposure atmosphere
  2     characterizations are not described in detail here but may be found in Appendix A.  A summary
  3     of the animal carcinogenicity studies and their results is presented in Table 7-1.
  4            Also included are studies that assessed the carcinogenic and tumorigenie effects of DPM
  5     and solvent extracts of these particles following dermal application, subcutaneous (s.c.) injection,
,  6     intraperitoneal (i.p.) injection, or intratracheal (itr.) instillation in rodents, as well as co-
  7     carcinogenicity studies. Individual chemicals present in the gaseous phase or adsorbed to the
  8     particle surface were not included in this review because adequate assessments of those of likely
  9     concern (i.e., formaldehyde, acetaldehyde, benzene, PAHs) have been published in other health
 10     assessment documents.
 11
 12     7.2. INHALATION STUDIES
 13     7.2.1.  Rat Studies
 14            Mauderly et al. (1987) provided data affirming the carcinogenicity of automotive diesel
 15     engine exhaust in F344/CM rats following chronic inhalation exposure. Male and female rats
 16     were exposed to diesel engine exhaust at nominal DPM concentrations of 0.35  (n = 366), 3.5
 17     (n = 367), or 7.1 (n = 364) mg/m3 for 7 h/day, 5 days/week for up to 30 mo. Sham-exposed
 18     (n = 365) controls breathed filtered room air. A total of 230,223,221, and 227 of these rats
 19     (sham-exposed, low-, medium-, and high-exposure groups, respectively) were examined for lung
 20     tumors. These numbers include those animals that died or were euthanized during exposure and
 21     those that were terminated following 30 mo of exposure. The exhaust was generated by 1980
 22     model 5.7-L Oldsmobile V-8 engines operated through continuously repeating  U.S. Federal Test
 23     Procedure (FTP) urban certification cycles. The engines were equipped with automatic
 24     transmissions connected to eddy-current dynamometers and flywheels simulating resistive and
 25     inertial loads of a midsize passenger car. The D-2 diesel control fuel (Phillips Chemical Co.) met
 26     U.S. EPA certification standards and contained approximately 30% aromatic hydrocarbons and
 27     0.3% sulfur.  Following passage through a standard automotive muffler and tail pipe, the exhaust
 28     was diluted 10:1 with filtered air in a dilution tunnel and serially diluted to the final
 29     concentrations. The primary dilution process was such that particle coagulation was retarded.
 30     Mokler et al. (1984) provided a detailed description of the exposure system. The gas-phase
 31     components of the diesel exhaust atmospheres are presented in Appendix A. No exposure-
 32     related changes in body weight or life span were noted for any of the exposed animals nor were
 33     there any signs of overt toxicity. Collective lung tumor incidence was greater (z statistic,.
 34     j?<0.05) in the high (7.1 mg/m3) and medium (3.5 mg/m3) exposure groups (12.8% and 3.6%,
 35     respectively) versus the control and low (0.35 mg/m3) exposure  groups (0.9% and 1.3%,

        2/1/98                                    7-2        DRAFT-DO NOT CITE OR QUOTE

-------
                     Table 7-1.  Summary of animal carcinogenicity studies
oo
O
O
H

O
h— *
H
W

o-
90
Study


Mauderly etal.
(1987)





Nikula et al.
(1995)
_




Heinrich et al.
(1986a,b)
Mohr et al.
(1986)


Heinrich et al.
(1986a,b)






Species/ Sex/total
strain number


Rat/F344 M + F, 230b
M + F, 223
M + F, 221
M + F, 227



Rat/F344 M + F,214b
M + F, 210
M + F.212
M + F, 213
M + F, 21 1


Rat/ F, 96
Wistar F, 92

F.9S


Mouse/ ' M + F, 84
NMRI M + F, 93

M + F, 76
Hamsters M + F, 96
/Syrian M + F, 96

M + F, 96
Particle
Exposure concentration
atmosphere (mg/m3)


Clean air
Whole exhaust
Whole exhaust
Whole exhaust



Clean air
Whole exhaust
Whole exhaust
Carbon black
Carbon black


Clean air
Filtered
exhaust
Whole exhaust


Clean air
Filtered
exhaust
Whole exhaust
Clean air
Filtered
exhaust
Whole exhaust


0
0.35
3.5
7.1



0
2.5
6.5
2.5
6.5


0
0

4.0


0
0

4.0
0
0

4.0
Other Exposure Posteiposure
treatment protocol observation


None
None
None
None



None
None'
None
None
None


None
None

None


None
None

None
None
None

None


7 h/day, NA
5 days/week
up to 30 mo




16 h/day, 6 weeks
5 days/week
for up to
24 mo



19 h/day, NA
5 days/week
• for up to
35 mo


19 h/day, NA
5 days/week
for up to
30 mo
19 h/day, NA'
5 days/week
for up to
30 mo

Adenomas .
(0)
. (0)
(2.3)
(0.4)


Adenomas
1/214 (<1)
7/210(3)
23/212(11)
3/213(1)
13/211 (6)

Adenomas
0/96(0)
0/92 (0)

8/95 (8.4)

Adenomas
9/84(11)
11/93(12)

11/76(15)
0/96 (0)
0/96 (0)

0/96 (0)
Tumor type and incidence (%)*
Adenoearcinoma
+ squamous cell
carcinomas
(0.9)
(1.3)
(0.5)
(7.5)


Adenocarcinomas
1/214 (<1)
4/210(2)
22/212(10)
7/213(3)
21/211(10)

Carcinomas
0/96 (0)
0/92 (0)

0/95 (0)

Adenoearcinoma
2/84 (2)
18/93(19)c

13/76(17)c
0/96 (0)
0/96 (0)

0/96 (0)
Squamous
cysts
(0)
(0)
(0.9)
(4.9)
Squamous
cell
carcinoma
1/214 (<1)
3/210(1)
3/212(1)
0/213(0)
3/211 (1)
Squamous
cell tumors
0/96 (0)
0/92 (0)

9/95 (9.4)
Squamous
cell tumors
. —
—

—
0/96
0/96

0/96

AM tumors
(0.9)
(1.3)
(3.6)c
(12.8)c
Adeno-
squamous
carcinoma
0/214(0)
0/210(0)
1/212 (<1)
0/213 (0)
2/211 (<1)

All tumors
0/96(0)
0/92 (0)

1 7/95(1 7. 8)c

All tumors
11/84(13)
29/93(31)°

24/76 (32)c
0/96(0)
0/96 (0)

0/96 (0)
Comments







Other
neoplasms
0/214(0)
0/210(0)
0/212(0)
1/213 (<1)
0/211 (0)
















O
H-
W

-------
                      Table 7-1.  Summary of animal carcinogenicity studies (continued)
OO
o

2
o
H

n
»— i
H
W
Species/
Study strain


Henrich et al. Rat/
(I989a) Wistar









Heinrich et al. Rat/
(1995) ( Wistar




Mouse/
C57BL/
6N



Mouse/
NMRI


Mouse/
NMRI


Sen/total
number


F,NS
F,NS
- F, NS

F.NS
F.NS
F.NS




F.220
. F.200
F, 200
F, 100
F, 100
F, 100
F, 120

F, 120

F, 120

F, 120
F, 120


F.120
F.120
F.I 20

Particle
Exposure concentration Other Exposure Postexposure
atmosphere (mg/m3) treatment protocol observation


Clean air
Whole exhaust
Filtered
exhaust
Clean air
Whole exhaust
Filtered
exhaust



Clean air
Whole exhaust
Whole exhaust
Whole exhaust
Carbon black
TiO2
Clean air

Whole exhaust

Particle-free
exhaust
Clean air
Whole exhaust
Carbon black
TiO2
Clean air
Whole exhaust
Particle-free
exhaust


0
4.2
0

0
4.2
0




0
0.8
2.5
7.0
11.6
10.0
0

4,5

0

0
4.5
11.6
10
0
4.5
0



DPNd
DPNd
DPNd

DPN*
DPNe
DPNe




None
None
None
None
None
None
None

None

None

None
None
None
None
None
None
None



19h/day,
5 days/week
for 24 to
30 mo







18h/day,
5 days/week,
for up to 24
mo


18h/day,
5 days/week,
for up to 2 1
mo


18h/day,
5 days/week
for up to
13.5 rno
18h/day,
5 days/week,
23 mo



NA









Adenomas
6 mo 0/217(0)
0/198 (0)
2/200(1)
4/100(4)
13/100(13)
4/100(4)
6 mo





9.5 mo (25)
(21.8)
(11.3)
(11.3)
None (25)
(18.3)
(31.7)

Tumor type and incidence (%)*












Adenoearcinomas
1/217 (<1)
0/198(0)
1/200 (<1)
4/100(4)
13/100(13)
13/100(13)






(15.4)
(15.4)
(10)
(2.5)
(8.8)
(5.0)
(15)

Squamous
cell
carcinoma
(4.4)
(46.8)'
(4.4)

(16.7)
(31.3?
(14.6)

Squamous
cell
carcinomas
0/217(0)
0/198(0)
0/200 (0)
2/100(2)
4/100(4)
3/100(3)







•






All lung
tumors
(84.8)
(83.0)
(67.4)

(93.8)
(89.6)
(89.6)

Benign
squamous cell
lurnonj
0/217(0)
0/198(0)
7/200 (3.5)
14/100(14)
20/100 (20)
20/100(20)














Comments
















Tumor
incidences
after 30 mo
5.1% tumor
rate
8.5% tumor
rate
3.5% tumor
rate





*i


O
H
W

-------
                      Table 7-1. Summary of animal careinogenieity studies (continued)
oo
o
x
o
H
n
)**mt
H
M
O
!*>
o
H
w
Study


Takakietal.
(1989)
Light-duty
engine




Ishinishi et at.
(1988a)
.
Heavy-duty
engine




Iwai et al.
(1986)



Takemoto et al.
(1986)










Species/ Sex/total
strain number


Rat/F344 M
M
M
M
M



Rat/F344 M
M
M
M
M




Rat/F344 F,
F,

F,
-
Rat/F344 F,
F,
F,
F,


Mouse/ "M
IRC M

Mouse/ M
C57BL M



+ F,
+ F,
+ F,
+ F,
+ F,



+ F,
+ F,
+ F,
+ F,
+ F,




24
24

24

12
21
15
18


+ F,
+ F,

+ F,
+ F,



123
123
125
123
124



123
123
125
123
124















45
69

12
38

Particle
Exposure concentration
atmosphere (mg/nt3)


Clean air
Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust



Clean air
Whole exhaust
Whole exhaust
Wfiole exhaust
Whole exhaust




Clean air
Filtered
exhaust
Whole exhaust

Clean air
Clean air
Whole exhaust
Whole exhaust


Clean air
Whole exhaust

Clean air
Whole exhaust



*0
0.1
0.4
1.1
2.3



0
0.5
1.0
1.8
3.7




0
0

4.9

0
0
2-4
2-4


0
2-4

0
2-4

Other
treatment


None
None
None
None
None



None
None
- None
None
None




None
None

None

None
DIPNh
None
DIPNh


None
None

None
None

Exposure Postexposure
protocol observation


16 h/day, NA
6 days/week,
for up to
30 mo




16 h/day, NA
6 days/week,
for up to
.30 mo





8 h/day, NA
7 days/week,
for 24 mo


4 h/day, NA
4 days/week,
18-24 rno



4 h/day, NA
4 days/week,
for 19-28 mo
4 h/day, NA
4 days/week
for 19-28 mo

Adenomas
1/23 (0.8)
1/23 (0.8)
1/25 (0.8)
0/23 (0)
1/24(8.1)


Adenomas
0/123 (0)
0/123 (0)
0/125 (0)
0/123 (0)
0/124 (0)



Adenomas
1/22 (4.5)
0/16 (0)

3/19 (0)













Tumor type and incidence (%)*
Adenosquamous
carcinomas
2/123(1.6)
1/23 (0.8)
0/125 (0)
5/123(4.1)
2/124(1.6)

Adenosquamous
carcinomas
1/123 (0.8)
0/123 (0)
0/125 (0)
4/123 (3.3)
6/124 (4.8)

Adenocarcinoma
and adenosquamous
carcinoma
0/22 (0)
0/16 (0)

3/19 (15.8)
Adenoma
0/12 (0)
10/21 (47.6)
0/15 (0)
12/18 (66.7)

Adenoma
3/45 (6.7)
6/69 (8.7)

1/12 (8.3)
8/38(21.1)

Squamous
cell
carcinomas
1/23 (0.8)
1/23 (0.8)
0/125 (0)
0/123 (0)
0/124 (0)
Squamous
cell
carcinomas
0/123 (0)
1/123 (0.8)
0/125 (0)
0/123 (0)
2/124(1.6)
Large cell
and squamous
cell
carcinomas
0/22 (0)
0/16 (0)

2/19 (10.5)
Carcinoma
"o/12 (0)
4/21 (19)
0/15 (0)
7/18 (38.9)
Adeno-
carcinoma
1/45 (2.2)
3/69 (4.3)

0/12 (0)
3/38 (7.9)

•Comments

All tumors
4/123 (3.3)
3/123 (2.4)
1/125(0.8)
5/123(4.1)
3/124 (2.4)


All tumors
1/123(0.8)
1/123 (0'.8)
0/125 (0)
4/123 (3.3)
8/124 (6.5)c



All.turn.ors
1/22 (4.5)f
0/16(0)

8/19














-------
Table 7-1.  Summary of animal carcinogenicity studies (continued)
oo


















-a
I
o\



O

3>
'"fl
H
6
o
z
M~~t
O
H
Q
H •
M
O
£3 •
tO
O
H
M

Species/ Sex/total
Study strain number

Brightwell et al. Rat/344 M + F, 260
(1989) M + F, 144



M + F, 143


M + F, 143
M + F, 144
M + F, 143




Hamster/ M + F,
Syrian M + F, 202
Golden M + F, 104



M + F, 104


M + F, 101
M + F, 102
M + F, 101
M + F, 204


M + F, 203

' ' Karagianes et Rat/ M, 40
al. (1981) Wistar M, 40






Exposure
atmosphere

Clean air
Filtered
exhaust
(medium
exposure)
Filtered
exhaust (high
exposure)
Whole exhaust
Whole exhaust
Whole exhaust




Clean air
Clean air
Filtered
exhaust
(medium
dose)
Filtered
exhaust
(high dose)
Whole exhaust
Whole exhaust
Whole exhaust
Filtered
exhaust
(high dose)
Whole exhaust

Clean air
Whole exhaust





Particle
concentration
(mg/m3)

0
0



0


0.7
2.2
6.6




0
0
0



0


0.7
2.2
6.6
0


6.6

0
8.3






Other Exposure Postexposure
treatment protocol observation

None 16 h/day, NA
None 5 days/week,
for 24 mo


None


None - •
None
None




None 16 h/day, NA
DENj 5 days/week,
DENJ for 24 mo



DENJ •


DEN1
DENJ
DENJ
None


None

None 6 h/day, NA
None 5 days/week,
for up to
20 mo





Tumor type and incidence (%)"
Primary lung tumors
3/260(1,2)
0/144 (0)



0/143 (0)


1/143 (0.7)
14/144 (9.7)c
55/143 (38.5)c



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)



Adenomas
0/6 (0)
1/6(16.6)







Comments

Tumor
incidence
for all rats
dying or
sacrificed



S 24/25
(96%) after
24 mo
d1 12/27
(44%)
after 24 rno

Respiratory
tract tumors
not related
to exhaust
exposure
for any of
the groups


















-------
Table 7-1. Summary of animal carcinogenicity studies (continued)
MD
OO


















-J
i
-J



O
hrt
3>
3
h_j
6
o
t-myl
jf
M~4
o
H
O

H
M


O
O •
H
W

Particle
Species/ Sex/total Exposure concentration Other Exposure Postexposure
Study strain number atmosphere (mg/m3) treatment protocol observation

Orlhoefer et al. Mouse/ M, 25 Clean air 0 None 20 h/day,
(1981) Strong A 7 days/week,
(Pepelko and ' for 7 weeks
Peirano, 1983) Whole exhaust 6.4 None 26 weeks


Whole exhaust 6.4 UV 26 weeks
irradiated


Mouse/ M + F, 40 Clean air 0 None 20 h/day, 8 weeks
Jackson 7 days/week,
A ' for 8 weeks
M + F, 40 Whole exhaust 6.4 None 8 weeks


Mouse/ F, 60 Clean air 0 None 20 ti/day,
Jackson 7 days/week,
A for approx.
F, 60 Clean air 0 Urethan1 7 mo.


F, 60 Whole exhaust 6,4 None


F, 60 Whole exhaust ' 6.4 Urethank



M, 429 Clean air 0 None


M, 430 Whole exhaust 6.4 None










Tumor type and incidence (%)°
Lung Tumors
3/22 (13.6)


7/19 (36.8)


6/22 (27.3)


Lung Tumors
16/36 (44.4)


11/34(32.3)


4/58 (6.9)


9/52 (17.3)


14/56 (25.0)


22/59 (37.3)
-


73/403 (18.0)


66/368 (17.9)










Comments

0.13
Tumors/
mouse
0.63
Tumors/
mouse
0.27
Tumors/
mouse

0.5
Tumors/
mouse
0.4
Tumors/
mouse
0.09
Tumors/
mouse
0.25
Tumors/
mouse
0.32
Tumors/
mouse
0.39
Tumors/
mouse

0.23
Tumors/
mouse
0.20
Tumors/
mouse






-------
                       Table 7-1. Summary of animal carcinogenicity studies (continued)
 NO
 co
 QO
•o
 o
 n
 i_^

 H
 W

 O
 JJO
Study

Pepelko and
Peirano (1983)
















Kaplan et al.
(1983)
White el al.
(1983)






Kaplan et ai.
(1982)

Particle
Species/ Sex/total Exposure concentration Other
strain number atmosphere (ing/m3) treatment

Mouse/ M + F, 260 Clean air
Sencar Clean air
Clean air
Whole exhaust
Whole exhaust
Whole exhaust

Mouse/ M + F, 90 Clean air
Strain A

Clean air


Whole exhaust
Whole exhaust
Clean air
Whole exhaust

Rat/F344 M, 30 Clean air
M, 30 Whole exhaust
M, 30 Whole exhaust
M, 30 Whole exhaust
'"
Mouse/ M, 388 Clean air
A/J M, 388 Whole exhaust
M, 399 Whole exhaust
M, 396 Whole exhaust

Mouse/ M, 458 Clean air
A/J M, 18 Clean air
M, 485 Whole exhaust

0
0
0
12
12
12

0


0


12
12
0
12

0
0.25
0.75
1.5

0
0.25
0.75
1.5

0
0
1.5

None
BHT1 "
Urethank
None
BHT1
Urethan1

None


Exposure
(darkness)

Exposure
(darkness)
Urethan1"
Urethan"1

None
None
None
None .

None
None
None
None

None
Urethank
None
Exposure Postexposure
protocol observation

Continuous NA
for 15 mo





NA










20 h/day, 8 mo
7 days/week, 8 mo
for up to 8 mo
15 mo 8 mo

20 h/day, NA
7 days/week,
for up to
8 mo

20 h/day, 6 mo
7 days/week,
for 3 mo
Tumor type and incidence (%)" Comments
Adenomas Carcinomas All tumors
(5.1) • (0.5) (5.6)
(12.2) (1.7) (2.8)
(8.1) (0.9) (9.0)
(10.2)° (1.0) (11.2)c
(5.4) (2.7) (8.1)
(8.7) (2.6) (11.2)
All tumors
21/87 (24) 0.29
Tumors/
mouse
59/237 (24.9) 0.27
Tumors/
mouse
10/8012.5) 0.14
22/250(0.10) 0.10
66/75(88) 2.80
42/75(0.95) .0.95
Broncho-alveolar carcinoma
0/30 (0)
1/30 (3.3)
3/30 (10.0)
1/30 (3.3)
Pulmonary adenoma
130/388 (33*5)
131/388 (33.8)
109/399 (27.3)
99/3% (25.0)
Pulmonary adenomas
144/458(31.4)
18/18(100)
165/485 (34.2)
 I

-------
                                 Table 7-1.  Summary of animal carcinogenicity studies (continued)
OO
T1
H
6
o
3
o
H
n
i— i
H
W
o
H
w
Study

Ishinishi et al.
(1988a)

Light duty


Heavy duty





Lewis et al.
(1989)




Mauderly et al.
(1996)


Species/ Sex/total
strain number

Rat/F344 NS, 5 -
NS, 8
NS, 11
NS, 5
NS, 9
NS, 11
NS,5
NS.9
NS 11
NS, 5
NS, 6
NS, 13
Rat/F344 M + F, 288"





Mouse/ M + F
CD-I M + F
M + F
M + F
Particle
Exposure concentration
atmosphere (mg/m3)

Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust
Whole exhaust
Clean air
Whole exhaust




Clean air
Whole exhaust
Whole exhaust
Whole exhaust

0.1
0.1
. 0.1
1.1
1.1
1.1
0.5
0.5
0.5
1.8
1.8
1.8
0
2.0




0
0.35
3.5
7.0
Other Exposure Postexposure
treatment protocol observation

None
None
None
None
None
None
None
None
None
None
None
None
None
None




None
None
None
None

16 h/day,
6 days/week,
for 12 mo



. 16 h/day,
6 days/week,
for 12 mo
V


7 h/day,
5 days/week,
24mo°



7 h/day, 5
days/week,
for up to 24
mo

6 mo
12 mo
18 mo
6 mo
12 mo
18 mo
6 mo
12 mo
18 mo
6 mo
12 mo
18 mo
NA





None



Adenomas
0/5 (0)
0/8 (0)
0/1 1 (0)
0/5 (0)
0/9 (0)
0/1 1 (0),
0/5 (0)
0/9 (0)
0/1 1 (0)
0/5 (0)
0/6 (0)
0/13(0)
No tumors



Multiple
adenomas
1/157(0.6)
2/171(1.2)
0/155 (0)
0/186(0)
Tumor type and
Carcinomas
0/5 (0)
0/8 (0)
0/1 1 (0)
0/5 (0)
0/9 (0)
0/1 1 (0)
0/5 (0)
0/9 (0)
0/1 1 (0)
0/5 (0)
0/6 (0)
1/13(0)




Multiple
carcinomas
2/157(1.3)
1/171 (0.6)
1/155(0.6)
0/186(0)
incidence (%)*
All tumors
0/5 (0)
0/8 (0)
0/1 1 (0)
0/5 (0)
0/9 (0)
0/11 (0)
0/5 (0)
0/9 (0.)
0/1 1 (0)
0/1 1 (0)
0/6 (0)
1/13(0)
0/192(0)
0/192(0)


Adenomas/
carcinoma
1/157(0.6)
1/171 (0.6)
0/155(0)
0/186(0)
Comments
















Alveolar/ Alveolar/
bronchiolar bronchiolar
adenoma carcinoma
10/157(6.4) 7/157(4.5)
16/171(9.4) 5/171(2.9)
8/155(5.2) 6/155(3.9)
10/186(5.4) 4/186(2.2)
*Table values indicate number with tumors/number examined (% animals with tumors).
'Number of animals examined for tumors.
'Significantly different from clean air controls.
dDiphenylnitrosamine; 6.25 mg/kg/week s,c. during first 25 weeks of exposure.
'Diphenylnitrasamine;  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%).
85.3% 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/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.
""12 mg/m'from 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.   .             •

-------
 1      respectively).  Bronchoalveolar adenomas, adenocarcinomas, and squamous cysts (considered
 2      benign, except for two that were classified as squamous cell carcinomas because of the presence
 3      of less differentiated cells and invasion of blood and lymph vessels) were identified.  Using the
 4      same statistical analysis of specific tumor types, adenocarcinoma plus squamous cell carcinoma
 5      and squamous cyst incidence was significantly greater in the high-exposure group, and the
 6      incidence of adenomas was significantly greater in the medium exposure group.  A significant
 7      (pO.OOl) exposure-response relationship was obtained for tumor incidence relative to exposure
 8      concentration and lung burden of DPM. These data are summarized in Table 7-1. A logistic
 9      regression model estimating tumor prevalence as a function  of time, dose (lung burden of DPM),
10      and sex indicated a sharp increase in tumor prevalence for the high dose level at about 800 days
11      after the commencement of exposure. A less pronounced, but definite, increase in prevalence
12      with time was predicted for the medium-dose level.  Significant effects were not detected at the
13      low concentration. DPM (mg per lung) of rats exposed to 0.35, 3.5, or 7.1  mg of DPM/m3 for 24
14      mo were 0.6, 11.5, and 20.8, respectively, and affirmed the greater than predicted accumulation
15      that was the result of decreased particle clearance following  high-exposure conditions.
16            In summary, this study demonstrated the pulmonary  carcinogenicity of high
17      concentrations of whole, diluted diesel exhaust in rats following chronic inhalation exposure.  In
18      addition, increasing lung particle burden resulting from this  high-level exposure and decreased
19      clearance was demonstrated. A logistic regression model presented by Mauderly et al. (1987)
20      indicated-that both lung DPM burden and exposure concentration may be useful for expressing
21      exposure-effect relationships.
22            A series of studies was conducted at the Fraunhofer Institute of Toxicology and Aerosol
23      Research in which female Wistar rats were exposed for 19 h/day, 5 days/week to both filtered
24      and unfiltered (total) diesel exhaust at an average particulate matter concentration of 4.24 mg/m3.
25      Animals were exposed for a maximum of 2.5 years. The exposure system  as described by
26      Heinrich et al. (1986b) used a 40 kw 1.6-L diesel engine operated continuously under the U.S. 72
27      FTP driving cycle. The engines used European Reference Fuel with a sulfur content of 0.36%.
28      Filtered exhaust was obtained by passing engine  exhaust through a Luwa FP-65 HT 610 particle
29      filter heated to 80°C and a secondary series of filters (Luwa FP-85, Luwa NS-30, and Drager CH
30      63302) at room temperature. The filtered and unfiltered exhausts were each diluted 1:17 with
31      filtered air and passed through respective 12m3 exposure chambers.  Mass median aerodynamic
32      diameter of DPM was 0.35 ± 0.10 um (mean ± SD).  The gas-phase components of the diesel
33      exhaust atmospheres are presented in Appendix A.
34             The effects of exposure to either filtered or unfiltered exhaust were described by Heinrich
35      et al. (1986a) and Stober (1986).  Exposure to unfiltered exhaust resulted in 8 bronchoalveolar

        2/1/98                                    7-10       DRAFT-DO NOT CITE OR QUOTE

-------
  1      adenomas and 9 squamous cell tumors in 15 of 95 rats examined for a 15.8% tumor incidence,
  2      Although statistical analysis was not provided, the increase appears to be highly significant. In
  3      addition to the bronchioalveolar adenomas and squamous cell tumors, there was a high incidence
  4      of bronchioalveolar hyperplasia (99%) and metaplasia of the bronchioalveolar epithelium (65%).
  5      No tumors were reported among female Wistar rats exposed to filtered exhaust (n = 92) or clean
  6      air(n = 96).
  7            Mohr et al. (1986) provided a more detailed description of the lung lesions and tumors
  8      identified by Heinrich et al. (1986a, 1986b) and Stober (1986). Substantial alveolar deposition of
  9      carbonaceous particles was noted for rats exposed to the unfiltered diesel exhaust. Squamous
10      metaplasia was observed in 65.3% of the rats breathing unfiltered diesel exhaust but not in the
11      control rats. Of nine squamous cell tumors, one was characterized as a Grade I carcinoma
12      (borderline atypia, few to moderate mitoses, and slight evidence of stromal invasion), and the
13.  remaining eight were classified as benign keratinizing cystic tumors.
14            The effect of chronic (19 h/day, 5 days/week, 2 to 2,5 years) diesel exhaust exposure on
15      the tumor-inducing effect of diphenylnitrosamine (DPN) was examined using female Wistar rats
16      (Heinrich et al., 1986a; Stober, 1986; Heinrich, 1989a). Groups of rats (45 to 48 per group) were
17      exposed to clean air or whole diesel exhaust (particle concentration of 4.24 mg/m3, as described
18      previously) and administered by s.c. injection 250 or 500 mg DPN/kg/week during the first 25
19      weeks of exposure. The total DPN dose administered equaled 6.25 or 12.5 g/kg of body weight.
20      The concentrations of B[a]P, benzo[e]pyrene (B[e]P), and chrysene in the diesel exhaust were 13,
21      21, and 76 ng/m3, respectively.
22            The overall tumor rate in the lungs of DPN-treated rats was not affected by the exposure
23      to either filtered or whole diesel engine exhaust. However, when only pulmonary squamous cell
24      carcinomas were considered, the exposure to whole diesel exhaust significantly (p<,0.05)
25      increased the tumor incidence (Table 7-1). Conversely, the high level of nasal tumors induced
26      by DPN was significantly decreased in the rats exposed to the diesel engine emissions.
27            Heinrich et al. (1986b) and Mohr et al. (1986) compared the effects of exposure to
28      particles having only a minimal carbon core but a much greater concentration of P AHs than
29      DPM does.  The desired exposure conditions were achieved by mixing coal oven  flue gas with
30      pyrolyzed pitch. The concentration of B[a]P and other PAHs per milligram of DPM was about
31      three orders of magnitude greater than that of diesel exhaust. Female rats were exposed to the
32      flue gas-pyrolyzed pitch for 16 h/day, 5 days/week at particle concentrations of 3  to 7 mg/m3 for
33      22 mo, then held in clean air for up to an additional 12 mo.  Among 116 animals exposed, 22
34      tumors were reported in 21 animals, for an incidence of 18.1 %. One was a bronchioloalveolar
35      adenoma, one was a bronchioloalveolar carcinoma, and 20 were squamous-cell tumors. Among

        2/1/98                                   7-11       DRAFT-DO NOT CITE OR QUOTE

-------
 1      the latter, 16 were classified as benign keratinizing cystic tumors, and 4 were classified as
 2      carcinomas. No tumors were reported in 115 controls. The tumor incidence in this study was
 3      comparable to that reported previously for the diesel exhaust-exposed animals.
 4            The importance of DPM and of insoluble respirable particles in tumorigenic responses of
 5      rats was investigated and reported by Heinrich et al. (1995).  In this chronic inhalation exposure
 6      study, female Wistar rats were exposed to whole diesel exhaust (0.8,2.5, or 7.0 mg/m3), 18
 7      h/day, 5 days/week for up to 24 mo. Groups of rats were also exposed to ultrafme Ti02 particles
 8      (10 mg/m3) and carbon black particles (11.6 mg/m3) using the same exposure regimen, except
 9      that after 4 mo the exposure concentrations of TiO2 and carbon black were increased to obtain
10      lung particle burdens similar to those observed in rats exposed to whole diesel exhaust. Controls
11      were exposed to clean air. Following exposure to the test atmospheres, the rats were maintained
12      in clean air atmospheres for an additional 6 mo.
13            In analyzing the studies of Heinrich et al. (1986a,b), Heinrich (1990), Mohr et al. (1986),
14      and Stober (1986), it must be noted that the incidence of lung tumors occurring following
15      exposure to whole diesel exhaust, coal oven flue gas, or carbon black (15.8%, 18.1%, and 8% to
16      17%, respectively) was very similar. This occurred despite the fact that the PAH content of the
17      P AH-enriched pyrolyzed pitch was more than three orders of magnitude greater than that of
18      diesel exhaust; carbon black, on the other hand, had only traces of PAHs. Based on these
19      findings, the organic fraction is not the sole cause of tumor induction by diesel exhaust. This
20      issue is discussed further in Chapter 10.       .
21      .      In the Heinrich et al. (1995) report, the cumulative exposures for the rats in the various
22      treatments groups were 61.7,21.8, and 7-4 g/m3 x h for the high, medium, and low whole-
23      exhaust exposures and 102.2 and 88.1 g/m3 x h for the carbon black and TiO2 groups,
24      respectively. For tumor incidence comparison (number of rats with tumors) among the diesel
25      exhaust exposure groups, significant increases were observed in the high (22/100;/K0.001) and
26      mid (11/200; /K0.01) exposure groups relative to clean air controls (Table 7-1). Only one tumor
27      (1/217), an adenocarcinoma, was observed in clean air controls. Relative to clean air controls,
28      significantly increased incidences were observed in the high exposure rats for benign squamous
29      cell tumors  (14/100; p<0.001), adenomas (4/100; /X0.01), and adenocarcinomas (5/100;
30      p<0.05). Only the incidence of benign squamous cell tumors (7/200; /?<0.01) was significantly
31      increased in the mid-exposure group relative to the clean air controls. In comparing the number
32      of rats with tumors (including benign squamous cell tumors), incidences in the TiO2 exposure
33      group (32/100) were similar to that of the high diesel exhaust exposure (22/100), but the .
34      incidence in the carbon black group (39/100) was significantly greater (p<0.01) than that of the
35      high diesel exhaust.  For the carbon black group, the incidences of adenomas (13/100) and

        2/1/98                                 .  7-12       DRAFT-DO NOT CITE OR QUOTE

-------
  1      adenoeareinomas (13/100) were significantly greater (p<0.05) than in the high diesel exhaust  •
  2      group.  The incidence of adenoeareinomas (13/100) was also significantly greater (p<0.05) in the
  3      TiO2 group than in diesel exhaust-exposed rats.
  4             Particle lung burden and alveolar clearance also were determined in the Heinrich et al,
  5      (1995) study. Even with adjustments to the exposure protocol to increase lung particle burden, at
  6      12 mo of exposure and beyond the particle lung burden in high diesel exhaust group (63,878
  7      fig/animal) was greater than that of the TiO2 (43,854 ^g/animal) and carbon black (39,287
  8      ng/animal) groups. Relative to clean air controls, alveolar clearance was significantly
  9      compromised by exposure to mid and high diesel exhaust, T1O2, and carbon black after 3 mo of
10      exposure.  Exposure to the high concentration of diesel exhaust resulted in a greater reduction of
11      alveolar clearance than exposure to either TiO2  or carbon black did. For the high diesel exhaust,
12      TiO2, and carbon black exposure groups, the 3-mo recovery time in clean air failed to reverse the
13      compromised alveolar clearance.
14             A study conducted at the Inhalation Toxicology Research Institute also compared the
15      lung tumor response of rats chronically exposed to diesel exhaust and carbon black (Nikula et al.
16      1995).  In this study, male (114-115 per exposure group) and female (114-116 per exposure
17      group) F344 rats were exposed to either carbon black or diesel exhaust for 16 h/day, 5 days/week
18      to particle  concentrations of 2.5 or 6.5 mg/m3 for up to 24 mo. Controls (118 males, 114
19      females) were exposed to  clean air. The progressive pulmonary accumulation of DPM tended to
20      be more rapid than that for carbon black and accumulations  of both tended to accelerate after 12
21      mo, a finding similar to that reported by Heinrich et al. (1994). At 23 mo, mean lung burdens of
2.2      females exposed to low carbon black or high carbon black were  17.3 and 36.9 mg, respectively;
23      and for males, the lung burdens were 24.7 and 40.1 mg, respectively. For low and high diesel
24      exhaust exposure, the lung burdens were 36.7 and 80.7 mg,  respectively, for females and 45.1
25      and 90.1 mg, respectively, for males. Both diesel exhaust and carbon black were pulmonary
26      carcinogens under the exposure conditions of the study.  The percentages of susceptible rats
27      (males and females combined) with malignant neoplasms were 0.9 (control), 3.8 (low carbon
28      black), 11.8 (high carbon black), 3.3 (low diesel exhaust), and 12.3 (high diesel exhaust). The
29      percentages of rats (males and females combined) with malignant or benign neoplasms were 1.4
30      (control), 4.7 (low carbon  black), 15.2 (high carbon black), 6.2 (low diesel exhaust), and 17.9
31      (high diesel exhaust). All primary neoplasms were associated with the parenchyma rather than
32      the conducting airways of the lungs. The first lung neoplasm was observed at 15 mo.  The
33      specific tumor types and incidences are shown in Table 7-1.  Analysis of the histopathologic data
34    -  suggested a progressive process from alveolar epithelial hyperplasia to adenomas and
35      adenoeareinomas. The neoplastic responses to carbon black and diesel exhaust were similar,

        2/1/98               •                    7-13       DRAFT-DO NOT CITE OR QUOTE

-------
 1      indicating that the organic fraction adsorbed to the DPM was not contributing significantly to the
 2      carcinogenic response. Although these data provide indirect evidence that the DPM-associated
 3      organic fraction of diesel exhaust did not play a significant role in the observed carcinogenic
 4      response, the data do not prove that the organic fraction has no role whatsoever. However, if
 5      DPM-associated organics are involved, the great difference in organic fraction content between
 6      carbon black and diesel exhaust (i.e., three orders of magnitude) suggests that its role was minor
 7      in the tumorigenic response of the rats in this study.
 8            A long-term inhalation study (Ishinishi et al., 1988a; Takaki et al., 1989) examined the
 9      effects of emissions from light-duty (LD) and heavy-duty (HD) diesel engines on male and
10      female Fischer 344/Jcl rats. The LD engines were 1.8-L, 4-cylinder, swirl-chamber-type power
11      plants, and the HD engines were 11-L, 6-cylinder, direct-injection-type power plants. The
12      engines were connected to eddy-current dynamometers and operated at 1,200 rpm (LD engines)
13      and 1700 rpm (HD engines). Nippon Oil Co. JIS No. 1  or No. 2 diesel fuel was used.  The 30-
14      mo whole-body exposure protocol (16 h/day, 6 days/week) used DPM concentrations of 0, 0.5, 1,
15      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
16      B[a]P concentrations were reported as 4.4 and 2.8 ug/g of particulate matter, and 1-nitropyrene
17      concentrations were 57.1 and  15.3 ug/g  of particulate matter for the LD and HD engines,
18      respectively. An analysis of gas-phase components is presented in Appendix A. The animals
19      inhaled the exhaust emissions from 1700 to 0900 h.  Sixty-four male rats and 59 to 61  female rats
20      from each exposure group were evaluated for carcinogenicity.
21            For the experiments using the LD series engines, the highest incidence of hyperplastic
22      lesions plus tumors (72.6%) was seen in the highest exposure (2.3 mg/m3) group.  However, this
23      high value was the result of the 70% incidence of hyperplastic lesions; the incidence of
24      adenomas was only 0.8% and that of carcinomas 1.6%.  Hyperplastic lesion incidence was
25      considerably lower for the lower exposure groups (9.7%, 4.8%, 3.3%, and 3.3% for the 1.1, 0.4,
26      and 0.1 mg/m3 and control groups,  respectively). The incidence of adenomas and carcinomas,
27      combining males and females, was not significantly different among exposure groups (2.4%,
28    .  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,
29      respectively).                                              .   -  '      •
30            For the experiments using the HD series engines, the total incidence of hyperplastic
31      lesions, adenomas, and carcinomas was highest (26.6%) in the 3.7 mg/m3 exposure group.  The
32      incidence of adenomas plus carcinomas for males and females combined equaled 6:5%, 3.3%,
33      .0%, 0.8%, and 0.8% at 3.7, 1.8, 1, and 0.4 mg/m3 and for controls,  respectively. A statistically
34      significant difference was reported  between the 3.7 mg/m3 and the control groups for the HD
        2/1/98                                    7-14       DRAFT-DO NOT CITE OR QUOTE

-------
 1      series engines.  A progressive dose-response relationship was not demonstrated. Tumor
 2      incidence data for this experiment are presented in Table 7-1.
 3            The Ishinishi et al. (1988a) study also included recovery tests in which rats exposed to
 4      whole diesel exhaust (DPM concentration of 0.1 or 1.1 mg/m3 for the LD engine and 0.5 or
 5      1.8 mg/m3 for the HD engine) for 12 mo were examined for lung tumors following 6-, 12-, or 18-
 6      mo recovery periods in clean air. The incidences of neoplastic lesions were low, and pulmonary
 7      DPM burden was lower than for animals continuously exposed to whole diesel exhaust and not
 8      provided a recovery period. The only carcinoma observed was in a rat examined 12 mo
 9      following exposure to exhaust (1.8 mg/m3) from the HD engine.
10            Iwai et al. (1986) also examined the long-term effects of diesel exhaust inhalation on
11      female F344 rats. The exhaust was generated by a 2.4 L displacement truck engine. The exhaust
12      was diluted 10:1 with  clean air at 20°C to 25 °C and 50% relative humidity. The engines were   .
13      operated at 1,000 rpm with an 80% engine load. These operating conditions were found to
14      produce exhaust with the highest particle concentration and lowest NO2 and SO2 content For
15      those chambers using filtered exhaust, proximally installed high-efficiency particulate air
16      (HEP A) filters were used. Three groups of 24 rats each were exposed to unfiltered diesel
17      exhaust, filtered diesel exhaust, or filtered room air for 8 h/day, 7 days/week for 24 mo. Particle
18      concentration was 4.9  mg/m3 for unfiltered exhaust. Concentrations of gas-phase exhaust
19      components were 30.9 ppm NOX, 1.8 ppm NO2,13.1 ppm SO2, and 7.0 ppm CO.
20            No lung tumors were found in the 2-year control (filtered room air) rats, although one
21      adenoma was noted in a 30-mo control rat, providing a spontaneous tumor incidence of 4.5%.
22      No lung tumors were observed in rats exposed to filtered diesel exhaust.  Four of 14 rats exposed
23      to unfiltered diesel exhaust for 2 years developed lung tumors, two of these were malignant.
24      Five rats of this 2-year exposure group were subsequently placed in clean room air for 3 to 6 mo
25      and four eventually (time not specified) exhibited lung tumors (three malignancies).  Thus, the
26      lung tumor incidence for total tumors was 42.1 % (8/19) and 26.3% (5/19) for malignant tumors
27      in rats exposed to whole diesel exhaust. The tumor types identified were adenoma (3/19),
28      adenocarcinoma (1/19), adenosquamous carcinoma (2/19), squamous carcinoma (1/19), and -
29      large-cell  carcinoma (1/19). The lung tumor incidence in rats exposed to whole diesel exhaust
30      was significantly greater than that of controls (p<0.01).  Tumor data are summarized in Table 7-
31      1. Malignant splenic lymphomas were detected in 37.5% of the rats in the filtered exhaust group
32      and in 25.0% of the rats in the unfiltered exhaust group, these values were significantly
33      (p^O. 05)  greater than  the 8.2% incidence noted in the control rats. The study demonstrates
34      production of lung cancer in rats following 2-year exposure to unfiltered diesel exhaust. In
35      addition, splenic malignant lymphomas occurred during exposure to both filtered and unfiltered

        2/1/98                                   7-15       DRAFT-DO NOT CITE OR QUOTE

-------
 1      diesel exhaust. This is the only report to date of tumor induction at an extrarespiratory site by
 2      inhaled diesel exhaust in animals.
 3            A chronic (up to 24 mo) inhalation exposure study by Takemoto et al. (1986) was
 4      conducted to determine the effects of diesel exhaust, di-isopropanol-nitrosamine (DIPN), and
 5      diesel exhaust following DIPN treatment of female F344/Jcl rats.  One mo after initiation of
 6      inhalation exposures, DIPN was administered i.p. at 1 mg/kg weekly for 3 weeks to clean air-
 7      and diesel-exposed groups of rats. Uninjected groups were also exposed to clean air and diesel
 8      exhaust. The treatment protocol consisted of exposure to diesel exhaust for 4 h/day, 4
 9      days/week.  The diesel exhaust was generated by a 269-cc engine operated at an idle state {1,600
10      rpm). Concentrations of the gas-phase components of the exhaust are presented in Appendix A.
11      The particle concentration of the diesel exhaust in the exposure chamber was 2 to 4 mg/m3.
12      B[a]P and 1-nitropyrene concentrations were 0.85 and 93 ug/g of particles, respectively.
13            In the Takemoto et al. (1986) study, no lung tumors were reported in either uninjected
14      controls or diesel-exposed animals. Among injected animals autopsied at 12 to 17 mo, 2
15      adenomas were reported in 8 rats exposed to clean air compared with 12 adenomas and 3
16      adenocarcinomas in 18 diesel-exposed rats.  Among injected rats autopsied at 18 to 24  mo, 10
17      adenomas and 4 adenocarcinomas were seen in 21 animals exposed to clean air compared with
18      12 adenomas and 7 adenocarcinomas in 18 diesel-exposed rats. According to the authors, the
19      incidence of malignant tumors was not significantly increased in either of the diesel exhaust-
20      exposed groups when compared with the appropriate control group. Tumor incidence data for
21      the various treatment protocols are presented in Table 7-1.  It was also noted that the diesel
22      engine employed in this study was originally used as an electrical generator and that its operating
23      characteristics (not specified) were different from those for a diesel-powered automobile.
24      However, the investigators deemed it suitable for assessing the of effects  diesel emissions.
25            Brightwell et al. (1986, 1989) studied the effects of filtered and unfiltered diesel exhaust
26      on male and female F344 rats. The diesel exhaust was generated by a 1.5-L Volkswagen engine
27      that was computer-operated  according to the U.S. 72 FTP driving cycle. The engine emissions
28      were diluted by conditioned air delivered at 800 m3/h to produce the high-exposure (6.6 mg/m3)
29      diesel exhaust atmosphere. Further dilutions of 1:3 and 1:9 produced the  medium- (2.2 mg/m3)
30      and low- (0.7 mg/m3) exposure atmospheres. Filtered diesel  exhaust was  generated by a similar
31      engine.  The CO and NOX concentrations (mean ± SD) were 32 ± 11 ppm and 8 ± 1 ppm for the
32      unfiltered diesel exhaust (high-exposure concentration chamber) and 32 ± 11 and 8 ± 1 for the
33      filtered diesel exhaust. The  inhalation exposures were conducted overnight to provide  five 16-h
34      periods per week for 2 years; surviving animals were maintained for an additional  6 mo.
        2/1/98                                    7-16       DRAFT-DO NOT CITE OR QUOTE

-------
 1            For males and females combined, a 9.7% (14/144) and 38,5% (55/143) incidence of
 2      primary lung tumors occurred in F344 rats following exposure to 2.2 and 6.6 mg of DPM/m3,
 3      respectively (Table 7-1).  The tumor incidence in the 0.7 mg/m3 exposure group was 0.7%
 4      (1/144) and that of controls was 1.2% (3/260). Diesel exhaust-induced tumor incidence in rats
 5      was dose-related and higher in females than in males (Table 7-1). These data included animals
 6      sacrificed at the interim periods (6, 12,18, and 24 mo); therefore, the tumor incidence does not
 7      accurately reflect the effects of long-term exposure to the diesel exhaust atmospheres.  When
 8      tumor incidence is expressed relative to the specific intervals, a lung rumor incidence of 96%
 9      (24/25), 76% (19/25) of which were malignant, was reported for female rats in the high dose
10      group exposed for 24 mo and held in clean air for the remainder of their lives.  For male rats in
11      the same group, the tumor incidence equaled 44% (12/27), of which 3 7% (10/27) were
12      malignant. It was also noted that many of the animals exhibiting tumors had more than one
13      tumor, often representing multiple histological types. The types of tumors identified in the rats
14      exposed to diesel exhaust included adenomas, squamous cell carcinomas, adenocarcinomas,
15      mixed adenoma/adenocarcinomas, and mesotheliomas. Similar to other studies, the tumor
16      incidence in rats occurred during exposure to whole exhaust rather than filtered exhaust.  It must
17      be noted, however, that the exposure during darkness (when increased activity would result in
18      greater respiratory exchange and greater inhaled dose) could account, in part, for the high
19      response reported for the rats.
20            Karagianes et al. (1981) exposed male Wistar rats (40 per group) to diesel engine exhaust
21     , diluted to a DPM concentration of 8.3 (± 2.0) mg/m3, room air, diesel engine exhaust (8.3
22      mg/m3) plus low-concentration coal dust (5.8 mg/m3), low-concentration coal dust only (6.6
23      mg/m3), or high-concentration coal dust (14.9 mg/m3) 6 h/day, 5 days/week for up to 20 mo. The
24      exhaust-generating system and exposure atmosphere characteristics are presented in Appendix A.
25      The type of engine used (3-cylinder, 43 bhp diesel) is normally used in mining situations and was
26      connected to an electric generator and operated at varying loads and speeds to simulate operating
27      conditions in an occupational situation.  To control the CO concentration at 50 ppm, the exhaust
28      was diluted 35:1 with clean air.
29            One broncbiolar adenoma was detected in the group exposed to diesel exhaust alone and
30      one in the rats receiving combined exposures. No lung tumors were reported in controls or
31      following exposure to either high or low concentrations of coal dust. The equivocal response
32      may have been caused by the relatively short exposure durations (20 mo).  In the Mauderly et al.
33      (1987) study, by comparison, most of the tumors were detected in rats exposed for more than 24
34      mo.                                     '
        2/1/98                                    7-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1            Lewis et al. (1989) also examined the effects of inhalation exposure of diesel exhaust
 2      and/or coal dust on tumorigenesis on F344 rats. Groups of 216 male and 72 female rats were
 3      exposed to clean air, whole diesel exhaust (2 mg soot/m3), coal dust (2 mg/m3 respirable
 4      concentration; 5 to 6 mg/m3 total concentration), or diesel exhaust plus coal dust (1 mg/m3 of
 5      each respirable concentration; 3.2 mg/m3 total concentration) for 7 h/day, 5 days/week for up to
 6      24 mo. Groups of 10 or more males were  sacrificed at intermediate intervals (3, 6, and 12 mo).
 7      The diesel exhaust was produced by a 7.0-L, 4-cycle, water-cooled Caterpillar Model 3304
 8      engine using No. 2 diesel fuel (<0.5% sulfur by mass). The exhaust was passed through a
 9      Wagner water scrubber, which lowered the exhaust temperature and quenched engine backfire.
10      An analysis of the exposure atmospheres is presented in Appendix A.
11            Histological examination was performed on 120 to 121 male and 71 to 72 female rats
12      terminated after 24 mo of exposure. The exhaust exposure did not significantly affect the tumor
13      incidence beyond what would be expected for aging F344 rats. There was  no postexposure
14      period, which may explain, in part, the lack of significant tumor induction. The particulate
15      matter concentration was also less than the effective dose in several other studies.
16            General Motors Research Laboratories sponsored chronic inhalation studies using male
17      Fischer 344 rats exposed to DPM concentrations of 0.25, 0.75, or 1.5 mg/m3 (Kaplan et al., 1983;
18      White et al., 1983).  The exposure protocol for this study conducted at the Southwest Research
19      Institute was 20 h/day, 7 days/week for 9 to 15 mo.  Some animals were sacrificed following
20      completion of exposure, while others were returned to clean air atmospheres for an additional 8
21      mo. Control animals received clean air. Exhaust was generated by  5.7-L Oldsmobile engines
22      (four different engines used throughout the experiment) operated at  a steady speed and  load
23      simulating a 40-mph driving speed of a full-size passenger car. Details of the exhaust-generating
24      system and exposure atmosphere are presented in Appendix A.
25            Five instances of bronchoalveolar carcinoma were observed in 90 rats exposed to diesel
26      exhaust for 15 mo and held an additional 8 mo in clean air. These included one tumor in the 0.25
27      mg/m3 group, three in the 0.75 mg/m3 group, and one in the 1.5 mg/m3 group. Rats kept in clean
28      air chambers for 23 mo did not exhibit any carcinomas. No tumors  were observed in any of the
29      180 rats exposed to diesel exhaust for 9 or 15 mo without a recovery period or in the respective
30      controls for these groups. Although the increases in tumor incidences  in the groups exposed for
31      15 mo and held an additional 8 mo in clean air were not statistically significant, they suggest an
32      effect because the background incidence for this specific lesion in this strain of rat is low.
33
        2/1/98               •  •'                 7-18       DRAFT-DO NOT CITE OR QUOTE

-------
  1      7.2.2. Mouse Studies
  2            Heinrich et al. (1986a) and Stober (1986), as part of a larger study, also evaluated the
  3      effects of diesel exhaust in mice. Details of the exposure conditions reported by Heinrich et al.
  4      (1986b) are given in Appendix A, Following lifetime (19 h/day, 5 days/week, for a maximum of
  5      120 weeks) exposure to filtered (n = 93) and unfiltered (n = 76) diesel exhaust (4.2 mg/m3),
  6      female NMRI mice exhibited a total lung tumor incidence of adenomas and adenocarcinomas
  7      combined of 31% (filtered) and 32% (unfiltered), respectively. Tumor incidences reported for
  8      control mice (n = 84) equaled 11% for adenomas and adenocarcinomas. combined. The effects
  9      are more dramatic when the incidences of only malignant tumors (adenocarcinomas) are
10      considered, 2.4% for controls, 19% for filtered exhaust, and 17% for unfiltered exhaust. This is
11      the only reported study in which filtered exhaust resulted in a definitive tumorigenic response in
12      the lungs of mice. These data are summarized in Table 7-1.
13            As part of the same study, groups of 64 female NMRI mice of 8 to 10 weeks of age were
14      dosed weekly with either 50 or 100 ug B[a]P intratracheally for 20 or 10 weeks, respectively, for
15      a total dose of 1 mg. Another group received 50 ug dibenz[a,/j]anthracene (DBA) intratracheally
16      for 10 weeks.  Additional groups of 96 newborn mice received one s.c. injection of 5 or 10 u-g of
17      DBA between 24 and 48 h after birth. The animals were concomitantly exposed to either diesel
18      exhaust or clean air. The mice receiving intratracheal instillations were observed throughout
19      their lifespan but the newborn mice were sacrificed after 6 mo. Although the chemical
20      treatments resulted in  large increases in lung tumor incidence, exposure to diesel exhaust did not
21      enhance this effect and in some cases even resulted in inhibition. For example, lung tumor rates
22      in clean air mice treated with 20 instillations of B[a]P equaled 71% compared with 41% for mice
23      similarly instilled but exposed to diesel exhaust. The decrease resulted from a smaller number of
24      adenocarcinomas, whereas the adenoma incidence remained unchanged. The high dose of DBA
25      injected into newborn mice also resulted in a greater tumor incidence in mice exposed to clean
26      air (81 %) than in the diesel exposed group (63%).  Effects of the other treatments were
27      apparently not inhibited by diesel exhaust exposure, although complete incidence data were not
28    .  reported.  The authors did not speculate on the reasons for this unexpected effect.
29            In addition to the studies using rats, investigators at the Fraunhofer Institute also
30      examined the effects in mice following long-term exposure to whole diesel exhaust, TiO2, and
31      carbon black particles (Heinrich et al., 1995). In these studies, NMRI mice were exposed (18
32      h/day, 5 days/week) to whole diesel exhaust (4.5 mg/m3 for 13.5 mo), TiO2 (7.0 mg/m3 for 4 mo
33      followed by 15.0 mg/m3 for 4 mo and 10 mg/m3 for 5.5 mo), or carbon black (7.0 mg/m3 for 4
34      mo followed by 12 mg/m3 for 9.5 mo). Following the 13.5-mo exposures, animals were kept in
35      clean air for a total experiment time of 23 mo. Controls were exposed to clean air. The lung

        2/1/98  .                                 7-19      DRAFT-DO NOT CITE OR QUOTE

-------
 1     burdens of the NMRI mice after 12 mo of exposure were 7.0, 7.4, and 5.2 mg/lung for DPM,
 2     carbon black, and TiO2, respectively. Adenomas and adenocarcinomas were the only tumor
 3     types observed in the mice.  Percentages for adenomas/adenocarcinomas were 21.8%/15.4% for
 4     whole diesel exhaust, 11.3%/10% for carbon black, 11.3%/2.5% for TiO2, and 25%/15.4% for
 5     clean air controls. The lung tumor rates for adenomas/adenocarcinomas were 32.1% (diesel
 6     exhaust), 20% (carbon black), 13.8% (TiO2), and 30% (clean air controls).
 7            A comparison between particle-free exhaust and whole exhaust was also conducted using
 8     NMRI and C57BL mice (Heinrich et al., 1995). In the experiments with NMRI mice, mice were
 9     exposed to whole diesel exhaust (4.5 mg/m3) and another group was exposed to particle-free
10     diesel exhaust for 23 mo.  The lung tumor rates for the groups of mice were 23% (whole
11     exhaust), 46.7% (particle-free exhaust), and 30% (clean air controls). The difference between the
12     controls and those exposed to the particle-free exhaust was not significant (/?=0.053). The
13     percentages of adenomas/adenocarcinomas were 18.3%/5% (whole exhaust), 31.7%/l 5%
14     (particle-free exhaust), and 25%/8.8% (clean air controls).  In the experiments using C57BL
15     mice, the exposure groups were similar except that the exposure duration was 24 mo. The tumor
16     rates (including an additional 6-mo exposure in clean air) were not significantly different among
17     the whole exhaust (8.5%), particle-free exhaust (3.5%), and clean air controls (5.1%).
18            The lack of a carcinogenic response in mice was reported by Mauderly et al.  (1996).  In
19     this study, groups of 540 to 600 CD-I male and female mice  were exposed to whole diesel
20     exhaust (7.0, 3.5, or 0.35 mg DPM/m3) for 7 h/day, 5 days/week for up to 24 mo.  Controls were
21     exposed to filtered air. DPM accumulation in the lungs of exposed mice was assessed at 6, 12,
22     and 18 mo of exposure and was shown to be progressive; DPM burdens were 0.2 ± 0.02, 3.7 ±
23     0.16, and 5.6 ± 0.39 mg for the low-, medium-, and high-exposure groups, respectively.  The
24     lung burdens in both the medium- and high-exposure groups  exceeded that predicted by exposure
25     concentration ratio to the low-exposure group.  Contrary to what was observed in rats (Heinrich
26     et al., 1986a; Stober, 1986; Nikula et al., 1995; Mauderly et al., 1987), an exposure-related
27     increase in primary lung neoplasms was not observed in the CD-I mice, supporting the
28     contention of a species difference in the pulmonary carcinogenic response to poorly  soluble
29     particles. The percentage incidence of mice (males and females combined) with one or more
30     malignant or benign neoplasms was 13.4, 14.6, 9.7, and 7.5 for controls and low-, medium-, and
31     high-exposure groups, respectively.
32            Takemoto et al. (1986) reported the effects of inhaled diesel exhaust (2 to 4 mg/m3, 4
33     h/day, 4 days/week, for up to 28 mo) in ICR and C57BL mice exposed from birth. Details of the
34     exposure conditions are presented in Appendix A. Among male and female ICR mice autopsied
35     at 13 to  18 mo, 4 adenomas and 1 adenocarcinoma were detected in 34 diesel exhaust-exposed

       2/1/98                                    7-20        DRAFT-DO NOT CITE OR QUOTE

-------
 1      mice compared with 3 adenomas among 38 controls. Among animals autopsied at 19 to 28 mo,
 2      6 adenomas and 3 adenocarcinomas were seen in 22 exposed animals compared with 3 adenomas
 3      and one adenoearcinoma in 22 controls. Among combined male and female C57BL mice
 4      autopsied at 13 to 18 mo, 4 adenomas and 2 adenocarcinomas were detected in 79 animals
 5      autopsied compared with none in 19 unexposed animals. Among males and females autopsied at
 6      19 to 28 mo, 8 adenomas and 3 adenocarcinomas were detected in 71 exposed animals compared
 7      with 1 adenoma among 32 controls. No' significant increases in either adenoma or
 8      adenoearcinoma incidences were reported for either strain of exposed mice. Although not tested
 9      by the authors, the combined incidence of adenomas and adenocarcinomas (11/71) in male and
10      female C57BL mice exposed to diesel exhaust for 19 to 28 mo versus that found in controls
11      (1/32), however, appears to be a significant increase. Although the results are not definitive,
12      there is the strong suggestion of an effect, especially since the C57BL strain has a low
13      background lung tumor incidence.  See Table 7-1 for details of tumor incidence.
14            Pepelko and Peirano (1983) summarized a series of studies on the health effects of diesel
15      emissions in mice. Exhaust was provided by two Nissan CN 6-33, 6-cylinder, 3.24-L diesel
16      engines coupled to a Chrysler A-272 automatic transmission and Eaton model 758-DG
17      dynamometer. Details of the exhaust-generating system and the exposure atmosphere are
18      presented in Appendix A. Sixty-day pilot studies were conducted at a 1:14 dilution, providing
19      DPM concentrations of 6 mg/m3' The engines were operated using the Modified California
20      Cycle. These 20-h/day, 7-days/week pilot studies using rats, cats, guinea pigs, and mice produced
21      decreases in weight gain and food consumption. Therefore, at the beginning of the long-term
22      studies, exposure time was reduced to 8 h/day, 7 days/week at an exhaust DPM concentration of
23      6 mg/m3:  During the final 12 mo of exposure, however, the DPM concentration was increased to
24      12 mg/m3. For the chronic studies, the engines were operated using the Federal Short Cycle.
25            Pepelko and Peirano (1983) described a two-generation study using Sencar mice exposed
26      to diesel exhaust alone or treated with either tumor initiators or promoters. Male and female
27      parent generation mice were exposed to diesel exhaust at a DPM concentration  of 6 mg/m3 prior
28      to (from weaning to sexual maturity) and throughout mating.  The dams continued exposure
29      through gestation, birth, and weaning. Groups of offspring (130 males and 130 females)
30      received i.p. injections of either butylated hydroxytoluene (BHT) (300 mg/kg for week 1, 83
31      mg/kg for week 2, and 150 mg/kg from  week 3 to 1 year), a single i.p. injection of 1 mg urethan
32      at 6 weeks of age, or no injections. The exhaust exposure was increased to a DPM concentration
33      of 12 mg/m3 when the offspring were 12 weeks of age and was maintained until termination of
34      the experiment when the mice were 15 mo old.
        2/1/98                                   7-21        DRAFT-DO NOT CITE OR QUOTE

-------
 1            The incidence of pulmonary adenomas (16.3%) was significantly increased in the non-
 2      injected female mice exposed to diesel exhaust compared with 6.3% in clean air controls. The
 3      incidence in males and females combined was 10.2% in 205 animals examined compared with
 4      5.1% in 205 clean air controls. This difference was also significant. The incidence of
 5      carcinomas was not affected by exhaust exposure in either sex.  Exhaust exposure reduced the
 6      adenoma incidence in female mice receiving BHT (3.9% vs. 16.7%). The response to BHT in
 7      males or urethan in both sexes was unaffected by diesel exposure. These results provided the
 8      earliest evidence for cancer induction following inhalation exposure to.diesel exhaust. The
 9      limited response may well have been influenced by the relatively early sacrifice times of the
10      mice.  On the other hand, an increase in the sensitivity of the study, allowing detection of tumors
11      at 15 mo, may have been the result of exposure from conception. It is interesting to note that in
12      this study diesel exposure appeared to inhibit effects of tumor promotion, whereas Stdber (1986)
13      reported diesel exposure inhibition of complete carcinogens. These data are summarized in
14      Table 7-1.
15            A series of inhalation studies using strain A mice was conducted by Orthoefer et al.
16      (1981). In assays with the strain A, mice are usually given a series of intraperitoneal injections
17      with the test agent; they are then sacrificed at about 9 mo of age and examined for lung tumors.
18      In the present series, inhalation exposure was substituted. In the current series, groups of 25
19      male Strong A mice were exposed to irradiated (to  simulate chemical reactions induced by
20      sunlight) or nonirradiated diesel exhaust (6 mg/m3) for 20 h/day, 7 days/week for 7 weeks.
21      Additional groups of 40 Jackson A (20 of each sex) were exposed similarly to either clean air or
22      diesel exhaust, then held in clean air until sacrificed at 9 mo of age.  No tumorigenic effects were
23      detected at 9 mo of age. Further studies were conducted in which male A/Strong mice were
24      exposed 8 h/day, 7 days/week until sacrifice (approximately 300 at 9 mo of age and
25      approximately 100 at 12 mo of age). With the exception of those treated with urethan, the
26      number of tumors per mouse did not exceed historical control levels in any of the studies.
27      Exposure to diesel exhaust, however, significantly inhibited the tumorigenic effects of the 5-mg
28      urethan treatment. Results are listed in Table 7-1.
29            Kaplan et al. (1982) also reported the effects of diesel exposure in strain A mice. Groups
30      of male strain A/J mice were exposed for 20 h/day, 7 days/week for 90 days and held until 9 mo
31      of age. Experimental conditions are described in Appendix A.  Briefly, the animals were
32      exposed to diesel exhaust at DPM concentrations of 0,0.25,0.75, or 1.5 mg/m3. Controls were
33      exposed to clean air.  Among 458 controls and 485 exposed animals, tumors were detected in
34      31.4% of those breathing clean air versus 34.2% of those exposed to diesel exhaust. The mean
35      number of tumors per mouse also failed to show significant differences.

        2/1/98                                    7-22      DRAFT-DO NOT CITE OR QUOTE

-------
  1             In a follow-up study, strain A mice were exposed to diesel exhaust for 8 mo (Kaplan et
  2      al, 1983; White et al., 1983). After exposure to the highest exhaust concentration (1.5 mg/m3),
  3      the percentage of mice with pulrrionary adenomas and the mean number of tumors per mouse
  4      were significantly less (p<0.05) than those for controls (25.0% vs. 33.5% and 0.30 ± 0.02 [S.E.]
  5      vs. 0.42 ±0.03 [S.E.]) (Table 7-1).
  6
  7      7.2.3. Hamster Studies
  8             Heinrieh et al. (1982) examined the effects of diesel exhaust exposure on the tumor
  9      frequency in female Syrian golden hamsters pretreated with the tumor initiators
10      dibenzo[a,h]anthracene (DBA) or diethylnitrosamine (DEN).  At the time of this work, it was
11      presumed that traditional inhalation exposure experiments would not result in definitive tumor
12      formation; thus a tumor  initiation animal model was used. Groups of 48 to 72 animals were
13      exposed to clean air, whole diesel exhaust at a mean DPM concentration of 3.9 mg/m3, or filtered
14      diesel exhaust with either no further treatment or administered DBA (itr. instillations of 0.1
15      mg/week for 20 weeks), DEN (1.5 or 4.5 mg/kg s.c.), or pyrene (itr. instillations of 0.1 rug/week
16      for 20 weeks), the last serving as  a noncarcinogenic PAH control.  Inhalation exposures were
17      conducted 7 to 8 h/day, 5 days/week for 2 years. The exhaust was produced by a 2.4-L Daimler-
18      Benz engine operated at 2,400 rpm.
19             Only two hamsters exhibited lung tumors, both having died during the exposure  period.
20      One tumor occurred in a hamster  receiving DBA and exposed to filtered diesel exhaust for 75
21      weeks; the other tumor occurred in a hamster receiving DEN and exposed to whole diesel
22      exhaust for 67 weeks. Compared with corresponding treatment groups, there was a higher
23      incidence of adenomatous proliferative changes in the lungs of hamsters exposed to whole diesel
24      exhaust. Hamsters exposed to filtered diesel exhaust also showed a greater incidence of
25      adenomatous proliferative changes than those of the respective clean air exposure groups did.
26      The incidence of proliferative changes in the lungs of hamsters receiving DEN or DBA was
27      greater than for those groups not treated with the initiators. Although not definitive, this study
28      provided information suggesting the possible involvement of whole diesel  exhaust and filtered
29      diesel exhaust in producing histologic alterations in the lungs of hamsters, though no increases in
30      tumors were observed.
31             In a more recent  study, Syrian hamsters were exposed 19 h/day, 5 days/week for a
32      lifetime to diesel exhaust diluted to a DPM concentration of 4.24 mg/m3 (Heinrieh et al., 1986a;
33      Stober, 1986). Details of the exposure conditions are reported in Appendix A. Ninety-six
34    ,  animals per group were exposed to clean air, whole exhaust, or filtered exhaust.  Additional
35      groups were treated with DEN (4.5 mg/kg, s.c.) or B[a]P (20 doses of 0.25 mg itr. instillation)

        2/1/98               •                     7-23       DRAFT-DO NOT CITE OR QUOTE

-------
 1      and exposed to the three experimental atmospheres. No lung tumors were seen in the uninjected
 2      clean air group or in either diesel exhaust-exposed group. Initial treatment with DEN or B [a]P
 3      resulted in lung tumor incidences of only 10% and 2%, respectively, which were not significantly
 4      changed by exposure to diesel exhaust.
 5            Heinrich et al. (1989b) reported results of experiments assessing the effects of DEN and
 6      diesel exhaust exposure in combination. Hamsters were exposed to exhaust from a Daimler-
 7      Benz 2.4-L engine operated at a constant load of about 15 kW and at a uniform speed of 2,000
 8      rpm.  The exhaust was diluted to an exhaust-clean air ratio of about 1:13, resulting in a mean
 9      particle concentration of 3.75 mg/m3.  The animals were exposed 19 h/day, 5 days/week
10      beginning at noon each day, under a 12-h light cycle starting at 7 a.m. DEN (3 or 6 mg/kg) was
11      given as a single s.c.  injection 2 weeks from the start of exposure to groups of 40 male and 40
12      female Syrian golden hamsters exposed to whole diesel exhaust, filtered diesel exhaust, or clean
13      air. Groups were also exposed to the exhaust without DEN or to only clean air. Exposures were
14      conducted in chambers maintained at 22° C to 24° C and 40% to 60% relative humidity for up to
15      18 mo.  Surviving hamsters were maintained in clean air for up to an additional 6 mo. The
16      concentrations of B[a]P and B[e]P in the whole exhaust atmospheres were 37.5 and 61.9 ng/m3,
17      respectively.
18            No lung tumors were detected in any of the treatment groups.  A nasal carcinoma was
19      detected in a female hamster treated with DEN (6 mg/kg) and exposed to filtered exhaust. A
20      tracheal carcinoma was detected in a male hamster exposed to whole diesel exhaust and
21      receiving DEN (3 mg/kg), and a laryngeal carcinoma was observed in a male hamster receiving
22      DEN (6 mg/kg) and exposed to whole diesel exhaust. Exposure of male hamsters to whole or
23      filtered diesel exhaust alone did not result in a significant increase in the tumors relative to clean
24      air controls.  Male hamsters receiving 6 mg DEN/kg plus whole diesel exhaust exposure and
25      dying before or after the 50% survival date, however, did show an increase in tumor rate
26      compared with DEN-treated animals exposed to clean air.  Using life-table analysis, a significant
27      (p<0.05) exposure-related increase in tumor rate was noted for this group (40.0% vs. 7.0% for
28     . filtered exhaust + DEN and 7.0% for clean air + DEN). No upper respiratory tract tumors were
29      detected in clean air controls or filtered-exhaust-exposed groups that did not receive the DEN
30      treatment.
31            In summary, diesel exhaust alone did not induce an increase in lung tumors in hamsters
32      of either sex.  Diesel exhaust significantly enhanced the tumorigenic effects of DEN in males
33      injected with 6 mg DEN/kg but not in females or in either sex given the 3 mg/kg dose.  The co-
34      carcinogenic effects of diesel with DEN therefore appear to be equivocal. It should be noted that
35      the tumors occurred in the upper airways as opposed to the alveolar region in rats.

        2/1/98                                    7-24       DRAFT-DO NOT CITE OR QUOTE

-------
  1            Brightwell et al, (1986, 1989) studied the effects of filtered or unfiltered diesel exhaust on
  2      male and female Syrian golden hamsters. Groups of 52 males and 52 females received no
  3      injections or s.c, injections of 4,5 mgDEN/kg 3 days before the start of exposure. The animals
  4      were 6 to 8 weeks old at the start of exposure to diesel exhaust at DPM concentrations of 0.7,
  5      2.2, or 6.6 mg/m3. They were exposed 16 h/day, 5 days/week for a total of 2 years and then
  6      sacrificed.  Exposure conditions  are described in Appendix A. Although the DEN-pretreated
  7      hamsters exhibited an increase in tracheal papillomas in all treatment groups when compared
  8      with non-DEN pretreated hamsters, there was no statistically significant (t test) relationship
  9      between tumor incidence and exhaust exposure. As noted in IARC (1989), however, the
10      reporting of tumor incidence and survival was incomplete.
11
12      7.2.4. Monkey Studies
13            Fifteen male cynomolgus monkeys were exposed to diesel exhaust (2 mg/m3) for 7 h/day,
14      5 days/week for 24 mo (Lewis et al., 1989).  The same numbers of animals were also exposed to
15      coal dust (2 mg/m3 respirable concentration; 5 to 6 mg/m3 total concentration), diesel exhaust
16      plus coal dust  (1  mg/m3 respirable concentration for each component; 3.2 mg/m3 total
17      concentration), or filtered air.  Details of exposure conditions were listed previously  in the
18      description of the Lewis et al. (1986) study with rats (Appendix A).
19            None of the monkeys exposed to diesel exhaust exhibited a significantly increased
20      incidence of preneoplastic or neoplastic lesions. It should be noted, however, that the 24-mo
21      time frame employed in this study may not have allowed the  manifestation of tumors in primates,
22      because this duration is only a small fraction of the monkeys' expected lifespan.  In fact, there
23      have been no near-lifetime exposure studies in nonrodent species.
24
25      7.3. LUNG IMPLANTATION OR INTRATRACHEAL  INSTILLATION STUDIES
26      7.3.1.  Rat Studies
27            Grimmer et al. (1987), using female Osborne Mendel rats (35 per treatment group),
28      provided evidence that the PAHs in diesel exhaust that consist of four or more rings have a  *
29      carcinogenic potential. Condensate was obtained from the whole exhaust of a 3.0 L passenger-
30      car diesel engine connected to a dynamometer operated under simulated city traffic driving
31      conditions.  This condensate was separated by liquid-liquid distribution into hydrophilic and
32      hydrophobic fractions representing 25% and 75% of the total condensate, respectively. The
33      hydrophilic, hydrophobic, or reconstituted hydrophobic fractions were surgically implanted into
34      the lungs of the rats. Untreated controls, vphicle (beeswax/trioctanoin) controls, and positive
35      (B[a]P) controls were also included in the protocol (Table 7-2). Fraction lib (made up of PAHs

        2/1/98                                   7-25       DRAFT-DO NOT CITE OR QUOTE

-------
oo
OS
o
' O
 O
 o
 H
 Q
 m
 o
 o
 a
 o
                 Table 7-2.  Tumor incidence and survival time of rats treated with fractions from diesel
                 exhaust 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
(la, b, c, d) (74.5)
Control, unrelated
Control (beeswax/trioctanoin)
Benzo[o]pyrene


Dose (mg)
6.70
20.00

19.22
0.21
0.29
0.19
19.91



0.3
0.1
0.03
Median survival
time in weeks
(range)
97(24-139)
99 (50-139)

103 (25-140)
102(50-140)
97 (44-138)
106 (32-135),
93 (46-136)

110(23-138)
103 (51-136)
69 (41-135)
98 (22-134)
97 (32-135)
Number of
carcinomas3
0
5

0
6
0
1
7

0
0
27
11
3
Number of
adenomas'*
1
0

1
0
0
0
1

0
1
0
0
0
Carcinoma
incidence (%)
0
14.2

0
17.1
0
2.8
20.0

0 -
0
77.1
31.4
8.6
     "Squamous cell carcinoma.
     bBronchiolar/alveolar adenoma.
     CPAC = polycyclic aromatic compounds.
     d PAH = polycyclic aromatic hydrocarbons.
      Source:  Adapted from Grimmer et al. (1987).

-------
  1      with four to seven rings), which accounted for only 0.8% of the total weight of DPM condensate,
  2      produced the highest incidence of lung carcinomas following implantation into the rat lungs. A
  3      carcinoma incidence of 17.1% was observed following implantation of 0.21 mg lib/rat, whereas
  4      the nitro-PAH fraction (lid) at 0.18 nig/rat accounted for only a 2.8% carcinoma incidence.
  5      Hydrophilie fractions of the DPM extracts, vehicle (beeswax/trioctanoin) controls, and untreated
  6      controls failed to exhibit carcinoma formation. Administration of all hydrophobic fractions (Ila-
  7      d) produced a carcinoma incidence (20%) similar to the summed incidence of fraction lib
  8      (17.1%) and lid (2.8%). The B[a]P positive controls (0.03, 0.1, 0.3 mg/rat) yielded a carcinoma
  9      incidence of 8.6%, 31.4%, and 77.1%, respectively. The study showed that the tumorigenic
10      agents were primarily four- to seven-ring PAHs and, to a lesser extent, nitroaromatics. However,
11      these studies demonstrated that simultaneous administration of various PAH compounds resulted
12      in a varying of the tumorigenic effect, thereby implying that the tumorigenic potency of PAH
13      mixtures may not depend on any one individual PAH. This study did not provide any
14      information regarding the bioavailability of the particle-associated PAHs that might be
15      responsible for earcinogemcity.
16            Kawabata et al. (1986) compared the effects of activated carbon and diesel exhaust on
17      lung tumor formation.  One group of 59 F344 rats was intratracheally instilled with DPM 1
18      mg/week for 10 weeks). A second group of 31 rats was instilled with the same dosing regime of
19      activated carbon. Twenty-seven rats received only the solvent (buffered  saline with 0.05%
20      Tween 80), and  53 rats were uninjected.  Rats dying after 18 mo were autopsied. All animals
21     . surviving 30 mo or more postinstillation were sacrificed and evaluated for histopathology.
22      Among 42 animals exposed to DPM surviving 18 mo or more, tumors were reported in 31,
23      including 20 malignancies. In the subgroup surviving for 30 mo, tumors were detected in 19 of
24      20 animals, including 10 malignancies.  Among the rats exposed to activated carbon, the
25      incidence of lung tumors equaled 11  of 23 autopsied, with 7 cases of malignancy. Data for those
26      dying between 18 and 30 mo and those sacrificed at 30 mo were not reported separately.
27      Statistical analysis indicated that activated carbon induced a significant increase in lung tumor
28      incidence compared with no tumors in 50 uninjected controls  and 1 tumor in 23 solvent-injected
29      controls. The tumor incidence increase was significant in the DPM-instilled group and was
30      significantly greater than the increase in the carbon-instilled group. This study provides evidence
31      for the carcinogenicity of DPM. It also shows, as Heinrieh et al. (1995) and Nikula et al. (1995)
32      did, that particles lacking organic constituents also can induce tumors.
33            Dasenbrock et al. (1996) conducted a study to determine the relative importance of the
34      organic constituents of diesel particles and particle surface area in the induction of lung cancer in
35      rats.  Fifty-two female Wistar rats were intratracheally instilled with 16-17 doses of diesel

        2/1/98                                    7-27       DRAFT-DO  NOT CITE OR QUOTE

-------
 1      particles (DPM), extracted DPM, printex carbon black (PR), lampblack (LB), benzo[a]pyrene
 2      (BaP), DPM + BaP, or PR + BaP. The animals were held for a lifetime or sacrificed when
 3      moribund. The lungs were necropsied and examined for tumors. Diesel particles were collected
 4      from a Volkswagen 1.6 L engine operating on a US FTP-72 driving cycle. The mass median
 5      aerodynamic diameter (MMAD) of the diesel particles was 0.25 um and the specific surface area
 6      wasl2m2/gm. Following extraction with toluene MMAD increased to 138 m2/gm. The
 7      MMAD for extracted PR was equal to 14 run, while the surface area equaled 271 m2/gm.  The
 8      MMAD for extracted lampblack was equal to 95 nm, with a surface area equal to 20 m2/gm.  The
 9      BaP content of the treated particles was  11.3 mg per gm diesel particles and 29.5 mg BaP per gm
10      PR. Significant increases in lung tumors were detected  in rats instilled with 15 mg unextracted
11      DPM and 30 mg extracted DPM, but not 15 mg extracted DPM. Printex CB was more potent
12      than lamplack CB for induction of lung tumors, while BaP was effective only at high doses.
13      Total dose and tumor responses are shown in Table 7-3.
14            A number of conclusions can be  drawn from these results.  First of all, particles devoid of
15      organics are capable of inducing lung tumor formation, as indicated by positive results in the
16      groups treated with high dose extracted diesel particles and printex. Nevertheless, extraction of
17      organics from diesel particles results in a decrease in potency, indicating that the organic fraction
18      does play a role in cancer induction.  A relationship between cancer potency and particle surface
19      area was also suggested by the  finding that printex with a large specific surface area was more
20      potent than either extracted DPM or lampblack, which have smaller specific areas.  Finally,
21      while very large doses of BaP are very effective in the induction of lung tumors, smaller doses
22      adsorbed to particles surfaces had little detectable effect, suggesting that other organic
23      components of DE may be of greater importance in the induction of lung tumors.
24
25      7.3.2. Syrian Hamster Studies
26            Kunitake et al. (1986) and Ishinishi et al. (1988b) conducted a study in which total doses
27      of 1.5, 7.5, or  15 mg of a dichloromethane extract of DPM were instilled intratracheally over "15
28      weeks into male Syrian hamsters that were then held for their lifetimes. The tumor incidences of
29      2.3% (1/44), 0% (0/56), and 1.7% (1/59) for the high-, medium-, and low-dose groups,
30      respectively, did not differ significantly  from the 1.7% (1/56) reported for controls.  Addition of
31      7.5 mg of B[a]P to a DPM extract dose of 1.5 mg resulted in a totaltumor incidence of 91.2%
32      and malignant tumor incidence of 88%.  B[a]P (7.5 mg over 15 weeks) alone produced a tumor
33      incidence rate of 88.2% (85% of these being malignant), which was not significantly different
34      from the DPM extract+ B[a]P  group. Intratracheal administration of 0.03 ug B[a]P, the
35      equivalent content in 15 mg of DPM extract, failed to cause a significant increase in tumors in

        2/1/98                .                    7-28       DRAFT-DO NOT CITE OR QUOTE

-------
90
-J
to
Tl
H

6
o
2
o
H
o
HH
H
W
       Table 7-3. Tumor incidences in rats following intratracheal instillation of

       diescl exhaust particles (DPM), extracted DPM, carbon black (CB),

       benzo[ff]pyrene (BflP), or particles plus BaP
Experimental
group
Control
DPM (original)
DPM (extracted)
DPM (extracted)
CB (printex)
CB (lampblack)
BaP
BaP
DEP + BaP
CB (printex) + BaP
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
BaP
15 mg443 ng
BaP
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
significance"
-
<0.01
< 0.001
NS
< 0.001
NS
< 0.001
< 0,001
NS
< 0.001 -
Tischer's exact test.
XD
C
O
H
W

-------
 1      rats. This study demonstrated a lack of detectable interaction between DPM extract and B[a]P,
 2      the failure of DPM extract to induce carcinogenesis, and the propensity for respiratory tract
 3      carcinogenesis following intratracheal instillation of high doses of B[a]P.  For studies using the
 4      DPM extract, some concern must be registered regarding the known differences in chemical
 5      composition between DPM extract and DPM. As with all intratracheal instillation protocols,
 6      DPM extract lacks the complement of volatile chemicals found in whole diesel exhaust.
 7            The effects on hamsters of intratracheally instilled DPM suspension, DPM with Fe2O3, or
 8      DPM extract with Fe2O3 as the carrier were studied by Shefneretal. (1982). The DPM
 9      component in each of the treatments was administered at concentrations of 1.25, 2.5, or 5.0
10      mg/week for 15 weeks to groups of 50 male Syrian golden hamsters.  The total volume instilled
11      was 3.0 mL (0.2 mL/week for 15 weeks). The DPM and dichloromethane extracts were
12      suspended in physiological saline with gelatin (0.5% w/v), gum arabic (0.5% w/v), and
13      propylene glycol (10% by  volume). The Fe2O3 concentration, when used, was 1.25 mg/0.2 mL
14      of suspension. Controls received vehicle and, where appropriate, carrier particles (Fe2O3)
15      without the DPM component. Two replicates of the experiments were performed. Adenomatous
16      hyperplasia was reported to be most severe in those animals treated with DPM or DPM plus
17      Fe2O3 particles and least severe in those animals receiving DPM plus Fe2O3. Of the two lung
18      adenomas detected microscopically, one was in an animal treated with a high dose of DPM and
19      the other was in an animal receiving a high dose of DPM extract. Although lung damage was
20      increased by instillation of DPM, there was no evidence of tumorigenicity.
21
22      7.3.3.  Mouse Studies
23            Ichinose et al. (1997a) intratracheally instilled 36 four-week-old male ICR mice per
24      group weekly for 10 weeks with sterile saline or 0.05, 0.1, or 0.2 mg DPM. Particles were
25      collected from a 2.74 L, four-cylinder Isuzu engine run at a steady speed of 1,500 rpm under a
26      load of 10 torque (kg/m). Twenty-four hours after the last instillation, six animals per group
27      were sacrificed for measurement of lung 8-hydroxydeoxyguanosine (8-OHdG).  The remaining
28      animals were sacrificed after 12 mo for histopathological analysis. Lung tumor incidence varied
29      from 4/30 (13.3%) for controls to 9/30 (30%), 9/29 (31%), and 7/29 (24.1%) for mice instilled
30      with 0.05, 0.1, and 0.2 mg/week, respectively.  The increase in animals with lung tumors
31      compared with controls was statistically significant for the 0.1  mg dose group, the only group
32      analyzed statistically.  Increases in 8-OHdG, an indicator of oxidative DNA damage, correlated
33      well with increases in tumor incidences, with the correlation coefficients r = 0.916, 0.765, and
34      0.677 for the 0.05, 0.10, and 0.20 mg DPM groups, respectively.
        2/1/98                                    7-30       DRAFT-DO NOT CITE OR QUOTE

-------
  1            In a similar study, 33 four-week-old male ICR mice per group were intratraeheally
  2      instilled weekly for 10 weeks with sterile saline. 0.1 mg DPM, or 0.1 mg DPM from which the
  3      organic constituents were extracted with hexane (Ichinose et al., 1997b).  Exhaust was collected
  4      from a 2.74 L four-cylinder Izuzu engine run at a steady speed of 2,000 rpm under a load of 6
  5      torque (kg/m).  Twenty-four hours after the last instillation, six animals per group were sacrificed
  6      for measurement of 8-OHdG.  Surviving animals were sacrificed after 12 mo. The incidence of
  7      lung tumors increased from 3/27 (11.1%) among controls to 7/27 (25.9%) among those instilled
  8      with extracted diesel particles and 9/26 (34.6%) among those instilled with unextracted particles.
  9      The increase in number of tumor-bearing animals was statistically significant compared with
10      controls (p<0.05) for the group treated with unextracted particles.  The increase in 8-OHdG was
11      highly correlated with lung tumor incidence, r = 0.99.
12
13  '    7.4. SUBCUTANEOUS AND INTRAPEMTONEAL INJECTION STUDIES
14      7.4,1. Mouse Studies
15            In addition to inhalation studies, Orthoefer et al. (1981) also tested the effects of i.p.
16      injections of DPM on male Strong A mice. Three groups of 30 mice were injected with 0.1 mL
17      of a suspension (particles in distilled water) containing 47, 117, or 235 ug of DPM collected
18      from Fluoropore filters in the inhalation exposure chambers. The exposure system and exposure
19      atmosphere are  described in Appendix A. Vehicle controls received injections of particle
20      suspension made up of particulate matter from control exposure filters, positive controls received
21      20 mg of urethan, and negative controls received no injections. Injections were made three times
22      weekly for 8 weeks, resulting in a total DPM dose of 1.1,2.8, and 5.6 mg for the low-, medium-,
23      and high-dose groups and 20 mg of urethan for the positive control group. These animals were
24.     sacrificed after 26 weeks and examined for lung tumors.  For the low-, medium-, and high-dose
25      DPM groups, the tumor incidence was 2/30,10/30, and 8/30, respectively. The incidence among
26      urethan-treated  animals (positive controls) was 100% (29/29), with multiple tumors per animal
27      The tumor incidence for the DPM-treated animals did not differ significantly from that of vehicle
28      controls (8/30) or negative controls (7/28).  The number of tumors per mouse was also unaffected
29      by treatment.
30            In further studies conducted by Orthoefer et al. (1981), an attempt was made to compare
31      the potency of DPM with that of other environmental pollutants. Male and female Strain A mice
32      were injected i.p. three times weekly for 8 weeks with DPM, DPM extracts, or various
33      environmental mixtures of known carcinogenicity, including cigarette smoke eondensate, coke
34      oven emissions, and roofing tar emissions.  Injection of urethan or dimethylsulfoxide (DMSO)
35      served as positive or vehicle controls, respectively. In addition to DPM from the Nissan diesel

        2/1/98                                   7-31       DRAFT-DO NOT CITE OR QUOTE

-------
 1     previously described, an 8-cylinder Oldsmobile engine operated at the equivalent of 40 mph was
 2     also used to compare emission effects from different makes and models of diesel engine. The
 3     mice were sacrificed at 9 mo of age and their lungs examined for histopathological changes. The
 4     only significant findings, other than for positive controls, were small increases in numbers of
 5     lung adenomas per mouse in male mice injected with Nissan DPM and in female mice injected
 6     with coke oven extract. Furthermore, the increase in the extract-treated mice was significant
 7     only in comparison with uninjected controls (not injected ones) and did not occur when the
 8     experiment was repeated. Despite the use of a strain of mouse known tp be sensitive to tumor
 9     induction, the overall findings of this study were negative. The authors provided several possible
10     explanations for these findings, the most likely of which were (1) the  carcinogens that were
11     present were very weak or (2) the concentrations of the active components reaching the lungs
12     were insufficient to produce positive results.
13            Kunitake et al. (1986) conducted studies using DPM extract obtained from a 1983 HD
14     MMC M-6D22P 11-L V-6 engine.  Five s.c. injections of DPM extract (500 mg/kg per injection)
15     resulted in a significant (p<0.01) increase in subcutaneous tumors for female C57BL mice (5/22
16     [22.7%] vs. 0/38 among controls). Five s.c. doses of DPM extract of 10,25,30,100, or 200
17     rng/kg failed to produce a significant increase in tumor incidence. One of 12 female ICR mice
18     (8.3%) and 4 of 12 male ICR mice (33.3%) developed malignant lymphomas following neonatal
19     s.c. administration of 10 mg of DPM extract per mouse. The increase in malignant lymphoma
20     incidence for the male mice was statistically significant at (p<0.05) compared with an incidence
21     of 2/14 (14.3%) among controls. Treatment of either sex with 2.5 or 5 mg of DPM extract per
22     mouse did not result in statistically significant increases in tumor incidence.
23            Additional studies using DPM extract from LD (1.8-L, 4-cylinder) as well  as HD engines
24     with female ICR and nude mice (BALB/c/cA/JCL-nu) were also reported (Kunitake et al., 1988).
25     Groups of 30 ICR and nude mice each were given a single s.c. injection of 10 mg HD extract, 10
26     mg HD + 50  ug 12-O-tetradecanoylphorbol 13-acetate (TPA), 10 mg LD extract + 50 ^g TPA,
27     or 50 ug TPA. No malignant tumors or papillomas were observed. One papillomatous lesion
28     was observed in an ICR mouse receiving LD extract + TPA, and acanthosis was observed in one
29    » nude mouse receiving only TPA.
30            In what appears to be an extension of the Kunitake et al. (1986) s.c. injection studies,
31     Takemoto et al. (1988) presented additional data for subcutaneously administered  DPM extract
32     from HD and LD diesel engines. In this report, the extracts were administered to 5-week-old and
33     neonatal (<24 h old) C57BL mice of both sexes. DPM extract from HD or LD engines was
34     administered weekly to the 5-week-old mice for 5 weeks at doses of 10,25, 50,100, 200, or 500
35     mg/kg, with group sizes ranging from  15 to 54 animals. After 20 weeks, comparison with a

       2/1/98                                   7-32       DRAFT-DO NOT CITE OR QUOTE

-------
  1      control group indicated a significant increase in the incidence of subcutaneous tumors for the 500
  2      mg/kg HD group (5 of 22 mice [22.7%J,/K0.01), the 100 mg/kg LD group (6 of 32 [18.8%],
  3      p<0.0l), and the $00 mg/kg LD group (7 of 32 [21.9%],/?<0.01) in the adult mouse experiments.
  4      The tumors were characterized as malignant fibrous histiocytomas. No tumors were observed in
  5      other organs. The neonates were given single doses of 2.5, 5, or 10 mg DPM extract
  6      subcutaneously within 24 h of birth. There was a significantly higher incidence of malignant
  7      lymphomas in males receiving 10 mg of HD extract and of lung tumors for males given 2.5 mg
  8      HD extract and for males given 5 mg and females given 10 mg LD extract. A dose-related trend
  9      that was not significant was observed for the incidences of liver tumors for both the HD extract-
10      and LD extract-treated neonatal mice. The incidence of mammary tumors in female mice and
11      multiple-organ tumors in male mice was also greater for some extract-treated mice but was not
12      dose related. The report concluded that  LD DPM extract showed greater carcinogenicity than did
13      HD DPM extract.
14
15      7.5. DERMAL STUDIES
16      7.5.1.  Mouse Studies
17            In one of the earliest studies of diesel emissions, the effects of dermal application of
18      extract from DPM were examined by Kotinetal. (1955). Acetone extracts were prepared from
19      the DPM of a diesel engine (type and size not provided) operated at warm-up mode and under
20      load. These extracts were applied dermally three times weekly to male and female C57BL and
21      strain A mice.  Results of these experiments are summarized in Table 7-4. In the initial
22      experiments using 52 (12 male, 40 female) C57BL mice treated with DPM extract from an
23      engine operated in a warm-up mode, two papillomas were detected after 13 mo. Four tumors in
24      8 surviving of 50 exposed male strain A mice treated with DPM extract from an engine operated
25      under full load were detected 16 mo after the start of treatment.  Among female strain A mice
26      treated with DPM extract from an engine operated under full load, 17 tumors were detected in 20
27      of 25 mice surviving longer than 13 mo. This provided a significantly increased tumor incidence
28      of 85%.  Carcinomas as well as papillomas were seen, but the numbers were not reported.
29            Depass et al. (1982) examined the potential of DPM and dichloromethane extracts of
30      DPM to act as  complete carcinogens, carcinogen initiators, or carcinogen promoters.  In skin-
31      painting studies, the DPM was obtained from an Oldsmobile 5.7 L diesel engine operated under
32      constant load at 65  km/h. The DPM was collected at a temperature of 100°C.  Groups of 40
33      C3H/HeJ mice were used because of their low spontaneous tumor incidence.  For the  complete
34    .  carcinogenesis experiments, DPM was applied as a 5% or 10% suspension in acetone.
35      Dichloromethane extract was applied as 5%, 10%, 25%, or 50% suspensions. Negative  controls

        2/1/98               •                    7-33       DRAFT-DO NOT CITE OR QUOTE

-------
oo
U)
O
o

25
O
H
O
HH
H
W
                 Table 7-4.  Tumorigenic effects of dermal application of acetone extracts of diesel particulate

                 matter (DPM)
Number of
animals
52

50

25

Strain/sex
C57BL/40 F
C57BL/12 M
, Strain A/M

Strain A/F

Sample material
Extract of DPM obtained during
warm-up
Extract of DPM obtained during
full load
Extract of DPM obtained during
full load
Time to first
tumor (mo)
13

15

13

Survivors at
time of first
tumor Total tumors
33 2

84

20 17

Duration of
experiment
(mo)
22

23

17

       Source:  Kotin et al. (1955).
/O
a
O

-------
 1      received acetone, and positive controls received 0.2% B[a]P.  For tumor-promotion experiments,
 2      a single application of 1.5% B[a]P was followed by repeated applications of 10% DPM
 3      suspension, 50% DPM extract, acetone only (vehicle control), 0.0001% phorbol  12-myristate 13-
 4      acetate (PMA) as a positive promoter control, or no treatment (negative control). For the tumor-
 5      initiation studies, a single initiating dose of 10% diesel particle suspension, 50% diesel particle
 6      extract, acetone, or PMA was followed by repeated applications of 0.0001% PMA. Following 8
 7      mo of treatment, the PMA dose in the initiation and promotion studies was increased to 0.01%.
 8      Animals were treated three times per week in the complete carcinogenesis and initiation
 9      experiments and five times per week in promotion experiments. All test compounds were
10      applied to a shaved area on the back of the mouse.
11            In the complete carcinogenesis experiments, one mouse receiving the high-dose (50%)
12      suspension of extract developed a squamous cell carcinoma after 714 days of treatment. Tumor
13      incidence in the B[a]P group was  100%, and no tumors were observed in any of the other groups.
14      For the promotion studies, squamous cell carcinomas with pulmonary metastases were identified
15      in one mouse of the 50% DPM extract group and in one in the 25% extract group. Another
16      mouse in the 25% extract group developed a grossly diagnosed papilloma.  Nineteen positive
17      control mice had tumors (11 papillomas, 8 carcinomas). No tumors were observed for any of the
18      other treatment groups. For the initiation studies, three tumors (two papillomas and one
19      carcinoma) were identified in the group receiving DPM suspension and three tumors (two
20      papillomas and one fibrosarcoma) were found in the DPM extract group. These findings were
21      reported to be statistically insignificant using the Breslow and Mantel-Cox tests.
22            The data from this study indicated that DPM and dichloromethane extracts of these
23      particles are not effective with regard to tumor promotion  or initiation.  Although these findings
24      were not consistent with those of Kotin et al. (1955) (Table 7-2), the occurrence of a single
25      carcinoma in a strain known to have an extremely low spontaneous tumor incidence may be of
26      importance. Furthermore, a comparison between studies employing different strains of mice
27      with varying spontaneous tumor incidences may result in erroneous assumptions.
28            Nesnow et al. (1982) studied the formation of dermal papillomas and carcinomas
29      following dermal application of dichloromethane extracts  from coke oven emissions, roofing tar,
30      DPM, and gasoline engine exhaust.  DPM from five different engines, including a preproduction
31      Nissan 220C, a 5.7-L Oldsmobile, a prototype Volkswagen Turbo Rabbit, a Mercedes 300D, and
32      a HD Caterpillar 3304, was used for various phases of the study. Male and female Sencar mice
33      (40 per group) were used  for tumor-initiation, tumor-promotion, and complete carcinogenesis
34      studies. For the rumor-initiation experiments, the DPM extracts were topically applied in single
35      doses of 100, 500, 1,000 or 2,000  (ig/mou'se. The high dose (10,000 ng/mouse) was applied in

        2/1/98                                   7-35       DRAFT-DO NOT  CITE OR QUOTE

-------
 1      five daily doses of 2.000 ug. One week later, 2 ug of the tumor promoter tetradecanoylphorbol
 2      acetate (TPA) was applied topically twice weekly. The tumor-promotion experiments used mice
 3      treated with 50.5 u-g of B[a]P followed by weekly (twice weekly for high dose) topical
 4      applications (at the aforementioned doses) of the extracts. For the complete carcinogenesis
 5      experiments, the test extracts were applied weekly (twice weekly for the high doses) for 50 to 52
 6      weeks. Only extracts from the Nissan, Oldsmobile, and Caterpillar engines were used in the
 7      complete carcinogenesis experiments.
 8            In the tumor-initiation studies, both B[a]P alone and the Nissan engine DPM extract
 9      followed by TPA treatment produced a significant increase in turner (dermal papillomas)
10      incidence at 7 to 8 weeks postapplication. By  15 weeks, the tumor incidence was greater than
11      90% for both groups. No significant carcinoma formation was noted for mice in the tumor-
12      initiation experiments following exposure to DPM extracts of the other diesel engines, although
13      the Oldsmobile engine DPM extract at 2,0 mg/mouse did produce a 40 percent papilloma
14      incidence in male mice at 6 mo. This effect, however, was not dose dependent.
15            B[a]P (50.5 ug/week), coke oven extract (at 1.0,2.0, or 4.0 mg/week), and the highest
16      dose of roofing tar extract (4.0 mg/week) all tested positive for complete carcinogenesis activity.
17      DPM extracts from only the Nissan, Oldsmobile, and Caterpillar engines were tested for
18      complete carcinogenic potential, and all three proved to be negative using the Sencar mouse
19      assay.
20            The results of the dermal application experiments by Nesnow et al. (1982) are presented
21      in Table 7-5. The tumor initiation-promotion assay was considered positive if a dose-dependent
22      response was obtained and if at least two doses provided a papilloma-per-mouse value that was
23      three times or greater than that of the .background value. Based on these criteria, only emissions
24      from the Nissan were considered positive.  Tumor initiation and complete carcinogenesis assays
25      required that at least one dose produce a tumor incidence of at least 20%. None of the DPM
26      samples yielded positive results based on this criterion.
27            Kunitake et al. (1986,1988) evaluated the effects of a dichloromethane extract of DPM
28      obtained from a 1983 MMC M-6D22P 11-L V-6 engine.  An acetone solution was applied in 10
29      doses every other day, followed by promotion with 2.5 |jg of TPA three times weekly for 25
30      weeks. Exposure groups received a total dose of 0.5, 5, 15, or 45 mg of extract.  Papillomas
31      were reported in 2 of 50  animals examined in the 45 mg exposure group and in 1 of 48 in the 15
32      mg group compared with 0 of 50 among controls. Differences, however, were not statistically
33      significant.
        2/1/98                                    7-36      DRAFT-DO NOT CITE OR QUOTE

-------
                 Table 7-5. Dermal tumorigenic and carcinogenic effects of various emission extracts
oo
o
O
O
o
H
G
Crt
O
Sample
Benzo[a]pyrene
Topside coke oven
Coke oven main
Roofing tar
Nissan
Oldsmobile
VW Rabbit
Mercedes
Caterpillar
Residential furnace
Mustang
Tumor initiation Complete carcinogenesis Tumor .promotion
Papillomas3 Carcinomasb Carcinomas'5 Papillomas3
+/+c +/+ +/+ +/H-
+/+ -/+ . NDd , ND
+/+ +/+ ' +/+ ' +/ +
+/+ +/+ • +/+ ' +/+.
+/+ +/+ -/- ' ND
+/+ -/- -/- ND
+/+ -/- !e * ND
+/=- -/- ND ND
-/- -/- • -/- ND
-/-'.' -/- ND ND
+/+ -/+ • ND ND
aScored at 6 mo.
 Cumulative score at 1 year.
cMale/female.
dND = Not determined.
el = Incomplete.

Source: Nesnow et al. (1982).
O
w

-------
 1      7.6.  SUMMARY AND CONCLUSIONS OF LABORATORY ANIMAL
 2           CARCINOGENICITY STUDIES
 3           As early as 1955, Kotin et al. (1955) provided evidence for tumorigenicity and
 4      earcinogenicity of acetone extracts of DPM following dermal application and also provided data
 5      suggesting a difference in this potential depending on engine operating mode. Until the early
 6      1980s, no chronic studies assessing inhalation of diesel exhaust, the relevant mode for human
 7      exposure, had been reported. Since then, inhalation studies have been emphasized.
 8           Studies using rats and an experimental protocol including long-term exposure at high
 9      exposure concentrations (up to 8 nig/rn3), resulting in large lung particle loads and a
10      postexposure observation period, were generally positive in demonstrating DPM-induced
11      increases in tumorigenicity. The highest incidences of tumors were reported by Brightwell et al.
12      (1986, 1989). Among female rats exposed for 24 mo and held for their lifetimes, tumors were
13      detected in 24 of 25 animals. This study points out the probable cumulative effects of high
14      exposure concentration (6.6 mg/m3),  lengthy daily exposures (16 h/day), exposure in the dark
15      resulting in a probable increase in ventilation and thereby DPM  intake, and maintenance of the
16      animals for their lifetimes. In two major studies reported by Heinrich et al. (1986a) and
17      Mauderly et al. (1987), significant but lower lung tumor incidences were observed at the high-
18      dose levels, 15.8% and 12.8%, respectively. Although exposure concentrations differed (7
19      mg/m3 for Mauderly et al. v$. 4 mg/m3 for Heinrich et al.), the longer daily exposure periods in
20      the Heinrich et al. study, 19 h versus 7 h, would probably result  in only slightly differing intakes.
21      More recently, Nikula et al. (1995) provided additional data for rats, and Heinrich et al. (1995)
22      reported on a study involving rats. In the Nikula et al. (1995) report, the percentages of rats
23      (males and females combined) with neoplasms following exposure up to 23 mo were 1%
24      (controls), 6% (2.5 mg/m3 diesel exhaust), and 18% (6.5 mg/m3 diesel exhaust). In the Heinrich
25      et al. (1995) report, the percentages of rats with tumors following exposure up to 23 mo were
26      <1% (controls), 0% (0.8 mg/m3), 6% (2.5 mg/m3), and 22% (7.0 mg/m3).
27           Ishinishi et al. (1988a, 1988b) reported a 6.5% incidence of lung tumors in rats exposed to a
28      concentration of 4 mg/m3 DPM from a HD engine. In this study, although the concentration was
29      relatively low, duration and length of daily exposure were long (16 h/day for 30 mo).  Iwai et al.
30      (1986) reported an increased lung tumor incidence (4/14) in Fischer rats exposed 8 h/day, 7
31      days/week for 24 mo to a DPM concentration of 4.9 mg/m3.  Four of five rats held in clean air for
32      an additional 3 to 6 mo, however, also developed tumors, pointing out again the importance of a
33      long study duration.  Iwai et al. (1986) reported the only diesel exhaust inhalation-induced tumor
34     • increase at a nonrespiratory site (splenic lymphoma).                       ,
        2/1/98                •                    7-38       DRAFT-DO NOT CITE OR QUOTE

-------
  1           Low exposure concentrations and/or short exposure durations were generally used in the
  2     negative studies (Karagianes et al. 1981; Lewis et al. 1989; White et al. 1983; Takemoto et al.,
  3     1986). The lowest DPM concentrations resulting in significant positive effects in rats were in the
  4     range of 2 to 3 mg/m3.
  5           Inhalation of diesel exhaust induced significant increases in lung tumors in female NMRJ
  6     mice (Heinrich et al., 1986a; Stober,  1986) and in female Sencar mice (Pepelko and Peirano,
  7     1983), An apparent increase was also seen in female C57BL mice (Takemoto et al., 1986). In a
  8     series of inhalation studies using strain A mice, no increases in lung tumor rates were detected
  9     (Orthoefer et al., 1981; Kaplan et al., 1982; Kaplan et al., 1983;  White et al., 1983).  The only
 10     study in which lung tumor incidences were increased in animals exposed to filtered exhaust was
 11     reported by Heinrich et al. (1986a) and Stober (1986) using NMRI mice. In a more recent study
 12     by Heinrich et al.  (1995) exposure of NMRJ and C57BL/6N mice to diesel exhaust (4.5 mg/m3
 13     for up to 23 mo) did not produce a tumorigenic response that was significantly different from that
 14     observed in clean air controls.
 15           Attempts to  induce significant increases in lung tumors in  Syrian hamsters were
 16     unsuccessful after inhalation (Heinrich et al., 1982; Heinrich et al., 1986a; Heinrich et al., 1989b;
 17     Brightwell et al., 1986) or itr. instillation (Kunitake et al., 1986;  Ishinishi et al., 1988b). Neither
 18     cats (Pepelko and Peirano, 1983 [see  Chapter 4J) nor monkeys (Lewis et al., 1986) developed
 19     tumors following  2-year exposure to  diesel exhaust. The duration of these exposures, however,
• 20     may well have been inadequate in these two longer-lived species. Exposure levels were also
 21     below the maximum tolerated dose (MTD) in the monkey studies and borderline  for detection of
 22     lung tumor increases in rats.
 23           Kawabata et al. (1986) demonstrated the induction of lung tumors in Fischer 344 rats
 24     folio wing intratracheal instillation of DPM.  Grimmer et al. (1987) showed not only that an
 25     extract of DPM was carcinogenic when instilled in the lungs of rats but also that most of the
 26     carcinogenicity resided in the portion containing PAHs with four to seven rings.
                                                                            i
 27           Alternative exposure routes including dermal exposure and s.c. injection in  mice provided
 28     additional evidence for tumorigenic effects of DPM. Particle extracts applied dermally to mice
 29     have been shown to induce significant skin tumor increases in two studies (Kotin et al., 1955;
 30     Nesnow et al., 1982).  Kunitake et al. (1986) also reported a marginally significant increase in
 31     skin papillomas in ICR mice treated with an organic extract from an HD diesel engine. Negative
 32     results were reported by Depass et al. (1982)  for skin-painting studies using mice and acetone
 33     extracts of DPM suspensions. However, in this study the exhaust particles were collected at
 34     temperatures of 100°C, a temperature that would minimize the condensation of vapor-phase
 35     organics and, therefore, reduce the availability of potentially carcinogenic compounds that might

        2/1/98      .                            7-39        DRAFT-DO NOT CITE OR QUOTE

-------
 1      normally be present on diesel exhaust particles. A significant increase in the incidence of  •
 2      sarcomas in female C57B1 mice was reported by Kunitake et al. (1986) following s.c.
 3      administration of LD DPM extract at doses of 500 mg/kg. Takemoto et al. (1988) provided
 4      additional data for this study and reported an increased tumor incidence in the mice following
 5      injection of LD engine DPM extract at doses of 100 and 500 mg/kg. Results of i.p. injection of
 6      DPM or DPM extracts in strain A mice were generally negative (Orthoefer et al.,  1981; Pepelko
 7      and Peirano,  1983), suggesting that the strain A mouse may not be a good model for testing
 8      diesel emissions.
 9           Experiments using tumor initiators such as DEN, B[a]P, DPN, or DBA (Brightwell et al.,
10      1986; Heinrich et al., 1986a; Takemoto et al., 1986) did not provide conclusive results regarding
11      the tumor-promoting potential of either filtered or whole diesel exhaust.  A report by Heinrich et
12      al. (1982), however, indicated that filtered exhaust may promote the tumor-initiating effects of  .
13      DEN in hamsters.
14  .         Several reports (Wong et al., 1986; Bond et al., 1990) affirm observations of the potential
15      carcinogenicity of diesel exhaust by providing evidence for DNA damage in rats.  These findings
16      are discussed in  more detail in Chapter 9. Evidence for the mutagenicity of organic agents
17      present in diesel engine emissions is also provided in Chapter 8.
18           It appears reasonably certain that with adequate exposures, inhalation of diesel exhaust will
19      induce lung cancer in rats.  The relationship between exposure levels and response, however, is
20      less clearcut. Although significant increases in lung tumors were not reported at concentrations
21      less than about 2 mg/m3, the response at higher concentrations varies considerably. A significant
22      percentage of this variation can probably be attributed to the exposure  regime. A  better method
23      than concentration alone for assessing exposure-response relationships could be achieved by
24      comparing cumulative exposure (concentration x daily exposure duration x  days of exposure).
25      Only those studies conducted for a sufficient length of time (2:24 mo) for expression of
26      carcinogenic responses have been included in this analysis. Examination of the rat data (shown
27      in Table 7-6 and plotted in Figure 7-1) reveals that most studies indicate a trend of increasing
28      tumor incidence at exposures exceeding 1 x I04mg-hr/m3.                    .
29           A similar comparison could not be adequately made for mice, because experimental
30      designs were not comparable. The study reported by Stober (1986), for example,  involved
31      lifetime exposure following weaning.  In the studies reported by Pepelko and Peirano (1986),
32      however, Sencar mice were exposed from conception, when they are presumably more sensitive
33      to tumor induction, and sacrificed at 15 mo, whereas strain A mice were sacrificed at 9 mo of age
34      because of the rapid increase in the incidence of lung tumors in controls. The successful
        2/1/98                                    7-40       DRAFT-DO NOT CITE OR QUOTE

-------
               Table 7-6.  Cumulative (concentration x time) exposure data for rats exposed to whole diesel exhaust
VO
oo
Study
Mauderly et al.
(1987)


Nikula et al. (1995)


Heinrich et al.
(1986b)
Heinrich et al.
(1995)
Ishinishi et al.
(1988a)
(Light-duty engine)


(Heavy-duty engine)




Exposure
rate/duration Total exposure
(h/week, mo) time (h)
35, 30
35,30
35,30
35,30
80, 23
80,23
80, 23
95, 35
95, 35
90,24
90,24
90,24
90,24
96,30
96, 30
96, 30
96,30
96,30
96,30
96,30
96, 30
96, 30
96,30
4,200
4,200
4,200
4,200
7,360
7,360
7,360
13,300
13,300
8,640
8,640
8,640
8,640
11,520
11,520
11,520
11,520
11,520
11,520
11,520
11,520
11,520
11,520
Particle
concentration
(mg/m3)
0
0.35
3.5
7.1
0
2.5
6.5
0
. 4.24
0
0.8
2.5
7.0
o
0.1
0.4
1.1
2.3
0
0.5
1.0
1.8
3.7
Cumulative exposure
(nig-h/m3)
Per week
0
12.25
122.5
248.5
o
200.0
520.0
0
402.8
0
72.0
225.0
630.0
0
9.6
38.4
105.6
220.8
0
48.0
96.0
172.8
355.2
Tumor incidence
Total (%)a
0
1,470
14,700
29,820
0
18,400
47,840
0
56,392
0
7,400b
21,800b
61,700b
0
1,152
4,608
12,672
26,496 •
0
5,760
11,520
20,736
42,624
0.9
1.3
3.6
12.8
1.0
7.0
18.0
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

-------
OO
               Table 7-6, Cumulative (concentration x time) exposure data for rats exposed to whole diesel
               exhaust (continued)


Exnosure
rate/duration
Study (h/week, mo)








-4
I
to


O
i
H
6 •
o
o
H
Q
H
W '
o
90
a
o
H
W
Brightwell et al.
(1989)


Kaplan et al. (1983)





Iwai et al. (1986)

Takemoto et al.
(1986)
Karagianes et al.
(1981)

"Combined data for male?
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

and females.

Total exposure
time (h)
7,680
7,680
7,680
7,680
8,400
8,400
8,400
8,400


5,376
5,376
1,152-1,536
1,152-1,536
2,400
2,400


Particle
concentration
(mg/m3)
0
0.7
2.2
6.6
0
0.25
0.75
1.5


o
4.9
0
2-4
0
8.3


Cumulative exposure
(mg-h/in3)
Tumor incidence
Per week
0
56.0
176.0
528.0
0
35.0
105.0
210.0


0
274.4
0
32-64
0
249


Total
. 0
5,376
16,896
50,688
0
2,100
6,300
12,600


0
26,342
0
3,456-4,608
0
19,920

.
(%)a
• 1.2
0.7
9.7
38.5
0
3.3
10.0
3.3


0
36.8
0
0
0
16.6


bAs reported in Heinrich et al. (1995).







































































-------
                 40
                 30
                          MaucUrty «til. (1987)
                        • Helnrich »t il. (19SBb)
                        V l»hlnl«hl it ll. (1988b) (LD)
                        T lihlnlahl »til. (1986b)(HD)
                        G Brl8htw.ll»tll. (198»)
                        A Iwti»til. (1986)
              A
              O
              9
              •o
              o
              i
                 20
                 10
rf
                                                    --V
                   0            1            23            45
                                     Cumulative Exposure (10 mg  h/m)

              Figure 7-1. Cumulative exposure data for rats exposed to whole diesel exhaust.

 1      induction of lung tumors in mice by Ichinose et al. (1997a,b) via intratracheal instillation may be
 2      the result of focal deposition of larger doses.
 3           Although the preceding analysis accounts for total inspired dose, it does not account for
 4      delayed particle clearance at higher exposure levels. The analysis might be improved further by
 5      comparing tumor response with lung burden of particulate matter.  This analysis could account
 6      for not only differences in exposure times but also for overload inhibition of clearance at high
 7      exposure levels. Although such data were not available for all studies, it could be estimated by
 8      using available estimates of respiration along with deposition and clearance models. The
 9      extrapolation models (Appendix A) and the qualitative/quantitative evaluations of Chapter 11
10      attempt this relative to human exposure.
11           To evaluate accurately the carcinogenic risk to humans from diesel engine emissions, it is
12      important to ascertain the fraction or fractions of exhaust responsible for inducing lung tumors.
13      Several of the previously discussed studies indicated that only whole (unfiltered) diesel exhaust
14      is tumorigenie or carcinogenic and that these responses are eliminated or greatly minimized in
15    ,  exposures to filtered diesel exhaust. Data for NMRI and C57BL/6N mice exposed to whole
16      diesel exhaust (4.5 m/m3), filtered exhaust, or clean air for up to 23 mo showed no significant
        2/1/98
         7-43
DRAFT-DO" NOT CITE OR QUOTE

-------
 1      increases in tumor incidences for mice exposed to filtered exhaust relative to clean air controls
 2      (Heinrich et al., 1995). In one study (Stober, 1986), however, a significant increase in lung
 3      rumors was seen in mice exposed to filtered exhaust. Heinrich et al. (1982) also provided some
 4      evidence suggesting that the gaseous fraction promoted the tumorigenic effects of DEN.
 5      Nevertheless, because of the lack of positive data in rats and the limited positive data in mice, the
 6      tumorigenicity of the gaseous fraction must be considered to be unresolved.
 7           The importance of DPM has been affirmed for the tumorigenic response observed in rats
 8      (Heinrich et al., 1995; Nikula et al., 1995). Evidence for the importance of the carbon core was
 9      initially provided by studies of Kawabata et al. (1986), which showed induction of lung tumors
10      following intratracheal instillation of carbon black that contained no more than traces of organics
11      and studies of Heinrich (1990) that indicated that exposure via inhalation to carbon black
12      (Printex 90) particles induced lung tumors at concentrations similar to those effective in DPM
13      studies. Induction of lung tumor by other particles of low solubility such as titanium dioxide
14      (Lee et al., 1986) confirmed the capability of particles in inducing lung tumors. Pyrolyzed pitch,
15      on the other hand, essentially lacking a carbon core but having much higher PAH concentrations
16      than DPM, also was effective in  tumor induction (Heinrich et al., 1986b; 1994).
17           The relative importance of the adsorbed  organics, however, remains to be elucidated and is
18      of some concern because of the known carcinogenic capacity of some of these chemicals, These
19      include polycyclic aromatics as well as nitroaromatics as described in Chapter 2.  Organic
20      extracts of particles also have been shown to induce tumors in a variety of injection, intratracheal
21      instillation, and skin-painting studies, and Grimmer et al. (1987) have, in fact, shown that the
22      great majority of the carcinogenic potential following instillation resided in the fraction
23      containing four- to  seven-ring PAHs.
24           In summary, based on positive inhalation exposure data in rats and mice, intratracheal
25      instillation in rats, and injection or skin painting in mice and supported by positive mutagenicity
26      studies, the evidence for carcinogenicity of diesel exhaust is considered to be adequate. The
27      contribution of the  various fractions of diesel exhaust to the carcinogenic response is less certain.
28    .  The effects of the gaseous phase are equivocal. The presence of known carcinogens adsorbed to
29      diesel particles and the demonstrated tumorigenicity of particle extracts in a variety of injection,
30      instillation, and skin-painting studies suggest carcinogenic potential for the organic fraction.
31      Studies showing that insoluble particles (e.g.,  carbon black, TiO2) can also  induce tumors, on the
32      other hand, have provided definitive evidence that the carbon core of the diesel particle is
33      instrumental in the  carcinogenic  response observed in rats.
34           A summary of studies assessing the tumorigenic and carcinogenic effects in laboratory
35      animals following inhalation exposure to diesel exhaust is presented in Table 7-1.

        2/1/98  .                                 7-44       DRAFT-DO NOT CITE OR QUOTE

-------
  1   •    7.7. REFERENCES
  2
  3       Bond, JA; Johnson, NF; Snipes, MB; Mauderly, JL, (1990) DNA adduct formation in rat alveolar type II cells; cells
  4  t     potentially at risk for inhaled diesel exhaust. Environ Mol Mutagen 16:64-69.
  5  '
  6       Brightwell, J; Fouillet, X; Cassano-Zoppi, A-L; Gate, R; Duchosal, F. (1986) Neoplastic and functional changes in
  7       rodents after chronic inhalation of engine exhaust emissions. In: Ishinishi, N; Koizumi, A; McClellan, R; Stober, W,
  8       eds. Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam; Elsevier; pp. 471-485.
  9
10       Brightwell, J; Fouillet, X; Cassano-Zoppi, A-L; Bernstein, D; Crawley, F; Duchosal, F; Gatz, R; Perczel, S; Pfeifer,
11       H. .(1989) Tumors of the respiratory tract in rats and hamsters following chronic inhalation of engine exhaust
12       emissions. JAppl Toxicol 9:23-31.
13
14       Carpenter, K; Johnson, JH. (1979) Analysis of the physical characteristics of diesel particulate matter using
15       transmission electron microscope techniques. SAE Transactions 88:2743.
16
17       Cuddihy, RG; Griffith, WC;  McClellan, RO. (1984) Health risks from light-duty diesel vehicles. Environ Sci
18       Technol 18:I4A-21A.
19
20       Dasenbrock, C; Peters, L; Creutzenberg, O; Heinrich, U. (1996) The carcinogenic potency of carbon particles with
21       and without PAH after repeated intratracheal administration in the rat. Toxicol Lett 88:15-21.
22
23       Depass, LR; Chen, KC; Peterson, LG. (1982) Dermal carcinogenesis bioassays of diesel particulates and
24       dichloromethane extract of diesel particulates in C3H mice. In: Lewtas, J, ed. Toxicological effects of emissions
25       from diesel engines. Amsterdam: Elsevier; pp. 321-326.
26                                                               .            -'
27       Grimmer, G; Brune, H; Deutseh-Wenzel, R; Dettbam, G; Jacob, J; Naujack, K-W; Mohr, U; Ernst, H. (1987)
28       Contribution of polycyclic aromatic hydrocarbons and nitro-derivatives to the carcinogenic impact of diesel exhaust
29       condensate evaluated by implantation into the lungs of rats. Cancer Lett 37:173-180.
30                              -                      .                                -   . '
31       Heinrich, U. (1990) Carcinogenicity of carbon black "Printex 90" in female Wistar rats. Presented at: U.S.
32       Environmental Protection Agency Diesel Emissions Peer Review Workshop; July 18-19; Chapel Hill, NC.
33
34       Heinrich, U; Peters, L; Funcke, W; Pott, F; Mohr, U; Stober, W. (1982) Investigation of toxic and carcinogenic
35       effects of diesel exhaust in long-term inhalation exposure of rodents. In: Lewtas, J,  ed.  Toxicological effects of
36       emissions from diesel emissions. Amsterdam: Elsevier; pp. 225-242.
37
38       Heinrich, U; Muhle, H; Takenaka, S; Ernst, H; Fuhst, R; Mohr, U; Potf, F; Stober, W. (1986a) Chronic effects on
39       the respiratory tract of hamsters, mice and rats after long-term inhalation of high concentrations of filtered and
40 '      unfiltered diesel  engine emissions. J Appl Toxicol 6:383-395.
41                                      .
42       Heinrich, U; Pott, F; Rittinghausen, S. (1986b) Comparison of chronic inhalation effects in rodents after long-term
43       exposure  to either coal oven  flue gas mixed with pyrolized pitch or diesel engine exhaust. In: Ishinishi, N; Koizumi,
44       A; McClellan, R; Stdber, W,  eds. Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam:
45       Elsevier; pp. 441-457.                                                                      .
46
47       Heinrich, U; Peters, L; Fuhst, R; Mohr, U. (1989a) The effect of automotive exhaust exposure on the
48       carcinogenicity of diphenylnitrosamine (DPN) in the respiratory tract of rats. Exp -Pathol 37:51 -55.
49                                                                                                             .
50  -    Heinrich, U; Mohr, U; Fuhst, R; Brockmeyer, C. (I989b) Investigation of a potential cotumorigenic effect of the
51       dioxides of nitrogen and sulfur, .and of diesel-engine exhaust, on the respiratory tract of Syrian  golden hamsters.
52       Research  Report No. 26. Cambridge, MA: Health Effects Institute; pp. 27.
53
         2/1/98                                           7-45         DRAFT-DO NOT CITE OR QUOTE

-------
  1      Heinrich, U; Dungworth, D; Pott, F; Peters, L; Dasenbrock, C; Levsen, K; Koch, W; Creutzenberg, O; Schulte, A.
  2      (1994) The carcinogenic effects of carbon black particles and tar-pitch condensation aerosol after inhalation
  3      exposure of rats. Am Occup Hyg Suppl 38:351-356.
  4
  5      Heinrich, U; Fuhst, R;  Rittinghausen, S; Creutzenberg, O; Bellamann, B; Koch, W; Levsen, K. (1995) Chronic
  6      inhalation exposure of Wistar rats and two different strains of mice to diesel engine exhaust, carbon black, and
  7      titanium dioxide. Inhal Toxicol 7:553-556.
  8
  9      Huisingh, J; Bradow, R; Jungers, R; Claxton, L; Zweidinger, R; Tejada, S; Baumgarner, J; Duffield, J; Waters, M;
 10      Simon, V; Hare, C; Rodriguez, C; Snow, L. (1978) Application of bioassay to the characterization of diesel particle
 11      emissions. In: Waters, MD; Nesnow, S; Huisingh, JL; Sandhu, SS; Claxton, L, eds. Application of short-term
 12      bioassay in the fractionation and analysis of complex mixtures. New York: Plenum; pp. 381-418.
 13
 14      Ichinose, T; Yajima, Y; Nagashima, M; Takenoshita, S; Nagamachi, Y;  Sagai, M. (1997a) Lung carcinogenesis
 15      and formation of 8-hydroxydeoxyguanosine in mice by diesel exhaust particles. Carcinogenesis 18:185-192.
 16.
 17      Ichinose, T; Yamanushi, T; Seto, H; Sagai, M. (1997b) Oxygen radicals in lung carcinogenesis accompanying
 18      phagocytosis of diesel  exhaust particles.  IntJOncol 11:571-575.
 19
 20      International Agency for Research on Cancer (IARC). (1989) IARC monographs on the evaluation of carcinogenic
 21      risks to humans, vol. 46. Diesel and gasoline engine  exhausts and some nitroarenes. Lyon, France: IARC; p. 104.,
 22
 23      Ishinishi, N; Kuwabara, N; Takaki, Y; Nagase, S; Suzuki, T; Nakajma, T; Maejima, K; Kato, A; Nakamura, M.
 24      (1988a) Long-term inhalation experiments on diesel exhaust. In: diesel exhaust and health risks. Results of the
 25      HERP studies. Ibaraki, Japan: Research Committee for HERP Studies; pp. 11-84.
 26
 27      Ishinishi, N; Inamasu, T; Hisanaga, A; Tanaka, A; Hirata, M; Ohyama, S. (1988b) Intratracheal instillation study of
 28      diesel paniculate extracts in hamsters. In: Diesel exhaust and health risk. Ibaraki, Japan: Research Committee for
 29      HERP Studies; pp. 209-216.
 30
 31      Iwai, K; Udagawa, T; Yamagishi, M; Yamada, H. (1986) Long-term inhalation studies of diesel exhaust on F344
 32      SPF rats. Incidence of lung cancer and lymphoma. In: Ishinishi, N; Koizumi, A; McClellan, R; Stober, W, eds.
 33      Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam: Elsevier; pp. 349-360.
 34
 35      Kaplan, HL; Mackenzie, WF; Springer, KJ; Schreck, RM; Vostal, JJ. (1982) A subchronic study of the effects of
 36      exposure of three species of rodents to diesel exhaust. In: Lewtas, J, ed.  Toxicological effects of emissions from
 37      diesel engines. New York: Elsevier Biomedical; pp.  161-220.
 38                     •      ,       •
 39      Kaplan, HL; Springer, KJ; MacKenzie, WF. (1983) Studies of potential health effects of long-term exposure to
 40      diesel exhaust emissions. Final Report No. 01 -0750-103 (SWRI) and No. 1239 (SFRE). San Antonio, TX:
 41      Southwest Research Institute.
 42
 43      Karagianes, RF; Palmer, RF; Busch, RH. (1981) Effects of inhaled diesel emissions and coal dust in rats. Am Ind
 44      Hyg AssocJ. 42:382-391.
 45
 46      Kawabata, Y; Iwai, K; Udagawa, T; Tukagoshi, K; Higuchi, K. (1986) Effects of diesel soot on unscheduled DNA
. 47      synthesis of tracheal epithelium and lung tumor formation. In: Ishinishi, N; Koizumi, A; McClellan, R; Stober, W,
 48      eds. Carcinogenic and mutagenic effects of diesel engine exhaust.  Amsterdam: Elsevier; pp. 213-232.
 49
 50      Kotin, P; Falk, H; Thomas, M. (1955) Aromatic hydrocarbons. III. Presence in the particulate phase of diesel engine
 51      exhausts and the carcinogenicity of exhaust extracts. Arch Ind Health 11:113-120.
 52
 53      Kunitake, E; Shimamura, K; Katayama, H; Takemoto, K; Yamamoto,'A; Hisanaga, A; Ohyama, S; Ishinishi, N.
 54      (1986) Studies concerning carcinogenesis of diesel particulate extracts following intratracheal instillation,


         2/1/98                                          7-46        DRAFT-DO NOT CITE  OR QUOTE

-------
  1       subcutaneous injection, or skin application. In: Ishinishi, N; Koizumi, A; McClellan, R; Stober, W, eds.
  2       Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam: Elsevier; pp. 235-252.
  3
  4       Kunitake, E; Imase, A; Shimamura, K; Ishinishi, N; Hisanga, A; Tanaka, A. (1988),Skin application and
  5   '    subcutaneous injection experiments of diesel particulate extracts using ICR mice and nude mice. In; Diesel exhaust
  6       and health risks. Results of the HERP studies. Ibaraki, Japan: Research Committee for HERP Studies; pp. 217-225,
  7                                                                               •
  8       Lee, KP; Henry, NW, III; Trochimowiez, HJ; Reinhardt, CF. (1986) Pulmonary response to impaired lung clearance
  9       in rats following excessive TiO2 dust deposition. Environ Res 41:144-167.
10
11       Lewis, TR; Green, FHY; Moorman, WJ; Burg, JA; Lynch, DW. (1989) A chronic inhalation toxicity study of diesel
12       engine emissions and coal dust, alone and combined. J Am Coll Toxicol 8:345-375.,
13
14       Mauderly, JL; Jones, RK; Griffith, WC; Henderson, RF; McClellan, RO. (1987) Diesel exhaust is a pulmonary
15       carcinogen in rats exposed chronically by inhalation.  Fundam Appl Toxicol 9:208-221.
16                                                                               '
17       Mauderly, JL; Banas, DA; Griffith, WC; Hahn, FR; Henderson, RF; McClellan, RO. (1996) Diesel exhaust is not a
18       pulmonary carcinogen in CD-1 mice exposed under conditions carcinogenic to F344 rats. Fundam Appl Toxicol
19       30:233-242.
20                    ,
21   ,    Mohr, U; Takenaka, S; Dungworth, DL. (1986) Morphologic effects of inhaled diesel engine exhaust on lungs of
22       rats: comparison with effects of coal oven flue gas mixed with pyrolized pitch. In: Ishinishi, N; Koizumi, A;
23       McClellan, R; Stober, W, eds. Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam: Elsevier;
24       pp. 459-470.
25                                                      '                                               :
26       Mokler, BV; Archibeque, FA; Beethe, RL; Kelly, CPJ; Lopez, JL; Stafford, DL. (1984) Diesel exhaust exposure
27       system for animal studies. Fundam Appl Toxicol 4:270-277,
28                                              '                                  '
29       Nesnow, S; Evans, C; Stead, A; Creason, J; Slaga, TJ; Triplett, LL. (1982) Skin carcinogenesis studies of emission
30       extracts. In: Lewtas, J, ed. Toxicological effects of emissions from diesel engines. Amsterdam: Elsevier; pp. 295-
31       320.
32
33       Nikula, KJ; Snipes, MB; Barr, EB; Griffith, WC; Henderson, RF; Mauderly, JL. (1995) Comparative pulmonary
34       toxicities of chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl Toxicol 25:80-94.
35
36       Orthoefer,  JG; Moore, W; Kraemer, D. (1981) Carcinogenicity of diesel exhaust as tested in strain A mice. Environ
37       Int 5:461-471.
38                                                                     •
39       Pepelko, WE; Peirano, WB.  (1983) Health effects of exposure to diesel engine emissions. J Am Coll Toxicol 2:253-
40       306.
41
42       Shefher, AM; Collins, SC; Dooley, L; Fiks, A; Graf, JL; Preache, MM. (1982) Respiratory careinogenieity of diesel
43       fuel emissions interim results. In: Lewtas, J, ed, Toxicological effects of emissions from diesel engines. Amsterdam:
44       Elsevier; pp. 329-350.
45
46       Stober, W. (1986) Experimental induction of tumors in hamsters, mice and rats after long-term inhalation of filtered
47       and  unfiltered diesel engine  exhaust. In: Ishinishi, N;  Koizumi, A; McClellan, R; Stober, W, eds. Carcinogenic and
48       mutagenic effects of diesel engine exhaust. Amsterdam: Elsevier; pp. 421-429.
49
50       Takaki, Y; Kitamura, S; Kuwabara, N; Fukuda, Y. (1989) Long-term inhalation studies of exhaust from diesel
51       engine in F-344 rats: the quantitative relationship between pulmonary hyperplasia and anthracosis. Exp Pathol
52       37:56-61.
53                                                   '                                                     '
         2/1/98                                          7-47        DRAFT-DO NOT CITE OR QUOTE

-------
 1      Takemoto, K; Yoshimura, H; Katayama, H. (1986) Effects of chronic inhalation exposure to diesel exhaust on the
 2      development of lung tumors in di-isopropanol-nitrosamine-treated F344 rats and newborn C57BL and ICR mice. In:
 3      Ishinishi, N; Koizumi, A; McClellan, R; Stober, W, eds. Carcinogenic and mutagenic effects of diesel engine
 4      exhaust. Amsterdam: Elsevier; pp. 311-327.
 5    '
 6      Takemoto, K; Katayama, H; Kuwabara, T; Hasumi, M. (1988) Carcinogenicity by subcutaneous administration of
 7      diesel particulate extracts in mice. In: Diesel exhaust and health risks. Ibaraki, Japan: Research Committee for
 8      HERP Studies; pp. 227-234.
 9
10      Vuk, CT; Jones, MA; Johnson, JH. (1976) The measurement and analysis of the physical character of diesel
11      particulate emissions. SAE Transactions 85:556.
12
13      White, HJ; Vostal, JJ; Kaplan, HL; MacKenzie, WF. (1983) A long-term inhalation study evaluates the pulmonary
14      effects of diesel emissions. JAppl Toxicol 1:332.
15
16      Wong, D; Mitchell, CE; Wolff, RK; Mauderly, JL; Jeffrey, AM. (1986) Identification of DNA damage as a result of
17      exposure of rats to diesel engine exhaust.  Carcinogenesis 7:1595-1597.
18
19      Zamora, PO; Gregory, RE; Brooks, AL. (1983) In vitro evaluation of the tumor-promoting potential of diesel-
20      exhaust-particle extracts. J Toxicol Environ Health 11: 188-197.
         2/1/98                                          7-48        DRAFT-DO NOT CITE OR QUOTE

-------
                   8. EPIDEMIOLOGIC STUDIES OF THE CARCINOGENICITY
                              OF EXPOSURE TO DIESEL EMISSIONS

 1      8.1. INTRODUCTION
 2            Emissions from diesel engine exhaust are made up of toxicants that include oxides of
 3      nitrogen and sulfur, carbon monoxide, and particulate matter consisting of a carbon core with
 4      many organic compounds, especially the polycyclic aromatic hydrocarbons adsorbed on the
 5      surface.  Diesel engine exhaust contains about 100 times more particulate matter than gasoline
 6      engine exhaust.
 7            In this chapter, various mortality and morbidity studies of the health effects of exposure
 8      to diesel engine emissions are reviewed.  Although an attempt was made  to cover all the relevant
 9      studies, a number of studies are not included for several reasons. First, the change from steam to
10      diesel engines in locomotives began in 1935 and was about 95% complete by 1959 (Garshick et
11      al., 1988). Diesel buses also were, introduced about the same time.  Therefore, exposure to diesel
12      exhaust was less common, and the follow-up period for studies conducted prior to 1959 (Raffle,
13      1957; Kaplan, 1959) was not long enough to cover the long latency period of lung cancer. The
14      usefulness of these studies in evaluating the carcinogenicity of diesel exhaust is greatly reduced;
15      thus, they are not considered here.
16            Second, hypothesis-generating studies were excluded from this review because their
17      findings need subsequent confirmation by definitive studies (Silverman et al., 1983; Schenker et
18      al., 1984; Buiatti et al.,  1985; Flodin et al., 1987; Siemiatycki et al., 1988; Swanson et al., 1993;
19      Cordier et al., 1993; Notanietal, 1993).
20            Third, studies in which exposure to diesel exhaust was uncertain or was defined as motor
21      exhaust (which includes both gasoline and diesel exhaust) were excluded because they would
22      have contributed little to the evaluation of the carcinogenicity of diesel exhaust (Waxweiler et al.,
23      1973; AMberg et al., 1981; Stern et al., 1981; Vineis and Magnani, 1985; Gustafsson et al., 1986;
24      Silverman et al., 1986; Jensen et al., 1987; Garland et al., 1988; Risch et al., 1988; Guberan et
25      al., 1992).
26            Fourth, a study by Coggon et al. (1984) was not included because the occupational
27      information abstracted from death certificates had not been validated; this would have resulted in
28      limited information.
29            Three types of studies of the health effects of exposure to diesel engine emissions are
30      reviewed in this chapter: (1) cohort studies, (2) case-control studies of lung cancer, and (3) case-
31      control studies of bladder  cancer. In the cohort studies, the cohorts of heavy construction
32      equipment operators, railroad and locomotive workers, and bus garage employees were studied

        2/1/98                                    8-1        DRAFT-DO NOT CITE OR QUOTE

-------
 1      retrospectively to determine increased mortality and morbidity resulting from exposures to
 2      varying levels of diesel emissions in the workplace. A total of nine cohort mortality (one of the
 3      mortality studies also included a nested lung cancer case-control study), ten lung cancer case-
 4      control, and seven bladder cancer case-control studies are considered in this section.
 5
 6      8.2. COHORT STUDIES
 7      8.2.1. Waller (1981): Trends in Lung Cancer in London in Relation to Exposure
 8            to Diesel Fumes
 9            A retrospective mortality study of a cohort of London transport workers was conducted to
10      determine if there was an excess of deaths from lung cancer that could be attributed to diesel
11      exhaust exposure.  Nearly 20,000 male employees aged 45 to 64 were followed for the 25-year
12      period between 1950 and 1974, constituting a total of 420,700 man-years at risk.  These were
13      distributed among five job categories: drivers, garage engineers, conductors, motormen or
14      guards, and engineers (works).  Most employees lived in the greater London area.  Lung cancer
15      cases occurring in this cohort were ascertained only from, death certificates of individuals who
16      died while still employed, or if retired, following diagnosis. Expected death rates were
17.      calculated by applying greater London death rates to the population at risk within each job
18      category. Data were calculated in 5-year periods and 5-year age ranges, finally combining the
19      results to obtain the total expected deaths in the required age range for the calendar period.  A
20      total of 667 cases of lung cancer was reported, compared with 849 expected, to give a mortality
21      ratio of 79%.  In each of the five job categories, the observed numbers were below those
22      expected. Engineers in garages had the highest mortality ratio (90%), but this did not differ
23      significantly from the other job categories. Environmental sampling was done at one garage, on
24      1 day in 1979, for benzo[a]pyrene concentrations and was compared with corresponding values
25      recorded in 1957.  Concentrations of benzo[a]pyrene recorded in 1957 were at least 10 times
26      greater than those measured in 1979.
27            This study has several methodologic limitations. The lung cancer deaths ascertained for
28     . the study occurred while the worker was employed (the worker either died of lung cancer or
29      retired after lung cancer was diagnosed).  Although man-years at risk were based on the entire
30      cohort, no attempt was made to trace or evaluate the individuals who had resigned from the
31      London transport company for any other reason. Hence, information on resignees who may have
32      had significant exposure to diesel exhaust, and lung cancer deaths among them, was not available
33      for analysis. This fact may have led to a dilution effect, resulting in underascertainment of
34      observed lung cancer deaths and underestimation of mortality ratios.  Eligibility criteria for
35      inclusion in the cohort, such as starting date and length  of service with the company, were not

        2/1/98   .                                 8-2        DRAFT-DO NOT CITE OR QUOTE

-------
  1      specified.  Because an external comparison group was used to obtain expected number of deaths,
  2      the resulting mortality ratios were less than 1; this may be a reflection of the "healthy worker
  3      effect."  Investigators also did not categorize the five job categories by levels of diesel exhaust
  4      exposure, nor did they use an internal comparison group to derive risk estimates.
  5            The age range considered for this study was limited (45 to 64 years of age) for the period
  6      between 1950 and 1964.  It is not clear whether this age range was applied to calendar year 1950
  7      or 1964 or at the mid-point of the 25-year follow-up period. No analyses were presented either
  8      by latency or by duration of employment (surrogate for exposure). The environmental survey
  9      based on benzofajpyrene concentrations suggests that the cohort in its earlier years was exposed
10      to much higher concentrations of environmental contaminants than currently exist.  It is not clear
11      when the reduction in benzo[a]pyrene concentration occurred because there are no environmental
12      readings available between 1957 and 1979. It is also important to note that the concentrations of
13      benzo[a]pyrene inside the garage in 1957 were not very different from those outside the garage,
14      thus indicating that exposure for garage workers was not much different from that of the general
15      population. Last, no data were collected on smoking habits.
16
17      8.2.2. Howe et al. (1983): Cancer Mortality (1965 to 1977) in Relation to Diesel Fume
18            and Coal Exposure in  a Cohort of Retired Railroad Workers
19            This is a retrospective cohort study of the mortality experience of 43,826 male pensioners
20      of the Canadian National Railroad (CNR) between 1965 and 1977, Members of this cohort
21      consisted of male CNR pensioners who had retired before 1965 and who were known to be alive
22      at the start of that year, as well  as those who retired between 1965 and 1977. The records were
23      obtained from a computer file that is regularly updated and used by the company for payment of
24      pensions. To receive a pension, each pensioner must provide, on a yearly basis, evidence that he
25      is alive.  Specific cause of death among members of this cohort was ascertained by linking these
26      records to the Canadian Mortality Data Base, which contains records of all deaths registered in
27      Canada since 195.0. Of the 17,838 deaths among members of the cohort between 1965 and 1977,
28    .  16,812 (94.4%) were successfully linked to a record in the mortality file. A random sample
29      manual check on unlinked data revealed that failure to link was due mainly to some missing
30      information on the death records.
31            Occupation at time of retirement was used by the Department of Industrial Relations to
32      classify workers into three diesel fume and coal dust exposure categories:  (1) nonexposed, (2)
33      possibly exposed, and (3) probably exposed. Person-years of observation were calculated and
34      classified by age at observation in 5-year age groups (35 to 39,40 to 44,..., 80 to 84, and ^85
35      years). The observed deaths  were classified by age at death for different cancers, for all cancers

        2/1/98   .                                 8-3        DRAFT-DO NOT CITE OR  QUOTE

-------
 1      combined, and for all causes of death combined. Standard mortality ratios (SMRs) were then  .
 2      calculated using rates of the Canadian population for the period between 1965 and 1977.
 3            Both total mortality (SMR = 95, p0.05)
 4      were close to that expected for the entire cohort. Analysis by exposure to diesel fume levels in
 5      the three categories (nonexposed, possibly exposed, and probably exposed) revealed  an increased
 6      relative risk for lung cancer among workers with increasing exposure to diesel fumes. The
 7      relative risk for nonexposed workers was presumed to be 1.0; for those possibly exposed, the
 8      relative risk was elevated to 1.2, which was statistically significant (p=0.013); and, for those
 9      probably exposed, it was elevated to 1,35, which was statistically highly significant (p=0,001),
10      The corresponding rates for exposure to varying levels of coal dust were very similar at  1.00,
11      1.21 (p=Q.O\2), and 1.35 (/?=0.001), respectively. The trend tests were highly significant for both
12      exposures (p<0.001). Analysis performed after the exclusion of individuals who worked in the
13      maintenance of steam engines, and hence were exposed to high levels of asbestos, yielded the
14      risk of lung cancer to be 1.00,1.21, and 1,33 for those nonexposed, possibly exposed, and
15      probably exposed to diesel exhaust, respectively, with a highly significant trend (p<0.001).
16            An analysis done on individuals who retired prior to 1950 showed the relative risk of lung
17      cancer among nonexposed, possibly exposed, and probably exposed to be 1.00,0,70, and 0.44,
18      respectively, based on fewer than 15 deaths in each category. A similar analysis of individuals
19      who retired after 1950 found the results in the same categories to be 1.00,1.23, and 1.40,
20      respectively. Although retirement prior to  1950 indicated exposure to coal dust alone, retirement
21      after 1950 shows the results of mixed exposure to coal dust and diesel fumes. As there was
22      considerable overlap between occupations involving probable exposure to diesel fumes and
23      probable exposure to coal dust, and as most members of the cohort were employed during the
24      years in which the transition from coal to diesel occurred, it was difficult to distinguish whether
25      lung cancer was associated with exposure to coal dust or diesel fumes or a mixture of both.
26            Although this study showed a highly significant dose-response relationship between
27      diesel fumes and lung cancer, it has some methodological limitations. There were concurrent
28      exposures to both diesel fumes and coal dust during the transition period; therefore,
29      misclassification of exposure may have occurred, because only occupation at retirement was
30      available for analysis.  It is possible that the elevated response observed for lung cancer was due
31      to the combined effects of exposure to both coal dust and diesel fumes and not just one or the
32      other. However, it should be noted that so  far coal dust has not been demonstrated to be a
33      pulmonary carcinogen in studies of coal miners. No information was provided on duration of
34      employment in either diesel work or the coal dust-related jobs for other than those jobs held at
35      retirement. Therefore, it was not possible to evaluate whether this omission would have led to an

        2/1/98                                    8-4       DRAFT-DO NOT CITE OR QUOTE

-------
  1      under- or overestimate of the true relative risk. Furthermore, a lack of information on potential
  2      confounders such as smoking makes the interpretation of the excess risk of lung cancer even
  3      more difficult. Information on cause of death was acquired from the mortality data linkage.
  4      There is a possibility that the cause of death may have been misclassified because of miscoding
  5      of the underlying cause of death.
  6
  7      8.2.3. Rushton et al. (1983):  Epidemiological Survey of Maintenance Workers in the
  8            London Transport Executive Bus Garages and Chiswick Works
  9            This is a retrospective mortality cohort study of male maintenance workers employed for
10      at least 1 continuous year between January 1,1967, and December 31,1975, at 71 London
11      transport bus garages (also known as rolling stock) and at Chiswick Works, For all men, the
12      following information was obtained from computer listings: surname with initials, date of birth,
13      date of joining company, last or present jobs, and location of work.  For those individuals who
14      left their job, date of and reason for leaving were also obtained.  For those who died in  service or,
15      after retirement and for men who had resigned, full name and last known address were  obtained
16      from an  alphabetical card index in the personnel department. Additional tracing of individuals
17      who had left was carried out through social security records. The area of their residence was
18      assumed to be close to their work; therefore their place of work was coded as their residence.
19      One hundred different job titles were coded into 20 broader groups. These 20 groups were not
20      ranked for diesel exhaust exposure, though. The reason for leaving was coded as died in service,
21      retired, or other. The underlying cause of death was coded using the eighth revision of the
22      International Classification of Diseases (ICD). Person-years were calculated from date of birth
23      and dates of entry to and exit from the study using  the man-years computer language program.
24      These were then subdivided into 5-year age and calendar period groups. The expected  number of
25      deaths was calculated  by applying the 5-year age and calendar period death rates of the
26      comparison population to the person-years of corresponding groups.  The mortality experience of
27      the male population in England and Wales was used as the comparison population.  Significance
28      values were calculated for the difference between the observed and expected deaths, assuming a
29      Poisson  distribution,
30            The number ,of person-years of observation totaled 50,008 and was contributed by 8,490
31      individuals in the study with a mean follow-up of 5.9 years. Only 2.2% (194) of the men were
32      not traced. Observed deaths from all causes were significantly lower than expected (observed =
33      495, /KO.OO1). The observed deaths  from all neoplasms and cancer of the lung were
34      approximately the same  as those expected. The only significant excess observed for cancer of
35      the liver and gall bladder at Chiswick Works was based on four deaths (p<0.05).  A few job

        2/1/98   -                                 8-5        DRAFT-DO NOT CITE OR QUOTE

-------
 1      groups showed a significant excess of risks for various cancers. All the excess deaths observed
 2      for the various job groups, except for the general hand category, were based on very small
 3      numbers (usually smaller than five) and merited cautious interpretation. Only a notable excess in
 4      the general hand category for lung cancer was based on 48 cases (SMR = 133,p<0.03).
 5      However, given the fact that there was no adjustment for confounding variables such as smoking,
 6      the result should be interpreted cautiously.
 7            This mortality study of London transport maintenance workers did not demonstrate any
 8      cancer excesses based on a large number of cases; this needs further exploration. Its limitations,
 9      including the small sample size, short duration of follow-up (average of only 6 years), and lack
10      of sufficient latency period, make this study inadequate to draw any conclusions. The number of
11      deaths by different causes and among the various job groups was too small to allow any
12      meaningful conclusions. Details of work history were not obtained to permit any analysis by
13      diesel exhaust exposure. Death information was ascertained from death certificates, with
14      inherent problems of inaccuracy, misdiagnosis, and errors in coding, and it was not known
15      whether a trained nosologist coded the death certificates. No adjustments were made for the
16      confounding effects of smoking and socioeconomic factors.
17
18      8.2.4. Wong et al. (1985): Mortality Among Members of a Heavy Construction
19            Equipment Operators Union With Potential Exposure to Diesel Exhaust Emissions
20            This is a retrospective mortality study conducted on a cohort of 34,156 male members of
21      a heavy construction  equipment operators union with potential exposure to diesel exhaust
22      emissions. Study cohort members were identified from records maintained at Operating
23      Engineers' Local Union No. 3-3 A in San Francisco, CA. This union has maintained both work
24      and death records on  all its members since 1964. Individuals with at least 1 year of membership
25      in this union between January 1, 1964, and December 31, 1978, were  included in the study.
26      Work histories of the cohort were obtained from job dispatch computer tapes. The study follow-
27      up period was from January 1964 to December 1978. Death information was obtained from a
28      trust fond, which provided information on retirement dates, vital status, and date of death for
29      those who were entitled to retirement and death benefits. Approximately 50% of the cohort had
30      been  union members  for less than 15 years, whereas the other 50% had been union members for
31      15 years or more. The average duration of membership was 15 years.  As of December 31,1978,
32      29,046 (85%>cohort  members were alive, 3,345 (9.8%) were dead, and 1,765 (5.2%) remained
33      untraced. Vital status of 10,505 members who had left the  union as of December 31,1978, was
34    ,  ascertained from the Social  Security Administration. Death certificates were  obtained  from
35      appropriate state health departments. Altogether, 3,243 deaths (for whom death certificates were

        2/1/98               .                     8-6       DRAFT-DO NOT CITE OR QUOTE

-------
  1      available) in the cohort were coded using the seventh revision of the ICD. For 102 individuals,
  2      death certificates could not be obtained, only the date of death; these individuals were included in
  3      the calculation of the SMR for all causes of death but were deleted from the cause-specific SMR
  4      analyses. Expected deaths and SMRs were calculated using the U.S. national age-sex-race
  5      cause-specific mortality rates for 5-year time periods between 1964 and 1978.  The entire cohort
  6      population contributed to 372,525.6 person-years in this 5-year study period.
  7             A total of 3,345 deaths was observed, compared with 4,109 expected.  The corresponding
  8      SMR for all causes was 81.4 (p=0.01), which confirmed the "healthy worker effect." A total of
  9      817 deaths was attributed to malignant neoplasms, slightly fewer than the 878.34 expected based
10      on U.S. white male cancer mortality rates (SMR = 93.0,/?=0.05). Mostly there were SMR
11      deficits for cause-specific cancers, including lung cancer for the entire cohort (SMR = 98.6,
12      observed = 309). The only significant excess SMR was observed for cancer of the liver (SMR =
13      166.7, observed = 23, p<0.05).
14             Analysis by length of union membership as a surrogate of duration for potential exposure
15      showed statistically significant increases in SMRs of cancer of the liver (SMR = 424, p<0.01) in
16      the 10- to 14-year membership group and of the stomach (SMR = 248, p
-------
 1             To analyze cause-specific mortality by job held (potential exposure to diesel exhaust
 2      emissions). 20 functional job titles were used, which were further grouped into three potential
 3      categories:  (1) high exposure, (2) low exposure, and (3) unknown exposure. A person was
 4      classified in a job title if he ever worked on that job.  Based on this classification system, if a
 5      person had ever worked in a high-exposure job title he was included in that group, even though
 6      he may have worked for a longer time in a low-exposure group or in an unknown exposure
 7      group. Information on length of work in any particular job, hence indirect information on
 8      potential length of exposure, was not available either.
 9             For the high-exposure group a statistically significant excess was observed for cancer of
10      the lung among bulldozer operators who had 15 to 19 years of membership and 20+ years of
11      follow-up (SMR = 343.4,p<0.05). This excess was based on 5 out of 495 deaths observed in
12      this group of 6,712 individuals, who contributed 80,327.6 person-years of observation.
13             The cause-specific mortality analysis in the low-exposure group revealed statistically
14      significant SMR excesses in individuals who had ever worked as engineers.  These excesses were
15      for cancer of the large intestine (SMR = 807.2,  observed = 3,p<0.05) among those with 15 to 19
16      years of membership and length of follow-up of at least 20 years, and cancer of the liver (SMR =
17      871.9, observed = 3,p<0.05) among those with 10 to 14 years of membership and length of
18      follow-up of 10 to 19 years. There were 7,032 individuals who contributed to 78,402.9 person-
19      years of observation in the low-exposure group.
20             For the unknown exposure group, a statistically significant SMR was observed for motor
21      vehicle accidents only (SMR = 173.3, observed = 21, j?<0.05). There were 3,656 individuals
22      who contributed to 33,388.1 person-years of observation in this category.
23             No work histories were available for those who started their jobs before 1967 and for
24      those who held the same job prior to and after 1967.  This constituted 9,707 individuals (28% of
25      the cohort) contributing to 104,447.5 person-years.  Statistically significant SMR excesses were
26      observed for all cancers (SMR = 112, observed = 339,/?<0.05) and cancer of the lung (SMR =
27      119.3, observed = 141, j?<0.01). A significant SMR elevation was also observed for cancer of
28      the stomach (SMR = 199.1, observed = 30, p<0.01).
29             This study demonstrates a statistically significant excess for cancer of the liver but also
30      shows statistically significant deficits in cancers of the large intestine and rectum. It may be, as
31      the authors suggested, that the liver cancer cases were actually cases resulting from metastases
32      from the large intestine and/or rectum, since tumors of these sites will frequently metastasize to
33      the liver. The excess in liver cancer mortality and the deficits in mortality that are due to cancer
34      of the large intestine and rectum could also^ as the authors indicate, be due to misclassification.
35      Both possibilities have been considered by the investigators in their discussion.

        2/1/98                                    8-8       DRAFT-DO NOT CITE OR QUOTE

-------
 1      .       Cancer of the lung showed a positive trend with length of membership as well as with
 2      latency, although none of the SMRs were statistically significant except for the workers without
 3      any work histories. The individuals without any work histories may have been the ones who
 4      were in their jobs for the longest period of time, because workers without job histories included
 5      those who had the same job before and after 1967 and thus may have worked 12 to 14 years or
 6      longer. If they had belonged to the category in which heavy exposure to diesel exhaust
 7      emissions was very common for this prolonged time, then the increase in lung cancer, as well as
 8      stomach cancer, might be linked to diesel exhaust. Further information on those without work
 9      histories should be obtained if possible because such information may be quite informative with
10      regard to the evaluation of the carcinogeniciry of diesel exhaust.
11             The study design is adequate, covers about a 15-year observation period, has a large
12      enough population, and is appropriately analyzed; however, it has too many limitations to permit
13      any conclusions. First, no exposure histories are available. One has to make do with job
14      histories, which provide limited information on exposure level. Any person who ever worked at
15      the job or any person working at the same job over any period of time is included in the same
16      category; this  would have a dilution effect, since extremely variable exposures were considered
17      in the study.  Second, the length of time worked in any particular job is not available. Third,
18      work histories were not available for 9,707 individuals, who contributed 104,447.5 person-years,
19      a large proportion of the study cohort (28%). These individuals happen to show the most
20      evidence of a carcinogenic effect. Confounding by alcohol consumption for cancer of the liver
21      and smoking for emphysema and cancer of the  lung was not ruled out.  Last, although 34,156
22      members were eligible for the study, the vital status of 1,765 individuals was unknown.
23      Nevertheless,  they were still considered in the denominator of all the analyses. The investigators
24      fail to mention how the person-year calculation for these individuals was handled. Also, some of
25      the person-years might have been overestimated, as people may have paid the dues for a
26      particular year and then left work. These two causes of overestimation of the denominator may
27      have resulted in  some or all the SMRs being underestimated.
28             As for the smoking survey, the investigators took a very small sample (133 out of 34;156,
29      which was not even 1%).  Of 133, only 107 (80%) participated. It was a systematic sample, but
30      the authors neglected to mention how the list was prepared. Hence, the sample may not be
31      representative of the study population and, with a small sample size, the results are not
32      generalizable. The questionnaire asked only for current smoking history. No detailed history
33      was obtained for the amount smoked or length of smoking history, both of which have a bearing
34      on emphysema as well as lung carcinoma. t
35

        2/1/98                                     8-9        DRAFT-DO NOT CITE OR QUOTE

-------
 1      8.2.5. Edling et al. (1987): Mortality Among Personnel Exposed to Diesel Exhaust
 2            This is a retrospective cohort mortality study of bus company employees, which
 3      investigated a possible increased mortality in cardiovascular diseases and cancers from diesel
 4      exhaust exposure. The cohort comprised all males employed at five different bus companies in
 5      southeastern Sweden between 1950 and 1959. Based on information from personnel registers,
 6      individuals were classified into one or more categories and could have contributed person-years
 7      at risk in more than one exposure category. The study period was from 1951 to 1983;
 8      information was collected from the National Death Registry, and copiejs of death certificates
 9      were obtained from the National Bureau of Statistics. Workers who died after age 79 were
10      excluded from the study because diagnostic procedures were likely to be more uncertain at
11      higher ages (according to investigators).  The cause-, sex-, and age-specific national death rates
12      in Sweden were applied to the 5-year age categories  of person-years of observation to determine
13      expected deaths for all causes, malignant diseases, and cardiovascular diseases. A Poisson
14      distribution was used to calculate p-values and confidence limits for the ratio of observed to
15      expected deaths.  The total cohort of 694 men (after loss of 5 men to follow-up) was divided into
16      three exposure categories: (1) clerks with the lowest exposure, (2) bus drivers with moderate
17      exposure, and (3) bus garage workers with highest exposure.
18            The 694 men provided 20,304 person-years of observation, with 195 deaths compared
19      with 237 expected. A deficit in  cancer deaths largely accounted for this lower-than-expected
20      mortality in the total cohort. Among subcohorts, no difference between observed and expected
21      deaths for total mortality, total cancers, or cardiovascular causes was observed for clerks (lowest
22      diesel exposure), bus drivers (moderate diesel exposure), and garage workers (high diesel
23      exposure). The risk ratios for all three categories were less than 1 except for cardiovascular
24      diseases among bus drivers, which was 1.1.
25            When the analysis was restricted to  members who had at least a 10-year latency period
26      and either any exposure or an exposure exceeding 10 years, similar results were obtained, with
27      fewer neoplasms than expected, whereas cardiovascular diseases showed risk around or slightly
28      above unity.
29            Five lung cancer deaths were observed among bus drivers who had moderate diesel
30      exhaust exposure, whereas 7.2 were expected. The only other  lung cancer death was observed
31      among bus garage workers who  had the highest diesel exhaust exposure. The small size of the
32      cohort and poor data on diesel exhaust exposure are among the major limitations of this study.
33      Although lifetime occupational histories were available, no industrial hygiene data were
34      presented to validate the classification of workers into low, moderate, and high exposure to diesel
35      exhaust based on job title. The power of the present study was estimated to  be 80% to detect a

        2/1/98                                    8-10       DRAFT-DO NOT CITE OR QUOTE

-------
  1      relative risk of 1.2 for cardiovascular diseases and 1.4 for cancers, but for specific cancer sites.
  2      the power was much lower than this. No information was available on confounding effects of
  3      smoking and asbestos exposure at the work sites,
  4
  5      8.2.6.  Boffetta and Stellman (1988): Diesel Exhaust Exposure and Mortality Among
  6             Males in the American Cancer Society Prospective Study
  7             Boffetta and Stellman conducted a mortality analysis of 46,981 males whose vital status
  8      was known at the end of the first 2 years of follow-up.  The analysis was restricted to males aged
  9      40 to 79 years in 1982 who enrolled in the American Cancer Society's prospective mortality
10      study of cancer. Mortality was analyzed in relation to exposure to diesel exhaust and to
11      employment in selected occupations related to diesel exhaust exposure. In 1982, more than
12      77,000 American Cancer Society volunteers enrolled over 1.2 million men and women from all
13      50 states, the District of Columbia, and Puerto Rico in a long-term cohort study, the Cancer
14      Prevention Study II (CPS-II). Enrollees were usually friends, neighbors, or relatives of the
15      volunteers; enrollment was by family groups with at least one person in the household 45 years
16      of age or older.  Subjects were asked to fill out a four-page confidential questionnaire and return
17      it in a sealed envelope. The questionnaire included history of cancer and other diseases; use of
18      medications and vitamins; menstrual and reproductive history; occupational history; and
19      information on diet, drinking, smoking, and other habits. The questionnaire also included three
20      questions on occupation: (1) current occupation, (2) last occupation, if retired, and (3) job held
21      for the  longest period of time, if different from the other two. Occupations were coded to an ad
22      hoc two-digit classification in 70 categories.  Exposures at work or in daily life to any of the 12
23      groups of substances were also ascertained.  These included diesel engine exhausts, asbestos,
24      chemicals/acids/solvents, dyes, formaldehyde, coal or stone dusts, and gasoline exhausts.
25      Volunteers checked whether their enrollees .were alive or dead and recorded the date and place of
26      all deaths every other year during the study. Death certificates were then obtained from state
27      health departments and coded according to a system based on the ninth revision of the ICD by a
28      trained nosologist.
29             The data were analyzed to determine the mortality for all causes and lung cancer in
30      relation to diesel exhaust exposure, mortality for all causes and lung cancer in relation to
31      employment in selected occupations with high diesel exhaust exposure, and mortality from other
32      causes  in relation to diesel exhaust exposure. The incidence-density ratio was used as a measure
33      of association, and test-based confidence limits were calculated by the Miettinen method.. For
34      stratified analysis, the Mantel-Haenszel method was used for testing linear trends. Data on
35      476,648 subjects comprising 939,817 person-years of risk were available for analysis. Three

        2/1/98                                     8-11       DRAFT-DO NOT CITE OR QUOTE

-------
  1     percent of the subjects (14,667) had not given any smoking history, and 20% (98,026) of them
  2     did not give information on diesel exhaust exposure and were therefore excluded from the main
  3     diesel exhaust analysis.  Among individuals who had provided diesel exhaust exposure history,
  4     62,800 were exposed and 307,143 were not exposed. Comparison of the population with known
  5     information on diesel exhaust exposure with the excluded population with no information on
  6     diesel exhaust exposure showed that the mean ages were 54.7 and 57.7 years, the nonsmokers
  7     were 72.4% and 73.2%, and the total mortality rates per 1,000 per year were 23.0% and 28.8%,
  8     respectively.
  9            The all-cause mortality was elevated among railroad workers (relative risk [RR] = 1.43,
 10     95% confidence interval [CI] = 1.2,1.72), heavy equipment operators (RR = 1.7, 95% CI = 1.19,
 11     2.44), miners (RR = 1.34, 95% CI = 1.06, 1.68), and truck drivers (RR = 1.19,  95% CI = 1.07,
 12     1.31).  For lung cancer mortality the risks were significantly elevated for miners (RR = 2.67,
 13     95% CI = 1.63, 4.37) and heavy equipment operators (RR = 2.60, 95% CI = 1.12, 6.06).  Risks
 14     were also elevated but not significantly for railroad workers (RR = 1.59, 95% CI = 0.94, 2.69)
 15     and truck drivers (RR =  1.24, 95% CI = 0.93,  1.66). These risks were calculated according to the
 16     Mantel-Haenszel method, controlling for age and smoking. Although the relative risk was
 17     nonsignificant for truck drivers, a small dose-response effect was observed when duration of
. 18     diesel exhaust exposure for them was examined.  For drivers who worked for 1 to 15 years, the
 19     relative risk was 0.87, while for drivers who worked for more than 16 years, the relative risk was
 20     1.33 (95% CI = 0.64, 2.75). Relative risks for lung cancer were not presented for other
 21     occupations. Mortality analysis for other causes and diesel exhaust exposure showed a
 22     significant excess of deaths (p<0.05) in the following categories: cerebrovascular disease,
 23     arteriosclerosis, pneumonia, influenza, cirrhosis of the liver, and accidents.
 24            The two main methodologic concerns in this study are the representativeness of the study
 25     population and the quality of information on exposure. The sample, though very large, was
 26     composed of volunteers.  Thus, the cohort was healthier and less frequently exposed to important
 27     risk factors such as smoking and alcohol. Self-administered questionnaires were used to obtain
 28     data on occupation and diesel exhaust exposure. None of this information was validated. Nearly
 29     20% of the individuals had an unknown exposure status to diesel exhaust, and they experienced a
 30     higher mortality for all causes and lung cancer than both the diesel exhaust exposed and
 31     unexposed groups. This could have introduced a substantial bias in the estimate of the
 32     association. Although only 0.8% of the subjects were lost to follow-up, the  use of death
 33     certificates alone as a source of medical information poses problems in accuracy and coding. But
 34     the authors report that cancer deaths are routinely checked by histological confirmation from
 35     physicians or cancer registries. Given the fact that all diesel exhaust exposure  occupations, such

        2/1/98                                    8-12        DRAFT-DO NOT CITE OR QUOTE

-------
  1      as heavy equipment operators, truck drivers, and railroad workers, showed elevated lung cancer
  2      risk, this study is suggestive of a causal association.
  3
  4      8.2.7.  Garshick et al. (1988):  A Retrospective Cohort Study of Lung Cancer and Diesel
  5             Exhaust Exposure in Railroad Workers
  6             An earlier case-control study of lung cancer and diesel exhaust exposure in U.S. railroad
  7      workers by these investigators had demonstrated a relative odds of 1.41  (95% CI = 1.06,1.88)
  8      for lung cancer with 20 years of work in jobs with diesel exhaust exposure. To confirm these
  9      results, a large retrospective cohort mortality study was conducted by the same investigators.
10      Data sources for the study were the work records of the U.S. Railroad Retirement Board (RRB).
11      The cohort was selected based on job titles in 1959, which was the year by which 95% of the
12      locomotives in the United States were diesel powered. Diesel exhaust exposure was considered
13      to be a dichotomous variable depending on yearly job codes between  1959 and death or
14      retirement through 1980. Industrial hygiene evaluations and descriptions of job activities were
15      used to classify jobs as exposed or unexposed to diesel emissions. A questionnaire survey of 534
16      workers at one of the railroads where workers were asked to indicate the amount of time spent in
17      railroad locations, either near or away from sources of diesel exhaust, was used to validate this
18      classification. Workers selected for this survey were actively employed at the time of the survey,
19      40 to 64 years of age, who started work between 1939 and 1949, in the job codes sampled in
20      1959, and were eligible for railroad benefits. To qualify for benefits, a worker must have 10
21      years or more of service with the railroad and should not have worked for more than 2 years in a
22      nonrailroad job after leaving railroad work. Workers with recognized asbestos exposure, such as
23      repair of asbestos-insulated steam locomotive boilers, passenger cars, and steam pipes, or
24      railroad building construction and repairs, were excluded from the job categories selected for
25      study. However, a few jobs with some potential for asbestos exposure were included in the
26      cohort, and the analysis was done both ways, with and without them.
27            The death certificates for all subjects identified in 1959 and reported by the RRB to have
28     , died through 1980 were searched.  Twenty-five percent of them were obtained from the RRB and
29      the remainder from the appropriate state departments of health.  Coding of cause of death was
30      done without knowledge of exposure history, according to the eighth revision of the ICD.  If the
31      underlying cause of death was not lung cancer, but was mentioned on the death certificate, it was
32      assigned as a secondary cause of death, so that the ascertainment of all cases was complete.
33      Workers not reported by the RRB to have died by December 31, 1980, were considered to be
34      alive.  Deceased workers for whom death certificates had not been obtained or, if obtained, did
35      not indicate cause of death, were assumed to have died of unknown causes.

        2/1/98        .                            8-13       DRAFT-DO NOT CITE OR QUOTE

-------
 1             Proportional hazard models were fitted that provided estimates of relative risk for death
 2      caused by lung cancer using the partial likelihood method described by Cox, and 95% confidence
 3      intervals were constructed using the asymptotic normality of the estimated regression
 4      coefficients of the proportional hazards model.  Exposure was analyzed by diesel exhaust-
 5      exposed jobs in 1959 and by cumulative number of years of diesel exhaust exposure through
 6      1980. Directly standardized rate ratios for deaths from lung cancer were calculated for diesel
 7      exhaust exposed compared with unexpo'sed for each 5-year age group in 1959.  The standardized
 8      rates were based on the overall 5-year person-year time distribution of individuals in each age
 9      group starting in 1959. The only exception to this was between 1979 and 1980, when a 2-year
10      person-year distribution was used.  The Mantel-Haenszel analogue for person-year data was used
11      to calculate 95% confidence intervals for the standardized rate ratios,
12             The cohort consisted of 55,407 workers, 19,396 of whom had died by the end of 1980.
13      Death certificates were not available for 11.7% of all deaths. Of the 17,120 deaths for whom
14      death certificates were obtained, 48.4% were attributable to diseases of the circulatory system,   .
15      whereas 21% were attributable to all neoplasms. Of all neoplasms, 8.7% (1,694 deaths) were due
16      to lung cancer. A higher proportion of workers in the younger age groups, mainly brakemen and
17      conductors, were exposed to diesel exhaust, while a higher proportion of workers in the older age
18      groups were potentially exposed to asbestos.  In a proportional hazards model, analyses by age in
19      1959 found a relative risk of 1.45 (95% CI = 1.11,1.89)  among the age group 40 to 44 years and
20      a relative risk of 1.33 (95% CI = 1.03,1.73) for the age group 45 to 49 years. Risk estimates in
21      the older age groups 50 to 54, 55 to 59, and 60 to 64 years were 1.2,1.18, and 0.99, respectively,
22      and were not statistically significant.  The two youngest  age groups in 1959 had workers with the
23      highest prevalence and longest duration of diesel exhaust exposure and lowest exposure to
24      asbestos.  When potential asbestos exposure was considered as a confounding variable in a
25      proportional hazards model, the estimates of relative risk for asbestos exposure were all near null
26      value and not significant. Analysis of workers exposed to diesel exhaust in 1959 (n = 42,535),
27      excluding the workers with potential past exposure to asbestos, yielded relative risks of 1.57
28      (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
29      45 to 49 years. Directly standardized rate ratios were also calculated-for each 1959  age group
30      based on diesel exhaust exposure in 1959. The results obtained confirmed those obtained by
31      using the proportional hazards model.
32             Relative risk estimates were then obtained using duration of diesel exhaust exposure as a
33      surrogate for dose.  In a model that used years of exposure up to and including exposure in the
34      year of death, no exposure duration-response relationship was obtained.  When analysis was done
35      by disregarding exposure in the year of death and 4 years prior to death, the risk of dying from

        2/1/98                                   8-14       DRAFT-DO NOT CITE OR QUOTE

-------
  1      lung cancer increased with the number of years worked in a diesel-exhaust-exposed job.  In this
  2      analysis, exposure to diesel exhaust was analyzed by exposure duration groups and in a model
  3      entering age in 1959 as a continuous variable. The workers with greater than 15 years of
  4      exposure had a relative risk of lung cancer of 1.72 (95% CI = 1.27, 2.33). The risk for 1 to 4
  5      years of cumulative exposure was 1.20 (95% CI = 1.01, 1.44); for 5 to 9 years of cumulative
  6      exposure, it was 1.24 (95% CI = 1.06, 1.44); and for 10 to 14 years of cumulative exposure, it
  7      was 1.32 (95% CI = 1.13, 1.56). Directly standardized rate ratios were also calculated for each
  8      1959 age group based on diesel exposure in 1959. The results obtained confirmed those obtained
  9      by using the proportional hazards model.
10             The results of this study, demonstrating a positive association between diesel exhaust
11      exposure and increased lung cancer, are consistent with the results of the case-control study
12      conducted by the same investigators in railroad workers dying of lung cancer from March 1981
13      through February 1982, This cohort study has addressed many of the weaknesses of the other
14      epidemiologic studies. The large sample size (60,000) allowed sufficient power to detect small   .
15      risks and also permitted the exclusion of workers with potential past exposure to asbestos. The
16      stability of job career paths in the cohort ensured that of the workers 40 to 44 years of age in
17      1959 classified as diesel exhaust-exposed, 94% of the cases were still in diesel exhaust-exposed
18      jobs 20 years later.
19             The main limitation of the study is the lack of quantitative data on exposure to diesel
20      exhaust. This is one of the few studies in which industrial hygiene measurements of diesel
21      exhaust were done. These measurements were correlated with job titles to divide the cohort in
22      dichotomous exposure groups of exposed and nonexposed.  This may have led to an
23      underestimation of the risk of lung cancer since exposed groups included individuals with low to
24      high exposure. The number of years exposed to diesel exhaust was used as a surrogate for dose.
25      The dose, based on duration of employment, may have been inaccurate because individuals were
26      working on steam or diesel locomotives during the transition period. If the categories of
27      exposure to diesel exhaust would have been set up as no, low, moderate, and high exposure, the
28      results would have been more meaningful and so would have been the dose-response
29      relationship. Another limitation of this study was the inability to examine the effect of years of
30      exposure and latency. No adjustment for smoking was made in this study. However, an earlier
31      case-control study done in the same cohort (Garshick et al., 1987) showed no significant
32      difference in the risk estimate after adjusting for smoking.  Despite these limitations, the results
33      of this study demonstrate that occupational exposure to diesel exhaust is associated with a
34      modest risk (1.5) of lung cancer.
35

        2/1/98                                    8-15       DRAFT-DO NOT CITE  OR QUOTE

-------
 1      8.2.8.  Gustavsson et al. (1990):  Lung Cancer and Exposure to Diesel Exhaust Among Bus
 2            Garage Workers
 3            A retrospective mortality study (from 1952 to 1986), cancer incidence study (from 1958
 4      to 1984), and nested case-control study were conducted among a cohort of 708 male workers  f
 5      from five bus garages in Stockholm, Sweden, who had worked for at least 6 months between
 6      1945 and 1970.  Thirteen individuals were lost to follow-up, reducing the cohort to 695.
 7            Information was available on location of workplace, job type, and beginning and ending
 8      of work periods. Workers were traced using a computerized register of the living population,
 9      death and burial books, and data from the Stockholm city archives.
10            For the cohort mortality analyses, death rates of the general population of greater
11      Stockholm were used. Death rates of occupationally active individuals, a subset of the general
12      population of greater Stockholm, were used as a second comparison group to reduce the bias
13      from "healthy worker effect." Mortality analysis was conducted using the "occupational
14      mortality analysis program" (OCMAP-PC). For cancer incidence analysis, the "epidemiology in
15      Linkoping" (EPILIN) program was used, with the incidence rates obtained from the cancer
16      registry.
17            For the nested case-control study, both dead and incident primary lung cancers, identified
18      in the register of cause of deaths and the cancer register, were selected as cases (20). Six controls
19      matched on age ±2 years, selected from the noncases at the time of the diagnosis of cases, were
20      drawn at random without replacements. Matched analyses were done to calculate odds ratios
21      using conditional logistic regression. The EGRET and Epilog programs were used for these
22      analyses.
23            Diesel exhaust and asbestos exposure assessments were performed by industrial
24      nygienists based on the intensity of exposure to diesel exhaust and asbestos, specific for
25      workplace, work task, and calendar time period.  A diesel exhaust exposure assessment was
26      based on (1) amount of emission (number of buses, engine size, running time, and type of fuel),
27      (2) ventilatory equipment and air volume of the garages, and (3) job types and work practices.
28      Based on detailed historical data and very few actual measurements, relative exposures were
29      estimated (these were not absolute exposure levels).  The scale was set to 0 for unexposed and 1
30      for lowest exposure, with each additional unit increase corresponding to a 50% increase in
31      successive intensity (i.e., 1.5, 2.25, 3.38, and 5.06).
32            Based on personal sampling of asbestos during 1987, exposures were estimated and time-
33      weighted annual mean exposures were classified on a scale of three degrees (0,  1, and 2).
34    ,  Cumulative exposures for both diesel exhaust and asbestos were calculated by multiplying the
35      level of exposure by the duration of every work period. An exposure index was calculated by

        2/1/98               -                     8-16       DRAFT-DO NOT CITE OR QUOTE

-------
  1      adding for every individual contributions from all work periods for both diesel exhaust and
  2      asbestos. Four diesel exhaust index classes were created: 0 to 10, 10 to 20, 20 to 30, arid >30,
  3      The four asbestos index classes were 0 to 20, 20 to 40,40 to 60, and >60. The cumulative
  4      exposure indices were used for the nested case-control study.
  5             Excesses were observed for all cancers and some other site-specific cancers using both
  6      comparison populations for the cohort mortality study, but none of them was statistically
  7      significant. Based on 17 cases, standardized mortality ratios (SMRs) for lung cancer were 122
  8      and 115 using Stockholm occupationally active and general population, respectively. No dose-
  9      response was observed with increasing cumulative exposure. The cancer incidence study
10      reportedly confirmed the mortality results (results not given).
11             The nested case-control study showed increasing risk of lung cancer with increasing
12      exposure.  Weighted linear regression gave RRs of 1.34 (95% CI = 1.09 to 1.64), 1.81 (95% CI =
13      1.20 to 2.71), and 2.43 (95% CI = 1.32 to 4.47) for the diesel exhaust indices 10 to 20, 20 to 30,
14      and >30, respectively, using 0 to  10 as the comparison group. The study was based on 17 cases
15      and six controls for each case matched on age ± 2 years.  The results from conditional logistic
16      regression were similar to those obtained by weighted linear regression, but none was statistically
17      significant. Adjustment for asbestos exposure did not change the  lung cancer risk for diesel
18      exhaust.
19             The main strength of this  study is the detailed exposure matrices  constructed for both
20      diesel exhaust and asbestos exposure, although they were based primarily on job tasks and very
21      few actual measurements. There are a few methodological limitations to this study. The cohort
22      is small and there were only 17 lung cancer deaths, thus the power is low. Exposure or outcome
23      may be misclassified, although any resulting bias in the relative risk estimates is likely to be
24      toward unity, because exposure classification was done independently of the outcome.  Although
25      the analysis by dose indices was done, no latency analysis was performed.  Finally, data on
26      smoking were missing, thus potentially confounding the lung cancer results.  The authors suggest
27      that even the heaviest smoking among individuals who were heavily exposed to diesel exhaust
28      will be unable to explain the excess relative risk of 2.4 observed in this group. This may be an
29      .overstatement, however, as cigarette smoking is a very strong risk factor for lung cancer.
30      Overall, this study provides some support to the excess lung cancer results found earlier among
31      populations exposed to diesel exhaust.
32-.                ^    •                                       '    f
33      8.2.9.  Hansen (1993): A Follow-Up Study on the Mortality of Truck Drivers
34             This is a retrospective cohort mortality study of unskilled male laborers, ages  15 to 74
35      years, in Denmark, identified from a nationwide census file of November 9,1970.  The exposed

        2/1/98              .                    8-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1      group included all truck drivers employed in the road delivery or long-haul business (14,225).
 2      The unexposed group included all laborers in certain selected occupational groups considered to
 3      be unexposed to fossil fuel combustion products and to resemble truck drivers in terms of work-
 4      related physical demands and various personal background characteristics (43,024).
 5            Through automatic record linkage between the 1970 census register (the Central
 6      Population Register 1970 to 1980) and the Death Certificate Register (1970 to 1980), the
 7      population was followed for cause-specific mortality or emigration up to November 9, 1980.
 8      Expected number of deaths among truck drivers was calculated by using the 5-year age group
 9      and 5-year time period death rates of the unexposed group and applying them to the person-years
10      accumulated by truck drivers. International Classification of Diseases Revision 8 was used to
11      code the underlying cause of death. Test-based confidence intervals (CI) were calculated using
12      Miettinen's method.  A Poisson distribution was assumed for the smaller numbers, and CI was
13      calculated based on exact Poisson distribution (Ciba-Geigy).  Total person-years accrued by
14      truck drivers were 138,302, whereas for the unexposed population, they were 407,780. There
15      were 627 deaths among truck drivers and 3,811 deaths in the unexposed group. Statistically
16      significant (SS) excesses were observed for all cancer mortality (SMR =121, 95% CI = 104 to .
17      140); cancer of respiratory organs (SMR = 160, 95% Cl = 128 to 198), which mainly was due to
18      cancer of bronchus and lung (SMR = 160, 95% CI = 126 to 200); and multiple myeloma (SMR =
19      439, 95% CI = 142 to 1,024). When lung cancer mortality was further explored by age groups,
20      excesses were observed in most of the age groups (30 to 39, 45 to 49, 50 to 54, 55 to 59, 60 to
21      64, and 65 to 74), but there were  small numbers of deaths in each group when stratified by age,
22      and the excesses were statistically significant for the 55 to 59  (SMR = 229, 95% CI = 138 to 358)
23      and 60 to 64 (SMR = 227, 95% CI= 142 to 344) age groups only.
24            As acknowledged by the author, the study has quite a few methodologic limitations. The
25      exposure to diesel exhaust is assumed in truck drivers based on diesel-powered trucks, but no
26      validation of qualitative or quantitative exposure is attempted. It is also not known whether any
27      of these truck drivers or any other laborers had changed jobs after the census of November 9,
28      1970, thus creating potential misclassification bias in exposure to diesel exhaust. The lack of
29      smoking data and a 36% rural population (usually consuming less tobacco) in the unexposed
30      group further confound the lung cancer results. The follow-up period is relatively short, and a
31      latency analysis was not attempted. At best, the findings of this study are consistent with the
32      findings of other truck driver studies.                                      .          •     .
33            Table 8-1 summarizes the foregoing cohort studies.
34
        2/1/98                                    8-18       DRAFT-DO NOT CITE OR QUOTE

-------
                           Table 8-1. Epidemiologic studies of the health effects of exposure to diesel exhaust:
                           cohort mortality studies
oo
         Authors
                  Population studied
                               Diesel exhaust exposure
                                     assessment
            Results
                                            Limitations
        Waller
        (1981)
            Approximately 20,000 male
            London transportation workers

            Aged 45 to 64 years

            25 years follow-up (1950-1974)
                               Five job categories used
                               to define exposure

                               Environmental
                               benzo[a]pyrene
                               concentrations measured
                               in 1957 and 1979
SMR = 79 for lung cancer for the
total cohort

SMRs for all five job categories
were less than 100 for lung cancer
                                Exposure measurement of
                                benzo[a]pyrene showed very little
                                difference between inside and outside
                                the garage

                                Incomplete information on cohort
                                members
                                                                                                          No adjustment for confounding such
                                                                                                          as other exposures, cigarette
                                                                                                          smoking, etc,

                                                                                                          No latency analysis	
oo
O
O
*
O
H
n
g
O
a
3
W
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
-------
                            Table 8-1.  Epidemiological studies of the health effects of exposure to diesel exhaust:
                            cohort mortality studies (continued)
 oo
           Authors
       Population studied
 Diesel exhaust exposure
       assessment
             Results
Limitations
         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
                               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
                               diesel exhaust emissions
                        SMR = 166 (p<0.05) for liver
                        cancer for total cohort

                        SMR = 343 (observed •= 5,p<0.05)
                        for lung cancer for high-exposure
                        bulldozer operators with 15-19
                        years of membership, 20+ years
                        of follow-up

                        SMR = 119 (observed = 141,
                        p
-------
                           Table 8-1. Epidemiological studies of the health effects of exposure to diesel exhaust:
                           cohort mortality studies (continued)
oo
         Authors
                   Population studied
Diesel exhaust exposure
      assessment
            Results
           Limitations
        Garshick    55,407 white male railroad
        etal. (1988) workers

                   Aged 40 to 64 years in 1959

                   Started work 10-20 years earlier
                   than 1959
                                           Industrial hygiene data   RR = 1.45 (40-44 year age group)   Years of exposure used as surrogate
                                           correlated with job
                                           titles to dichotomize
                                           the jobs as "exposed"
                                           or "not exposed"
                       RR = 1.33 (45-49 year age group)   for dose
                       Both SS
                                                        Not possible to separate the effect of
                       After exclusion of workers exposed  time since first exposure and duration
                       to asbestos                       of exposure
                       RR = 1.57 (40-44 year age group)
                       RR = 1,34 (45-49 year age group)
                       Both SS
                                                                         Dose response indicated by
                                                                         increasing lung cancer risk with
                                                                         increasing cumulative exposure
oo
O
o
•z
o
H
O
frmmt
H
W
a
o
H
W
 Gustav.sson  695 male workers from 5 bus
 et al, (1990) garages in Stockholm, Sweden,
            who had worked for 6 months
            between 1945 and 1970
Four diesel exhaust
indices were created:
OtolO, 10 to 20, 20-30,
and >30 based on job
tasks and duration of
                   34 years follow-up (1952-1986)   work

                   Nested case-control study
                   17 cases, six controls for each case
                   matched on age ± 2 years
SMRs of 122 and 115 (OA and
GP), respectively, SNS

Case-control study results
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
 Hansen     Cohort of 57,249 unskilled        Diesel exhaust exposure
 (1993)     laborers, ages 15 to 74, in        assumed based on diesel-
            Denmark (nationwide census file)  powered trucks
            November 9, 1970

            Follow-up through November 9,
            1980
                       SS SMR = 160 for bronchus and
                       lung for total population
                                No actual exposure data available

                                Lack of smoking data

                                Job changes may have occurred from
                                laborer to driver

                                Short follow-up period
Abbreviations: RR = relative risk; SMR = standardized mortality ratio; SNS = statistically nonsignificant; SS = statistically significant;
OA = occupationally active; GP = general population.

-------
 1      8.3. CASE-CONTROL STUDIES OF LUNG CANCER
 2      8.3.1.  Williams etal. (1977): Associations of Cancer Site and Type With
 3            Occupation and Industry From the Third National Cancer Survey Interview
 4            This paper reports findings of the analysis of the Third National Cancer Survey (TNCS).
 5      The lifetime histories, occupations, and industries were studied for associations with specific
 6      cancer sites and types after controlling for age, sex, race, education, use of cigarettes or alcohol,
 7      and geographic location.  Of 13,179 cancer patients, a 10% random sample of all incident
 8      invasive cancers in eight areas, a total of 7,518 were successfully interviewed in the 3 years
 9 .     surveyed by the TNCS. These comprised 57% of those eligible to participate. The interview
10      included items on use of tobacco and alcohol (by type, amount, and duration), family income,
11      patient education, and employment history.  Actual descriptions of the occupation and industry
12      were recorded by interviewers and were coded separately for main lifetime employment, recent
13      employment, and other jobs held according to the 1970 Census Coding Scheme.  Occupations or
14      industries were combined to form larger groups. Coding of occupational and industrial labels in
15      meaningful job categories was done by one of the authors.  Of the 3,539 interviewed males and
16      3,937 interviewed females, 95% and 84%, respectively, listed some main employment. The
17      basic analysis consisted of an intercancer comparison and involved comparing the proportions of
18      specific main lifetime industries and occupations among patients with cancer at one site with
19      those of patients having cancer at other sites combined as a control group; this was done using a
20      series of Mantel-Haenszel stratified contingency table analyses to yield odds ratios and chi-
21      square values. Odds ratios (ORs) were computed separately for males and females, controlling
22      for age, race, education, tobacco, alcohol,  and geographic location.
23            A total of 432 and 128 lung cancers were present in males and females, respectively. For
24      males, an excess risk of lung cancer was observed for the following main industrial groups:
25      mines (OR =1.21), construction (OR = 1.24), transportation (OR =1.17), utility and sanitary
26      services (OR = 2.79, p<0.05), and professional (OR = 1.41).  An excess of bladder cancer was
27      reported for the mining industry (OR = 1.61). For females, an excess of lung cancer was
28     • detected for the transportation industry (OR  = 1.96); finance and retail industry (OR = 1.73); and
29      the business, car repair, and miscellaneous service industry (OR = 2.29).  None of these excesses
30      were statistically significant.  All of these odds ratios were adjusted for age, race, education,
31      tobacco, alcohol, and geographic location. The transportation industry for males and females
32      also showed a nonsignificant excess risk for cancers of the liver and gall bladder ducts. When
33      the analysis was done for specific lifetime industries, the odds ratio for lung cancer in males was
34      1.40 for railroad workers and 1.34 for truck drivers.  Both of these excesses were statistically
35      nonsignificant.

        2/1/98   .                                 8-22       DRAFT-DO NOT CITE OR QUOTE

-------
 1             The strengths of the TNCS interview data set are its large size, histological confirmation
 2      of nearly 95% of diagnoses, availability of information on occupation, and details of
 3      confounding variables obtained by personal interview and ability to control for them.  Among its
 4      weaknesses are a 47% nonresponse rate and the fact that the population surveyed came from
 5      predominantly urban areas and did not represent many industries. Also, most of the associations
 6      observed did not achieve statistical significance because they were based on small numbers of
 7      patients who had both specific cancers and specific types of employment. The control group was
 8      the combined "other cancers," which may have diluted the association because diesel exhaust is
 9      also suspected of being associated with bladder cancer, and this category was included in the
10      control group when the comparison was made with lung cancer.  The study presented several
11      tables, but the total population in each table was different and never added up to the initial
12      number interviewed.  The authors failed to explain these omissions. Further, when multiple
13      comparisons are made, some significant associations arise by chance.  This analysis suggests an
14      association with lung cancer for three industries with potential diesel exhaust exposure: trucking,
15      railroading, and mining.
16
17      8.3.2.  Hall and Wynder (1984): A Case-Control Study of Diesel Exhaust Exposure and
18             Lung Cancer
19             Hall and Wynder conducted a case-control study of 502 male lung cancer cases and 502
20      controls without tobacco-related diseases that examined an association between occupational
21      diesel exhaust exposure and lung cancer. Histologically confirmed primary lung cancer patients
22      who were 20 to 80 years old were ascertained from 18 participating hospitals in six U.S. cities,
23      12 months prior to the interview.  Eligible controls, patients at the same hospitals without
24      tobacco-related diseases, were matched to cases by age (±5 years), race, hospital, and hospital
25      room status. The number of male lung cancer cases interviewed totaled 502, which was 64% of
26      those who met the study criteria for eligibility. Of the remaining 36%, 8% refused, 21% were
27      too ill or had died, and 7% were unreliable. Seventy-five percent of eligible controls completed
28      interviews. Of these interviewed controls, 49.9% were from the  all-cancers category, whereas
29      50.1% were from the all-noncancers category. All interviews were obtained in hospitals to
30      gather detailed information on smoking history, coffee consumption, artificial sweetener use,
31      residential history, and abbreviated medical history as well as standard demographic variables.
32      Occupational information was elicited by a question on the usual lifetime occupation and was
33      coded by the abbreviated list of the U.S. Bureau of Census Codes.  The odds ratios were .
34      calculated to evaluate the association between'diesel exhaust exposure and risk of lung cancer
35      incidence. Summary odds ratios were computed by the Mantel-Haenszel method after adjusting

        2/1/98                                    8-23       DRAFT-DO NOT CITE OR QUOTE

-------
 1      for potential confounding by age, smoking, and socioeconomic class. Two-sided, 95%
 2      confidence intervals were computed by Woolf s method. Occupational exposure to diesel
 3      exhaust was defined by two criteria. First, occupational titles were coded "probably high
 4      exposure" as defined by the industrial hygiene standards established for the various jobs. The
 5      job titles included under this category were warehousemen, bus and truck drivers, railroad
 6      workers, and heavy equipment operators and repairmen. The second method used the National
 7      Institute for Occupational Safety and Health (NIOSH) criteria to analyze  occupations by diesel
 8      exposure. In this method, the estimated proportion of exposed workers was computed for each
 9      occupational category by using the NIOSH estimates of the exposed population as the numerator
10      and the estimates of individuals employed in each occupational category  from the 1970 census as
11      the denominator. Occupations estimated to have at     20% of their employees exposed to
12      diesel exhaust were defined as "high exposure," those with 10 to 19% of their employees
13      exposed were defined as "moderate exposure," and those with less than 10% of their employees
14      exposed were defined as "low exposure."
15            Cases and controls were compared with respect to exposure.  The relative risk was 2.0
16      (95% CI = 1.2, 3.2) for those workers who were exposed to diesel exhaust versus those who were
17      not.  The risk, however, decreased to a nonsignificant 1.4 when the data were adjusted for
18      smoking. Analysis by NIOSH criteria found a nonsignificant relative risk of 1.7 in the high-
19      exposure group. There were no significantly increased cancer risks by occupation either by the
20      first method or by the NIOSH method. To assess any possible synergism between diesel exhaust
21      .exposure and smoking, the lung cancer risks were calculated for different smoking categories.
22      The relative risks were 1.46 among nonsmokers and ex-smokers, 0.82 among current smokers of
23      <20 cigarettes/day, and 1.3 among current smokers of 20+ cigarettes/day, indicating a lack of
24      synergistic effects.
25            The major strength of this study is the availability of a detailed smoking history for all the
26      study subjects. However, this is offset by the lack of diesel exhaust exposure measurements, use
27      of a poor surrogate for exposure, and lack of consideration of latency period. Information was
28      collected on only one major lifetime occupation, and it is likely that those workers who had more
29      than one major job may not have reported the occupation with the heaviest diesel exhaust
30      exposures. Further, occupational histories were obtained from self-reports and were not
31      validated with work records. This could have resulted in recall bias and misclassification of
32      exposure status.
33                                                          ,
        2/1/98                                    8-24       DRAFT-DO NOT CITE OR QUOTE

-------
  1     8.3.3.  Damber and Larsson (1987): Occupation and Male Lung Cancer, a Case-Control
  2            Study in Northern Sweden
  3            A case-control study of lung cancer was conducted in northern Sweden to determine the
  4     occupational risk factors that could explain the large geographic variations of lung cancer
  5     incidence in that country. The study region comprised the three northernmost counties of
  6     Sweden, with a total male population of about 390,000. The rural municipalities with 15% to
  7     20% of the total population have forestry and agriculture as dominating industries, and the urban
  8     areas have a variety of industrial activities (mines, smelters, steel factories, paper mills, and
  9     mechanical workshops). All male cases of lung cancer reported to the Swedish Cancer Registry
 10     during the 6-year period between 1972 and 1977, who had died before the start of the study, were
 11     selected. Of 604 eligible cases, 5 did not have microscopic confirmation and in another 5 the
 12     diagnosis was doubtful, but these cases were included nevertheless.  Cases were classified as
 13     small carcinomas, squamous cell carcinomas, adenocarcinomas, and other types. For each case a
 14     dead control was drawn from the National Death Registry matched by sex, year of death, age,
 15     and municipality.  Deaths in controls classified as lung cancer and suicides were excluded. A
 16     living control matched to the case by sex, year of birth, and municipality was also drawn from
 17     the National Population Registry. Postal questionnaires were sent to close relatives of cases and
. 18     dead controls, and to living controls themselves to collect data on occupation, employment, and
 19     smoking habits. Replies were received from 589 cases (98%), 582 surrogates of dead controls
 20     (96%), and 453 living controls (97%).
 21            Occupational data were collected on occupations or employment held for at least 1 year
 22     and included type of industry, company name, task, and duration of employment.
 23     Supplementary telephone interviews were performed if occupational data were lacking  for any
 24     period between age 20 and time of diagnosis. Data analysis involved calculation of the odds
 25     ratios by the exact method based on the hypergeometric distribution and the use of a linear
 26     logistic regression model to adjust for the potential confounding effects of smoking. Separate
 27     analyses were performed with dead and living controls, and on the whole there was good
 28     agreement between the two control groups. A person who had been active for at least 1 year  in a
 29     specific occupation was in the analysis assigned to that occupation.
 30            Using dead controls, the odds ratios adjusted for smoking were 1.0 (95% CI = 0.7,1.5)
 31     and 2.7 (95% CI = 1.0, 8.1) for professional drivers (> 1 year of employment) and underground
 32     miners (> 1  year of employment), respectively. For 20 or more years of employment in those
 33     occupations, the odds ratios adjusted for smoking were 1.2 (95% CI = 0.6,2.2) and 9.8  (95% CI
 34     = 1.5,414).  These were the only two occupations listed with potential diesel exhaust exposure.
 35     An excess significant risk was detected for copper smelter workers, plumbers, and electricians,  as

        2/1/98       .                           8-25        DRAFT-DO NOT CITE OR QUOTE

-------
 1      well as concrete and asphalt workers. Occupational asbestos exposure was also associated with
 2      an elevated odds ratio of 2.6 (95% CI = 1.6, 3.6) for > 1 year of employment and 3.6 (95% CI ='
 3      1.9, 7.2) for >20 years of employment.  All the odds ratios were calculated by adjusting for age,
 4      smoking, and municipality. After comparison with the live controls, the odds ratios were found
 5      to be lower than those observed with dead controls.  None of the odds ratios were statistically
 6      significant in this comparison.
 7             This study did not detect any excess risk of lung cancer for professional drivers, who,
 8      among all the occupations listed, had the most potential for exposure to motor vehicle exhaust.
 9      However, it is not known whether these drivers were exposed exclusively to gasoline exhaust,
10      diesel exhaust, or varying degrees of both.  An excess risk was detected for underground miners,
11      but it is not known if this was due to diesel emissions from engines or from radon daughters in
12      poorly ventilated mines.  Although a high response rate (98%) was obtained by the postal
13      questionnaires, the use of surrogate respondents is known to lead to misclassification errors that
14      can bias the odds ratio to 1.
15
16      8.3.4.  Lerchen et al. (1987): Lung Cancer and Occupation in New Mexico
17             This is a population-based case-control study conducted in New Mexico that examines
18      the association between occupation and occurrence of lung cancer in Hispanic and non-Hispanic
19      whites.  Cases involved residents of New Mexico, 25 through 84 years of age and diagnosed
20      between January 1,  1980, and December 31, 1982, with primary lung cancer, excluding
21      bronchioalveolar carcinoma.  Cases were ascertained through the New Mexico Tumor Registry,
22      which is a member of the Surveillance Epidemiology and End Results (SEER) Program of the
23      National Cancer Institute. Controls were chosen by randomly selecting residential telephone
24      numbers and, for those over 65 years of age, from the Health Care Financing Administration's
25      roster of Medicare participants.  They were frequency-matched to cases for sex, ethnicity, and
26      10-year age category with a ratio of 1.5 controls per case. The 506 cases (333 males and 173
27      females) and 771 controls (499 males and 272 females) were interviewed, with a nonresponse
28     . rate of 11% for cases. Next of kin provided interviews for 50% and 43% of male and female
29      cases, respectively.  Among controls, only 2% of the interviews were provided by next of kin for
30      each sex. Data were collected by personal interviews conducted by bilingual interviewers in the
31      participants' homes. A lifetime occupational history and a self-reported history of exposure to
32      specific agents were obtained for each job held for at least 6 months since age 12. Questions
33      were asked about the title of the position, duties performed, location and nature of industry, and
34      time at each job title.  A detailed smoking history was also obtained. The variables  on
35      occupational exposures were coded according to the Standard Industrial Classification scheme  by

        2/1/98                                   8-26       DRAFT-DO NOT CITE OR QUOTE

-------
  1      a single person and reviewed by another. To test the hypothesis about the high-risk jobs for lung
  2      cancer, an a priori listing of suspected occupations and industries was created by a two-step
  3      process involving a literature review for implicated industries and occupations by the principal
  4      investigator.  The appropriate Standard Industrial Classification and Standard Occupational
  5      Codes associated with job titles were also determined by the principal investigator. For four
  6      agents—asbestos, wood dust, diesel exhaust, and formaldehyde—the industries and occupations
  7      determined to have exposure were identified, and linking of specific industries and occupations
  8      was based on literature review and consultation with local industrial hyjienists.
  9             The relative odds were calculated for suspect occupations and industries, classifying
10      individuals as ever employed for at least 1 year in an industry or occupation and defining the
11      reference group as those subjects never employed in that particular industry or occupation,
12      Multiple logistic regression models were used to control simultaneously for age, ethnicity, and
13      smoking status. For occupations with potential diesel exhaust exposure, the analysis showed no
14      excess risks for diesel engine mechanics and auto mechanics. Similarly, when analyzed by
15      exposure to specific agents, the odds ratio adjusted for age, smoking, and ethnicity was not
16      elevated for diesel exhaust fumes (OR = 0.6,95% CI = 0.2,1.6).  Elevated odds ratios were
17      found for uranium miners (OR = 2.8, 95% CI = 1.0,  7.7), underground miners (OR = 2.4, 95% CI
18      =1.2/4.4), construction painters (OR = 2.4,  95% CI = 0.6, 9.6), and welders (OR = 4.3,95% CI
19      =1.6,11.0). No excess risks were detected for the following industries: shipbuilding, petroleum
20      refining, construction, printing, blast furnace, and steel mills; or for the following occupations:
21      construction workers, painters, plumbers, paving equipment operators, roofers, engineers and
22      firemen, woodworkers, and shipyard workers.  Females were excluded from detailed analysis
23      because none of the Hispanic female controls had been employed in high-risk jobs; among the
24      non-Hispanic white controls, employment in a high-risk job was recorded for at least five
25      controls for only two industries, construction and painting, for which the odds ratios were riot
26      significantly elevated. Therefore, the analyses were  presented for males only.
27             Among the many strengths of this study are its population-based design, high
28      participation rate, detailed smoking history, and the separate analysis done for the two ethnic
29      groups, southwestern Hispanic and non-Hispanic white males.  The major limitations pertain to
30      the occupational exposure date.  Job titles obtained from occupational histories were used as
31      proxy for exposure status, but these were not validated. Further, for nearly half the cases, next of
32      kin provided occupational histories.  The authors acknowledge the above sources of bias but state
33      without substantiation that these biases would not strongly affect their results.  They also did not
34      use a job exposure matrix to link occupations to exposures and did not provide details on the
35      method they used to classify individuals as diesel exhaust exposed based on reported

        2/1/98                                     8-27       DRAFT-DO NOT CITE OR QUOTE

-------
 1      occupations.  The observed absence of an association for exposure to asbestos, a well-established
 2      lung carcinogen, may be explained by the misclassification errors in exposure status or by
 3      sample size constraints (not enough power). Likewise, the association for diesel exhaust
 4      reported by only 7 cases and  17 controls also may have gone undetected because of low power.
 5      In conclusion, there is insufficient evidence from this study to confirm or refute an association
 6      between lung cancer and diesel exhaust exposure.
 7
 8      8.3.5.  GarsMek et al. (1987): A Case-Control Study of Lung Cancer and Diesel Exhaust
 9            Exposure in Railroad Workers
10            An earlier pilot study of the mortality of railroad workers by the same investigators
11      (Schenker et al., 1984) found a moderately high risk of lung cancer among the workers who were
12      exposed to diesel exhaust as compared to those who were not. This study was designed to
13      evaluate the feasibility of conducting a large retrospective cohort study.  On the basis of these
14      findings the investigators conducted a case-control study of lung cancer in the same population.  .
15      The population base for this case-control study was approximately 650,000 active and retired
16      male U.S. railroad workers with 10 years or more of railroad service who were born in 1900 or
17      later. The U.S. Railroad Retirement Board (RRB), which operates the retirement system, is
18      separate from the Social Security System, and to qualify for the retirement or survivor benefits
19      the workers had to acquire 10 years or more of service. Information on deaths that occurred
20      between March 1,1981, and February 28,1982, was obtained from the RRB. For 75% of the
21      deceased population, death certificates were obtained from the RRB, and, for the remaining 25%,
22      they were obtained from the appropriate state departments of health. Cause of death was coded
23      according to the eighth revision of the ICD. The cases were selected from deaths with primary
24      lung cancer, which was the underlying cause of death in most cases. Each case was matched to
25      two deceased controls whose dates of birth were within 2.5 years of the date of birth of the case
26      and whose dates of death were within 31 days of the date of death noted in the case.  Controls
27      were then selected randomly from workers who did not have cancer noted anywhere on their
28      death certificates and who did not die of suicide or of accidental or unknown causes.
29            Each subject's work history was determined from a yearly job report filed by his
30      employer with the RRB from 1959 until death or retirement. The year 1959 was chosen as the
31      effective start of diesel exhaust exposure for this study, since by this time 95% of the
32      locomotives in the United States were diesel powered.  Investigators acknowledge that because
33      the transition to diesel-powered engines took place in the early 1950s, some workers had
34      additional exposure prior to 1959; however, if a worker had died or retired prior to 1959, he was
35      considered unexposed. Exposure to diesel exhaust was considered to be dichotomous for this

        2/1/98                                    8-28       DRAFT-DO NOT CITE OR QUOTE

-------
  1      study, which was assigned based on an industrial hygiene evaluation of jobs and work areas.
  2      Selected jobs with and without regular diesel exhaust exposure were identified by a review of job
  3      title and duties.  Personal exposure was assessed in 39 job categories representative of workers
  4      with and without diesel exhaust exposure. Those jobs for which no personal sampling was done
  5      were considered exposed or unexposed on the basis of similarities in job activities and work
  6      locations and by degree of contact with diesel equipment. Asbestos exposure was categorized on
  7      the basis of jobs held in 1959, or on the last job held if the subject retired before 1959, Asbestos
  8      exposure in railroads occurred primarily during the steam engine era and was related mostly to
  9      the repair of locomotive steam boilers that were insulated with asbestos.  Smoking history
10      information was obtained from the next of kin.
11             Death certificates were obtained for approximately 87% of the  15,059 deaths reported by
12      the RRB, from which 1,374 cases of lung cancer were identified. Fifty-five cases of lung cancer
13      were excluded from the study for either incomplete data (20) or refusal by two States to use
14      information on death certificates to contact the next of kin. Successful matching to at least one
15      control with work histories was  achieved for 335 (96%) cases <64 years of age at death and 921
16      (95%) cases >65 years of age at death. In both age groups, 90% of the cases were matched with
17      two controls. There were 2,385 controls in the study, 98% were matched within ±31 days of the
18      date of death, whereas the remaining 2% were matched within 100 days. Deaths from diseases
19      of the circulatory system predominated among controls. Among the younger workers,
20      approximately 60% had exposure to diesel exhaust, whereas among older workers, only 47%
21      were exposed to diesel exhaust.
22             Analysis by a regression model, in which years of diesel exhaust exposure were the sum
23      total of the number of years in diesel-exposed jobs, used as a continuous exposure variable,
24      yielded an odds ratio of lung cancer of 1.39 (95% CI = 1.05,1.83) for over 20 years of diesel
25      exhaust exposure in the ^64 years of age group. After adjustment for asbestos exposure and
26      lifetime smoking (pack-years), the odds ratio was 1.41 (95% CI = 1.06,1.88).  Both crude odds
27      ratio and asbestos exposure as well as lifetime smoking adjusted odds ratio for the >65 years of
28      age group were not significant.  Increasing years of diesel exhaust exposure, categorized as > 20
29      diesel years and 5 to 19 diesel years, with 0 to 4 years as the referent group, showed significantly
30      increased risk in the <64 years of age group after adjusting for asbestos exposure and pack-year
31      category of smoking. For individuals who had 2:20 years of diesel exhaust exposure, the odds,
32      ratio was 1.64 (95% CI = 1.18, 2.29), whereas among individuals who had 5 to 19 years of diesel
33      exhaust exposure, the odds ratio was 1.02 (95% CI = 0.72,1.45). In the >65 years of age group,
34     , only 3% of the workers were exposed to diesel exhaust for more than 20 years.  Relative odds for
        2/1/98                •                    8-29       DRAFT-DO NOT CITE OR QUOTE

-------
 1      5 to 19 years and >20 years of diesel exposure were less than 1 (p>0.01) after adjusting for
 2      smoking and asbestos exposure.
 3             Alternate models to explain post-asbestos exposure were tested. These were variables for
 4      regular and intermittent exposure groups and an estimate of years of exposure based on estimated
 5      years worked prior to 1959. No differences in results were seen.  The interactions between diesel
 6      exhaust exposure and the three pack-year categories (<50,  >50, and missing pack-years) were
 7      explored. The cross-product terms were not significant. A model was also tested that excluded
 8      recent diesel exhaust exposure occurring within the 5 years before death and gave an odds ratio
 9      of 1.43 (95% CI = 1.06, 1.94) adjusted for cigarette smoking and asbestos exposure, for workers
10      with 15 years of cumulative exposure. For workers with 5 to 14 years of cumulative exposure,
11      the relative odds were not significant.
12             The many strengths of the study are consideration of confounding factors such as
13      asbestos exposure and smoking; classification of diesel exhaust exposures by job titles and
14      industrial hygiene sampling; exploration of interactions between smoking, asbestos exposure,
15      and diesel exhaust exposure; and good ascertainment (87%) of death certificates from  the 15,059
16      deaths reported by the RRB.
17             The investigators also recognized and reported the  following limitations: overestimation
18      of cigarette consumption by surrogate respondents, which may have exaggerated the contribution
19      of smoking to lung cancer risk, and use of the Interstate Commerce Commission (ICC) job
20      classification as a surrogate for exposure, which may have led to misclassification of diesel
21      exhaust exposure jobs with low intensity and intermittent exposure, such as railroad police and
22      bus drivers, as unexposed.  These two limitations would result in the underestimation of the lung
23      cancer risk. This source of error could have been avoided  if diesel exhaust exposures were
24      categorized by a specific dose range associated with a job title that could have been classified as
25      heavy, medium, low, and zero exposure instead of a dichotomous variable. The use of death
26      certificates to identify cases and controls may have resulted in misclassification. Controls may
27      have had undiagnosed primary lung cancer, and lung cancer cases might have been secondary
28      lesions misdiagnosed as primary lung  cancer.  However, the investigators quote a third National
29      Cancer Survey report in which the death certificates for lung cancer were coded appropriately in
30      95% of the cases. Last, as in all previous studies, there is a lack of data on the contribution of
31      unknown occupational or environmental exposures and passive smoking.  In conclusion, this
32      study, compared with previous studies (on diesel exposure and lung cancer risk), provides the
33      most valid evidence that occupational diesel exhaust emission exposure increases the risk of lung
34    -  cancer.
35

        2/1/98               •                     8-30       DRAFT-DO NOT CITE OR QUOTE

-------
 1      8.3.6.  Benhamou et al. (1988): Occupational Risk Factors of Lung Cancer in a French
 2             Case-Control Study
 3             This is a case-control study of 1,625 histologically confirmed cases of lung cancer and
 4      3,091 matched controls, conducted in France between 1976 and 1980.  This study was part of an
 5      international study to investigate the role of smoking and lung cancer.  Each case was matched
 6      with one or two  controls whose diseases were not related to tobacco use, sex, age at diagnosis
 7      (±5 years), hospital of admission, or interviewer. Information was obtained from both cases and
 8      controls on place of residence since birth, educational level, smoking, and drinking habits.  A
 9      complete lifetime occupational history was obtained by asking participants to give their
10      occupations from the most recent to the first.  Women were excluded because most of them had
11      listed no occupation. Men who smoked cigars and pipes were excluded because there were very
12      few in this category.  Thus, the study was restricted to nonsmokers and cigarette smokers.
13      Cigarette smoking exposure was defined by age at the first cigarette (nonsmokers, <20 years, or
14      >20 years), daily consumption of cigarettes (nonsmokers, <20 cigarettes a day, and >20
15      cigarettes a day), and duration of cigarette smoking (nonsmokers, <35 years, and a 35 years).
16      The data on occupations were coded by a panel of experts according to their own chemical or
17      physical exposure determinations. Occupations were recorded blindly using the International
18      Standard Classification of Occupations. Data on 1,260 cases and 2,084 controls were available
19      for analysis.  The remaining 365 cases and 1,007 controls were excluded because they did not
20      satisfy the required smoking status criteria.
21             A matched logistic regression analysis was performed to estimate the effect of each
22      occupational exposure after adjusting for cigarette status. Matched relative risk (RR) ratios were
23      calculated for each occupation with the baseline category, which consisted of patients who had
24      never been engaged in that particular occupation. The matched relative risk ratios adjusted for
25     'cigarette smoking for the major groups of occupations showed that the risks were significantly
26      higher for production and related workers, transport equipment operators, and laborers (RR =
27      1.24,95% CI = 1.04,1.47). On further analysis of this group, for occupations with potential
28      diesel emission exposure, significant excess risks were found for motor vehicle  drivers (RR -
29      1.42, 95% CI = 1.07,1.89) and transport equipment operators (RR =1.35, 95%  CI = 1.05,1.75).
30      No interaction with smoking status was found in any of the occupations.  The only other
31      significant excess was observed for miners and quarrymen (RR = 2.14, 95% CI  = 1.07,4.31).
32      None of the significant associations showed a dose-response relationship with duration of
33      exposure.
34             This study was designed primarily t,o investigate the relationship between smoking (not
35      occupations or environmental exposures) and lung cancer. Although an attempt was made to

        2/1/98                                    8-31       DRAFT-DO NOT CITE OR QUOTE

-------
 1      obtain complete occupational histories, the authors did not clarify whether, in the logistic
 2      regression analysis, they used the subjects' first occupation, predominant occupation, last
 3      occupation, or ever worked in that occupation as the risk factor of interest.  The most important
 4      limitation of this study is that the occupations were not coded into exposures for different
 5      chemical and physical agents, thus precluding the calculation of relative risks for diesel
 6      exposure. Using occupations as surrogate measures of diesel exposure, an excess significant risk
 7      was obtained for motor vehicle drivers and transport equipment operators, but not for motor
 8      mechanics. However, it is not known if subjects in these occupations worked with diesel engines
 9      or nondiesel engines.
10
11      8.3.7.  Hayes et al. (1989): Lung Cancer in Motor Exhaust-Related Occupations
12             This study reports the findings from an analysis of pooled data from three lung cancer
13      case-control studies that examine in detail the association between employment in motor
14      exhaust-related (MER) occupations and lung cancer risk adjusted for confounding by smoking
15      and other risk factors.  The three studies were carried out by the National Cancer Institute in
16      Florida (1976 to 1979), New Jersey (1980 to 1981), and Louisiana (1979 to 1983). These three
17      studies were selected because the combined group would provide a sufficient sample to detect a
18      risk of lung cancer in excess of 50% among workers in MER occupations.  The analyses were
19      restricted to males who had given occupational history. The Florida study was hospital based,
20      with cases ascertained through death certificates.  Controls were randomly selected from hospital
21      records and death certificates, excluding psychiatric diseases, matched by age and county. The
22      New Jersey study was population based, with cases ascertained through hospital records, cancer
23      registry, and death certificates.  Controls were selected from among the pool of New Jersey
24      licensed drivers and death certificates.  The Louisiana study was  hospital based (it is not
25      specified how the cases were ascertained), and controls were randomly selected from hospital
26      patients,  excluding those with lung diseases and tobacco-related cancers.
27             A total of 2,291 cases of male lung cancers and 2,570 controls were eligible, and the data
28      on occupations were collected by next-of-kin interviews for all jobs held for 6 months or more,
29      including the  industry, occupation, and number of years employed. The proportion of next-of-
30      kin interviews varied by site between 50% in Louisiana to 85% in Florida.  The coding schemes
31      were reviewed to identify MER occupations, which included truck drivers and heavy equipment
32      operators (cranes,  bulldozers, and graders); bus drivers, taxi drivers, chauffeurs,  and other motor
33      vehicle drivers;  and automobile and truck mechanics. Truck drivers were classified as routemen
34      and delivery men and other truck drivers. AH jobs were also classified with respect to potential
35      exposure to known and suspected lung carcinogens.  Odds ratios were calculated by the

        2/1/98                                    8-32        DRAFT-DO NOT CITE OR QUOTE

-------
  1      maximum likelihood method adjusting for age by birth year, usual amount smoked, and study
  2      area. Logistic regression models were used to examine the interrelationship of multiple
  3      variables.
  4             A statistically significant excess risk was detected for employment of 10 years or more
  5      for all MER occupations (except truck drivers) adjusted for birth cohort, usual daily cigarette use,
  6      and study area. The odds ratio for lung  cancer using data gathered by direct interviews  was 1.4
  7      (95% CI - 1.1, 2.0), allowing for multiple MER employment, and 2.0 (95% CI - 1.3, 3.0),
  8      excluding individuals with multiple MER employment. Odds ratios for all MER employment,
  9      except truck drivers who were employed for less than 10 years, were 1.3 (95% CI = 1.0,1.7) and
10      1.3 (95% CI = 0.9,1.8) including and excluding multiple MER employment, respectively. Odds
11      ratios were then derived for specific MER occupations and, to avoid the confounding effects of
12      multiple MER job classifications, analyses were also done excluding subjects with multiple
13      MER job exposures. Track drivers employed for more than 10 years had an odds ratio of 1.5
14      (95% CI = 1.1, 1.9). A similar figure was obtained excluding subjects with multiple MER
15      employment. An excess risk was not detected for truck drivers employed less than 10 years.
16      The only other job category that showed a statistically  significant excess for lung cancer was the
17      one that included taxi drivers and chauffeurs who worked multiple MER jobs for less than 10
18      years (OR — 2.5, 95% CI =1.4,4.8). For the same category, the risk for individuals working in
19      that job for more than 10 years was 1.2 (95% CI = 0.5,2.6). A statistical significant positive
20      trend (p<0.05) with increasing employment of <2 years, 2 to 9 years, 10 to 19 years, and 20+
21      years was observed for track drivers but not for other MER occupations.  A statistically
22      nonsignificant excess risk was also observed for heavy equipment operators, bus drivers, taxi
23      drivers and chauffeurs, and mechanics employed for 10 years or more. All of the above-
24      mentioned odds ratios were derived adjusted for birth cohort, usual daily cigarette use, and State
25      of residence. Exposure to other occupational suspect lung carcinogens did not account for the
26      excess risks detected.
27             Results of this large study provide evidence that workers in MER jobs are at an excess
28      risk of lung cancer that is not explained  by their smoking habits or exposures to other lung
29      cancers. Because no information on type of engine had been collected, it was not possible to
30      determine if the excess risk was due to exposure to diesel exhaust or gasoline exhaust or the
31      mixture of the two.  Among the study's  limitations are  possible bias due to misclassiflcation of
32      jobs reported by the large proportion of next-of-kin interviews and the problems in classifying
33      individuals into uniform occupational groups based on the pooled data in the three studies, that
34      used different occupational classification schemes.
35

        2/1/98                                    8-33       DRAFT-DO NOT CITE OR QUOTE

-------
 1      8.3.8. Steenlandetal. (1990): A Case-Control Study of Lung Cancer and
 2            Truck Driving in the Teamsters Union
 3            Steenland et al. conducted a case-control study of lung cancer deaths in the Teamsters
 4      Union to determine the risk of lung cancer among different occupations. Death certificates were
 5      obtained from the Teamsters Union files in the central States for 10,485 (98%) male decedents
 6      who had filed claims for pension benefits and who had died in 1982 and 1983. Individuals were
 7      required to have 20 years tenure in the union to be eligible to claim benefits.  Cases comprised all
 8      deaths (n = 1,288) from lung cancer, coded as ICD 162 or 163 for underlying or contributory
 9      cause on the death certificate. The 1,452 controls comprised every sixth death from the entire
10      file, excluding deaths from lung cancer, bladder cancer, and motor vehicle accidents. Detailed
11      information on work history and potential confounders such as smoking, diet, and asbestos
12      exposure was obtained by questionnaire. Seventy-six percent of the interviews were provided by
13      spouses and the remainder by some other next of kin.  The response rate was  82% for cases and
14      80% for controls.  Using these interview data and the 1980 census occupation and industry
15      codes, subjects were classified either as nonexposed or as having held other jobs with potential
16      diesel exhaust exposure. Data on job categories were missing for 12% of the study subjects.  A
17      second work history file was also created based on the Teamsters Union pension application that
18      lists occupation, employer, and dates of employment.  A three-digit U.S. census code for
19      occupation and industry was assigned to each job for each individual. This Teamsters Union
20      work history file did not have information on whether men drove diesel or gasoline trucks, and
21      the four principal occupations were long-haul drivers,  short-haul or city drivers, truck mechanics,
22      and dock workers. Subjects were assigned the job category in which they had worked the
23      longest.
24            The case-control analysis was done using unconditional logistic regression. Separate
25      analyses were conducted for work histories from the Teamsters Union pension file and from
26      next-of-kin interviews.  Covariate data were obtained from next-of-kin interviews. Analyses
27      were also performed for two time periods: employment after 1959 and employment after 1964.
28      These two cut-off years reflect years of presumed dieselization; 1960 for most trucking
29      companies and 1965 for independent driver and nontrucking firms. Data for analysis could be
30      obtained for 994 cases and 1,085 controls using Teamsters Union work history and for 872 cases
31      and 957 controls using next-of-kin work history. When exposure was considered as a
32      dichotomous variable, for both Teamsters Union and next-of-kin work history, no single job
33      category had an elevated risk. From the next-of-kin data, diesel truck drivers had an odds ratio of
34      1.42  (95% CI = 0.74, 2.47) and diesel truck mechanics had an odds ratio of 1.35  (95% CI = 0.74,
35      2.47). Odds ratios by duration of employment as a categorical variable were  then estimated.  For

        2/1/98                                    8-34       DRAFT-DO NOT CITE OR QUOTE

-------
  1      the Teamsters Union work history data and when only employment after 1959 was considered,
  2      both long-haul (p<0.04) and short-haul drivers (not significant) showed an increase in risk with
  3      increased years of exposure. The length of employment categories for which the trends were
  4      analyzed were I to 11 years, 12 to 17 years, and 18 years or more. Using 1964 as the cutoff date,
  5      long-haul drivers continued to show a significant positive trend (p=0.04), with an odds ratio of
  6      1.64 (95% CI = 1.05, 2.57) for those who worked for 13+ years, the highest category. Short-haul
  7      drivers, however, did not show a positive trend when 1964 was used as the cutoff date. Similar
  8      trend analysis was done for most next-of-kin data. A marginal increase in risk with increasing
  9      duration of employment as a truck driver (p=0.12) was observed. For truck drivers who
10      primarily drove diesel trucks for 35 years or longer, the odds ratio for lung cancer was 1.89 (95%
11      CI = 1,04, 3.42). The odds ratio was 1.34 (95% CI = 0.81,2.22) for gasoline truck  drivers and
12      1.09 (95% CI = 0.44, 2.66) for truck mechanics.  No significant interactions between age and
13      diesel exhaust exposure or smoking and diesel exhaust exposure were observed.  All the odds
14      ratios were adjusted for age, smoking, and asbestos in addition to various exposure  categories.
15             The authors acknowledge several limitations of this study, which include possible
16      misclassifications of exposure and smoking habits, as information was provided by next of kin;
17      lack of sufficient latency to observe lung cancer excess; and a small nonexposed group (n = 120).
18      Also, concordance between Teamsters Union and next-of-kin job categories could not be easily
19      evaluated because job categories were defined differently in each data set. No  data were
20      available on levels of diesel exposure for the different job categories.  Given these limitations,
21      the positive findings of this study  are probably underestimated.
22
23      8.3.9.  Boffetta et al. (1990):  Case-Control Study on Occupational Exposure to
24             Diesel Exhaust and Lung Cancer Risk
25             This is an ongoing (since 1969) case-control study of tobacco-related diseases in 18
26      hospitals (six U.S. cities).  Cases comprise 2,584 males with histologically confirmed primary
27      lung  cancers.  Sixty-nine cases were matched to one control, whereas 2,515 were matched to two
28      controls. Controls were individuals who were diagnosed with non-tobacco-related diseases.  The
29      matching was done for sex, age (±2 years), hospital, and year of interview. The interviews were
30      conducted at the hospitals at the time of diagnosis. In 1985, the occupational section of the
31      questionnaire was modified to include the usual occupation and up to five other jobs as well as
32      duration (in years) worked in those jobs. After 1985, information was also obtained on exposure
33      to 45 groups of chemicals, including diesel exhaust at the workplace or during hobby activities.
34      A priori aggregation of occupations was categorized into low probability of diesel exhaust
35      exposure (reference group), possible exposure (19 occupations), and probable exposure (13

        2/1/98        /    .                       8-35      DRAFT-DO NOT  CITE OR QUOTE

-------
 1      occupations). Analysis was conducted based on "usual occupation" on all study subjects, and
 2      any occupation with sufficient cases was eligible for further analysis. In addition, cases enrolled
 3  .    after 1985 for which there were self-reported diesel exhaust exposure and detailed work histories
 4      were also analyzed separately.       •   .
 5             Both matched and unmatched analyses were done by calculating the adjusted (for
 6      smoking and education) relative odds using the Mantel-Haenzael method and calculating the test-
 7      based 95% confidence interval using the Miettinen method. Unconditional logistic regression
 8      was used to adjust for potential confounders (the PROC LOGIST of SAS). Linear trends for risk
 9      were also tested according to Mantel.
10             Adjusted relative odds for possible and probable exposure groups as well as the truck
11      drivers were slightly below unity, none being statistically significant for the entire study
12      population. Although slight excesses were observed for the self-reported diesel exhaust exposure
13      group and the subset of post-1985 enrollees for highest duration of exposure (for self-reported
14      exposure, occupations with probable exposure and for truck drivers), none was statistically
15      significant. Trend tests for the risk of lung cancer among self-reported diesel exhaust exposure,
16      probable exposure, and truck drivers with increasing exposure (duration of exposure used as
17      surrogate for increasing dose) were nonsignificant too. Statistically significant lung cancer
18      excesses were observed for cigarette smoking only.
19             The major strength of this study is availability of detailed smoking history. Even though
20      detailed information was obtained for the usual and five other occupations (1985), no effort was
21      made to estimate or verify the actual exposure to diesel exhaust; instead, duration of employment
22      was used as a surrogate for dose.  The numbers of cases and controls were large; however, the
23      number of individuals exposed to diesel exhaust was relatively few, thus reducing the power of
24      the study. This study did not attempt latency analysis either. Given these limitations, the
25      findings of this study are unable to provide either positive or negative evidence for a  causal
26      association between diesel exhaust and occurrence of lung cancer.
27                     .
28    .  8.3.10. Emmelin et al. (1993): Diesel Exhaust Exposure and Smoking: A Case-Referent
29             Study of Lung Cancer Among Swedish Dock Workers
30             This is a case-control study of lung cancer drawn from the cohort defined as all-male
31      workers who had been employed as dock workers for at least 6 months between 1950 and 1974.
32      In the population of 6,573 from 20 ports, there were 90 lung cancer deaths (cases), identified
33      through Swedish death and cancer registers, during the period of 1960 to 1982. Of these 90
34      deaths, the 54 who were workers at the 15  ports for which exposure surrogate information was
35      available were chosen for the case-control  study. Four controls, matched on port and age, were

        2/1/98   .                                  8-36        DRAFT-DO NOT CITE OR QUOTE

-------
  1      chosen for each case from the remaining cohort who had survived to the time of diagnosis of the
  2      case. Both live and deceased controls were included. The final analyses were done on 50 cases
  3      and 154 controls who had complete information on employment dates and smoking data. The
  4      smoking strata were created by classifying ex-smokers as nonsmokers if they had not smoked for
  5      at least 5 years prior to the date of diagnosis of the case; otherwise they were classified as
  6      smokers.
  7             Relative odds and regression coefficients were calculated using conditional logistic
  8      regression models.  Comparisons were made both with and without smoking included as a
  9      variable, and the possible interaction between smoking and diesel exhaust was tested.  Both
10      weighted linear regressions of the adjusted relative odds, and the regression coefficients were
11      used to test mortality trends with all  three exposure variables.
12             Exposure to diesel exhaust was assessed indirectly by initially measuring (1) exposure
13      intensity based on exhaust emission, (2) characteristics of the environment in terms of
14      ventilation, and (3) measures of proportion of time in higher exposed jobs. For exhaust
15      emissions, annual diesel fuel consumption at a port was used as the surrogate.  For ventilation,
16      the annual proportion of ships with closed or semiclosed holds was used as the surrogate. The
17      proportion of time spent below decks was used as the surrogate for more exposed jobs. Although
18      data were collected for all three measures, only the annual fuel consumption was used for
19      analysis. Because every man was likely to rotate through the various jobs, the authors thought
20      using annual consumption of diesel fuel was the appropriate measure of exposure.
21      Consequently, in a second analysis, the annual fuel consumption was divided by the number of
22      employees in the same port that year to come up with the fuel-per-person measure, which was
23      further  used to create a second measure, "exposed time." The "annual fuel" and exposed-time
24      data were entered in a calendar time-exposure matrix for each port, from which individual
25      exposure measures were created.  A third measure, "machine time" (years of employment from
26      first exposure)', was also used to compare the results with other studies.  All exposure  measures
27      were accumulated from the first year of employment or first year of diesel machine use,
28      whichever came later.  The last year of exposure was fixed at 1979. All exposures within 2 years
29      prior to the date of lung cancer diagnosis were omitted from both cases and matched controls. A
30      priori classification into three categories of low, medium, and high exposure was done for all
31      three exposure variables, machine time, fuel, and exposed time.
32             Conditional logistic regression models, adjusting for smoking status and using low
33      exposures and/or nonsmokers as a comparison group, yielded positive trends for all exposure
34      measures, but no trend test results were reported, and only the relative odds for the exposed-time
35      exposure measure in the high-exposure group (OR = 6.8, 90% CI = 1.3 to 34.9) was reported as

        2/1/98                                    8-37       DRAFT-DO NOT CITE OR QUOTE

-------
 1      statistically significant. For smokers, adjusting for diesel exhaust exposure level, the relative
 2      odds were statistically significant and about equal for all the three exposure variables—machine
 3      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,
 4      OR = 6.2 (90% CI = 2.6 to 14.6). Interaction between diesel exhaust and smoking was tested by
 5      conditional logistic regression in the exposed-time variable. Although there were positive trends
 6      for both smokers and nonsmokers, the trend for smokers was much steeper—low, OR = 3.7 (90%
 7      CI = 0.9 to 14.6); medium, OR =10.7 (90% CI = 1.5 to 78.4); and high, OR = 28.9 (90% CI =
 8      3.5 to 240)—indicating more than additive interaction between these two variables.
 9             In the weighted linear regression model with the exposed-time variable, the results were
10      similar to those using the logistic regression model. The authors also explored the smoking
11      variable further in various analyses, some of which suggested a strong interaction between diesel
12      exhaust and smoking. However, with just six nonsmokers and no further categorization of
13      smoking amount or duration, these results are of limited value.
14             The diesel exhaust exposure matrices created using three different variables are intricate.
15      Analyses by  any of these  variables essentially yield the same positive results and positive trends,
16      providing consistent support for a real effect of diesel exhaust exposure, at least in smokers.
17      However, there are some  methodological limitations to this study that prevent a more definitive
18      conclusion. The numbers of cases and controls are small.  There are very few nonsmokers, thus
19      testing the effects of diesel exhaust exposure in them is futile. Lack of information on asbestos
20      exposure, to which dock workers are usually exposed, may also confound the results. Also, no
21      latency analyses are presented. Overall, despite these limitations, this study supports the earlier
22      findings of excess lung cancer mortality among individuals exposed to diesel exhaust.
23             Table 8-2 summarizes the above lung cancer case-control studies.
24
25      8.4. CASE-CONTROL STUDIES OF BLADDER CANCER
26      8.4.1.   Howe et al. (1980): Tobacco Use, Occupation, Coffee, Various Nutrients,
27             and Bladder Cancer
28             This is a Canadian population-based case-control study conducted in the provinces of
29      British Columbia, Newfoundland, and Nova Scotia.  These areas were selected because they had
30      cancer registries and were believed not to have concentrations of high-risk industries.  All
31    .  patients with newly diagnosed bladder cancer occurring in the three provinces between April
32      1974 and June  1976 were identified, and 77% of them were interviewed at home. A total of 480
33      male and 152 female case-control pairs were available for analysis.  For each case, one
34      neighborhood control, matched by age (±5 years) and sex, was also interviewed at home to
35      obtain data on smoking, occupation, dietary sources of nitrites and nitrates that convert to

        2/1/98                                     8-38       DRAFT-DO NOT CITE OR QUOTE

-------
SO
OG
                           Table 8-2. Epidemiologic studies of the health effects of exposure to diesel exhaust:
                           case-control studies of lung cancer
         Authors
                  Population studied
Diesel exhaust exposure
      assessment
            Results
Limitations
        Williams    7,518 (3,539 males and
        et al. (1977) 3,979 females) incident invasive
                   cancers from the Third National
                   Cancer Survey

                   Lung cancer cases:
                   32 in males
                   28 in females

                   Combined other cancer sites
                   were used as controls
                                           Main lifetime, recent,
                                           and other employment
                                           information obtained at
                                           the time of survey

                                           1970 Census Coding
                                           Scheme for Employment
                                           was used to code the
                                           occupations by one of
                                           the authors
                       SNS elevated relative odds were
                       observed among occupations of
                       trucking, railroading, and mining
                                 Exposure estimation based on self-
                                 report that was not validated

                                 47% nonresponse

                                 Control group consisted of other
                                 cancers, probably diluting the risk
                                 estimation

                                 Small numbers in cause-specific
                                 cancers and individual occupations
oo
•O
o
2
o
H
O
M
H
M
O
*3
XD
a
o
H
W
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 Complete lifetime employment
smoking for lung cancer:           history not available
RR= 1.4 (1st criteria)
and                              Self-reported occupation history not
RR = 1.7 (NIOSH criteria)          validated

                                 No analysis by dose, latency, or
                                 duration of exposure

                                 No information on nonoccupatibnal
                                 diesel exposure

-------
SO
oo
                           Table 8-2.  Epidemiologic studies of the health effects of exposure to diesel exhaust:
                           case-control studies of lung cancer (continued)
         Authors
                  Population studied
                                                  Diesel exhaust exposure-
                                                        assessment
Results
Limitations
        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
                                           least 1 year or more

                                           Using a 5-digit code the
                                           occupations were
                                           classified according to
                                           Nordic Classification of
                                           Occupations
                                                                         SS OR = 2.7 (a 1 year
                                                                         of employment)

                                                                         SS OR = 9.8 (;>20 years of
                                                                         employment)
                    Uncertain diesel exhaust exposure

                    No validation of exposure done

                    Underground miners data not
                    adjusted for other confounders such
                                                                         Adjustment for smoking was done  as radon, etc.

                                                                         SNS OR = 1.2 for professional
                                                                         drivers (>20 years of employment)
                                                                         with dead controls

                                                                         SNS OR = 1.1 ^20 years of
                                                                         employment) with living controls
oo
w
2
O
o
•2,
o
H
O
I****
H
W
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 was
                                                                         observed for diesel exhaust
                                                                         exposure
                    Exposure based on occupational
                    history and self-report, which was not
                    validated

                    50% occupational history provided by
                    next of kin

                    Absence of lung cancer association
                    with asbestos suggests
                    misclassification of exposure
o
H
W

-------
                           Table 8-2. Epidemiologic studies of the health effects of exposure to diesel exhaust:
                           case-control studies of lung cancer (continued)
00
         Authors
                  Population studied
 Diesel exhaust exposure
       assessment
            Results
           Limitations
        Garshick     1,319 lung cancer cases who died
        etal. (1987)  between March 1, 1981,
                    and February 28, 1982

                    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
                                           Personal exposure
                                           assessed for 39 job
                                           categories

                                           This was corrected with
                                           job titles to dichotomize
                                           the exposure into:
                                           Exposed
                                           Not exposed
                        SS OR = 1.41 (<64 year age group)

                        SS OR = 1,64 (^ 64 year age group)
                        for a20 years diesel exhaust
                        exposure group when compared to
                        0- to 4-year exposure group

                        All ORs adjusted for lifetime
                        smoking and asbestos exposure
                                 Probable misclassification of diesel
                                 exhaust exposure jobs

                                 Years of exposure used as surrogate
                                 for dose

                                 13% of death certificates not
                                 ascertained

                                 Overestimation of smoking history
oo
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) histories not validated
and transport equipment operators
(RR = 1.35) (smoking adjusted)     Exposures classified as chemical and
                                 physical exposure, not specific to
                                 diesel exhaust
O
O
z
O
H
O
^••l
H
m
O
JO
/O
d
o
H
W
Hayes et al.  Pooled data from three different
(1989)      studies consisting of 2,291 male
            lung cancer cases

            2,570 controls
Occupational information SS OR =1.5 for truck drivers (>10   Exposure data based on job
 from next of kin for all
jobs held

 Jobs classified with
 respect to potential
 exposure to known and
 suspected pulmonary
 carcinogens
years of employment)

SS positive trend with increasing
employment as truck driver
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

-------
                           Table 8-2.  Epidemiologic studies of the health effects of exposure to diesel exhaust:
                           case-control studies of lung cancer (continued)
oo
          Authors
                   Population studied
 Diesel exhaust exposure
       assessment
Results
Limitations
        Steenland   1,058 male lung cancer deaths
        et al. (1990) between 1982 and 1983
                                            Longest job held:  diesel  As 1964 cut-off point:
                                            truck driver, gasoline
                                            truck driver, both types
             1,160 every sixth death from entire of trucks, truck
             mortality file sorted by social     mechanic, and dock
             security number (excluding lung   workers
             cancer, bladder cancer, and motor
             vehicle accidents)
                                                                           SS OR = 1.64 for long-haul drivers
                                                                           with 13+ years of employment

                                                                           Positive trend test for long-haul
                                                                           drivers (p=0.04)
                                                         Exposure based on job titles not
                                                         validated

                                                         Possible misclassification of exposure
                                                         and smoking, based on next-of-kin
                                                         information

                                                         Lack of sufficient latency
                    Cases and controls were from
                    Central State Teamsters who
                    had filed claims (requiring 20-year
                    tenure).
                                                                    SS OR = 1.89 for diesel truck
                                                                    drivers of 35+ years of employment
oo
k
 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 diesel
 exhaust
                                                                           OR slightly below unity SNS
                    No verification of exposure

                    Duration of employment used as
                    surrogate for dose

                    Number of individuals exposed to
                    diesel exhaust was small
O
O
2
O
H
O
3
w
§
O
C3
O
H
W
 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 diesel exhaust    SS OR for high-exposure group =
 exposure assessment done 6.8
 based on (1) exposure
 intensity, (2)
 characteristics of
 ventilation, (3) measure
 of proportion of time in
 higher exposure jobs
                    Numbers of cases and controls are
                    small

                    Very few nonsmokers

                    Lack of exposure information on
                    asbestos

                    No latency analysis
Abbreviations: OR = odds ratio; RR = relative risk; SNS = statistically nonsignificant; SS = statistically significant.

-------
  1      nitrosamines (nonpublic water supply and preserved meat products), and beverage consumption,
  2      including a detailed history of coffee consumption. A detailed smoking history was obtained.
  3      The occupational history included a chronological account of all jobs and the number of years
  4      and months during which the respondent had worked in each job, experience in industries that
  5      were suspected a priori to increase the risk of bladder cancer, and exposure to any jobs that
  6      involved exposure to dust and fumes at the workplace. Relative risk estimates  were computed
  7      using the linear logistic model applied to individually matched case-control pairs.
  8             A baseline comparison of cases and controls showed that, whereas male patients were
  9      similar to controls on income and education, there was an excess of female cases  with low family
10      incomes and low levels of educational attainment. For both sexes, the mean ages for cases and
11      controls did not differ, and the times required for the interview were similar. An analysis by the
12      a priori suspect industries showed elevated risks for a number of industries for males. These
13      included the chemical (RR = 7.5, 95% CI = 1.7,67.6), rubber (RR = 5.0, 95% CI  = 0.6, 236.5),
14      petroleum (RR - 5.3,95% CI - 1.5, 28.6), medicine (RR = 2.6, 95% CI - 0.9, 9.3), and spray
15      painting (RR = 1.8, 95% CI = 0.7,4.6) industries. The excess risks were statistically significant
16      only for the petroleum and chemical industries. The estimates did not change when the analysis
17      was done separately for subjects who reported only one exposure and for those  who reported
18      exposure to more than one suspect industry.  The estimates also remained unchanged after
19      controlling for smoking.  Too few females reported working in the a priori suspect industries to
20      make any meaningful contribution to the analysis. Among males, statistically nonsignificant
21      excess risks were observed for tanning, electric cable, photographic, commercial paint, tailoring,
22      medicine, food processing, and agricultural industries. The analysis by exposure to dust and
23      fumes in occupations other than those in the a priori suspect list detected the relative risks for
24      diesel and traffic fumes (RR = 2.8,95% CI - 0.8,11.8).  Statistically significant excess risks
25      were observed for railroad workers (RR = 9.0, 95% CI = 1.2, 394.5) and welders (RR = 2.8, 95%
26      CI = 1.1, 8.8). For occupations other than those on the a priori list for males and females,
27      statistically significant excesses were detected for metal machinists (RR = 2.7, 95% CI = 1.1,
28      7.6), metal recorders (RR = 2.6,95% CI = 1.0, 7.3), and nursery men (RR = 5.5, 95% CI = 1.2,
29      51.1). Statistically nonsignificant excesses were also detected for exposure to two chemicals:
30      benzidine and its salts, RR = 1.3, and fe-ehloromethyl ether, RR = 5.0.  A detailed analysis was
31      done for cigarette smoking, which demonstrated statistically significant increasing bladder
32      cancer risk with increasing duration of smoking, total lifetime consumption of packs of
33      cigarettes, and average frequency of cigarettes per day. In males the highest significant risk was
34      observed for latency of less than 35 years; after that time the risk reduced slightly with increasing
35      latency. In females the highest significant risk was for more than 35 years of latency. Risks

        2/1/98                                    8-43        DRAFT-DO NOT CITE OR QUOTE

-------
 1      were elevated for males consuming all types of coffee and for females consuming instant coffee.
 2      Hair dye usage in females and phenacetin usage in males and females carried no risk. Significant
 3      risks for use of artificial sweeteners and use of nonpublic water supplies (nitrates and nitrites)
 4      were found among males only.
 5            This study was mainly designed to evaluate the various risk factors for bladder cancer
 6      such as smoking, coffee consumption, nitrates and nitrites in diet, etc. The major limitation of
 7      this study, as the authors noted, was that the three selected provinces did not have high
 8      concentrations of industries suspected to be linked to bladder cancer. An excess risk was,
 9      however, detected for railroad workers and for those in the "exposed to diesel and traffic fumes
10      category." Risks for those exposed to "diesel fumes only" were not available, nor do we know
11      the exact job title of the railroad workers and the type of engines they were operating. The
12      authors also did not detail the method by which they coded the information given by respondents
13      in response to questions on exposure to dust and fumes into the various categories they used in
14      the analysis. These analyses were done for subjects who reported having "ever been exposed"
15      versus "never been exposed" to these fumes, and although detailed chronological work histories
16      were obtained, no attempt was made to develop a lifetime cumulative exposure index to diesel
17      fumes.  In multiple logistic regression models, the authors used the a priori high-risk
18      occupations; hence, nothing can be concluded about exposure to diesel exhaust for occupations
19      that were not part of that list.  The authors provided no explanation on possible selection bias, as
20      only 77% of the eligible population was included in the study.
21
22      8.4.2. Wynder et al. (1985): A Case-Control Study of Diesel Exhaust Exposure and
23            Bladder Cancer
24            A case-control study of diesel exhaust exposure and bladder cancer risk was conducted by
25      Wynder etal. (1985). Cases and controls were obtained from 18 hospitals located in six U.S.
26      cities between January 1981 and May 1983. Cases were individuals with histologically
27      confirmed primary cancer of the bladder, diagnosed within 12 months prior to the interview.
28      Controls were individuals with non-tobacco-related diseases who were matched to the cases by
29      age (within 8 years), race, year of interview, and hospital of admission.  Women were excluded
30      from the study because the focus was on male-dominated occupations.  A structured
31      questionnaire was administered in the hospital to cases and controls to elicit  information on usual
32      occupation, smoking history,  alcohol and coffee consumption, as well as other demographic
33      factors.
34            Two methods were used to define occupational exposure to diesel exhaust. First,
35      occupational titles defined by the industrial hygiene standards as probable high exposure were

        2/1/98                                    8-44       DRAFT-DO NOT CITE OR QUOTE

-------
  1      classified as exposed or not exposed to diesel exhaust. The probable high-exposure category
  2      consisted of bus and truck drivers, heavy equipment operators and repairmen, railroad workers,
  3      and warehousemen. In the second method, guidelines set by NIOSH were used to classify
  4      occupations based on exposure to diesel exhaust. In this method, the estimated proportion of
  5      exposed workers was computed for each occupational category by using the NIOSH estimates of
  6      the exposed population as the numerator and the estimates of individuals employed in each
  7      occupational category from the 1970 census as the denominator. Occupations estimated to have
  8      at least 20% of their employees exposed to diesel exhaust were defined as "high exposure," those
  9      with 10% to 19% of their employees exposed as "moderate exposure," and those with less than
10      10% of their employees exposed as "low exposure." The odds ratio was used as a measure of
11      association to assess the relationship between bladder cancer and diesel exhaust exposure. The
12      overall participation among those eligible and available for interview was 75% and 72% in cases
13      and controls, respectively.
14             A total of 194 bladder cancer cases and 582 controls were examined, and the two groups
15      were found to be comparable by age and education.  Except for railroad workers, who had
16      relative odds of 2.0 based on two cases and three controls (95% CI = 0.34,11,61), the relative
17      odds were less than 1 for other diesel exhaust exposure occupations such as bus and track
18      drivers, warehousemen, material  handlers, and heavy equipment workers. When the risk was
19      examined using the NIOSH criteria for high, moderate, and low exposure, relative odds were
20      1.68 and 0.16 for high and moderate, respectively, with low as the referent group; neither was
21      statistically significant.  Cases and controls were compared by smoking status. Cases were more
22      likely to be current cigarette smokers than were controls. Current smokers of 1 to 20
23      cigarettes/day had relative odds of 3.64 (95% CI = 2.04, 6.49), current smokers of 21+
24      cigarettes/day had relative odds of 3.51 (95% CI = 2.00,6.19), while ex-smokers had relative
25      odds of 1.72 (95% CI = 1.01, 2.92). After controlling for smoking, there was no significant
26      increase in the risk of bladder cancer for occupations with diesel exhaust exposure compared to
27      occupations without diesel exhaust exposure. A synergistic effect between the two also was not
28     . detected.
29             This study has two major methodologic limitations, both pertaining to exposure
30      classification. First, the use of "usual" occupation may have led to misclassification of those
31      individuals who had held a previous job with diesel exhaust exposure that was not their usual
32      occupation; this may have resulted in reduced power to detect weak associations.  Second, since
33      there was no information on amount or duration of diesel exhaust exposure, no analysis of dose-
34      response relationships could be done. Also, no information was available on other confounding
35      risk factors of bladder cancer such as urinary retention, amphetamine abuse, and smoking within

        2/1/98                                    8-45       DRAFT-DO  NOT CITE OR QUOTE

-------
  1      the confined space of a truck cab, all of which are lifestyle factors specific to the truck-driving
  2      occupation.
  3              .                                                          .
  4      8.4.3.  Hoar and Hoover (1985): Truck Driving and Bladder Cancer Mortality in
  5             Rural New England
  6             This study investigated the relationship between the occupation of truck driving and
  7      bladder cancer mortality in a case-control study in New Hampshire and Vermont. Cases
  8      included all white residents of New Hampshire and Vermont who died from bladder cancer
  9      (eighth revision of the ICD) between 1975 and 1979. Death certificates were provided by the
10      vital records and health statistics office of the two States, and the next of kin were traced and
11      interviewed in person. Two types of controls were selected for each case. One control was
12      randomly selected from all other deaths, excluding suicides, and matched on State, sex, race, age
13      (±2 years), and year of death. The second for control was selected with the additional matching
14      criteria of county of residence.  Completed interviews were obtained from 325 (out of 410) next
15      of kin for cases and 673 (out of 923) for controls. Information on demographic characteristics,
16      lifetime occupational and residential histories, tobacco  use, diet, and medical history was
17      obtained on each subject.  The odds ratio was calculated to ascertain a measure of association
18      between truck driving and bladder cancer. Because separate analyses of the two control series
19      gave similar results, the two control series were combined.  Also, because matched analyses
•20      yielded results similar to those provided by the unmatched analyses, results of the unmatched
21      analyses were presented.
22             Sixteen percent (35) of the cases and  12% (53) of the controls had been employed as
23      truck drivers, yielding an odds ratio of 1.5 (95% CI = 0.9, 2.6) after adjustment for county of
24      residence and age at death. For New Hampshire, the odds ratio was 1.3  (95% CI = 0.7, 2.3), and
25      for Vermont, the odds ratio was 1.7 (95% CI = 0.8, 3.4).  For a large number of subjects, the next
26      of kin were unable to give the durations of truck driving,  and there was an inconsistent positive
27      association with years of truck driving.  Crude relative  odds were not altered after adjustment for
28      coffee drinking, cigarette  smoking, and education as a surrogate for social class.  Little variation
29      in risks was seen when the data were analyzed by the industry in which the men had driven
30      trucks. No relationship was seen between age at which employment as a truck driver started and
31      occurrence of bladder cancer. Analysis by duration of employment as a truck driver and bladder
32      cancer showed a positive trend of increasing  relative odds with increasing duration of
33      employment. The trend test was statistically significant (p=0.006). The odds ratio was
34      statistically significant for the 5 to 9 years of employment category only (OR = 2.9, 95% CI =
35      1.2, 6.7). Similarly, analysis by calendar year first employed showed a statistically significant

        2/1/98                                    8-46       DRAFT-DO NOT CITE OR QUOTE

-------
  1      odds ratio for 1930 to 1949 (OR = 2.6, 95% CI = 1,3, 5.1), whereas relative odds were not
  2      significant if subjects were employed prior to 1929 or after 1950.
  3             The effects of reported diesel exhaust exposure from fuel or engines in truck driving or
  4      other occupations were then analyzed. An odds ratio of 1.8 (95%  CI = 0.5, 7.0) was derived for
  5      those who were exposed to diesel exhaust during their track-driving jobs as compared to an odds
  Q      ratio of 1.5 (95% CI = 0.8,2.7) for those not reporting diesel exhaust exposure. Analysis by
  7      duration of exposure (0,1 to 19 years, 20 to 29 years, 30 to 39 years, and 40+ years) to diesel
  8      fuel or engines in other occupations, which were "self-reported" by participants, showed a
  9      statistically significant positive trend (p=0.024) for bladder cancer, although none of the
10      individual odds ratios in these duration categories were statistically significant.
11             This study investigated an association between truck driving and bladder cancer in an
12      attempt to understand the reasons for the high rates of bladder cancer in rural areas of New
13      Hampshire and Vermont. Although an elevated odds ratio for bladder cancer (not statistically
14      significant) was observed for reported truck-driving occupations, there was insufficient evidence
15      to conclude that the excessjisk of bladder cancer was due to exposure  to diesel emissions. This
16      is because the excess bladder cancer risk was observed for all truck drivers irrespective of their
17      exposure to diesel emissions. Also, no information was provided on the confounding effects of
18      other aspects of the road environment such as urinary retention, amphetamine abuse, arid
19      concentrated cigarette smoke within the track cab.  Other limitations of this study include the use
20      of next of kin for occupational histories, who may either under- or overreport occupations, and
21      the use of death certificates with their inherent problems of misclassification.
22
23      8.4.4.  Steenland et al. (1987): A Case-Control Study of Bladder Cancer Using City
24             Directories as a Source of Occupational Data
25             The primary objective of the study was to test the usefulness of city directories as a
26      source of occupational data in epidemiologic studies of illness and occupational exposure.
27      Commercial city directories provide data on occupations and employers and are compiled from a
28      door-to-door yearly census of all residents 18 years old and older.  The directories are available
29      in most medium-size cities in the United States. A  unique feature  of city directory data is that
30      they  identify specific employers, and as the authors suggest, this information may be better than
31  .    death certificates for rapid, inexpensive, record-based epidemiologic studies.
32             A case-control study was conducted of male bladder cancer deaths in Hamilton County
33      (including Cincinnati), OH. This county was selected because it is in an industrialized area with
34      high  bladder cancer rates and also because city directories cover approximately 85% of the
35      people  living in the county. A computerized list of all male bladder cancer deaths (n = 731) and

        2/1/98                                     8-47        DRAFT-DO NOT CITE OR QUOTE

-------
 1      all other male deaths (n = 95,057), with the exclusion of deaths from urinary tract tumors and
 2      pneumonia, that occurred between 1960 and 1982 was obtained from the Ohio Department of
 3      Vital Statistics.  Death certificates had been coded by a nosologist according to the ICD code in
 4      use at the time of death.  A pool of six controls was created for each case matched on sex,
 5      residence in Hamilton County at time of death, year of death, age at death (±5 years), and race.
 6             Two types of analysis were performed, one based on city directory data and the other
 7      based on death certificate data. In the former, cases and controls were restricted to individuals
 8      who had at least one yearly directory listing with some occupational date. The first two controls
 9      from the pool of six who met the requirements were selected. This analysis involved 648 cases
10      (627 cases had 2 controls and 21  cases had only 1 control) and 1,275 controls.
11             The death certificate analysis involved all 731 cancer deaths, with two controls per case.
12      In most cases, the same two controls were used in this analysis too. The usual lifetime
13      occupation and industry on the death certificate was abstracted from them. Data on occupation
14      and industry were coded with a three-digit U.S. census code using the method adopted by the
15      U.S.  Bureau of the Census.  Five of the occupational data were recorded for occupation and
16      industry by a second coder, with a high degree of reproducibility.  Odds ratios were calculated
17      for bladder cancer using a Mantel-Haenszel procedure.
18           '  The city directory data identified four employers significantly associated with bladder
19      cancer deaths; only one of them was identified by the death certificate data which provided only
20      lifetime type of industry rather than the name of a specific employer.  The industries represented
21      by the four employers were a chemical plant, printing company, valve company, and machinery
22      plant. The city directory data analysis demonstrated significant positive associations for quite a
23      few occupations. The occupations that had at least  10 cases or more were engineers (OR = 3.00,
24      ;?=0.01), carpenters (OR = 2.36,/><0.01),  tailors (OR = 2.56,/><0.01), and furnace operators (OR
25      = 2.5, p=Q.03). Relative odds were  increased significantly with increased duration of
26      employment (>20 years) for truck drivers (OR = 12,/>=0.01) and furnace operators (based on
27      four cases and no controls, /»=0.05).  For occupations in which subjects had ever been  employed,
28      a significant increase  in the relative  odds with increased duration of employment was observed
29      for the  railroad industry (> 20 years  of employment, OR = 2.21, /K0.05).  Both truck driving and
30      railroad industry occupations involve diesel emission exposures.
31             The analysis of death certificate data yielded associations in the  same direction for most
32      of the occupations. A check of the validity of city directory data indicated that 77% of the
33      listings agreed with the Social Security earnings report for the employer in any given year. A
34      comparison of city directory and death certificate information on occupations indicated a match
        2/1/98                                    8-48       DRAFT-DO NOT CITE OR QUOTE

-------
  1      for occupation between at least one city directory listing and occupation on death certificates for
  2      68,3% of the study subjects.                                 •                     '
  3             This study demonstrated that city directories are a relatively inexpensive and accessible
  4      source of occupational data for epidemiologic studies.  Limitations of this study include the
  5      problem in tracing women because of the change from maiden to married name and the
  6      availability of data for only the year of residence in the city. They are superior to death
  7      certificates in being able to identify high-risk employers in specific geographic sites. Although
  8      death .certificate data reflect usual lifetime occupation, city directories yield data on short-term
  9      jobs, some of which may involve critical exposure. Thus,  a combination of the two approaches
10      may be most productive in record-based hypothesis-generating studies.  The occupational data
11      were missing for 15%, whereas employer data were missing for 36%  in the city directory.  In the
12      context of the mentioned pros and cons of using city directories, this study found an excess risk
13      for bladder cancer among two occupations with potential diesel exposure: truck drivers and
14      railroad workers.  Two sources of bias that may have influenced these findings are selection bias
15      arising from the use of deaths instead of incident cases, because survival for bladder cancer is
16      high, and the absence of data on confounding factors such  as smoking, beverage consumption,
17      and medication use.
18
19      8.4.5. Iscovich et al. (1987): Tobacco Smoking, Occupational Exposure, and Bladder
20            Cancer in Argentina
21            This is a hospital-based case-control study of bladder cancer conducted in La Plata,
22      Argentina, to estimate the risk of bladder cancer associated with different types of tobacco,
23      beverages, and occupational exposures. Bladder cancer is  one of the most common cancers
24      among males in the La Plata area,
25            Cases were selected from patients with a histologically confirmed diagnosis of bladder
26      cancer (transitional, squamous-cell, or nonspecific cell type) admitted to the 10 general hospitals
27      in the greater La Plata area (population in 1980 — 580,000) between March 1983 and December
28      1985. Cases with true bladder papilloma and individuals who were residents of greater La Plata
29      for less than 5 years were excluded. Of the 120 cases eligible to participate, 1 died prior to the
30      interview, 2 refused to participate, and the remaining 117 cases, representing 60% of the incident
31      cases registered in the registry, were interviewed. Two control groups (117 neighborhood and
32      117 hospital controls) were matched by sex and age (±5 years).  Of the 117 cases,  99 were males
33      and 18 were females. Hospital controls, selected from the  same hospital as the cases, were
34     , hospitalized for the first time within 3 months of diagnosis of the illness of the cases. Twelve
35      percent of the hospital controls had illnesses known to be associated with tobacco  smoking.

        2/1/98               •                     8-49       DRAFT-DO NOT CITE OR QUOTE

-------
 1      Neighborhood controls were sampled from among persons living in the same block. The
 2      interviewer proceeded north in the block where the case resided and selected the first control who
 3      met the matching criteria. Seven hospital controls could not be interviewed because of their poor
 4      physical health and were replaced. Three neighborhood controls refused to participate and were
 5      replaced. Cases and hospital controls were interviewed at the hospital and the neighborhood
 6      controls at their homes to collect data on demographic, socioeconomic, and medical variables,
 7      detailed smoking habits, and alcoholic and other beverages consumed.
 8            The interviews were done by trained interviewers, two physicians, and a social worker.
 9      The cases and hospital controls were interviewed in the hospital by the physicians; hence, the
10      interviews could not be  conducted "blind." A detailed occupational history was obtained for the
11      three occupations of longest duration and the most recent one. For each job title, the activity of
12      the plant and type of production were also ascertained. Job titles were coded according to the
13      International Labor Union (ILO) 1970 classification. Plant activity and type of production were
14      coded according to the United Nations 1975 classification categories. Information was also
15      collected on exposure to 33 chemical and physical agents, which included confirmed or
16      suspected bladder carcinogens.  A detailed history of smoking habits was also obtained, and
17      individuals were categorized as current smokers if they were smoking at present or if they had
18      stopped smoking less than 2 years previously. Ex-smokers were those who ceased smoking at
19      least for 2 years or more than 2 years previously. For each subject a cumulative lifelong
20      consumption of cigarettes by type was estimated, and an average consumption of cigarettes/day
21      was computed.                                         .
22             Relative risks were computed for occupational factors using the unconditional logistic
23      regression method, adjusting for age and tobacco smoking.  These risks were derived for those
24      who were ever employed in that occupation versus those who were never employed in that
25      occupation. Socioeconomic status of cases and neighborhood controls was similar, but there
26      were fewer professionals and more manual workers among hospital controls compared with
27      cases. Occupational variables included job title and type  of activity of the plant. Significant
28      excess risks were observed for truck and railroad drivers (RR = 4.31, p<0.002) and oil refinery
29      workers (RR = 6.22, p
-------
  1      average daily amount as well as for duration of smoking. A strong dose-response relationship
  2      was also observed for coffee drinking in males, with a relative risk of 12 (95% CI = 4.3, 33.31)
  3      for those drinking more than three cups of coffee per day after adjusting for the effect of
  4      smoking. No association was found with use of saccharin in males. No results were presented
  5      for females for these risk factors.
  6             This case-control study  was conducted primarily to determine the reasons for the high
  7      rates of bladder cancer in the La Plata region of Argentina.  Only 60% of the cases registered in
  8      the cancer registry were interviewed, and no information was provided, for the remaining 40%
  9      eligible nonrespondents to determine if the study sample was selectively biased in any way.  The
10     . sample size of 117 was small, and the analysis of males reduced it to 99. Although the use of
11      two different types of control groups is a strength of this study, none of the interviews were done
12      blind, and it appears that the hospital interviews were done by the physicians and the
13      neighborhood interviews were done by the social worker.  Job titles were used as surrogates of
14      exposure, but the authors state that although they attempted to analyze by an exposure index
15      derived from a job exposure matrix (details not provided), they found no difference in exposure
16      between cases and controls. This explanation is ambiguous. The authors also grouped truck and
17      railroad drivers together for reasons not mentioned and did not present separate risk estimates.  A
18      table showing the distribution of cases and controls for selected activities or professions did not
19      indicate if the data pertain to both sexes or males only, and the text did not clarify that either.
20      The reported significant excess  risks for track and railroad drivers were reduced after adjusting
21      for smoking, but it was not known if the statistical significance persisted.  No analysis was
22      provided for the data collected in the interviews on exposures to the 33 chemical and physical
23      agents, and it was not known if the truck and railroad drivers were operating diesel engines.
24      Although rare in the La Plata area, the occupations known to be associated with bladder cancer
25      (dye, rubber, leather, and textile workers) are acknowledged by the authors.
26
27      8.4.6.  Iyer et al. (1990): Diesel Exhaust Exposure and Bladder Cancer Risk
28             This study is a hospital-based case-control study of bladder cancer and potential exposure
29      to diesel exhaust using data from a large ongoing case-control study of tobacco-related
30      neoplasms conducted by the American Health Foundation.  An earlier study by Wynder et al.
31      (1985) looked at the relationship between occupational exposure to diesel exhaust and the risk of
32      bladder cancer.  For this study, the objective was  to evaluate the relationship between the       .
33      different measures of exposure to diesel exhaust,  occupational and self-reported, and the risk of
34      bladder cancer.  Cases comprised 136 patients with histologically confirmed primary cancer of
35      the urinary bladder interviewed at 18 hospitals in six U.S. cities. Two controls were selected per

        2/1/98                                   8-51       DRAFT-DO NOT CITE OR QUOTE

-------
 1      case, matched for sex, age (within 2 years), race, hospital, and year of interview. A total of 160
 2      controls had non-tobacco-related malignancies distributed as follows:  stomach cancer (6%),
 3      colorectal cancer (20%), prostate cancer (6%), and leukemia or lymphoma (8%). Among the  112
 4      controls with nonmalignant diseases, 3% had benign neoplasms, 6% had hyperplasia of the
 5      prostate, and 6% had dorsopathies. Distribution of the other nonmalignant illnesses was not
 Q      provided. Occupational history included information on usual occupation and up to five other
 7      jobs. Exposure to diesel exhaust in hobby activities also was collected. For the purpose of this
 8      analysis, occupations were aggregated a priori into three categories: low probability of exposure
 9      (reference group), possible exposure, and probable exposure. Analyses were also done for self-
10      reported exposure to diesel exhaust. Risk estimates were obtained by unconditional logistic
11      regression using PROC LOGIST of SAS.  Cases and controls were first compared by age, race,
12      education, and smoking habit. Cases were found to be less educated than controls (p<0.05).
13      Crude odds ratios for diesel exhaust exposure, based on occupational or self-reported exposure,
14      were not significantly elevated after controlling for smoking and educational status (OR = 1.2,
15      95% CI = 0.8, 2.0).  When individual occupations were analyzed separately, truck drivers
1Q      showed no excess risk (OR = 0.48, 95% CI = 0.15, 1.56).
17             The authors concluded that their study does not support the hypothesis of an association
18      between exposure to diesel exhaust and bladder cancer. Several significant limitations of
19      exposure assessment and analysis are evident in this study. In the introduction, the authors stated
20      that they refined the definition of exposure to diesel exhaust by obtaining a lifetime occupational
21      history, but in the methods section they stated that they restricted analysis to usual occupational
22      history and five other jobs, which was not that different from their earlier study (Wynder et al.,
23      1985).  The terms, low probability of exposure, possible exposure, and probable exposure, also
24      were not clearly defined. Information on duration of employment or exposure was not presented,
25      and no attempts were made to validate occupational history! No information was available on
26      calendar years of employment in the truck-driving industry or the  locomotive occupations.
27      Because diesel trucks and locomotives were introduced in the mid-1950s and the dieselization
28      was completed by 1960, it would be important to use 1960 as a cutoff date and to restrict analysis
29      to subjects who worked in these industries after that date.  No information on nonrespondent
30      cases and controls was provided.  The authors indicated in the methods section that cases were
31      individually matched to controls, but they performed an unmatched analysis to calculate the odds
32      ratios and did not address why they did not preserve the matching in the analysis, especially
33      because such an analysis could bias the risk estimates to unity.
34
        2/1/98                                     8-52       DRAFT-DO NOT CITE OR QUOTE

-------
  1      8.4.7.  Steineck et al. (1990): Increased Risk of Urothelial Cancer in Stockholm From 1985
  2             to 1987, After Exposure to Benzene and Exhausts
  3             This study was conducted to investigate the association between benzene, diesel, and
  4      petrol exhausts as well as some other industry-related chemicals and the risk of urothelial cancer.
  5      Designed as a population-based case-control study, it was conducted among all men born
  6      between 1911 and 1945 and living  in the County of Stockholm for all or part of the observation
  7      period (September 15,1985, to November 30,1987). All incident cases of urothelial cancer and
  8      squamous-cell carcinoma of the lower urinary tract were contacted for inclusion in the study.
  9      Controls were selected by stratified random sampling during the observation period from a
10      computerized register for the population of Stockholm. A postal questionnaire was sent to study
11      subjects at their homes to collect information on occupational history. The questions on
12      occupation included exposure to certain specified occupations/indusrries/chemicals and lists of
13      all jobs held and duration in each job.  An industrial hygienist, unaware of case-control status,
14      classified subjects as being exposed or unexposed to 38 agents and groups of substances,.
15      including 17 exposure categories with aromatic amines.  Using all the exposure information, the
16      exposure period was defined and the annual dose was rated as low, moderate, or high based on
17      the accumulated dose (exposure duration multiplied by intensity of exposure) during the course
18      of 1 average year for the defined exposure period. Swedish and international data were used to
19      classify subjects as exposed, based  on air concentrations in the work environment that were
20      higher than for the general public, or skin contact with liquids of low volatility. To allow for
21      latency, the authors ignored exposures after 1981. Data were gathered from 256 cases and 287
22      controls.  Controls were selected by stratified random sampling four times from the computerized
23      register during the observation period of the population of the County of Stockholm. These
24      subjects comprised 80% of eligible cases and 79% of eligible controls. Nine of the cases and
25      16% of the controls refused to participate in the study. The  distribution of urothelial cancers was
26      as follows: 5 tumors in the renal pelvis, 243 in the urinary bladder, 5 in the ureter, none in the
27      urethra, and 3 at multiple sites. Two cases who were exposed to a high annual  dose of aromatic
28      amines were omitted from all further analysis to eliminate their confounding effects. Crude
29      relative risks were calculated for men classified as exposed or not exposed to several substances,
3D      Twenty-five cases and 19 controls reported having been exposed to diesel exhaust, yielding an
31      odds ratio of 1.7 (95% CI = 0.9, 3.3). The corresponding relative odds for petrol exhausts, based
32      on 24 cases and 24 controls, were 1.0 (95% CI = 0.5,1.9). Odds ratios were then calculated for
33      low, moderate, and high levels of the annual dose adjusted for smoking and year of birth. jFor
34      diesel exhausts,  the odds ratio was  1.3 (95% CI = 0.6, 3.1) for low levels, 2.2 (95% CI = 0.7, 6.6)
35      for moderate levels, and 2.9 (95% CI = 0.3, 30.0) for high levels, indicating a dose response.

 ;      2/1/98                                   8-53       DRAFT-DO NOT CITE OR QUOTE

-------
 1      The corresponding odds ratios for petrol exhausts were 0.6 (95% CI = 0.3, 1.3), 1.4 (95% CI =
 2     .0.5, 3.7), and 3.9 (95% CI = 0.4, 35.5).  Restricting the analysis to only moderate or high annual
 3      doses of exposure adjusted for year of birth and smoking showed a sevenfold increased risk for'
 4      subjects exposed to both diesel and petrol exhausts (OR = 7.1, 95% CI = 0.9, 58.8). For
 5      exposure to diesel (OR =1.1) and petrol (OR = 1.0) exhausts alone, no excess risk was detected
 Q      in this analysis.  Odds ratios were calculated for low, moderate, and high exposure to benzene,
 7      with rates of 1.7 (95% CI = 0.6, 5.1) for low annual doses, 1.1 (95% CI = 0.3, 4.5) for moderate
 8      annual doses, and 3.0 (95% CI = 1.0, 8.7) for high annual doses.
 9            The authors discuss misclassification and confounding as sources of bias in this study.
10      To examine misclassification they compared hygienist-assessed  exposure and self-reported
11      exposure for printing ink and found a higher relative risk and fewer exposed subjects for
12      hygienist-assessed exposure, indicating that specificity was a problem for self-reported exposure.
13      It is not known to what extent this may have affected the risk estimates for diesel exhausts since
14      data on self-reported exposure to diesel are not presented. They also mention the possibility of
15      exposure misclassification from using an average annual dose in which a person exposed to an
16      agent at a high level for a few working days and a person exposed to a low level for many days
17      are both rated as exposed to low annual doses. Although statistically nonsignificant elevated
18      odds ratios of 1.3, 2.3, and 2.9 were derived for low, moderate, and high levels of diesel
19      exposure, the authors state that some of their subjects may have  later worked in jobs with
20      benzene exposure, and because an elevated risk was detected for benzene exposure, this
21      confounding effect may explain some of the excess risk. An almost statistically significant
22      interaction was observed for exposure to combined diesel and petrol exhausts (OR = 7.1, 95% CI
23      = 0.9, 58.8), which changed to 5.1 (95% CI = 0.6, 43.3) after adjustment for benzene exposure,
24      again providing evidence for the confounding role of benzene exposure in explaining some of the
25      observed results.
26            Table 8-3 summarizes the bladder cancer case-control studies.
27
28      8.5. DISCUSSION AND SUMMARY
29             Certain extracts of diesel exhaust have been demonstrated to be both mutagenic and
30      carcinogenic in animals and in humans. Animal data suggest that diesel exhaust is a pulmonary
31      carcinogen among rodents exposed by inhalation to high doses over long periods of time.
32      Because large working populations are currently exposed to diesel exhaust and because
33      nonoccupational ambient exposures currently are of concern as well, the possibility that exposure
34      to this complex mixture may be carcinogenic ito humans has become an important public health
35      issue.

        2/1/98                                    8-54        DRAFT-DO NOT CITE OR QUOTE

-------
\o
oo
                           Table 8-3. Epidemiologic studies of the health effects of exposure to diesel exhaust:
                           case-control studies of bladder cancer
         Authors
                   Population studied
Diesel exhaust exposure
      assessment
Results
Limitations
        Howe et al. 480 male case-control pairs
        (1980)
                   152 female case-contrql pairs

                   Cases diagnosed between April
                   1974 and June 1976 in three
                   Canadian provinces

             	Matched on age and sex	
                                           Based OB occupational
                                           history of jobs involving
                                           exposure to dust and
                                           fumes

                                           A priori suspect
                                           industries
                       SNS RR = 2.8 for diesel and
                       traffic fumes
                       SS RR = 9,00 for railroad workers
                    Exposure based on occupational
                    history, which was not validated

                    Diesel exhaust and traffic fumes
                    were combined

                    Only 77% of eligible population
                    included in the study
oo
Wynder et   194 histologically confirmed male
al, (1985)   cases between the ages of 20 and
            80 years

            582 matched controls (age-,
            race, year of interview, and
            hospital of admission); diseases
            not related to tobacco use

            From 18 hospitals located in
            six U.S. cities between
	January 1981 and May 1983
Occupational titles were
defined by Industrial
Hygiene Standard into
dichotomous "exposed"
and "not exposed"

Also defined by NIOSH
standards into "high
exposure,"  "moderate
exposure,"  and "low
exposure"
                                                                          SNS ORs were 1.68 and 0.16 for
                                                                          high and moderate exposure,
                                                                          respectively, as compared to low
                                                                          exposure
                    Exposure based on usual occupation
                    may have led to misclassification

                    Dichotomous classification made
                    dose-response analysis unattainable

                    No data on other confounders such as
                    smoking
d
O
2
O
H
n
I«H
H
w
/o
H
tn

-------
 VO
 oo
oo
u<
ON
                   Table 8-3.  Epidemiologic studies of the health effects of exposure to diesel exhaust:
                   case-control studies of bladder cancer (continued)
Authors Population studied
Hoar and Population-based, case-control
Hoover study
(1985)
Diesel exhaust exposure
assessment Results Limitations
Lifetime occupational SS OR = 2.9 for 5 to 9 years of Exposure defined as occupation of
history obtained from employment as truck driver but not "truck driver" (i.e., it could have
next of kin for ^ 10 years of employment been diesel or gasoline or both)
            325 cases from the residents of
            New Hampshire and Vermont
            who died of bladder cancer
            between 1975 and 1979

            A total of 673 controls were
            chosen from other deaths during
            the same time period

            Two matched controls (age, sex,
            race, state, year of death, and
            second one matched also on
            county of residence)	
                                                                          Positive trend (p=0.006) observed
                                                                          with increasing duration
                                                                          of employment as truck driver
                                No histological confirmation of
                                bladder cancer diagnosis

                                No data on other confounders such as
                                other exposures,  smoking, etc.
o
o
a:
o
H
n
H—4
H
W
O
jo
/O
a
o
H
m
Steenland    648 male bladder cancer deaths
et al, (1987) from Hamilton County, OH

            1,275 matched controls from
            other deaths (pool of six controls
            for each case, excluding urinary
            tract rumors and pneumonias
            matched on sex, age at death,
            year of death, race)
                                                   Occupation or industry
                                                   listed in city directory
                                                   and on death certificates
OR =  12 (p=0.01) for truck
drivers with 220 years of
employment

OR = 2.21 (psO.05) for railroad
workers with  >20 years of
employment
Exposure based on city directory or
death certificate listing that was not
validated

Lack of controlling for confounders

City directory usually has short-term
job listing

Missing data on 15% of occupations
and 36% for employers in the
directory     	

-------
                   Table 8-3.  Epidemioiogic studies of the health effects of exposure to diesel exhaust:
                   case-control studies of bladder cancer (continued)
 CX3
 l
 u\
 -a
Authors Population studied
Iscovich 1 17 histologically confirmed
et al, (1987) bladder cancer cases (60% of
all incident cases)
1 17 hospital controls and
Diesel exhaust exposure
assessment
Past and present SS OR =
occupational data drivers
were collected by
questionnaire SS RR =
workers
Results
4.3 for truck and railway
6.2 for oil refinery
Limitations
Exposure based on job held that was
not validated
40% of eligible cases were
nonrespondent
                    117 neighborhood controls
                    (matched on age and sex)

                    Cases and hospital controls from
                    10 general hospitals in greater
                    La Plata between March 1983
                    and December 1985
                               An exposure index based
                               on a job exposure matrix
                               was generated
                                                        Small sample size

                                                        Interviewers were not "blind" to the
                                                        status of an individual, and this could
                                                        have biased the Findings

                                                        Truck and railroad drivers were
                                                        grouped together

                                                        Not adjusted for other confounding
                                                        exposures such as dye, rubber, etc.
 o
 o
 •z
 o
"H
 n
 »-H
 H
 M
        Iyer et al.   136 histologically confirmed
        (1990)      bladder cancer cases
                               Lifetime occupational
                               history
                       No excess found
                    272 controls, two each matched   Self-reported diesel
                    on sex, age, race, hospital, and   exhaust exposure
year of interview
(160 malignant, 112
normal ignant)

From 18 hospitals in six U.S.
cities
Exposure aggregated a
priori into:
Low probability
Possible
Probable
Exposure based on self-report, which
was not validated
                                                                                                                   was
Although lifetime occupational
history was obtained, analysis '
restricted to usual occupation

A priori classification was ambiguo
 a
 o
 H

-------
            Table 8-3. Epidemiologic studies of the health effects of exposure to diesel exhaust:
            case-control studies of bladder cancer (continued)
  Authors
                          Population studied
Diesel exhaust exposure
      assessment
Results
Limitations
oo
I
oo
Steineck    Population-based study from
et ai. (1990) County of Stockholm

            Men born between 1911 and
            1945

            256 (243 bladder) urinary tract
            cancer incident cases (80% of
            eligibles)

            287 controls (79% of eligibles)
            from population of Stockholm

            Observation period September 15,
            1985, to November 30, 1987
                                                  Occupational history
                                                  classified into exposed
                                                  and nonexposed by
                                                  industrial hygienist
                                                  "blind" toward case or
                                                  control status

                                                  Using all exposure
                                                  information, annual
                                                  dose rated as "low,"
                                                  "moderate," and "high"
                       SNS OR = 1.3 for low, OR = 2.2
                       for moderate, and OR = 2.9 for
                       high exposure were observed
                       for diesel exposure

                       SNS OR = 7.1 observed for diesel
                       and gasoline exhaust combined
                       exposure
                    Elaborate exposure history
                    classification not used to advantage
                    by simultaneous adjustment

                    Misclassification in exposure may
                    have occurred

                    Small sample size of only 25 cases
                    and 19 controls were exposed to
                    diesel exhaust

                    Confounding by other exposures not
                    accounted for, except benzene
Abbreviations: OR = odds ratio; RR = relative risks; SNS = statistically nonsignificant; SS = statistically significant.

-------
  1             Because diesel emissions become diluted in the ambient air, it is difficult to study the
  2      health effects in the general population.  Nonoccupational exposure to diesel exhaust is
  3      worldwide in urban areas.  Thus, "unexposed" reference populations used in occupational cohort
  4      studies are likely to contain a substantial number of individuals who are nonoccupationally
  5      exposed to diesel exhaust.  Furthermore, the "exposed" group in these studies is based on job
  6      titles, which in most instances are not verified or correlated with environmental hygiene
  7      measurement. The issue of health effect measurement gets further complicated by the fact that
  8      occupational cohorts tend to be healthy and have below-average mortality, usually referred to as
  9      the "healthy worker effect." Hence, the  usual standard mortality ratios observed in cohort
10      mortality studies are underestimations of real risk.
11             A major difficulty with the occupational studies considered here was the measurement of
12      the actual diesel exhaust exposure. Because all the cohort mortality studies were retrospective,
13      the assessment of health effects from exposure to diesel exhaust was naturally indirect. In these
14      occupational settings, no systematic quantitative records of ambient air were available. Most
15      studies compared men in job categories with presumably some exposure to diesel exhaust with
16      either standard populations (presumably no exposure to diesel exhaust)  or with men in other job
17      categories from industries with little or no potential for diesel exhaust exposure. A few studies
18      have included measurements of diesel fumes, but there is no standard method for the
19      measurement. No attempt is made to correlate these exposures with the cancers observed in any
20      of these studies, nor is it clear exactly which extract should have been measured to assess the
21      occupational exposure to diesel exhaust. All studies have relied on the job categories or self-
22      report of exposure to diesel exhaust.  In the studies by Garshick et al. (1987, 1988), the diesel
23      exhaust-exposed job categories were verified on the basis of an industrial hygiene survey done
24      by Woskie et al. (1988a,b). It was found by the investigators that in most cases the job titles
25      were good surrogates for diesel exhaust exposure.  Also, in the railroad industry where only
26      persons who had at least 10 years of work experience were included in the study, the workers
27      tended not to change job categories over the years. Thus, a job known only at one point in time
28      was a reasonable marker of past diesel exhaust exposure. Unfortunately, the exposure was only
29      qualitatively verified.  The  quantitative use of this  information would have been much more
30      meaningful.  The occupations involving  potential exposure to diesel exhaust are miners, truck
31      drivers, transportation workers, railroad workers, and heavy equipment operators.
32             With the exception of the study by Waxweiler et al. (1973), there have been no known
33      studies of miners to assess whether diesel exhaust is associated with lung cancer.  Currently,
34    ,  there are about 385 underground metal mines in the United States.  Of these, 250 have been
35      permanently operating, whereas 135 have been intermittently operating  (Steenland, 1986).

        2/1/98                .                    8-59       DRAFT-DO NOT CITE OR QUOTE

-------
 1      Approximately 20,000 miners are employed, but not all of them are currently working in the
 2      mines.  Diesel engines were introduced in the metal mines in the early to mid-1960s.  Although
 3      all these mines use diesel equipment, it is difficult to estimate how many of these miners were
 4      actually exposed'to diesel fumes.
 5             Diesel engines were introduced in coal mines at.an even later date, and their use is still
 6      quite limited. In 1983, approximately 1,000 diesel units were in place in underground coal
 7      mines, up from about 200 units in 1977 (Daniel, 1984). The number of units per mine varies
 8      greatly; one mine may account for more than 100 units.
 9             Even if it were possible to estimate how many miners (metal and coal) were exposed to
10      diesel exhaust, it would be very difficult to separate out the confounding effects of other potential
11      pulmonary carcinogens, such as radon decay products, heavy metals (such as arsenic,
12      chromium), etc.  Furthermore, the relatively short latency period limits the usefulness of these
13      cohorts of miners.
14
15      8.5.1. The Cohort Mortality Studies
16             The cohort studies mainly demonstrated an increase in lung cancer.  Studies of bus
17      company workers by Waller (1981), Rushton et al. (1983), and  Edling et al. (1987) failed to
18      demonstrate  any statistically significant excess risk of lung cancer, but these studies have certain
19      methodological problems, such as small sample sizes, short follow-up periods (just 6 years in the
20      Rushton et al. study), lack of information on confounding variables, and lack of analysis by
21      duration of exposure, duration of employment, or latency that preclude their use in determining
22      the carcinogenicity of diesel exhaust. Although the Waller (1981) study had a 25-year follow-up
23      period, the cohort was restricted to employees (ages 45 to 64) currently in service. Employees
24      who left the job earlier, as well as those who were still employed after age 64 and who may have
25      died from cancer, were excluded.                   '
26             Wong et al. (1985) conducted a mortality study of heavy equipment operators that
27      demonstrated a significant increased risk of liver cancer in total and in various subcohorts.  The
28      same analysis also showed statistically significant deficits in cancers of the large intestine and
29      rectum.  Metastases from the cancers of the large intestine and rectum in the liver probably were
30      misclassified as primary liver cancer, which led to  an observed  excess risk.  This study did
31      demonstrate  a nonsignificant positive trend for cancer of the lung with length of membership and
32      latency. Analysis of deceased retirees showed a significant excess of lung cancer. Individuals
33      without work histories who started work prior to 1967, when records were not kept, may have
34    .  been in the same jobs for the longest period of time. Workers without job histories included
35      those who had the same job before and after 1967 and thus may have worked about 12 to 14

        2/1/98                                     8-60      DRAFT-DO NOT CITE OR QUOTE

-------
  1     years longer; these workers exhibited significant excess risks of lung cancer and stomach cancer.
  2     If this assumption about duration of jobs is correct, then these site-specific causes can be linked
  3     to diesel exhaust exposure. One of the methodological limitations of this study is that most of
,  4     these men worked outdoors; thus, this cohort might have had relatively low exposure to diesel
  5     exhaust. The authors did not present any environmental measurement data either. Because of
  6     the absence of detailed work histories for 30% of the cohort and the availability of only partial
  7     work histories for the remaining 70%, jobs were classified and ranked according to presumed
  8     diesel exposure. Information is lacking regarding duration of employment in the job categories
  9     (used for surrogate of exposure) and other  confounding factors (alcohol consumption, cigarette
 10     smoking, etc.).  Thus, this study cannot be used to support a causal association or to refute the
 11     same between exposure to diesel exhaust and lung cancer.
 12            A 2-year mortality analysis by Boffetta and Stellman (1988) of the American Cancer
 13     Society's prospective study, after controlling for age and smoking, demonstrated an excess risk
 14     of lung cancer in certain occupations with potential exposure to diesel exhaust.  These excesses
 15     were statistically significant among miners (RR = 2.67,95% CI - 1,63,4.37) and heavy
 16     equipment operators (RR = 2.6,95% CI =  1,12, 6.06). The elevated risks were nonsignificant in
 17     railroad workers (RR = 1.59) and track drivers (RR = 1.24).  A dose response was also observed
 18     for truck drivers. With the exception of miners, exposure to diesel exhaust occurred in the three
 19     other occupations showing an increase in the risk of lung cancer.  Despite methodologic
 20     limitations, such as the lack of representiveness of the study population (composed of volunteers
 21     only, who were probably healthier than the general population), leading to an underestimation of
 22     the risk and the questionable reliability of exposure data based on self-administered
 23     questionnaires that were not validated, this study is suggestive of a causal association between
 24     exposure to diesel exhaust and excess risk of lung cancer.
 25            Two mortality studies were conducted by Gustavsson et al. (1990) and Hansen (1993)
 26     among bus garage workers (Stockholm, Sweden) and truck drivers, respectively. An SMR of
 27     122 was found among bus garage workers  based on 17 cases. A nested case-control study was
 28     also conducted in this cohort. Detailed exposure matrices based on job tasks were assembled'for
 29     both diesel exhaust and asbestos exposures. Statistically significant increasing lung cancer
 30     relative risks  of 1.34,1.81, and 2.43 were observed for diesel exhaust indices of 10 to 20,20 to
 31     30, and >30, respectively, using 0 to 10 as  a comparison group. Adjustment for asbestos
 32     exposure did  not change the results. The main strength of this study is the detailed exposure
 33     matrices; some of the limitations are lack of smoking histories and low power (small cohort).
 34            Hansen (1993), on the other hand, found  statistically significant SMR of 160 due to
 35     cancer of bronchus and lung.  No dose response was observed, although the excesses were

        2/1/98                                    8-61       DRAFT-DO NOT CITE OR QUOTE

-------
 1      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
 2      74). There are quite a few methodologic limitations to this study.  Exposure to diesel exhaust
 3      was assumed in truck drivers for diesel-powered trucks, but no validation of exposure was
 4      attempted. Smoking data were lacking, follow-up period was short, and no latency analysis was
 5      done.  The findings of both these studies are consistent with the findings of other truck driver
 6      studies.
 7            Two mortality studies of railroad workers were conducted, by Howe et al. (1983) in
 8      Canada and Garshick et al. (1988) in the United States. The Canadian study found relative risks
 9      of 1.2 (p<0.01) and 1.35 (p<0.001) among "possibly" and "probably" exposed groups,
10      respectively. The trend test showed a highly significant dose-response relationship with
11      exposure to diesel exhaust and the risk of lung cancer. The main limitation of the study was the
12      inability to separate overlapping exposures of coal dust and  diesel fumes. Information on jobs
13      was available at retirement only. There was also insufficient detail on the classification of jobs
14      by diesel exhaust exposure.  The exposures could have been nonconcurrent or concurrent, but
15      because the data are lacking, it is possible that the observed  excess could be due to the effect of
16      both coal dust and diesel fumes and not due to just one or the other. However, it should be noted
17      that, so far, coal dust has not been demonstrated to be a pulmonary carcinogen in studies of coal
18      miners, but lack of data on confounders such as asbestos and smoking makes interpretation of
19      this study difficult.  When three diesel exhaust exposure categories were examined for smoking-
20      related diseases such as emphysema, laryngeal cancer, esophageal cancer, and buccal cancer,
21      positive trends were observed, raising a possibility that the dose-response demonstrated for diesel
22      exposure may have been due to smoking. The findings of this study are at best suggestive of
23      diesel exhaust being a lung carcinogen.
24            The most definitive evidence for linking diesel exhaust exposure to lung cancer conies
25      from the Garshick et al. (1988) railroad worker study conducted in the United States. Relative
26      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
27      and 45 to 49, respectively, after the exclusion of workers exposed to asbestos. This study also
28      found that the risk of lung cancer increased with increasing duration of employment. As this was
29      a large cohort study with lengthy follow-up and adequate analysis, including dose response
30      (based on duration of employment as a surrogate) as well as adjustment for other confounding
31      factors such as asbestos, the observed association between increased lung cancer and exposure to
32      diesel exhaust is more meaningful.
33
        2/1/98                                     8-62        DRAFT-DO NOT CITE OR QUOTE

-------
  1      8.5,2.  Case-Control Studies of Lung Cancer
  2             Among the 10 lung cancer case-control studies reviewed in this chapter, only 2 studies
  3      did not find any increased risk of lung cancer. Lerchen et al. (1987) did not find any excess risk
  4      of lung cancer, after adjusting for age and smoking, for diesel fume exposure.  The major
  5      limitation of this study was a lack of adequate exposure data derived from the job titles obtained
  6      from occupational histories.  Next of kin provided the occupational histories for 50% of the cases
  7      that were not validated. The power of the study was small (analysis done on males only, 333
  8      cases). Similarly, Boffeta et al. (1990) did not find any excess of lung cancer after adjusting for
  Q      smoking and education. This study had a few methodological limitations.  The lung cancer cases
10      and controls were drawn from the ongoing study of tobacco-related diseases.  It is interesting to
11      note that the leading risk factor for lung cancer is cigarette smoking. The exposure was not
12      measured. Instead, occupations were used as surrogates for exposure.  Furthermore, there were
13      very few individuals in the study who were exposed to diesel exhaust.  On the other hand,
14      statistically nonsignificant excess risks were observed for diesel exhaust exposure by Williams et
15      al. (1977) in railroad workers (OR = 1,4) and truck drivers (OR = 1.34), by Hall and Wynder
16      (1984) in workers who were  exposed to diesel exhaust versus those who were not (OR = 1.4 and
17      1.7 with two different criteria), and by Damber and Larsson (1987) in professional drivers (OR =
18      1.2). These rates were adjusted for age and smoking. Both Williams et al.  (1977) and Hall and
19      Wynder (1984) had high nonparticipation rates of 47% and 36%, respectively.  Therefore, the
20      positive results found in these studies are underestimated at best. In addition, the self-reported
21      exposures used in the study by Hall and Wynder (1984) were not validated. This study also had
22      low power to detect excess risk of lung cancer for specific occupations.
23            The study by Benhamou et al. (1988), after adjusting for smoking, found significantly
24      increased risks of lung cancer among French motor vehicle drivers (RR = 1.42) and transport
25      equipment operators (RR = 1.35).  The main limitation of the study was the inability to separate
26      the exposures to diesel exhaust from those of gasoline exhaust because both motor vehicle
27      drivers and transport equipment operators probably were  exposed to the exhausts of both types of
28      vehicles.
29            Hayes et al. (1989) combined data from three studies (conducted in  three different states)
30      to increase the power to detect an association between lung cancer and occupations with a high
31      potential for exposure to diesel exhaust. They found that track drivers  employed for more than
32      10 years had a significantly increased risk of lung cancer  (OR =1.5,95% CI = 1.1, 1,9).  This
33      study also found a significant trend of increasing risk of lung cancer with increasing duration of
34      employment among truck drivers.  The relative odds were computed by adjusting for birth
35      cohort, smoking, and State of residence. The main limitation of this study is again the mixed

        2/1/98                                    8-63        DRAFT-DO NOT CITE OR QUOTE

-------
 1      exposures to diesel and gasoline exhausts, because information on type of engine was lacking.
 2      Also, potential bias may have been introduced because the way in which the cause of death was
 3      ascertained for the selection of cases varied in the three studies.. Further, the methods used in
 4      these studies to classify occupational categories were different, probably leading to
 5      incompatibility of occupational categories.
 6            The most convincing evidence comes from the Garshick et al. (1987) case-control study
 7      of railroad workers and the Steenland et al. (1990) case-control study of truck drivers in the
 8      Teamsters Union. Garshick et al. found that after adjustment for asbestos and smoking, the
 9      relative odds for continuous exposure were 1.39 (95% CI = 1.05,1.83).  Among the younger
10      workers with longer diesel exhaust exposure, the risk of lung cancer increased with the duration
11      of exposure after adjusting for asbestos and smoking.  Even after the exclusion of recent diesel
12      exhaust exposure (5 years before death), the relative odds increased to 1.43 (95% CI = 1.06,
13      1.94). This study appears to be a well-conducted and well-analyzed case-control study with
14      reasonably good power. Potential confounders were controlled adequately, and interactions
15      between diesel exhaust and other lung cancer risk factors were tested.
16            Steenland et al. (1990), on the other hand, created two separate work history files, one
17      from Teamsters Union pension files and the other from next-of-kin interviews.  Using duration of
18      employment as a categorical variable and considering employment after 1959 (when presumed
19      dieselization occurred) for long-haul drivers, the risk of lung cancer increased with increasing
20      years of exposure. Using  1964 as the cutoff, a. similar trend*was observed for long-haul drivers.
21      .For short-haul drivers, the trend was positive with a 1959 cutoff but not when 1964 was used as
22      the cutoff. For truck drivers who primarily drove diesel trucks and worked for 35 years, the
23      relative odds were 1.89. The limitations of this study include  possible misclassifications of
24      exposure and smoking, lack of levels of diesel exposure, smaller nonexposed group, and
25      insufficient latency period. Given these limitations, the findings of this study are probably
26      underestimated.
27            Emmelin et al. (1993) in their Swedish dock workers from 15 ports  found increased
28      relative odds of 6.8 (90% GI = 1.3 to 34.9). Intricate exposure matrices were created using three
29      different variables, but no direct exposure measurement was done. Of 50 cases and 154 controls,
30      only 6 individuals were nonsmokers. A strong interaction between smoking and diesel exhaust
31      was observed in this study.
32
33      8.5.3. Case-Control Studies of Bladder Cancer
34            Of the seven bladder cancer case-control studies, four  studies found increased risk in
35      occupations with a high potential diesel exhaust exposure. A  significantly increased risk of

        2/1/98                                     8-64       DRAFT-DO NOT CITE OR QUOTE

-------
  1     bladder cancer was found in Canadian railroad workers (RR = 9,0, 95% CI = 1,2, 349.5; Howe et
  2     al., 1980), truck drivers from New Hampshire and Vermont (OR = 2.9, p<0.05; Hoar and
  3     Hoover, 1985), and in Argentinean truck and railroad drivers (RR = 4.31, /?<0.002; Iscovich et
  4     al., 1987). A positive trend with increasing employment as truck driver (p=O.Q06) was observed
  5     by Hoar and Hoover (1985) in their study of truck drivers from New Hampshire and Vermont.
  6     Significantly increased risks also were observed with increasing duration of employment of ^20
  7     years in truck drivers (OR = 12,/?=0.01) and railroad workers (OR = 2.21, p<0.05; Steenland et
  8     al., 1987). No significant increased risk was found for any diesel-related occupations in studies
  9     by Wynder et al. (1985), Iyer et al. (1990), or Steineck et al. (1990).  All these studies had several
 10     limitations, including inadequate characterization of diesel exhaust exposure, lack of validation
 11     of surrogate measures of exposure, and presence of other confounding factors (cigarette smoking,
 12     urinary retention, concentrated smoke within the truck cab, etc.); most of them had small sample
 13     sizes and none presented any latency analysis.
 14
 15     8.5.4.  Relevant Methodologic Issues
 16            Throughout this chapter various methodologic limitations of individual studies have been
 17     discussed, such as small sample size, short follow-up period, lack of latency analysis, and lack of
. 18     data on confounding variables. However, two of the major methodologic concerns in these
 19     studies are use of death certificates to determine the cause of death and lack of data on cigarette
 20     smoking, which is a strong risk factor for both lung cancer and bladder cancer.  Death certificates
 21     were used by all  of the seven cohort mortality studies, two case-control studies of lung cancer,  .
 22     and one case-control study of bladder cancer to determine cause of death.  Use of death
 23     certificates could lead to misclassification bias. Studies of autopsies done between  1960 and
 24     1971 demonstrated that lung cancer was overdiagnosed when compared with hospital discharge,
 25     with no incidental cases found at autopsy (Rosenblatt et al., 1971). Sehottenfeld et al. (1982)
 26     also found an overdiagnosis of lung cancer among autopsies conducted in  1977 and 1978.  On
 27     the other hand, Percy et al. (1981) noted 95% concordance when comparing 10,000 lung cancer
 28     deaths observed in the Third National Cancer Survey from 1969 to 1971 (more than 90% were
 29     confirmed histologically) to death certificate-coded cause of death. For bladder cancer,  the
 30     concordance rate was 91%. These more recent findings suggest that the diagnosis of lung cancer
 31     as well as bladder cancer on death certificates is better than anticipated.  Furthermore, an
 32     overdiagnosis of lung cancer or bladder cancer on death certificates would reduce the ability of
 33     the study to detect an effect of diesel exhaust exposure.
 34            All the cohort studies  considered for this report are retrospective mortality studies.  It is
 35     usually difficult to obtain smoking history in such instances. The smoking histories obtained

        2/1/98        .                            8-65       DRAFT-DO NOT CITE OR QUOTE

-------
  1      from surrogates (next of kin being either a spouse or an offspring) were found to be accurate by
  2      Lerchen and Samet (1986) and McLaughlin et al. (1987). Lerchen and Samet (1986) did not
  3      detect any consistent bias in the report of cigarette consumption. In contrast, overreporting of
  4      cigarette smoking by surrogates was observed by Rogot and Reid (1975), Kolonel et al. (1977),
  5      and Humble et al. (1984). Kolonel et al. (1977) found that the age at which an individual started
  6      smoking was reported within 4 years of actual age 84% of the time. The indication from these
  7      studies is that surrogates were able to provide fairly credible information on the smoking habits
  8      of the study subjects. If the surrogates of the cases were more likely to overreport cigarette
  9      smoking as compared to the controls, then it might be harder to  find an effect of diesel exhaust
10      because most of the increase in lung cancer would be attributed  to smoking rather than to the
11      effect of exposure to diesel exhaust.
12
13      8.5.5.  Criteria of Causal Inference
14             In most situations epidemiologic data are used to delineate the causality of certain health
15      effects.  Several cancers have been causally associated with exposure to agents for which there  is
16      no direct biological evidence.  Insufficient knowledge about the biological basis for diseases in
17      humans makes it difficult to identify exposure to an agent as causal, particularly for malignant
18      diseases when the exposure was in the distant past. Consequently, epidemiologists and biologists
19      have provided a set of criteria that define a causal relationship between exposure and the health
•20      outcome. A causal interpretation is enhanced for studies that meet these criteria. None of these
21      criteria actually proves causality;  actual proof is rarely attainable when dealing with
22      environmental carcinogens. None of these criteria should be considered either necessary (except
23      temporality of exposure) or sufficient in itself.  The absence of any one or even several of these
24      criteria does not prevent a causal interpretation. However, if more criteria apply, it provides
25      more credible evidence for causality.
26             Thus, applying the criteria of causal inference to the seven cohort mortality and eight
27      case-control studies in which risk of lung cancer was assessed resulted in the following:
28
29             •   Temporality: There is a temporality of exposure to diesel exhaust prior to the
30                 occurrence of lung cancer.
31
32             •   Strength of association: The strength of association between exposure and the
33                 occurrence of lung cancer in the cohort studies showed a 30% to 57% higher risk
34                 among exposed persons as compared to nonexposed (Howe et al., 1983; Wong et al.,
35                 1985; Boffetta and Stellman, 1988;  Garshick et al., 1988). In case-control studies, the
36                 risk varied from 20% to 89% higher among exposed as compared to nonexposed
37                 (Williams et al., 1977; Hall and Wynder, 1984; Damber and Larsson, 1987; Garshick

        2/1/98                                    8-66       DRAFT-DO NOT CITE OR QUOTE

-------
  1                et al, 1987; Benhamou et al., 1988; Hayes et al., 1989; Steenland et al, 1990;
  2                Gustavsson et al., 1990; Emmelin et al., 1993).  Some of these studies did adjust for
  3                the confounding effects of smoking, asbestos, and other exposures. Furthermore, a
  4                recent publication by HEI (1995) demonstrates this strength of association in graphic
  5                presentation (Figures 8-1 and 8-2).  Although the studies had smaller increases in
  6                lung cancer risk and only some of the studies considered by HEI (1995) are
  7                considered in this chapter, it demonstrates the lung cancer excesses consistently all
  8                across the various populations.
  9                                          '
10             *   Consistency: Five cohort studies and nine (including one nested case-control) case-
11                control studies of lung cancer conducted in several populations in the United States
12                and Europe consistently found the same effect (i.e., lung cancer).
13
14             •   Specificity: All of the above-mentioned studies found the same specific effect (i.e.,
15                lung cancer).
16
17             •   Biological gradient: The biological gradient, which refers to the dose-response
18                relationship, was observed in the cohorts of Canadian railway workers (Howe et al.,
19                1983), heavy bulldozer operators (Wong et al., 1985), and truck drivers who had
20                enrolled in the American Cancer Society's prospective mortality study (Boffetta and
21                Stellman, 1988). In the case-control studies, a dose response was observed in railroad
22                workers (Garshick et al., 1988;  Hayes et al., 1989; Steenland et al., 1990).  Although
23                other studies failed to observe a dose response, these studies were methodologically
24                limited due to confounding by other exposures and lack of either quantitative data on
25                exposure or surrogate data on dose.
26
27             •   Biological plausibility:  Because diesel exhaust consists of a carbon core particle
28                with surface layers of organics and gases, the tumorigenic activity may reside in one,
29                some, or all of these components. As explained in Chapter 9, there is clear evidence
30                that the organic constituents have the capacity to interact with DNA and give rise to
31                mutations, chromosomal aberrations, and cell transformations, all well-established
32                steps in the process of carcinogenesis. Furthermore, these organic chemicals include
33                a variety of polycyclic aromatic hydrocarbons and nitroaromatics, many of which are
34                known to be pulmonary carcinogens. Alternatively, Vostal (1986) suggests that
35                "diesel" particles themselves induce lung cancer, most likely via an epigenetic
36                mechanism, if they are present at sufficiently high doses; This makes a convincing
37                argument for biological plausibility of lung cancer occurrence under some condition
38                of exposure.
39
40             When the same causal inference criteria were applied to the seven case-control
41      studies in which risk of bladder cancer was assessed, the results were:
42
43             •   Temporality: There is temporality of exposure to diesel exhaust prior to the
44                occurrence of bladder cancer.
        2/1/98                                     8-67       DRAFT-DO NOT CITE OR QUOTE

-------
                               Hall-i
                         Garshick (88) J
                          Boffetta (90) -j
                            Schenker -j
                         Garshick (87) 1
                              Bums-
                           Siemiatycki -
                            Williams -
                          Boffetta (88) -
                                  0.1
       1
Relative Risk (95% Cl)
                                                                   10
   Figure 8-1.  Lung cancer and exposure to diesel exhaust in railroad workers.  • = Relative
               risk adjusted for cigarette smoking; O = relative risk not adjusted for cigarette
               smoking. For the two studies by Howe and Williams, confidence intervals were
               not reported and could not be calculated.

   Source: HEI, 1995.
Menck-
Hall -
Leupker -
Ahlberg -
Siemiatycki -
Boffetta (88) -
Hayes -
Boffetta (90) -
Steenland -
Bums-
Williams-
0

• 	
•—

•

•—

•-
• —


1 1
»-o-«
	 O 	 • '
-O 	 •
•-O-.
0 B

-• 	 •
• • •
• m
"
— • 	 •
•— • 	 •
•
1
                                           Relative Risk (95% Cl)


  Figure 8-2.  Lung cancer and exposure to diesel exhaust in truck drivers.  • = Relative risk
              adjusted for cigarette smoking; O = relative risk not adjusted for cigarette
              smoking.  For the study by Williams, confidence intervals were not reported and
              could not be calulated. For the Steenland study, the data were gathered from
              union reports of long-haul truck drivers; for the Boffetta (1988) study, the data
              were self-reported by diesel truck drivers; and for the Siemiatycki study, they
              were self-reported by heavy-duty truck drivers (personal communication).

  Source: HEI, 1995.                                                                       .
2/1/98
8-68
DRAFT-DO NOT CITE OR QUOTE

-------
  1             •  Strength of association:  The relative odds of getting bladder cancer among exposed
  2                as compared to nonexposed ranged from 2 to 12 times higher (Howe et al., 1980;
  3                Hoar and Hoover, 1985; Iscovieh et al., 1987; Steenland et al., 1987). None of these
  4                studies adjusted for other confounding effects such as cigarette smoking, exposures to
  5                other chemicals, urinary retention, etc.
  6    •
  7             •  Consistency:  Four out of seven bladder case-control studies conducted in the United
  8                States and abroad found an increased relative odds of bladder cancer in the exposed
  9                population.  None of the cohort studies showed increased bladder cancer mortality;
10                however, people rarely die from bladder cancer, so bladder cancer excess is unlikely
11                to be detected in mortality studies.
12
13             •  Specificity: Four out of seven case-control  studies found an excess of bladder cancer.
14                The specificity criterion, per se, does not apply in this particular instance because
15                these are case-control studies.
16                                                                           .
17             •  Biological gradient: Dose response was observed in two out of four studies showing
18                increasing relative odds with increasing length of employment (Hoar and Hoover,
19                1985; Steenland etal., 1987).
20
21             •  Biological plausibility: It has been demonstrated that motor exhaust emissions
22                contain PAHs and nitro-PAHs (Stenberg et al., 1983; Rosenkranz and Mermelstein,
23                1983). There is some evidence that nitro-PAHs may be responsible for the induction
24                of human bladder cancer.  Nitro-PAHs can be metabolized to aromatic amine
25                derivatives, and some of these agents are known to be capable of inducing urinary
26                bladder cancer (Clayson and Garner,  1976).  Furthermore, 1-nirropyrene (1-NP) has
27                been reported to be carcinogenic in the rat mammary gland (Hirose et al., 1984); the
28                structurally related 4-aminobiphenyl, which  induces bladder cancer in humans, also
29                induces mammary gland tumors in rats (Hirose et al., 1984).  Although the
30                applicability of these experimental results to humans is unknown, the laboratory
31                evidence certainly suggests the biological plausibility of diesel exhaust to be a urinary
32                bladder carcinogen.                                                   .
33
34             In summary, although some of the causality inference criteria do apply to bladder cancer,
35      the evidence for bladder cancer in populations exposed  to diesel exhaust is inadequate. On the
36      other hand, all the causality inference criteria apply well to lung cancer.  An excess risk of lung
37      cancer was observed in four out of seven cohort studies and seven out of eight case-control
38      studies. Dose response was found in three cohort studies and three case-control studies using
39      duration of employment as a surrogate for dose.  A recent meta-analysis (Bhatia, 1998) shows
40      the consistency of elevated risks in 23 of 29 diesel exposure epidemiology studies, with
41      statistically significant relative risks of 1.33. However, because of lack of actual data on
42     ' exposure to diesel exhaust in these studies and other subtle methodologic limitations, the human
        2/1/98                •                    8-69       DRAFT-DO NOT CITE OR QUOTE

-------
 1       evidence falls just short of being sufficient to call diesel exhaust a human carcinogen.  Using

 2       EPA's 1986 Guidelines for Carcinogen Risk Assessment, the human evidence alone is "limited,"

 3       however, the "limited" classification doesn't indicate how close this is to being a known human

 4       carcinogen. Although diesel exhaust exposure is classified as "Limited," the human evidence is

 5       highly suggestive of a causal association between lung cancer and occupational exposure.  Based

 .6       on the human evidence alone, diesel exhaust is close to being a known human carcinogen.

 7

 8       8.6. REFERENCES
 9
10       Ahlberg, J; Ahlbom, A; Lipping, H; Norell, S; Osterblom, L. (1981) [Cancer among professional drivers—a
11       problem-oriented register-based study]. Lakartidningen 78:1545-1546.
12
13       Benhamou, S; Benhamou, E; Flamant, R. (1988) Occupational risk factors of lung cancer in a French case-control
14       study. BrJIndMed 45:231-233.
15
16       Bhata,  R; Lopipero, P; Smith A. (1998) Diesel exhaust exposure and lung cancer. Epidemiol9(l):84-91.
17
18       Boffetta, P; Stellman, SD. (1988) Association between diesel exhaust exposure and multiple myeloma: an example
19       of confounding. PrevMed 17:236-237.
20
21       Boffetta, P; Harris, RE; Wynder, EL. (1990) Case-control study on occupational exposure to diesel exhaust and lung
22       cancer risk. Am JIndMed 17:577-591.
23
24       Buiatti, E; Kriebel, D; Geddes, M; Santucci, M; Pucci, N. (1985) A case control study of lung cancer in Florence,
25       Italy. I. Occupational risk factors. J Epidemiol Commun Health 39:244-250.
26
27       Clayson, DB; Garner, RC. (1976) Carcinogenic aromatic amines and related compounds. In: Searle, CE, ed.
28       Chemical carcinogens. Washington, DC: American Chemical Society. (ACS monograph no. 173).
29
30       Coggon, D; Pannett, B; Acheson, ED. (1984) Use of job-exposure matrix in an occupational analysis of lung and
31       bladder cancers on the basis of death certificates. JNatl Cancer Inst 72:61-65.
32
33       Cordier, S; Clavel, J; Limasset, JC; Boccon-Gibod, L; Le Moual, N; Mandereau, L; Hemon, D. (1993) Occupational
34       risks of bladder cancer in France: a multicenter case-control study. IntJ Epidemiol 22:403-411.
35
36       Damber, LA; Larsson, LG. (1987) Occupation and male lung cancer: a case-control study in northern Sweden. Br J
37       IndMed 44:446-45 3.
38
39       Daniel, JH. (1984) The use of diesel-powered equipment in U.S. underground coal operations. Presented at:
40       American Mining Congress international coal show;  May; Chicago, IL. Washington, DC: Department of the
41       Irtterior, Bureau of Mines.
42
43       Edling, C; Anjou, C-G; Axelson, O; Kling, H. (1987) Mortality among personnel exposed to diesel exhaust. Int
44       Arch Occup Environ Health 59:559-565.
45
46       Emmelin, A; Nystrom, L; Wall, S. (1993) Diesel exhaust exposure and smoking: a case-referent study of lung
47       cancer among Swedish dock workers. Epidemiology 4:237-244.
48
49       Flodin, U; Fredriksson, M; Persson, B. (1987) Multiple myeloma and engine exhausts, fresh wood, and creosote: a
50       case-referent study. Am JIndMed 12:519-529.

         2/1/98                 ,                     8-70        DRAFT-DO NOT CITE OR QUOTE

-------
  1      Garland, FC; Gorham, ED; Garland, CF; Ducatman, AM. (1988) Testicular cancer in US Navy personnel. Arn J
  2      Epidemiol 127:411-414.
  3
  4      Garshick, E; Sehenker, MB; Munoz, A; Segal, M; Smith, TJ; Woskie, SR; Hammond, SK; Speizer, FE. (1987) A
  5  %   case-control study of lung cancer and diesel exhaust exposure in railroad workers. Am Rev Respir Dis 135:1242-
  6      1248.
  7
  8      Garshick, E; Schenker, MB; Munoz, A; Segal, M; Smith, TJ; Woskie, SR; Hammond, SK; Speizer, FE. (1988) A
  9      retrospective cohort study of lung cancer and diesel exhaust exposure in railroad workers. Am Rev Respir Dis
10      137:820-825.
11
12      Guberan, E; Usel, M; Raymond, L; Bolay, J; Fioretta, G; Puissant, J. (1992) Increased risk for lung cancer and For
13      cancer of the gastrointestinal tract among Geneva professional drivers. Br J Ind Med 49:337-344.
14
15      Gustafsson, L; Wall, S; Larsson, LG; Skog, B, (1986) Mortality and cancer incidence among Swedish dock
16      workers—a retrospective cohort study. Scand J Work Environ Health  12:22-26.
17
18      Gustavsson, P; Plato, N;  LidstrSm, EB; Hogstedt, C. (1990) Lung cancer and exposure to diesel exhaust among bus
19      garage workers. Scand J  Work Environ Health 16:348-354.
20
21      Hall, NEL; Wynder, EL. (1984) Diesel exhaust exposure and lung cancer: a case-control study. Environ Res 34:77-
22      86.
23
24      Hansen, ES. (1993) A follow-up study on the mortality of truck drivers. Am J Ind Med 23:811-821.
25
26      Hayes, RB;  Thomas, T; Silverman, DT; Vineis, P; Blot, WJ; Mason, TJ; Pickle, LW; Correa, P; Fontham, ETH;
27      Schoenberg, JB. (1989) Lung cancer in motor exhaust-related occupations. Am J Ind Med 16:685-695.
28
29      Health Effects Institute. (1995) Diesel exhaust: a critical analysis of emissions, exposure, and health effects.
30      Cambridge, MA: Related HEI Publications; p. 27.
31
32      Hirose, M; Lee, MS.; Wang, CY; King, CM. (1984) Induction of rat mammary gland tumors by 1-nitropyrene, a
33      recently recognized environmental mutagen. Cancer Res 44:1158-1162.
34
35      Hoar, SK; Hoover, R (1985) Truck driving and bladder cancer mortality in rural New England. J Natl Cancer Inst
36      74:771-774.
37
38      Howe, GR; Burch, JD; Miller, AB; Cook, GM; Esteve, J; Morrison, B; Gordon, P;  Chambers, LW; Fodor, G;
39      Winsor,  GM. (1980) Tobacco use, occupation, coffee, various nutrients, and bladder cancer, J Natl Cancer Inst 64:
40      701-713.          '
41
42      Howe, GR; Fraser, D; Lindsay, J; Presnal, B; Yu, SZ. (1983) Cancer mortality (1965-77) in relation to diesel fume
43      and coal exposure in a cohort of retired railway workers. J Natl Cancer Inst 70: 1015-1019.
44
45      Humble, CG; Samet, JM; Skipper, BE. (1984) Comparison of self- and surrogate-reported dietary information. Am
46      J Epidemiol 119:86-98.
47
48      Iscovich, J; Castelleto, R; Esteve, J; Muftoz, N; Colanzi, R; Coronel, A; Deamezola, I; Tassi, V; Arslan, A. (1987)
49      Tobacco smoking, occupational exposure and bladder cancer in Argentina. Int J Cancer 40:734-740.
50
51      Iyer, V;  Harris, RE; Wynder, EL. (1990) Diesel exhaust exposure and bladder cancer risk. Eur J Epidemiol 6:49-54.
52
53      Jensen, OM; Wahrendorf, J; Knudsen, JB; S0rensen, BL. (1987) The Copenhagen case-referent study on bladder
54      cancer: risks among drivers, painters, and certain other occupations. Scand. J Work Environ Health 13:129-134.

         2/1/98                      "                    8-71        DRAFT-DO NOT CITE OR QUOTE

-------
 '1       Kaplan, I. (1959) Relationship of noxious gases to carcinoma of the lung in railroad workers. J Am Med Assoc
 2       171:2039-2043,                      '                                               •
 3
 4       [Colonel, LN; Hirohata, T; Nomura, AMY. (1977) Adequacy of survey data collected from substitute respondents.
 5   '    Am JEpidemiol 106:476-484.
 6
 7       Lerchen, ML; Samet, JM. (1986) An assessment of the validity of questionnaire responses provided by a surviving
 8       spouse. Am JEpidemiol  123:481-489.
 9
10       Lerchen, ML; Wiggins, CL; Samet, JM. (1987) Lung cancer and occupation in New Mexico. J Natl Cancer Inst
11       79:639-645.
12                                                                               .
13       McLaughlin, JK; Dietz, MS; Mehl, ES; Blot, WJ, (1987) Reliability of surrogate information on cigarette smoking
14       by type of informant. Am J Epidemiol 126:144-146.
15
16       Notani, PN; Shah, P; Jayant, K; Balakrishnan, V. (1993) Occupation and cancers of the lung and bladder: a case-
17       control study in Bombay. Int JEpidemiol 22:185-191.
18
19       Percy, C; Stanek, E, III; Gloeckler, L. (1981) Accuracy of cancer death certificates and its effect on cancer mortality
20       statistics. Am J Public Health 71:242-250.
21
22       Raffle, PAB. (1957) The health of the worker. Br J Ind Med 14:73-80.
23
24       Risch, HA; Bureh, JD; Miller, AB; Hill, GB; Steele, R; Howe, GR. (1988) Occupational factors and the incidence of
25       cancer of the bladder in Canada. BrJ Ind Med 45:361-367.
26
27       Rogot, E; Reid, DD. (1975) The validity of data  from next-of-kin in studies of mortality among migrants. Int J
28       Epidemiol 4:51-54.
29
30       Rosenblatt, MB; Teng, PK; Kerpe, S; Beck, I. (1971) Causes of death in 1,000 consecutive autopsies. NY State J
31       Med 71:2189-2193.
32
33       Rosenkranz, HS; Mermelstein, R,  (1983) Mutagenicity and genotoxicity of nitroarenes: all nitro-containing
34       chemicals were not created equal. Mutat Res 114:217-267.
35
36       Rushton, L; Alderson, MR; Nagarajah, CR. (1983) Epidemiological survey of maintenance workers in London
37       Transport Executive bus  garages at Chiswick Works. Br J Ind Med 40:340-345.
38                                                   ...'•••
39       Schenker, MB; Smith, T; Munoz,  A; Woskie, S;  Speizer, FE. (1984) Diesel exposure and mortality among railway
40       workers: results of a pilot study. BrJ Ind Med 41:320-327.
41
42       Schottenfeld, D; Eaton, M; Sommers, SC; Alonso, DR; Wilkinson, C. (1982) The autopsy as a measure of accuracy
43       of the death certificate. Bull NY Acad Med 58:778-794.
44
45       Siemiatycki, J; Gerin, M; Stewart, P; Nadon, L; Dewar, R; Richardson, L.  (1988) Associations between several sites
46       of cancer and ten types of exhaust and combustion products: results from a ease-referent study in Montreal. Scand J
47       Work Environ Health 14:79-90.
48
49       Silver-man, DT; Hoover,  RN; Albert, S;  Graff, KM. (1983) Occupation and cancer of the lower urinary tract in
50       Detroit. J Natl Cancer Inst 70:237-245.
51
52       Silverman, DT; Hoover,  RN; Mason, TJ; Swanson,.GM. (1986) Motor exhaust-related occupations and bladder
53       cancer. Cancer Res 46:2113-2116.  .
54

         2/1/98                                     '     8-72        DRAFT-DO NOT CITE OR QUOTE

-------
  1      Steenland, K. (1986) Lung cancer and diesel exhaust: a review. Am J Ind Med 10:177-189.
  2
  3      Steenland, K; Burnett, C; Osoria, AM. (1987) A Case-control study of bladder cancer using city directories as a
  4      source of occupational data. Am JEpidemiol 126:247-257,
  5
  6      Steenland, NK; Silverman, DT; Hornung, RW. (1990) Case-control study of lung cancer and truck driving in the
  7      Teamsters Union. Am J Public Health 80:670-674.
  8
  9      Steineck, G; Plato, N; Gerhardsson, M; Norell, SE; Hogstedt, C. (1990) Increased risk of urothelial cancer in
10      Stockholm during 1985-87 after exposure to benzene and exhausts. IntJ Cancer 45:1012-1017,
11
12      Stenberg, U; Alsberg, T; Westerholm, R. (1983) Emission of carcinogenic components with automobile exhausts.
13      Environ Health Perspect 47:53-63.
14
15      Stern, FB; Lemen, RA; Curtis, RA. (1981) Exposure of motor vehicle examiners to carbon monoxide: a historical
16      prospective mortality study. Arch Environ Health 36:59-66.
17
18      Swanson, GM; Lin, CS; Burns, PB. (1993) Diversity in the association between occupation and lung cancer among
19      black and white men. Cancer Epidemiol Biomark Prev 2:313-320.
20
21    '  Vineis, P; Magnani, C. (1985) Occupation and bladder cancer in males: a case-control study. Int J Cancer 35:599-
22      606.
23
24      Vostal, JJ. (1986) Factors limiting the evidence for chemical carcinogenicity of diesel emissions in long-term
25      inhalation experiments. In: Ishinishi, N; Koizumi, A; McClellan, RO; Stdber, W, eds. Carcinogenic and mutagenic
26      effects of diesel engine exhaust: proceedings of the international satellite symposium on toxicological effects of
27      emissions from diesel engines; July; Tsukuba Science City, Japan. (Developments in toxicology and environmental
28      science: v. 13.) Amsterdam: Elsevier  Science Publishers BV; pp. 381-396.
29
30      Waller, RE (1981) Trends in lung cancer in London in relation to exposure to diesel fumes. Environ Int 5:479-483.
31
32      Waxweiler, RJ; Wagoner, JK; Archer, VE (1973) Mortality of potash workers. J Occup Med 15:486-489.
33
34      Williams, RR; Stegens, NL; Goldsmith, JR (1977) Associations of cancer site and type with occupation and industry
35      from the Third National Cancer Survey interview. JNatl Cancer Inst 59:1147-1185.
36
37      Wong, O; Morgan, RW; Kheifets, L; Larson, SR; Whorton, MD (1985) Mortality among members of a heavy
38      construction equipment operators union with potential exposure to diesel exhaust emissions. Br J Ind Med 42:435-
39      448.
40
41      Woskie, SR; Smith, TJ; Hammond, SK; Schenker, MB; Garshick, E; Speizer, FE (1988a) Estimation of the diesel
42      exhaust exposures of railroad workers: I. Current exposures. Am J Ind Med 13:381-394.
43
44      Woskie, SR; Smith, TJ; Hammond, SK; Schenker, MB; Garschick, E; Speizer, FE (1988b) Estimation of the diesel
45      exhaust exposures of railroad workers: II. National and historical exposures. Am J Ind Med 13:395-404.
46
47      Wynder, LE; Dieck, G; Hall, NE (1985) A case control study of diesel exhaust exposure and bladder cancer.
48      Environ Res 34:475-489.
         -2/1/98                                          8-73         DRAFT-DO NOT CITE OR QUOTE

-------

-------
                                         9.  MUTAGENICITY

  1             Since 1978, more than 100 publications have appeared in which genotoxicity assays were
  2      used with diesel emissions, the volatile and particulate fractions (including extracts), or
  3      individual chemicals found in diesel emissions. Although most of the studies deal with the
  4      question of whether particulate extracts from diesel emissions possess mutagenie activity in
  5      microbial and mammalian cell assays, a' number of studies in recent years have employed
  6      bioassays (most commonly Salmonella TA98 without S9) to evaluate (1) extraction procedures,
  7      (2) fuel modifications, (3) bioavailability of chemicals from diesel particulate matter (DPM), and
  8      (4) exhaust filters or other modifications and other variables associated with diesel emissions.
  9      This chapter will focus on the application of the available data to issues of genetic risk
10      assessment; reports dealing with mutagenie activity associated with the metabolism of particular
11      chemicals of DPM are discussed in Chapter 10. Also, because of the large number of reports,
12      this discussion will focus on key references. An International Agency for Research on Cancer
13      (IARC) monograph (International Agency for Research on Cancer, 1989) contains an exhaustive
14      description of the available studies and other review articles (Claxton, 1983; Pepelko and
15      Peirano, 1983); the proceedings of several symposia on the health effects of diesel emissions
16      (U.S. Environmental Protection Agency, 1980; Lewtas, 1982; Ishinishi et al., 1986;  International
17      Agency for Research on Cancer, 1989) are also available.
18
19   '   9.1.  GENE MUTATIONS
20            Huisingh et al. (1978) demonstrated that dichloromethane extracts from DPM were
21      mutagenie in strains TA1537, TA1538, TA98, and TA100 of S. typhimurium, both with and
22      without rat liver S9 activation. This report contained data from several fractions as well as DPM
23      from different vehicles and fuels. Similar results with diesel extracts from various engines and
24      fuels have been reported by a number of investigators using the Salmonella frameshift-sensitive
25      strains TA1537, TA1538, and TA98 (Siak et al., 1981; Claxton, 1981; Dukovich et al., 1981;
26      Brooks et al., 1984). Similarly,  mutagenie activity was observed in Salmonella forward mutation
27      assays measuring 8-azaguanine resistance (Claxton and Kohan, 1981) and in E. colt mutation
28      assays (Lewtas,  1983).
29            One approach to identifying significant mutagens in chemically complex environmental
30      samples such as diesel exhaust or ambient particulate extracts is the combination of short-term
31      bioassays with chemical fractionation (Scheutzle and Lewtas, 1986). The analysis is most
32      frequently carried out by sequential extraction with increasingly polar or binary solvents.
33      Prefractionation by silica-column chromatography separates compounds by polarity or into

        2/1/98                                     9-1        DRAFT-DO NOT CITE OR QUOTE

-------
 1      acidic, basic, and neutral fractions.  The resulting fractions are too complex to characterize by
 2      chemical methods, but the bioassay analysis can be used to determine fractions for further
 3  .    analysis. In most applications-of this concept, Salmonella strain TA98 without the addition of S9
 4      has been used as the indicator for mutagenic activity. Generally, a variety of nitrated polynuclear
 5      aromatic compounds have been found that account for a substantial portion of the mutagenicity
 6      (Liberti et al., 1984; Schuetzle and Frazer, 1986; Schuetzle and Perez, 1983). However, not all
 7      bacterial mutagenicity has been identified in this way, and the identity of the remainder of the
 8      mutagenic compounds remains unknown. The nitrated aromatics thus far identified in diesel
 9      exhaust were the subject of review in the I ARC monograph on diesel exhaust (International
10      Agency for Research on Cancer, 1989). In addition to the simple qualitative identification of
11      mutagenic chemicals, several investigators have used numerical data to express mutagenic
12      activity as activity per distance driven or mass of fuel consumed. The^e types of calculations
13      have been the basis for estimates that the nitroarenes (both mono- and dinitropyrenes) contribute
14      a significant amount of the total mutagenic activity of the whole extract (Nishioka et al., 1982;
15      Salmeen et al., 1982; Nakagawa et al.,  1983). However, as noted by Claxton (1983), because
16      most of these studies used only strain TA98 without exogenous activation, several classes of  .
17      mutagenic chemicals may have gone undetected.
18            Matsushita et al. (1986) tested particle-free diesel exhaust gas and a number of benzene
19      nitro-derivatives and polycyclic aromatic hydrocarbons (PAHs) (many of which have been
20      identified as components of diesel exhaust gas). The particle-free exhaust gas was positive in
21      both TA100 and TA98, but only without S9 activation. Of the 94 nitrobenzene derivatives
22      tested, 61 were mutagenic, and the majority showed greatest activity in TA100 without S9.
23      Twenty-eight of 50 PAHs tested were mutagenic,  all required the addition of S9 for detection,
24      and most appeared to show a stronger response  in TA100. When 1,6-dinitropyrene was mixed
25      with various PAHs or an extract of heavy-duty (HD) diesel exhaust, the mutagenic activity in
26      TA98 was greatly reduced when S9 was absent but was increased significantly when S9 was
27      present.  These latter results  suggested that caution should be used in estimating mutagenicity (or
28      other toxic effects) of complex mixtures from the  specific activity of individual components."
29            Mitchell et al. (1981) reported mutagenic activity of DPM extracts of diesel emissions in
30      the mouse lymphoma L5178Y mutation assay.  Positive results were seen both with and without
                                   /
31      S9 activation in extracts from several different vehicles, with mutagenic activity only slightly
32      lower in the presence of S9.  These findings have  been confirmed in a number of other
33      mammalian cell systems using several different genetic markers. Casto et al. (1981), Chescheir
34      et al. (1981), Li and Royer (1982), and Brooks et al. (1984) all reported positive responses at the
35      HGPRT locus in Chinese hamster ovary (CHO) cells.  Morimoto et al. (1986) used the APRT

        2/1/98                                     9-2       DRAFT-DO NOT CITE OR QUOTE

-------
  1      and Ouar loci in CHO cells; Curren et al, (1981) used Ouar in BALB/c 3T3 cells.  In all of these
  2      studies, mutagenic activity was observed without S9 activation.  Liber et al. (1981) used the
  3      thymidine kinase (TK) locus in the TK6 human lymphoblast cell line and observed induced
  4      mutagenesis only in the presence of rat liver S9 when testing a methylene chloride extract of
  5      diesel exhaust. Barfhecht et al. (1982) also used the TK6 assay to identify some of the chemicals
  6      responsible for this activation-dependent mutagenicity.  They suggested that fluoranthene, 1 -
  7      methylphenanthrene, and 9-methylphenanthrene could account for over 40% of the observed
  8      activity.
  9            Morimoto et al. (1986) injected DPM extracts (250 to 4,000 mg/kg) into pregnant Syrian
10      hamsters and measured mutations at the APRT locus in embryo cells cultivated 11 days after i.p.
11      injection. Neutral fractions from both light-duty (LD) and HD tar samples resulted in increased
12      mutation frequency at 2,000 and 4,000 mg/kg. Belisario et al. (1984) applied the Ames test to
13      urine from Sprague-Dawley rats exposed to single applications of DPM administered by gastric
14      intubation, i.p. injection, or s.c. gelatin capsules. In all cases, dose-related increases were seen in
15      TA98 (without and with S9) from urine concentrates taken 24 h after particle administration.
16      Urine from  Swiss mice exposed by inhalation to filtered exhaust (particle concentration 6 to 7
17      mg/m3) for 7 weeks (Pereira et al., 1981 a) or Fischer 344 rats exposed to DPM (2 mg/m3)- for 3
18      months to 2 years was negative in Salmonella strains. Because of the large differences in
19      dosages, these findings should not be construed as conflicting.
20            Schuler and Niemeier (1981) exposed DrosopMa males .in a stainless steel chamber
21      connected to the 3 m3 chamber used for the chronic animal studies at EPA (see Hinners et al.,
22      1980 for details). Flies were exposed for 8 h and mated to untreated females 2 days later.
23      Although the frequency of sex-linked recessive lethals from treated males was not different from
24      that of controls, the limited sample size precluded detecting  less than a threefold increase over
25      controls. The authors noted that, because there were no signs of toxicity, the flies might tolerate
26      exposures to higher concentrations for longer time periods.
27            Specific-locus mutations were not induced in (C3H * 101)F, male mice exposed to diesel
28      exhaust 8 h/day, 7 days/week for either 5 or 10 weeks (Russell et at, 1980). The exhaust was a
29      1:18 dilution and the average particle concentration was 6 mg/m3. After exposure, males were
30      mated to T-stock females and matings continued for the reproductive life of the males. The
31      results were unequivocally negative; no mutants were detected in 10,635 progeny derived from
32      postspermatogonial cells or in 27,917 progeny derived from spermatogonial cells.
33
        2/1/98               •                      9-3        DRAFT-DO NOT CITE OR QUOTE

-------
 1      9.2. CHROMOSOME EFFECTS
 2            Mitchell et al. (1981) and Brooks et al. (1984) reported increases in sister chromatid
 3      exchanges (SCE) in CHO cells exposed to DPM extracts of emissions from b'oth LD and HD
 4      diesel engines. Morimoto et al. (1986) observed increased SCE from both LD and HD DPM
 5      extracts in PAH-stimulated human lymphocyte cultures. Tucker et al. (1986) exposed human
 6      peripheral lymphocyte cultures from four donors to direct diesel exhaust for up to 3 h. Exhaust
 7      was cooled by pumping through a plastic tube about 20 feet long; airflow was 1.5 L/min.
 8      Samples were taken at 16, 48, and  160 min of exposure. Cell cycle del.ay was observed in all
 9      cultures; significantly increased SCE levels were reported for two of the four cultures. Structural
10      chromosome aberrations were induced in CHO cells by DPM extracts from a Nissan diesel
11      engine (Lewtas, 1983) but not by similar extracts from an Oldsmobile diesel engine (Brooks et
12      al., 1984).
13            Pereira et al. (198la) exposed female Swiss mice to diesel exhaust 8 h/day, 5 days/week
14      for 1, 3, and 7 weeks.  The incidence of micronuclei and structural aberrations was similar in
15      bone marrow cells of both control and exposed mice.  Increased incidences of micronuclei, but
16      not SCE, were observed in bone marrow cells of male Chinese hamsters after 6 months of
17      exposure to diesel exhaust (Pereira etal., 1981 b).
18            Guerrero et al. (1981) observed a linear concentration-related increase in SCE in lung
19      cells cultured after intratracheal instillation of DPM at doses up to 20 ing/hamster. However,
20      they did not observe any increase in SCE after. 3 months of inhalation exposure to diesel exhaust
21     . particles (6 mg/m3).
22            Pereira et al. (1982) measured SCE in embryonic liver cells of Syrian hamsters.  Pregnant
23      females were exposed to diesel exhaust (containing about 12 mg/m3 particles) from days 5 to 13
24      of gestation or injected intraperitoneally with diesel particles or particle extracts on gestational
25      day 13 (18 h before sacrifice).  Neither the incidence of SCE nor mitotic index was affected by
26      exposure to diesel exhaust.  The injection of DPM extracts but not DPM resulted in a dose-
27      related increase in SCE; however, the toxicity of the DPM was about twofold greater than the
28      DPM extract.
29            In the only studies with mammalian germ cells, Russell et al. (1980) reported no increase
30      in either dominant lethals or heritable translations in males of T-stock mice exposed by
31      inhalation to diesel emissions.  In the dominant lethal test, T-stock males were exposed for 7.5
32      weeks and immediately mated to females of different genetic backgrounds (T-stock; [C3H x
33      101]; [C3H x C57BL/6]; [SEC x C57BL/6]). There were no differences from controls in any of
34      the parameters measured in this assay. For heritable translocation analysis, T-stock males were
35      exposed for 4.5 weeks and mated to (SEC x C57BL/6) females, and the Ft males were tested for

        2/1/98                                     9-4        DRAFT-DO NOT CITE OR QUOTE

-------
  1      the presence of heritable translocations. Although no translocations were detected among 358
  2      progeny tested, the historical control incidence is less than 1/1,000.
  3
  4      9.3. OTHER GENOTOXIC EFFECTS
  5             Pereira et al. (1981b) exposed male strain A mice to diesel exhaust emissions for 31  or 39
  6      weeks using the same exposure regimen noted in the previous section. Analyses of caudal sperm
  7      for sperm-head abnormalities were conducted independently in three separate laboratories.
  8      Although the incidence of sperm abnormalities was not significantly above controls in any of the
  9      three laboratories, there were extremely large differences in scoring among the three (control
10      values were 9.2%, 14.9%, and 27.8% in the three laboratories).  Conversely, male Chinese
11      hamsters exposed for 6 mo (Pereira et al., 1981c) exhibited almost a threefold increase in sperm-
12      head abnormalities. It is noted that the control incidence in the Chinese hamsters was less than
13      0.5%. Hence, it is not clear whether the differing responses reflect true species differences or
14      experimental artifacts.
15
16      9.4. SUMMARY
17             Extensive studies with Salmonella have unequivocally demonstrated mutagenic activity
18      in both particulate and gaseous fractions of diesel exhaust.  In most of the studies using
19      Salmonella, DPM extracts and individual nitropyrenes exhibited the strongest responses in strain
20      TA98 when no exogenous activation was provided. Gaseous fractions reportedly showed greater
21      response in TA100, whereas benzo[a]pyrene and other unsubstituted PAHs are mutagenic only in
22      the presence of S9 fractions. The induction of gene mutations has been reported in several in
23      vitro mammalian cell lines after exposure to extracts of DPM. Note that only the TK6 human
24      cell line did not give a positive response to DPM extracts in the absence of S9 activation.
25      Mutagenic activity was recovered in urine from animals treated with DPM by gastric intubation
26      and i.p. and s.c. implants, but not by inhalation of DPM or diluted diesel exhaust. Dilutions of
27      whole diesel exhaust did not induce sex-linked recessive lethals in Drosophila or specific-locus
28      mutations in male mouse germ cells.
29             Structural chromosome aberrations and SCE in mammalian cells have been induced by
30      particles and extracts.  Whole exhaust induced micronuclei but not SCE or structural aberrations
31      in bone marrow of male Chinese hamsters exposed to whole diesel emissions for 6 mo. In a
32      shorter  exposure (7 weeks), neither micronuclei nor structural aberrations were increased in bone
33      marrow of female Swiss mice.  Likewise, whole diesel exhaust did not induce dominant lethals
34      or heritable translocations in male mice exposed for 7.5 and 4.5 weeks, respectively.
        2/1/98                                    9-5        DRAFT-DO NOT CITE OR QUOTE

-------
  1             Mutagenicity data have been applied both to issues of heritable genetic risk and somatic
  2      cell effects, most notably cancer. For heritable genetic effects, the U.S. Environmental
  3      Protection Agency's Guidelines for Mutagenicity Risk Assessment (U.S. EPA, 1987) are
  4      applicable here. The mammalian germ-cell studies measuring defined genetic endpoints
  5      conducted on diesel emissions have shown negative results, but the sample size in the heritable
  6      translocation test is too small for a meaningful conclusion. In the absence of definitive
  7      mammalian germ-cell results, the guidelines recommend that mutagenic activity and the ability
  8      to interact with mammalian germ cells be evaluated separately. As stated, the presence of a large
  9      number of mutagenic chemicals in diesel emissions is unambiguous.  Sperm abnormality  assays
10      are presumably the only other source of data on the interaction of diesel emissions with
11      mammalian germ cells.  The negative response in the mouse is in apparent conflict with the
12      positive observation in the hamster, and there is not sufficient information to resolve this
13      discrepancy.  Hence, the questions of germ-cell interaction and the potential for human germ-cell
14      mutagenic risk of diesel emissions remain unanswered.
15             The application of genotoxicity information to the question of the potential
16      carcinogenicity of chemical agents was initially based on the premise that somatic mutation is an
17      integral step in the carcinogenic process.  However, unlike the situation for germ-cell
18      mutagenicity, assays are not weighted strictly by their biological relationship to the particular
19      species, sex, or tissue site of cancer. The size of the database and the degree of correlation of
20      genotoxicity test results with animal cancer bioassays are frequently given great weight. Indeed,
21      a common conclusion of the efforts of the National Toxicology Program on the use of in vitro
22      assays is that no single in vitro genotoxicity test or battery of tests (among the four assays in their
23      program) improves on the performance of the Salmonella assay in predicting rodent
24      carcinogenicity of an untested chemical.  When rodent carcinogenicity data are available,
25      phylogenetic and other biological aspects of the genotoxicity data are important considerations in
26      the weight-of-evidence process. With diesel emissions, additional complications arise because of
27      the chemical complexity of the material being tested. Although it is clear that several of the
28    . individual chemical constituents of diesel exhaust have been demonstrated to be both mutagenic
29      and carcinogenic, it is likely that the constituents responsible for the mutational increases
30      observed in bacteria are different from those responsible for the observed increases in CHO cells
31      (Li and Dutcher, 1983) or in human hepatoma-derived cells (Eddy et al., 1986). Chapter  10 deals
32      more thoroughly with metabolism and mechanisms of carcinogenesis.
        2/1/98  .                                   9-6        DRAFT-DO NOT CITE OR QUOTE

-------
  1      9.5. REFERENCES
  2
  3      Barfknecht, TR; Kites, RA; Cavaliers, EL; Thilly, WG. (1.982) Human cell mutagenicity of polycyclic aromatic
  4      hydrocarbon components of diesel emissions. In: Lewtas, J, ed, Toxicological effects of emissions from diesel
  5      engines: proceedings of the Environmental Protection Agency 1981 diesel emissions symposium; October 1981;
  6      Raleigh, NC.  (Developments on toxicology and environmental science: v, 10.) New York: Elsevier Biomedical; pp.
 '• 7      277-294.
  8
  9      Belisario, MA; Buonocore, V; De Marinis, E; De Lorenzo, F, (1984) Biological availability of mutagenic
 10      compounds adsorbed onto diesel exhaust paniculate. Mutat Res 135:1-9.
 11
 12      Brooks, A; Li, AP; Dutcher, JS; Clark, CR; Rothenberg, SJ; Kiyoura, R; Bechtold, WE; McClellan, RO. (1984) A
 13      comparison of genotoxicity of automobile exhaust particles from laboratory and environmental sources. Environ
 14      Mutagen 6:651-668.                                                                     '
 15                              '
 16      Casto, BC; Hatch, GG; Huang, SL; Huisingh, JL; Nesnow, S; Waters, MD. (1981) Mutagenic and carcinogenic
 17'      potency of extracts of diesel and related environmental emissions: in vitro mutagenesis and oncogenic
 18      transformation. Environ Int 5:403-409.
 19
 20      Chescheir, GM, III; Garrett, NE; Shelburne, JD; Huisingh, JL; Waters, MD. (1981) Mutagenic effects of
 21      environmental particulates in the CHO/HGPRT system. In: Waters, MD; Sandhu, SS; Huisingh, JL; Claxton, LC;
 22      Nesnow, S, eds. Application of short-term bioassays in the fractionation and analysis of complex environmental
 23      mixtures. New York: Plenum Press; pp. 337-350.
 24
 25      Claxton, LD. (1981) Mutagenic and carcinogenic potency of diesel and related environmental emissions:
 26      Salmonella bioassay. Environ Int 5:389-391.
 27
 28      Claxton, LD. (1983) Characterization of automotive emissions by bacterial mutagenesis bioassay: a review. Environ
 29      Mutagen 5:609-631.
 30                 .                                               '
 31      Claxton, L; Kohan, M. (1981)  Bacterial mutagenesis and the evaluation of mobile-source emissions. In: Waters,
 32     • MD; Sandhu, SS; Huisingh, JL; Claxton, L; Nesnow, S, eds. Short-term bioassays in- the analysis of complex
 33      environmental mixtures II:  proceedings of the second symposium on the application of short-term bioassays in the
 34      fractionation and analysis of complex environmental mixtures; March 1980; Williamsburg, VA. (Hollaender, A;
 35      Welch, BL; Probstein, RF, eds. Environmental science research series: v. 22.) New York: Plenum Press; pp. 299-
 36      317.
 37
 38      Curren, RD; Kouri, RE; Kim, DM; Schectman, LM. (1981) Mutagenic and carcinogenic potency of extracts from
 39      diesel related environmental emissions: simultaneous morphological transformation and mutagenesis in BALB/c
40     '3T3 cells. Environ Internal 5:411-415.
41
42      Dukovich, M; Yashin, RE;  Lestz, SS; Risby, TH; Zweidinger, RB. (1981)  The mutagenic and SOS-inducing
43      potential of the soluble organic fraction collected from diesel paniculate emissions. Environ Mutagen 3: 253-264.
44
45      Eddy, EP; McCoy, EC; Rosenkranz, HS; Mermelstein, R. (1986) Dichotomy in the mutagenicity and genotoxicity
46      of nitropyrenes: apparent effect of the number of electrons involved in nitroreduction. Mutat Res 161:109- 111.
47
48      Guerrero, RR; Rounds, DE; Orthoefer, J. (1981) Sister chromatid exchange analysis of Syrian hamster lung cells
49      treated in vivo with diesel exhaust particulates. Environ Int 5:445-454.
 50
 51      Hinners, RG; Burkart, JK; Malanchuk, M. (1980) Facilities for diesel exhaust studies. In: Pepelko, WE; Danner,
 52      RM; Clarke, NA, eds.  Health effects of diesel engine emissions: proceedings of an international symposium;
 53      December 1979. Cincinnati, OH: U.S. Environmental Protection Agency, Health Effects Research Laboratory; pp.
54      681-697; EPA report no. EPA-600/9-80-057b. Available from: NTIS, Springfield, VA; PB81-173817.


         2/1/98                                          9-7         DRAFT-DO NOT CITE OR QUOTE

-------
  1       Huisingh, J; Bradow, R; Jungers, R; Claxton, L; Zweidinger, R; Tejada, S; Bumgarner, J; Duffield, F; Waters, M;
  2       Simmon, VF; Hare, C; Rodriguez, C; Snow, L. (1978) Application of bioassay to the characterization of diesel
  3       particle emissions. In: Waters, MD; Nesnow, S; Huisingh, JL; Sandhu, SS; Claxton, L, eds. Application of short-
  4       term bioassays in the fractionation and analysis of complex environmental mixtures: [proceedings of a symposium;
  5       February; Williamsburg, VA].  (Hollaender, A; Probstein, F; Welch, BL, eds. Environmental science research: v.
  6       15.) New York: Plenum Press; pp. 383-418.
  7                      .
  8       International Agency for Research on Cancer. (1989) Diesel and gasoline engine exhausts and some nitroarenes.
  9       (IARC monographs on the evaluation of carcinogenic risks to humans: v. 46).  Lyon, France: World Health
10       Organization; pp. 41-185.
11
12       Ishinishi, N; Koizumi, A; McClellan, RO; Stober, W, eds. (1986) Carcinogenic and mutagenic effects of diesel
13       engine exhaust: proceedings of the international satellite symposium on toxicological effects of emissions from
14       diesel engines; July; Tsukuba Science City, Japan. (Developments in toxicology and environmental science: v. 13.)
15       Amsterdam: Elsevier Science Publishers BV.
16.
17       Lewtas,  J. (1982) Mutagenic activity of diesel emissions. In: Lewtas, J, ed. Toxicological effects of emissions from
18       diesel engines: proceedings of the Environmental Protection Agency 1981 diesel emissions symposium; October
19       1981; Raleigh, NC. (Developments in toxicology and environmental science: v. 10.)  New York: Elsevier
20       Biomedical; pp. 243-264.
21
22       Lewtas,  J. (1983) Evaluation of the mutagenicity and carcinogenicity of motor vehicle emissions in short-term
23       bioassays. Environ Health Perspect 47:141-152.
24             '
25       Li, AP; Dutcher, JS. (1983) Mutagenicity of mono-, di-, and tri-nitropyrenes in Chinese hamster ovary cells. Mutat
26       Res 119:387-392.
27
28       Li, AP; Royer, RE. (1982) Diesel-exhaust-particle extract enhancement of chemical-induced mutagenesis in
29       cultured Chinese  hamster ovary cells: possible interaction of diesel exhaust with environmental chemicals. Mutat
30       Res 103:349-355.
31
32       Liber, HL; Andon, BM; Kites, RA; Thilly, WG. (1981) Diesel soot: mutation measurements in bacterial and human
33       cells. Environ Int 5:281-284.
34
35       Liberti, A; Ciccioli, P; Cecinato, A; Brancaleoni, E; Di Palo, C. (1984) Determination of nitrated-polyaromatic
36       hydrocarbons (nitro-PAHs) in environmental samples by high resolution chromatographic techniques. J High
37       Resolut  Chromatogr Commun 7:389-397.
38
39       Matsushita, H; Goto, S;  Endo, O; Lee, JH; Kawai, A. (1986) Mutagenicity of diesel exhaust and related chemicals.
40       In: Ishinishi, N; Koizumi, A; McClellan, RO;  Stober, W, eds. Carcinogenic and mutagenic effects of diesel engine
41       exhaust: proceedings of the international satellite symposium on toxicological effects of emissions from diesel
42       engines; July; Tsukuba Science City, Japan. (Developments on toxicology and environmental science: v. 13.)
43       Amsterdam: Elsevier  Science Publishers BV;  pp. 103-118.
44
45       Mitchell, AD; Evans,  EL; Jotz, MM; Riccio, ES; Mortelmans, KE; Simmon, VF. (1981) Mutagenic and
46       carcinogenic potency  of extracts of diesel and related environmental emissions: in vitro mutagenesis and DNA
47       damage. Environ Int 5:393-401.
48
49       Morimoto, K; Kitamura, M; Kondo, H; Koizumi, A. (1986) Genotoxicity of diesel exhaust emissions in a battery of
50       in-vitro  short-term bioassays. In: Ishinishi, N; Koizumi,  A; McClellan, RO;  Stober, W, eds. Carcinogenic and
51       mutagenic effects of diesel engine exhaust: proceedings of the international  satellite symposium on toxicological
52       effects of emissions from diesel engines; July; Tsukuba Science City, Japan. (Developments in toxicology and
53       environmental science: v. 13.) Amsterdam: Elsevier Science Publishers BV; pp. 85-102.
54


         2/1/98                                            9-8         DRAFT-DO NOT  CITE OR QUOTE

-------
  1      Nakagawa, R; Kitamori, S; Horikawa, K; Nakashima, K; Tokiwa, H. (1983) Identification of dinitropyrenes in
  2      diesel-exhaust particles: their probable presence as the major mutagens. Mutat Res 124:201-211.
  3
  4      Nishioka, MG; Petersen, BA; Lewtas, J. (1982) Comparison of nitro-aromatic content and direct-acting
  5      m'utagenicity of diesel emissions. In: Cooke, M; Dennis, AJ; Fisher, GL, eds. Polynuclear aromatic hydrocarbons:
  6    .  physical and biological chemistry. Columbus, OH: Battelle Press; pp. 603-613.
  7
  8   .   Pepelko,  WE; Peirano, WB. (1983) Health effects of exposure to diesel engine emissions: a summary of animal
  9      studies conducted by the U.S. Environmental Protection Agency's Health Effects Research Laboratories at
10      Cincinnati, Ohio. J Am Coll Toxicol 2:253-306.
11
12      Pereira, MA; Connor, TH; Meyne, J; Legator, MS, (198la) Metaphase analysis, micronucleus assay and urinary
13      mutagenicity assay of mice exposed to diesel emissions. Environ Int 5:435-438.
14
15      Pereira, MA; Sabharwal, PS; Gordon, L; Wyrobek, AJ. (1981b) The effect of diesel exhaust on sperm-shape
16      abnormalities in mice. Environ Int 5:459-460.
17
18      Pereira, MA; Sabharwal, PS; Kaur, P; Ross, CB; Choi, A; Dixon, T, (1981c) In Vivo detection of mutagenie effects
19      of diesel exhaust by short-term mammalian bioassays. Environ Int 5:439-443.
20
21      Pereira, MA; McMillan, L; Kaur, P; Gulati, DK; Sabharwal, PS. (1982) Effect of diesel exhaust emissions,      ^  •
22      particulates, and extract on sister chromatid exchange in transplacentally exposed fetal hamster liver. Environ
23      Mutagen  4:215-220.
24
25      Russell, LB; Generoso, WM; Oakberg, EF; Russell, WL; Bangham, JW; Stelzner, KF, (1980) Tests for heritable
26      effects induced by diesel exhaust in the mouse. Martin Marietta, Energy Systems, Inc., Oak Ridge National
27      Laboratory; report no. ORNL-5685.
28                                                                                                  .
29      Salmeen, I; Durisin, AM; Prater, TJ; Riley, T; Schuetzle, D. (1982) Contribution of 1-nitropyrene to direct-acting
30      Ames assay mutagenicities of diesel paniculate extracts. Mutat Res 104:17-23.
31
32      Schuetzle, D; Frazier, JA. (1986) Factors influencing the emission of vapor and paniculate phase components from
33      diesel engines. In: Ishinishi, N; Koizumi, A; McClellan, RO; StOber, W, eds. Carcinogenic and mutagenie effects of
34      diesel engine exhaust: proceedings of the international satellite symposium on toxicological effects of emissions
35      from diesel engines; July; Tsukuba Science City, Japan. (Developments in toxicology and environmental science: v.
36      13.) Amsterdam: Elsevier Science Publishers BV; pp. 41-63.
37
38      Schuetzle, D; Lewtas, J. (1986) Bioassay-directed chemical analysis in environmental research. Anal Chem
39      58:1060A-1076A.
40
41      Schuetzle, D; Perez, JM. (1983) Factors  influencing the emissions of nitrated-polynuclear aromatic hydrocarbons
42      (nitro-PAH) from diesel engines. J Air Pollut Control Assoc 33:751-755.
43
44    "  Schuler, RL; Niemeier, RW. (1981) A study of diesel emissions on Drosophila. Environ Int 5:431-434.
45
46      Siak, JS; Chan, TL; Lees, PS. (1981) Diesel paniculate extracts in bacterial test systems. Environ Int 5:243-248.
47
48      Tucker, JD; Xu, J; Stewart, J; Baciu, PC; Ong, T-M. (1986) Detection of sister chromatid exchanges induced by
49      volatile genotoxicants. Teratog Carcinog Mutagen 6:15-21.
50
51      U.S. Environmental Protection Agency.  (1980) Health effects of diesel engine emissions: proceedings of an
52      international symposium. Cincinnati, OH: Office of Research and Development; EPA 600/9-80-057b.
53
         2/1/98        .                                  9-9         DRAFT-DO NOT CITE OR QUOTE

-------
1      U.S. Environmental Protection Agency. (1987) The risk assessment guidelines of 1986. Washington, DC: Office of
2      Health and Environmental Assessment; EPA/600/8-87/045.
       2/1/98                '                      9-10        DRAFT-DO NOT CITE OR QUOTE

-------
                   10.  METABOLISM AND MECHANISM OF ACTION IN DIESEL
                             EMISSION-INDUCED CARCINOGENESIS

  1            Considerable research has been directed toward assessing the carcinogenic potential of
  2      diesel engine emissions. As indicated in Chapter 7, whole diesel exhaust (DE) is a pulmonary
  3      carcinogen in rats subjected to chronic exposures at high concentrations. This response to date
  4      has been clearly demonstrated only in rats, and apparently involves particle overload resulting in
  5      inflammation and proliferation of alveolar epithelial cells and subsequent rumor formation.
  6      Studies assessing diesel exhaust-induced carcinogenicity in hamsters were negative, and studies
  7      in mice were equivocal (depending on the strain). The organic components that are adsorbed to
  8      the diesel exhaust particle do not appear to be required for expression of the high-dose
  9      tumorigenic response in rats.  Positive responses are also observed following exposure to carbon
10      black (CB) and Ti02 particles that lack the organic components.  The organic components,
11      however, are likely to play a greater role in tumorigenic responses at lower exposure
12      concentrations. In susceptible mouse strains and in humans exposed at low concentrations,
13      genotoxic mechanisms induced by the organic components may even play a predominant role.
14      Epidemiologic data suggest that there is a small increased cancer risk in humans following long-
15      term occupational exposure to diesel exhaust.  In examining mechanisms of diesel exhaust-
16      induced carcinogenicity, it is  necessary to address several areas, including (1) the carcinogenic
17      potential of the carbon particle and the particle-overload effect, (2) the metabolism and
18      mechanism of action of known carcinogenic components such as benzo[a]pyrene (B[a]P) and
19      various nitroarenes, (3) the role of pulmonary leukocytes, and (4) DNA adduct formation in
20      diesel exhaust exposures.
21                      •                  '                            •    .
22      10.1. PARTICLE-INDUCED CARCINOGENIC RESPONSE
23            DE is,a pulmonary carcinogen in rats chronically exposed to high concentrations
24      (Heinrich et al, 1986; Mauderly et al., 1987).  Additional studies (Vostal, 1986; Kawabata et ah,
25      1986; Heinrich, 1990;  Wolff et al., 1990; Oberddrster and Yu, 1990) provided data indicating
26      that the carbonaceous core was a major factor in this response. In addition to diesel exhaust
27      paniculate matter (DPM), particle-specific pulmonary carcinogenesis in rats has been
28      demonstrated for particles with virtually no adsorbed organics, such as CB (Heinrich et al.,  1994,
29      1995; Nikula et al., 1995) and particles with no organic component,  such as TiO2 (Lee et al.,
30      1985, 1986; Heinrich et al., 1995). Results of studies (see Chapter 7) from both the Inhalation
31      Toxicology Research Institute (Nikula et al., 1995) and the Fraunhofer Institute of Toxicology
32      and Aerosol Research (Heinrich et al., 1995) have affirmed the instrumental role of the carbon
33      core in producing a carcinogenic response in rats chronically exposed to high concentrations (>2
       •2/1/98                                    10-1       DRAFT-DO NOT CITE OR QUOTE

-------
 1      mg/nr) of whole diesel exhaust.  The results of these studies clearly indicate that a particle
 2      overload effect is primarily responsible for this response. The rat alveolar epithelium may be
 3      predisposed to proliferative, metaplastic, and neoplastic responses, but the underlying mechanism
 4      for this effect is not clear. This section briefly reviews data affirming the particle effect, as well
 5      as data that provide some insight into possible mechanisms for this response.
 6             A preliminary report by Heinrich (1990) provided evidence for a particle effect. In this
 7      study, female Wistar rats (72 per group) were exposed to Printex 90 CB particles for 10 mo
 8      followed by a 20-mo exposure-free observation period or for 20 months followed by a 10-mo
 9      exposure-free observation period. A particle concentration of 6.09 mg/m3 was used in both
10      protocols.  The Printex 90 particles had an extremely low organic content (~ 1,000-fold less than
11      that of DPM). The tumor rates for the 10- and 20-mo exposure durations were 17% (14%
12      malignant) and 8% (all malignant), respectively.  Although the lower tumor incidence for the
13      longer exposure period was not consistent, the results demonstrate that the tumor incidences for
14      CB particles with an organic content 1,000-fold less than DPM are equivalent to those reported
15      for diesel exhaust exposures.  The fact that these particles were able to exert a significant
16      tumorigenic response implicates the carbon core of diesel exhaust particles as the possible tumor
17      initiator in diesel exhaust-induced carcinogenicity at high particle concentrations.
18             More recently, an extensive study at the Fraunhofer Institute of Toxicology and Aerosol
19      Research assessed the tumorigenic potential of DPM, the carbon core of DPM, CB, and TiO, in
20      rats and two strains of mice (Heinrich et al., 1995). In this study, Wistar rats and NMRI mice
21      were exposed to diesel engine whole exhaust, CB particles (Printex 90), and ultrafine TiO,
22      particles for 2 years and to clean air for an additional 6 months. The results showed that when
23      incidence of either all tumors or benign keratinizing cystic squamous-cell tumors was excluded,
24      CB was at least as potent as DE when lung particle burdens were comparable (see Chapter 7 for
25      details). DE was more potent on the basis of administered dose. The latter relationship can be at
26      least partially explained by the slower pulmonary clearance of DPM and suggests that DE may be
27      more toxic to phagocytic cells responsible for particle clearance.
28             Research efforts at the Inhalation Toxicology Research Institute (ITRJ) have also
29      provided data regarding the carcinogenic potential of whole diesel exhaust and of CB particles.
30      In a long-term study,  rats were exposed 16 h/day, 5 days/week for 24 mo to whole DE or CB
31      (free of adsorbed organics) at particle concentrations of 2.5 or 6.5 mg/m3 (Mauderly et al., 1991;
32      Nikulaet al.,  1995). Controls were exposed to clean air.  Al though-the CB particles were not
33      totally devoid of organic components, the solvent-extractable fraction was small and the CB
34      mutagenicity per unit of particle mass was three orders of magnitude less than that of diesel
35      exhaust soot.  Lung weights were increased in rats exposed to the highest concentrations of both
36      diesel exhaust or CB but were slightly higher for the  diesel exhaust group. The lung burdens of
        2/1/98                                    10-2        DRAFT-DO NOT CITE OR QUOTE

-------
  1      participate matter were significantly greater for the diesel exhaust-exposed rats at 18 and 23 mo.
  2      A substantial transfer of particles from the lungs to lung-associated lymph nodes was observed.
  3      but no difference was noted between the diesel exhaust and CB exposure groups.  Inflammation
  4      and cytotoxicity detected in lavage fluid were greater for diesel exhaust-exposed rats, but the
  5      difference was proportional to the higher lung burden of retained particles noted for these
  6      animals.  Tumor incidences and prevalence were similar for diesel soot-exposed and CB-exposed
  7      rats.  Tumor types observed included squamous cysts, squamous cell carcinomas, papillary
  8      adenocarcinomas, tubular adenocarcinomas, and solid carcinomas.  The growth of tumors
  9      transplanted into athymic mice also has been similar for diesel exhaust and CB exposures (74%
10      and 73%, respectively),
11            Additional support for the particle-overload effect has been provided by data showing
12      similar lung tumor rates in rats following intratracheal instillation of diesel exhaust soot or CB
13      particles (Pott et al., 1994). Total primary lung tumors (expressed as percent and including
14      adenomas, adenocarcinomas, cystic keratinizing squamous cell tumors, and squamous cell
15      carcinomas) for the three DE groups were 65%, 60%, and 66%, whereas total tumors for the CB
16      group were 65%. In addition to similar tumor rates, tumors induced by DPM and CB exhibited
                                         (
17      similar histopathologic patterns. Finally, Kawabata et al. (1993)  induced lung tumors in rats
18      intratracheally instilled with DPM from which the organic components had been extracted.
19            Wolff et al. (1990) reported the results of a study comparing pulmonary inflammation and
20      DMA adduct formation in rats exposed 7 h/day, 5 days/week for 12 weeks to diesel exhaust or
21      CB at concentrations of 10 mg/m3. Although the level of lung DNA adducts was slightly higher
22      for DE exposure, both exposures resulted in inflammatory responses, as determined by increased
23      numbers of neutrophils and macrophages and increased acid proteinase in the bronchoalveolar
24      lavage fluid.
25            Oberdorster and Yu (1990) evaluated the relationship between tumorigenic response of
26      the lung and physical characteristics of various insoluble particles. Based on data from studies
27      examining the effects of long-term inhalation exposure to DE TiO2 particles, CB, or toner
28      particles,  they found that only the surface area of retained particles in the lung showed a
29      reasonable concentration-response relationship relative to tumor incidence. Based upon this
30      information, particle overload (retained mass or volume of particles) may not be the only
31      determining factor in lung tumor formation for insoluble particles. The investigators
32      hypothesized that a tumorigenic effect would probably require that a "critical" surface area of
33      retained particles be attained for the manifestation of any mechanisms of tumorigenicity.
34            In comparing the potency of DE and CB, therefore, the particle surface area requires
35    '  consideration. The Printex particles used in the study by Heinrich et al. (1995) have a large
36      surface area (=227 nr/g), whereas the surface area of DPM varies from 18 (unextracted organics)
        2/1/98              .                     10-3       DRAFT-DO NOT CITE OR QUOTE

-------
 1      to 107 nr/g (organies fully extracted). Although the actual mean surface area of DPM in the
 2      lungs is uncertain, it is less than that of the Printex particles.  It could be hypothesized that
 3      particle effects from diesel exposure are less than those of CB because of a smaller specific
 4      surface, but the difference in potency is made up by surface-associated organies.  Similar
 5      conclusions might also be derived from the intratracheal studies reported by Pott et al. (1994).
 6      Proof for this hypothesis may be difficult to achieve because surface area of the diesel particle
 7      would be expected to increase as organies are eluted from the particle surface during residence
 8      time in the lungs.
 9            The involvement of persistent alveolar epithelial hyperplasia appears  to be associated
10      with the induction of nedplasia by particles. Oberdorster et al. (1995) reported an inflammatory
11      influx of neutrophils in rats exposed for 3 mo to CB at concentrations of 7 or 50 mg/m3 but not at
12      1 mg/m3. Data showing exposure-response relationships for cytokine expression, neutrophilic
13      inflammation, and dose-dependent alveolar Type II epithelial cell mutagenesis in rat lungs
14      following 13-week exposure to CB (7 or 50 mg/m3) were reported by Driscoll et al.  (1995).
15      Although these effects are seen at lower exposure concentrations as particle size decreases and
16      specific surface area increases, the effects of ultrafine particles at less than particle overload
17      conditions is uncertain.
18            In summary, results of these studies indicate that little difference exists in the type or
19      incidence of lung tumors in rats following long-term exposure to DE or CB at high
20      concentrations, that particle-associated organies are likely to play a minor role in lung particle
                                                             •             •
21      overload-induced pulmonary carcinogenicity of DE in rats, and that particle effects are clearly
22      relevant to the  carcinogenic response observed.  Persistent inflammation and hyperplasia of the
23      alveolar epithelium, cytokine release, and release of oxidant reactants by alveolar rnacrophages
24      (see Section 10.3) appear to be key elements leading to the species-specific neoplastic condition.
25      The mechanistic basis for the surface-area-associated effects on cancer potency is as yet
26      unknown.                                             •
27
28      10.2. METABOLISM AND MECHANISM OF ACTION OF ORGANIC
29           CARCINOGENIC COMPONENTS OF DIESEL EXHAUST
30            DE is a complex mixture containing gaseous-phase components as well as soot particles
31      to which more  than 450 organic compounds are adsorbed (Opresko et al., 1984; Nikula et al.,
32      1995). Although the involvement of particle-adsorbed organies in pulmonary carcinogenic
33      responses appears to be minor in rats under particle overload conditions, the contribution of a
34      subtle direct-mutagen effect should not be summarily dismissed, especially for humans.
35            In other species, such an effect may be more relevant, although currently available animal
36      data and epidemiologic data do not support such a contention. A direct mutagen effect may be
        2/1/98                                    10-4       DRAFT-DO NOT CITE OR QUOTE

-------
  1      inconsequential in rats relative to particle overload effects because bioavailability of the organic
  2      components is limited by the high-energy binding of these components with the carbon core,
  3      resulting in low desorption in the biological environment. The long residence times of particles
  4      in human lungs, however, may allow for greater desorption of organics,
  5            The mechanism of action of many PAH carcinogens has been attributed to the reactivity
  6      of certain metabolic intermediates with cellular macromolecules and the subsequent formation of
  7      DMA adducts. The organics adsorbed to DPM may become available for biotransformation to
  8      known reactive intermediates, and macromolecular binding of these metabolites has been
  9      demonstrated.
10            Except for some of the DMA adduct studies, the available database does not allow a
11      definitive discussion of the specific mechanism of carcinogenic action for these compounds
12      relative to diesel exhaust specifically but rather is approached from the standpoint of the
13      chemicals per se.  Some of the data are derived primarily from in vitro studies that were not  ,
14      specifically concerned with the potential carcinogenicity of diesel exhaust but may be relevant
15      because the compounds of concern are known components of diesel emissions.
16            More than 100 carcinogenic or potentially carcinogenic components have been
17      specifically identified in diesel emissions, including various PAHs and nitroarenes such as 1-
18      nitropyrene (1 -NP) and dinitropyrenes (DNPs). These compounds are adsorbed to the carbon
19      core of the paniculate phase of the exhaust and, if desorbed, may become available for biological
20      processes such as metabolic activation to mutagens. Among compounds identified from diesel
21      exhaust are B[a]P, dibenz[a,/i]anmracene, pyrene, chrysene, and nitroarenes such as 1-NP, 1,3-
22      DNP,  1,6-DNP, and 1,8-DNP, all of which are mutagenic, carcinogenic, or implicated as
23    » procarcinogens or cocarcinogens (Stenback et al., 1976; Weinstein and Troll, 1977; Thyssen et
24      al., 1981; Pott and Stober, 1983; Howard et al., 1983; Hirose et al., 1984; Nesnow et al.,  1984;
25      El-Bayoumy et al., 1988).
26            There is evidence  supporting a carcinogenic role for organics in the combustion process.
27      Mumford et al. (1989) reported greatly increased lung cancer mortality in Chinese communes
28      burning so-called  "smoky coal," but not wood or smokeless coal, in unvented open-pit fires used
29      for heating and cooking.  Particle concentrations ranged from 10 to 25 mg/m3 in communes
30      burning either smoky coal or wood, but PAH levels were five to six times greater in the air of
31      communes burning smoky coal.  Thus cancer mortality correlated more closely with
32      concentrations of PAHs than with particles. In the case of smokeless coal, both particle and PAH
33      concentrations were low.  Demonstration of the carcinogenicity of coke oven emissions in
34      humans (Lloyd, 1971) also provided evidence for the role of organics, because coke oven
35      particulate matter lacks an insoluble carbon core.  It should be recognized, however, that PAH
36      concentrations in  these cases are much greater than can be expected from inhalation of DE.,
        2/1/98                                    10-5       DRAFT-DO NOT CITE OR QUOTE

-------
 1            Diesel particles may well enhance the activity of adsorbed organics. Adsorption of PAHs
 2      to a carrier particle such as hematite, CB, aluminum, or titanium dioxide enhances their
 3      carcinogenic potency (Farrell and Davis, 1974). In a more recent report, adsorption to carbon
 4      particles greatly enhanced the tumprigenicity of pyrolyzed pitch condensate containing B[a]P and
 5      other aromatic carcinogens (Heinrich et al., 1995).  The increased effectiveness can be partly
 6      explained by more efficient transport to the deep lung. Slow release also enhances residence
 7      time in the lungs and prevents overwhelming of activating pathways. Even though the  organic
 8      constituents may be tightly bound to the particle surface, clearance half-times of nearly 1 year in
 9      humans (Bohning et al., 1982) allow time for elution to occur. Furthermore, Gerde et al. (1991)
10      presented a model demonstrating that large aggregrates of inert dust containing crystalline PAHs
11      are unlikely to form at doses typical of human exposure. This allows the particles to deposit and
12      react with the surrounding lung medium, without interference from other particles. Particle-
13      associated PAHs can then be expected to be released more rapidly from the particles. Bond et al.
14      (1984) provided evidence that alveolar macrophages from beagle dogs metabolized B[a]P  coated
15      on diesel particles to proximate carcinogenic forms. Unless present on the particle surface,
16      BfaJP is more likely to pass directly into the bloodstream and escape activation by phagocytic
17      cells.
18            The importance of DE-associated PAHs in the induction of lung cancer in humans  may
19      be enhanced because of the possibility that the human lung is more sensitive to these compounds
                                   i
20      than rat lungs.  Rosenkranz (1996) summarized information indicating that in humans and mice,
21      large proportions of lung cancers contain both mutated p53 suppressor genes and K-ras genes.
22      Induction of mutations in these genes by genotoxins, however, is much lower in rats than in
23      humans or mice.  It could be even be hypothesized that lung tumors in diesel-exposed mice,
24      which apparently could be induced only when the animals were exposed from conception to
25      increase sensitivity (Pepelko and Peirano, 1983), may be due primarily to the organic fraction.
26                                     .
27      10.2.1. Metabolism and Disposition of B|a]P Relative to Diesel Exhaust
28            It is generally recognized that B[a]P is an activation-dependent carcinogen, with the
29   '   activated metabolites forming covalent DNA adducts (Boyland,  1980). The. reactions responsible
30      for this activation are mediated by the cytochrome P-450 monooxygenases and are known to
31      occur in multiple tissues and in different species.  The activation proceeds through Phase I
32      oxidative and hydrolytic reactions, which result in the formation of the ultimate carcinogenic
33      metabolite, B[a]P 7,8-dihydrodiol 9,10-epoxide. Specifically, B[ojP undergoes a mixed function
34      oxidase (MFO)-mediated epoxidation to form B[a]P-7,8-oxide which, in turn, is subjected to an
35      epoxide hydrolase-mediated hydrolysis resulting in the stereoisomeric diols (+)-B[a]P 7,8-

        2/1/98      '.                              10-6       DRAFT-DO NOT CITE OR QUOTE

-------
  1     dihydrodiol and (- )-B[a]P 7,8-dihydrodiol.  The diasterioisomeric forms of B[a]P 7.8-diol 9.10-
  2     epoxide are derived following another P-450-mediated reaction.
  3            Relative to DE carcinogenicity, several studies have examined the metabolism and
  4     disposition of constituents such as B[a]P. Mitchell (1982) subjected 24 male F344 rats to nose-
•  5     only inhalation oPH-B[a]P aerosol (500 mg/mj) for 60 min. High levels of radiolabel were
  6     detected in the trachea, lungs, and turbinates. Based on measurement of the radiolabel, biphasic
  7     clearance was noted with half-time (t!/2) values of 2 to 3 h and 25 to 56 h. Absorption by the
  8     lungs and systemic distribution were demonstrated by the presence of radiolabel in soft tissues,
  9     such as the liver, kidney, gastrointestinal tract, spleen, brain, and testes. The majority of the
10     radiolabel in these tissues was removed after 2 days, and the major route  of excretion was in the
11     feces.  The significance of this study is the demonstration of rapid absorption and systemic
12     distribution of B[a]P and potential metabolites following inhalation exposure.
13            Metabolization of intratracheally instilled B[a]P (1.0 ug) in strain A/J mice exposed to
14     diluted diesel exhaust (8 h/day, 7 days/week for 9 mo) was reported by Tyreretal. (1981). The
15     radiolabel (14C or 3H) was rapidly distributed throughout the body within 2 h. The highest levels
16     were detected in the lungs, liver, and gastrointestinal tract. Only trace levels were detected in the
17     gastrointestinal tract 168 h after administration. A companion study (Cantrell et al., 1980)
18     examined the effects of prior DE  exposure on in vivo B[a]P metabolism in the aforementioned
19     mice.  Homogenates of lung, liver, and testes were obtained from five mice sacrificed at 2, 24, or
20     168 h after B[a]P instillation. High-performance liquid chromatography (HPLC) analysis
21     detected free B[a]P and nonconjugated primary metabolites and sulfate, glucuronide, and
22     glutathione conjugates in each of the tissues. The occurrence of primary and secondary B[a]P
23     metabolites in all  three tissues was verified.  The major hepatic metabolite was 3-hydroxy-B[a]P.
24     The investigators  concluded that diesel exhaust exposure may qualitatively affect the metabolism
25     of B[a]P but does not significantly affect its distribution.
26            Sun et al. (1984) provided additional information comparing the disposition of particle-
27     adsorbed B[a]P (0.1 wt %) and pure B[a]P following 30 min of nose-only inhalation by F344
28     rats. Long-term lung retention (percentage retained after 7 days) of particle-adsorbed 3H-B[a]P
29     was approximately 230-fold greater than that for pure 3H-B[a]P. Pulmonary clearance of
30     particle-associated 3H was biphasic, with an initial t\/2 of 1 h and a second-phase ty2 of 18 days,
31     the latter representing clearance of 50% of the initially deposited radiolabel.  Clearance of pure
32     B[a]P  aerosol was >99% within 2 h and was apparently caused by pulmonary and mucous
33     membrane absorption into the blood rather than by mucociliary clearance and subsequent
34     ingestion (Sun et al.,  1982).  Of the radiolabel retained in the lungs, 65%  to 76% was B[a]P, 13%
35     to 17% was B[a]P-phenol, and 5% to 18% was B[a]P-quinone, Although the Sun et al. (1984)
36     study demonstrated the biotransformation of B[a]P to several metabolites, the epoxide
        2/1/98                                    10-7       DRAFT-DO NOT CITE OR QUOTE

-------
 1      intermediates known to be carcinogenic (Sims et al., 1974; Siaga et al., 1976) were not identified.
 2      However, B[a]P-phenol metabolites are reported to be mutagenic (Glatt and Oesch, 1976;
 3      Wislocki et al., 1986; Wood et al., 1976).
 4  .          Leung et al. (1988) studied the role of mierosomes in the removal and metabolism of
 5      B[a]P from DPM. Hepatic and lung microsomal preparations were made from 3-
 6      methylcholanthrene-induced F344 rats. l4Garbon-B[a]P was adsorbed to DPM (0.49 u£i/mg)
 7      and incubated with the microsomal preparations. Results indicated that both lung and liver
 8      mierosomes were capable of removing B[a]P from these modified exhaust particles and that this
 9      capacity was dependent on the lipid content of the mierosomes.  Only small (<3%)  amounts of
10      B[a]P were transferred from the particles, with only 1% to 2% of this being metabolized. Free
11      B[a]P, however, was extensively metabolized by the mierosomes to B[a]P-9-10-diol.  Relative to
12      the liver mierosomes, the lung mierosomes exhibited an approximate twofold greater efficiency
13      in the transfer of particle-associated B[a]P.
14            Bond et al. (1984) demonstrated metabolism of particle-associated B[a]P and free B[a]P
15      by alveolar macrophages (AMs). B[a]P-9,lO-diol and B[a]P-7,8-diol were identified in the
16      culture media, and B[a]P-7,8-dipl and B[a]P-4,5-diol were detected in the cellular extracts.
17      Additionally", small amounts of B[a]P phenols and B[a]P quinones were detected in both the cells
18      and the media.  The total amount of metabolites from both the cells and media increased with y,
19      increasing incubation time up to 48 h. However, use of B[a]P in solution or B[a]P  coated onto
20      DPM did not alter the total amount of metabolites produced by the macrophages over a 24-h
21      incubation period.
22            Because maerophages are instrumental in sequestering and transporting DPM matter in
23      the lungs, AM-mediated metabolism of particle-associated B [a]P has been studied.  Although
24      some data show the ability of the AMs to metabolize B[a]P associated with DPM, Chen and
25      Vostal (1982) have reported that aryl hydrocarbon hydroxylase (AHH) in AMs is decreased after
26      in vivo exposure to diesel exhaust.- Whether such diesel-associated decreases in AM enzymatic
27      activity are counterbalanced by increases in the AM population size in response to diesel particle
28      deposition (White and Garg, 1981) is unknown. Although it is known that human AMs contain
29      AHH activity (McLemore et al., 1981) and that they can metabolize B[a]P (Harris,  1985),
30      comparative studies of the AHH activities in rat, hamster, and human AMs could contribute
31      toward determining the relationship such activity may have on the development of lung tumors.
32            Even though the AMs appear to contain the bulk of diesel particles deposited in the lung
33      during chronic exposures, other cell types may also participate in the sequestration  and/or
34      metabolic activation  of carcinogenic agents.  The ability of lung epithelial cells to sequester
35      diesel exhaust particles was reported by White and Garg (1981).  Furthermore, significant
36      metabolism of B[a]P by rat Type II alveolar epithelial cells was reported by Bond et al. (1983).

        2/1/98                                   10-8       DRAFT-DO NOT CITE OR QUOTE

-------
  1      In this .study, a lung epithelial cell line (LEG) was shown to metabolize B[a]P to B[a]P-7,8-diol
  2      and B[a]P-9,10-diol, the latter accounting for 80% of the total B[a]P metabolites.  Compared   .
  3      with the AMs that were examined in the aforementioned study, the rat Type II cells showed
  4      approximately 10 times greater ability to metabolize B[a]P.
  5             Under healthy conditions, the Type II cells represent about 12% to 16% of all cells in the
  6      pulmonary epithelium of mammalian lungs and account for approximately 4% to 9% of the cells
  7      in the lungs (Crapoet al., 1983).  Alveolar macrophages, on the other hand, account for
  8      approximately 4% to 9% of the cells in the pulmonary region (Crapo et al., 1983). In terms of
  9      their relative abilities to metabolize B[a]P, the Type II cells may play an even more important
10      role than the AMs in metabolically activating PAH, assuming PAH as a substrate is available to
11      them (e.g., extraction of PAH from diesel particles by AMs and the subsequent release of PAH or
12      metabolically susceptible metabolites of PAH at Type II  cell sites). The Type II cell hyperplasia
13      observed after the deposition of diesel and other types of particles (White and Garg, 1981; Lee et
14      al., 1986; Lee et al., 1988; Plopper et al.,  1983) seemingly would favor a prominent role for these
1.5      cells in producing activated PAH metabolites.
16             Another cell type that may be important in the metabolism of PAH to ultimate
17      carcinogens "is the nonciliated bronchiolar cell. These cells are relatively rich in chemical
18      metabolizing enzymes and, being also in a region of the respiratory tract where clearance of
19      material would be relatively fast, may receive exposure via mucus to  organics that have desorbed
20      in the pulmonary region. The respiratory tract cytochrome P-450 system, for example, is present
21      in Type II  cells, but it is not as concentrated in this epithelial cell type as it is in the nonciliated
22      bronchiolar cell (Boyd, 1984). It is worthy to note that bronchoalveolar adenomas that develop
23      following diesel exposure have been found to resemble both Type II and nonciliated bronchiolar
24      cells (Mauderly et al., 1987). Like the Type II cells, the nonciliated bronchiolar cells are not
25      especially  important in phagocytosis of particles deposited in the lung, although some particles
26      may enter  these cells. As previously indicated, any metabolism of procarcinogens by these cells
27      probably involves the preextraction of carcinogen(s) in the extracellular lining fluid and/or in
28      other endocytic cells.
29    "         Although the preceding studies indicate that particle-adsorbed B[a]P may be distributed
30      throughout much of the organism via absorption from the lung and transport by the mucociliary
31      escalator to the gastrointestinal tract, it is  imperative to note that particles containing additional
32      deposits of B[a]P exhibit greater potential for elution of organics than is observed for actual
33      DPM. Therefore, the dissociation rates for the exhaust particles containing thermally deposited
34      B[a]P do not realistically represent the dissociation of combustion-adsorbed organics. Even
35      though particle-associated B[a]P can ultimately be metabolized by AMs and/or Type II cells to
36      reactive intermediates, the contribution of this process to  carcinogenic potential is uncertain and,
        2/1/98      .                              10-9       DRAFT-DO NOT CITE OR QUOTE

-------
 1      in rats, is probably of questionable significance. The relevant importance may be different in
 2      humans, however, where particle clearance rates have half-times of ~ I year, thereby allowing
 3      greater time for elution of organics.
 4
 5      10.2.2. Metabolism and Disposition of Nitroarenes
 6            Diesel engine emissions contain a large number of components, including numerous
 7      nitroarenes. Quantitatively, the nitroarenes represent a relatively small contribution to the overall
 8      PAH component of diesel engine emissions. However, with respect to carcinogenic potential,
 9      some of the nitroarenes (e.g., 1-NP, 4-NP, 6-nitroehrysene, and some DNPs) are of concern
10      because  of their known or suspected carcinogenic activity and their high mutagenic activity in
11      some test systems (Manabe et al., 1985; International Agency for Research on Cancer, 1989).
12      Within the scope of this document, it is inappropriate to review all  of the studies regarding the
13      carcinogenicity, metabolism, and mechanism of action of these various nitroarenes.  Therefore,
14      emphasis has been placed on those nitroarenes considered by the International Agency for
15      Research on Cancer (1989).
16            1 -Nitropyrene, a genotoxic and carcinogenic nitrosubstituted organic, is a particle-
17      associated component of diesel exhaust (Pitts et al., 1982; Schuetzle et al,, 1982; King, 1988).
18      As with  B[a]P, several investigators have studied the metabolism and disposition of 1-NP both in
19      free form and in association with DPM.
20            Bond and Mauderly (1984) made quantitative measurements of 1-NP metabolism and
21      maeromolecular covalent binding in the isolated perfused rat lung.  The study verified oxidation,
22      reduction, acetylation, and conjugation biotransformation of 1-NP by the lung, with oxidation
23      being the major process. The major metabolites were 3-, 6-, and 8-hydroxynitropyrene.  The
24      overall metabolism of 1-NP was increased by prior exposure of the rats to the mixed-function
25      oxygenase (MFO) inducer 3-methylcholanthrene (3-MC) but not to phenobarbital. This  3-MC-
26      induced  increase in 1-NP metabolism and a parallel increase in maeromolecular covalent binding
27      suggest that this pathway may be responsible for the observed covalent binding.
28            Exposure of rats to diesel exhaust (7.4 mg/m3) for 7 h/day, 5 days/week for 4 weeks
29      resulted  in twofold increases in the rates of nitropyrene metabolism in nasal tissue and in isolated
30      perfused lungs from these animals (Bond etal.,  1986). HPLC analysis of ethyl acetate-
31      extractable 1-[!4C]NP metabolites indicated that the major metabolites were 3-, 6-, and 8-
32      hydroxy-1 -aminopyrene and 4,5-dihydro-4,5-dihydroxy-1 -nitropyrene. Furthermore, a fourfold
33      increase in 14C covalently bound  in the lungs of these rats was detected. The increase in  1-NP
34      metabolism was not observed for rats among lower exposure (0.35 or 3.3 mg/m3) groups or clean
35   '   air controls.  The data from this study indicate that exposure to  DPM matter at concentrations of
36      7.4 mg/m3 significantly alters the metabolism and subsequent covalent binding of nitropyrene.
        2/1/98             .                      10-10      DRAFT-DO NOT CITE OR QUOTE

-------
  1             Bond et al. (1986) also examined the metabolism and deposition of free and partiele-
  2      associated 1-NP in F344 rats.  Results of the work indicated that the urinary and fecal excretion
  3      of UC-1-NP was not altered by exposure to the pure form or to that adsorbed on DPM. Pure 1-NP
  4      was more efficiently absorbed in the lung than was 1-NP coated onto DPM and, therefore, greater
  5      lung retention was noted for particle-adsorbed 1-NP. However, no significant difference between
  6      the two forms of 1-NP was noted for extrapulmonary tissue distribution or metabolic profiles.
  7      Analysis of excreta and tissues indicated that 1-NP is rapidly metabolized by the lungs or
  8      metabolized by other tissues following translocation from the lungs. For both 1-NP forms, small
  9      amounts of 6- and 8- hydroxyaeetylaminopyrene were detected in the lungs, suggesting
10      pulmonary oxidation, reduction, and conjugation of the parent compound.  The demonstration of
11      pulmonary metabolism of 1-NP and greater retention of 1-NP when adsorbed to DPM may be
12      significant relative to the dose to the lungs of both parent compound and metabolites.
13             Ball and King (1985) administered [14C] 1 -NP to rats intraperitoneally, orally, or by
14      intratracheal instillation of vapor-phase-coated diesel exhaust particles (380 ug [14C] 1 -NP/g; 5
15      mg/rat). More than 50% of the radiolabel was recovered (20% to 30% in the urine and 40% to
16      60% in the feces) within 24 h, regardless of the route of administration. The metabolic profile
17      and elimination kinetics were similar for all routes of administration. The principal urinary
18      metabolite (representing 15% to 25% of the total urinary 14C) was 6-hydroxy-Ar-acetyl~l-
19      aminopyrene (6-OH-NAAP),  a compound with demonstrated S-9-dependent mutagenic activity
20      in Salmonella strain TA98. Gut flora was shown to be necessary for the formation of 6-OH-
21      NAAP, for the observed enterohepatie circulation of metabolites excreted in the bile, and for
22      excretion of mutagenic activity in the urine.  That intestinal microorganisms may alter the
23      metabolites of 1-NP and facilitate their reabsorption was also reported by Medinsky et al. (1985).
24      Accumulation of I4C and diesel exhaust particles was detected in the lungs and gastrointestinal
25      tract 24 h after intratracheal administration, thereby attesting to the importance of mucociliary
26      transport and distribution of particles and their adsorbed components. Based on these results and
27      previous in vitro studies (King et al., 1983) demonstrating 1-NP binding to macromolecules, the
28      authors note the possible risk to the gastrointestinal tract and lungs relative to 1-NP.
29             Howard et al. (1986) studied the binding of intratracheally instilled nitropyrenes and
30      B[a]P to mouse lung DNA following preexposure to intratracheally instilled doses of the putative
31      inducing agents B[a]P, dichloromethane extract of DPM, or 1-NP. The results indicated that 1-
32      NP was a potent DNA-binding agent even in the absence of enzyme induction and  that this
33      potency was increased following B[a]P  exposure. Dinitropyrene (a mixture of the  1,3-, 1,6-, and
34      1,8- isomers) was also a potent lung DNA-binding agent, with and without the inducers.
35      Benzo[a]pyrene was not as potent a binding agent. Preexposure to the DPM extract but not to
36      B[a]P resulted in increased DNA binding of B[a]P. Pretreatrnent with the dichloromethane
        2/1/98                                    10-11       DRAFT-DO NOT CITE OR QUOTE

-------
 1      extract of DPM failed to increase the DNA binding of the nitropyrenes. The significance of this
 2      report is the demonstration that exposure to DE may potentiate the DNA binding of some of its
 3      components.
 4            The preceding studies have shown that some of the nitroarenes known to be constituents
 5      of DE may undergo biotransformation to various metabolites, some of which are known to be
 6      carcinogenic to animal species.  Such data may become more relevant as a complete
 7      understanding is obtained regarding the desorption of these compounds from DPM and their
 8      subsequent availability for biotransformation processes.
 9            Nitroarenes quantitatively represent a relatively small portion of the PAH component of
10      DE and, at least in rats, may play a minor role in tumorigenic responses compared to the particle
11      overload effect.  However, their contribution to the potential carcinogenicity of diesel engine
12      emissions deserves some consideration. In the previous section, information was presented
13      regarding the in vivo and in vitro metabolism of various nitroarenes considered to be possible
14      human carcinogens.  The fact that some of these metabolites have been shown to form DNA
15      adducts in animal studies and are mutagenic  in several test systems warrants their inclusion in
16      assessing possible mechanisms of diesel-exhaust-induced carcinogenicity. In fact, Gallagher et
17      al. (1994) reported results suggesting that DNA adducts are formed from nitro-PAHs present in
18      DE and may play a role in the carcinogenic process.
19            However, the question remains as to  why animals exposed to DE do not develop tumors
20      characteristic of those induced by dinitropyrene.  Rosenkranz (1995, 1996) provided evidence in
21      support of a hypothesis that highly carcinogenic dinitropyrenes present on DE are bioactivated
22      only at low exposure concentrations.  Specifically, diesel exhaust exposure-mediated oxidative
23      stress may prevent the reduction of dinitropyrenes to arylhydroxylamines and the subsequent
24      formation of reactive arylnitrenium species that form 8C-DNA adducts. Such an oxidative stress
25      may also result in the oxidation of the arylhydroxylamines to nitrosoarenes that are incapable of
26      reacting with DNA (Boldt et al., 1991). Additionally, the inflammatory response resulting  from
27      particle overload also may induce oxidative stress-like conditions. This hypothesis suggests that
28      there are other potential carcinogenic mechanisms that could be expressed in other species or
29      under different exposure conditions.
30                                                                               .
31      10,2.3,  Formation of Reactive Oxygen Species From Organic Constituents of Diesel
32              Exhaust and Their Involvement in the Induction of Lung Cancer
33              Sagai et al. (1993) reported that diesel exhaust paniculate matter (DEP) could produce
34      superoxide and hydroxyl radicals without any biological activating systems. DEP washed with
35      methanol could no longer produce these radicals, indicating that the active components were
36      extractable with organic solvents.  The likely involvement of reactive oxygen species in the
        2/1/98                                   10-12      DRAFT-DO NOT CITE OR QUOTE

-------
  1      induction of toxic and potentially carcinogenic effects in the lungs was suggested by greatly
  2      reduced lung injury due to DEP following extraction of the organic fraction prior to intratracheal
  3      instillation in mice. Additional support for the involvement of radicals in tissue damage was also
  4      provided by the finding that pretreatment with superoxide dismutase (SOD), an antioxidant,
  5      markedly reduced lung injury and death due to instillation of DEP.  Similarly Hirafuji et al.
  6      (1995) found that catalase, deferoxamine, and MK-447 inhibited the toxic effects of DEP on
  7      guinea pig tracheal cells and tissues in vitro.  Nagashima et al. (1995) demonstrated that the
  8      production of 8-hydroxydeoxyguanosine (8-OHdG), a product of the reaction of guanine with
  9      oxygen radicals, is induced in mouse lungs by intratracheal instillation of DEP. Ichmose et al.
10      (1997) reported further that while intratracheal instillation of washed DEP in mice induced a
11      small but statistically significant increase in lung tumor incidence, unwashed DEP induced a
12      larger increase. Moreover, increases in 8-OHdG correlated very well with increases in tumor
13      rates.  It thus appears likely that generation of reactive oxygen species resulting from exposure to
14      DEP may be involved in the carcinogenic process. It should be noted, however, that this
15      nucleoside is efficiently repaired and that proof of a causal relationship in humans is still lacking.
16      It is also unlikely that superoxide or hydroxyl radicals chemically generated by DEP alone
17      promote 8-OHdG production in vivo and induce lung toxicity, because  SOD is extensively
18      located in mammalian tissues.  A recent study indicates that the major route by which active
19      oxygen species are generated from DEP components is via the P450 reductase system (Kumagai
20     . etal., 1997).
21
22      10.3.  POTENTIAL INVOLVEMENT OF PULMONARY LEUKOCYTES IN THE
23            DEVELOPMENT OF LUNG TUMORS
24            Phagocytic leukocytes have been shown by numerous investigators to be toxic to tumor
25      cells in vivo, and increasing evidence suggests that cells of the mononuclear phagocyte series in
26      particular may be of pivotal importance in providing protection against malignancy in situ. This
27      protective function may, at least in part, result from these cells' ability to produce a tumor
28      necrosis factor (TNF) (Urban et al., 1986). Whether the tumor surveillance and tumoricidal
29      activities of AMs (Hengst et al., 1978; Sone et al.-, 1983; Sone, 1986; Kan-Mitchell et al., 1985) .
30      are compromised or otherwise modified when they are engorged with even relatively benign
31      particles has not been experimentally evaluated. The possibility  remains that'diesel and other
32      types of particles at high lung burdens result in decreases in natural killer (NK) cell functional
33      activities in providing defense against tumor formation, either by direct particle-cell interactions
34      or by altering the ability of AMs to influence NK cell-mediated host defense against metastatic
35      tumor cells (Sone, 1986). These cells are subpopulations of lymphocytes that possess
36      spontaneous cytolytic activity toward neoplastic cells but not toward normal cells.  Moreover,  the
        2/1/98                                   10-13      DRAFT-DO NOT CITE OR QUOTE

-------
 1      tumoricidal function of cytotoxic T lymphocytes (Sone, 1986) may be directly or indirectly
 2      compromised by the presence of high lung burdens of particles in.the lungs.
 3            Phagocytes from a variety of species produce elevated levels of oxidant reactants in
 4      response to challenges such as phagocytic stimuli, with the physicochemical characteristics of a
 5      phagocytized particle  being a major factor in determining the magnitude of the oxidant-
 6      producing response. Hatch and co-workers (1980) have demonstrated that interactions of guinea
 7      pig AMs with a wide variety of particles, such as silica, metal oxide-coated fly ash,
 8      polymethylmethacrylate beads, chrysotile asbestos, fugitive dusts, polybead carboxylate
 9      microspheres, glass and latex beads, uncoated fly ash, and fiberglass increase the production of
10      reactive oxygen species.  Similar findings have been reported by numerous investigators for
11      human, rabbit, mouse, and guinea pig AMs (Drath and Kamovsky, 1975; Allen and Loose, 1976;
12      Beall et al, 1977; Lowrie and Aber, 1977; Miles et al, 1977; Rister and Baehner,  1977; Hoidal et
13      al., 1978). As well, polymorphonuclear leukocytes (PMNs) are also known to increase
14      production of superoxide radicals, hydrogen peroxide, and hydroxyl radicals in response to
15      membrane-reactive agents and particles (Goldstein et al., 1975; Weiss et al., 1978; Root and
16      Metcalf, 1977).
17            It  is well recognized that the deposition of particles in the lung can result in the efflux of
18      PMNs from the vascular compartment into the alveolar space compartment in addition to
19      expanding the AM population size. Following acute exposures, the influx of the PMNs is
20      transient,  lasting only  a few days (Adamson and Bowden,  1978; Bowden and Adamson,  1978;
21      Lehnert et al., 1988).  Strom (1984) has reported that PMNs become abnormally abundant
22      following chronic exposures to DE. In the study by Strom (1984), the numbers of PMNs lavaged
23      from the lungs of diesel-exposed rats generally increased with increasing exposure duration and
24      inhaled DPM concentration. Strom (1984) also found that PMNs in diesel-exposed lungs
25      remained persistently  elevated for at least 4 mo after cessation of exposure, a potential
26      mechanism that may be related to an ongoing release of previously phagocytized particles by
27      AMs that engulfed them shortly after deposition. Evidence in support of this possibility has been
28      obtained by Lehnert et al. (1989) in a study in which rats were intratracheally instilled with 0,85,
29      1.06, or 3,6 mg of polystyrene particles. The PMNs were not found to be abnormally abundant
30      during the clearance of the two lower lung burdens, but they did become progressively elevated
31      in the lungs of the animals in which alveolar-phase clearance was restored.  Moreover, the
32      particle burdens in the PMNs became progressively greater over time. Such findings are
33      consistent with an ongoing particle relapse process, given the relatively short lifespan of PMNs.
34      As previously indicated, lung tumors develop in the rat at lesser lung burdens of DPM than with
35      particles such as TiO2. Polymorphonuclear leukocytes characteristically are increased
36      abnormally in the lung by DE exposure, but their presence in the lungs does not appear to be
        2/1/98                                   10-14      DRAFT-DO NOT CITE OR QUOTE

-------
  1      excessive following the pulmonary deposition of even high lung burdens of TICK (Strom, 1984;
  2      Lee et al,, 1986), Thus, the generation of reactive oxygen species by both AMs and PMNs
  3      should be considered as one potential factor of what probably is a multistep process that
  4      culminates in the development of lung tumors in response to chronic deposition of DPM.
  5            As previously indicated, the production of oxygen species may afford protection against
  6      emerging tumor cells by killing the cells, whereas under other conditions the production of
  7      reactive oxygen products conceivably may contribute to the development of neoplastic cells. The
  8      potential involvement of AMs and PMNs in the development of lung tumors in laboratory rats
  9      administered high lung burdens of DPM (Mauderly et al., 1987) or having inhaled particles that
10      are generally considered to have low to no cytotoxic potential (e.g., TiO2 [Lee et al., 1986]) over
11      a prolonged period of time may be related to the ability of the lung-free cells to produce reactive
12      oxygen metabolites during phagocytic oxidative metabolism (Hatch et al., 1980). Even though
13      products of phagocytic oxidative metabolism, including superoxide anions, hydrogen peroxide,
14      and hydroxyl radicals, can kill tumor cells (Klebanoff and Clark, 1978), and the reactive oxygen
15      species can peroxidize lipids to produce cytotoxic metabolites such as malonyldialdehyde, some
16      products of oxidative metabolism apparently can also interact with DNA to produce mutations.
17      Along this line, Weitzman and Stossel (1981) found that human peripheral leukocytes were
18      mutagenic in the Ames assay. This mutagenic activity was related to PMNs and blood
19      monocytes; blood lymphocytes alone were not mutagenic.  These investigators speculated that
20      the mutagenic activity of the phagocytes was a result of their ability to produce reactive oxygen
21      metabolites, inasmuch as blood leukocytes from a patient with chronic granulomatous diseases,
22      in which neutrophils have a defect in the NADPH oxidase generating system (Klebanoff and
23      Clark, 1978), were less effective in producing mutations than were normal leukocytes. Of related
24      significance, Phillips et al. (1984) demonstrated that the incubation of Chinese hamster ovary
25      cells with xanthine plus xanthine oxidase (a system for enzymatically generating active oxygen
26      species) resulted in genetic damage hallmarked by extensive chromosomal breakage and sister
27      chromatid exchange and produced an increase in the frequency of thioguanidine-resistant cells
28      (HGPRT test).  Aside from interactions of oxygen species with DNA, increasing evidence also
29      points to an important role of phagocyte-derived oxidants and/or oxidant products in the
30      metabolic activation of procarcinogens to their ultimate carcinogenic form (Kensler. et al., 1987).
31            Another characteristic  of AMs and PMNs that may contribute to the pathologic process
32      leading to lung tumor development following DPM deposition is that these phagocytes are
33      known to release a variety of potentially destructive hydrolytic enzymes, a process known to
34      occur simultaneously with the phagocytosis of particles (Sandusky et al., 1977).  The essentially
35      continual release of such enzymes during'chronic particle deposition and phagocytosis in the lung
36      may be detrimental to the alveolar epithelium, especially to Type I cells. Evans et al. (1986)
        2/1/98                                    10-15       DRAFT-DO NOT CITE OR QUOTE

-------
 1      showed that injury to Type I cells is followed shortly thereafter by a proliferation of Type II cells.
 2      Type II cell hyperplasia is a generally common feature observed in the lungs of animals that have
 3      received high lung burdens of various types of particles, including unreactive polystyrene
 4      microspheres.  Exaggerated proliferation as a repair or defensive response to DPM deposition
 5      may have the effect of amplifying the likelihood of neoplastic transformation in the presence of
 6      carcinogens beyond that which would normally occur with lower rates of proliferation, assuming
 7      an increase in the cell cycling of target cells and the probability of a neoplastic-associated
 8      genomic disturbance.
 9            The proliferative response of Type II cells following the deposition of DPM or other types
10      of particles, however, has yet to be directly related to a Type I cell destruction by proteolytic
11      enzymes released by lung phagocytes or to a direct action of particles on the proliferation kinetics
12      of the Type II cells.  The production of reactive oxygen species or their products could also be
13      involved in the process. Whatever the stimulus, it remains possible that the lung's AM
14      population may play a role aside from any responsibility for Type I cell damage. Alveolar
15      macrophages have the ability to release several other effector molecules or cytokines that can
16      regulate numerous functions of other lung cells, including their rates of proliferation (Bitterman
17      et al., 1983; "Jordana et al, 1988; Denholm and Phan, 1989). The AM-derived mediators that
18      may stimulate Type II cell hyperplasia following particle deposition in the lung remain to be
19      identified, if in fact the AMs play a regulatory role in the Type II  cell proliferative response.
20        '    Driscoll (1995) and Oberdorster and Yu (1991) outlined a proposed mechanism for the
21      carcinogenicity of DE at high doses that emphasizes the role of phagocytic cells. Following
22      exposure, phagocytosis of particles acts as a stimulant for oxidant production and inflammatory
23      cytokine release by lung phagocytes.  It was hypothesized that at high particle exposure
24      concentrations the quantity of mediators released by particle-stimulated phagocytes exceeds the
25      inflammatory defenses of the lung (e.g., antioxidants, oxidant metabolizing enzymes, protease
26      inhibitors, cytokine inhibitors, etc.), resulting in tissue injury and inflammation.  With continued'
27      particle exposure and/or the persistence of excessive particle burdens, there then develops an
28      environment of'phagocytic activation, excessive mediator release-tissue injury and, consequently,
29   *   more tissue injury, inflammation, and tissue release. This is accompanied by cell proliferation.
30      As discussed in a review by Cohen and Ellwein (1991), conceptually cell proliferation can
31      increase the likelihood that any oxidant-induced or spontaneously occurring genetic damage
32      becomes fixed in a dividing cell and is clonally expanded.  The net result of chronic particle
33      exposures sufficient to elicit inflammation and cell proliferation in the rat lung is an increased
34      probability that the genetic changes necessary for neoplastic transformation will occur. In
35      support of this hypothesis, it was reported that concentrations  of inhaled CB resulting in lung
36      inflammation increased mutation rates, an effect ameliorated by treatment with antioxidants
        2/1/98  .                                   10-16       DRAFT-DO NOT CITE OR QUOTE

-------
  1      (Driscoil. 1995).  Although these responses can occur in the absence of organic carcinogens.
  2      those present in DE may still contribute to the process,
  3             The possibility that particles induce carcinogenic effects only at high doses still lacks
  4      conclusive proof.  For example, it has not been definitively shown that inflammatory responses
  5      are a prerequisite for tumor induction.  Furthermore, direct effects of ultrafine particles such.as
  6      DPM taken up by epithelial cells cannot be ruled out. In fact, Riebe-Imre et al. (1994) reported
  7      that CB is taken up by lung epithelial cells in vitro, inducing chromosomal damage and
  8      disruption of the cytoskeleton, lesions that closely resemble those present in tumor cells.
  9      Johnson et al. (1993) reported that 20 nm polytetrafluoroethylene particles are taken up by
10      pulmonary epithelial cells as well as polymorphonuclear leucocytes, inducing an approximate 4-,
11      8-, and 40-fold increase  in the release of interleukin-1 alpha and beta, inducible nitric oxide
12      synthetase, and macrophage inflammatory protein, respectively.
13
14      10.4. DNA ADDUCT CONSIDERATIONS
15             Although  DNA adduct formation by particle-adsorbed organics was originally thought to
16      be a mechanistic component of the diesel exhaust-induced pulmonary carcinogenesis observed in
17      rats, currently available data do not support this hypothesis. The following section provides a
18      brief overview of studies investigating DNA adduct formation following exposure to PAHs, DE,
19      and other particulate matter.
20             On the assumption that DNA adduct formation is a critical step in the initiation of
21      carcinogenesis (Harris, 1985), it was hypothesized that increased residence time of PAHs in the
22      lung would increase the  opportunity for metabolism and subsequent adduct formation. This
23      would be especially important if association of the PAHs with soot particles and their slow
24      release from those particles contributed to the increased residence time. Therefore, adduct
25      formation by B[a]P alone compared with that of particle-associated B[a]P was  investigated
26      regarding possible mechanisms of diesel exhaust carcinogeniciry.
27             An experiment was undertaken to test the hypothesis that inhalation of B[a]P associated
28      with CB particles would increase the levels of DNA adducts compared with inhalation of pure
29      B[a]P (Wolff et al., 1989). DNA modification was measured using the 32P-postlabeling method
30      developed by Randerath et al. (1985). The high sensitivity (~1 adduct in 10'° bases) of this
31      technique (Reddy and Randerath, 1986) made possible measurement of the low levels of DNA
32      adducts resulting  from repeated inhalation exposures to !4C-B[a]P aerosols (2 mg/m3), !4C-B[a]P
33      (2 mg/m3) adsorbed to CB particles (97 mg/m3) (B[a]P/CB), or filtered air. Total 14C levels in
34      the lung (a nonspecific indicator of reactive and nonreactive B[a]P metabolites, free B[a]P, and
35      particle-bound B[a]P) were 100-fold greater following exposure to B[a]P/CB than following
36      exposure to  B[a]P alone,
        2/1/98                                   10-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1      .      The levels of total DNA adducts or the B[a]P diol-epoxide(BPDE)-DNA adduct in the
 2      lung were not significantly different whether the rats were exposed to pure B[a]P or B[a]P/CB,
 3      However, association of B[a]P with CB resulted in the formation of unidentified lung adducts
 4      that were not seen in DNA from lungs of rats exposed to pure B[a]P. It is possible that the
 5      adducts seen only in the B[a]P/CB exposures may play a role in the potential tumorigenic effect
 6      of particle-associated B[a]P. Reasons for the discrepancy between particle effects on total DNA
 7      adducts and retention of 14C include the possibility that the kinetics for formation and decline of
 8      DNA adducts are different from those of total bound 14C. As a consequence, long-term retention
 9      of total B[a]P and metabolites in the lung may not be a good marker for adduct formation.
10            There were clear differences in the kinetics of the buildup and decline of DNA adduct
11      levels and total I4C for rats exposed to B[a]P/CB,  The t1/2 for the decline of total 14C was
12      approximately 10-fold faster than that for the decline in levels of DNA adducts for rats exposed
13      to B[a]P/CB.  Previous work has shown that at 1 day or later after the end of single exposures to
14      B[a]P or B[a]P/CB, most of the 14C present was bound to total macromolecules (Sun et al.,
15      1988), presumably largely.non-DNA protein.  This information in combination with the current
16      data suggests that decline or repair of DNA adducts is considerably faster than that of protein
17      turnover.  Following repeated exposures, this would be expected to lead to increased buildup of
18      14C in the lung relative to DNA adducts. The t,/2 values for decline in DNA adducts observed in
19      the current work are similar to the t!/2  values of approximately 4 weeks reported for B[a]P
20      rnetabolite-DNA  adducts in the lungs of A/HeJ and C57BL/6J mice (Stowers and Anderson,
21      1985). Protein turnover is generally on longer time scales than the aforementioned t,,2 values.
22            It appears that long-term retention of 14C radiolabel in the lung may not be as important as
23      previously suspected, at least with respect to indicating DNA damage. 14C binding levels and
24      DNA adducts were not closely related, and it is clear from these results that DNA adduct levels
25      cannot be predicted from total 14C levels. This observation is consistent with the work of Morse
26      and Carlson (1985), who observed that binding levels of 3H with lung protein were greater than
27      levels of 3H to lung DNA 6 h after administration of oral H-B[a]P to mice. They also found that
28      3H binding to  protein was more persistent than 3H binding to DNA.
29            Caution should be used in interpreting the results from short-term exposures in regard to
30      possible implications for long-term exposures when carcinogenicity might be observed. The
31      pattern of results  seen after 12 weeks might not continue after many months of exposure.  The
32      adduct levels were higher in the rats exposed to B[a]P/CB than B[a]P after 12 weeks of
33      exposure, so it is possible that this  difference might become greater with continued exposure. In
34      addition, the different adduct patterns between the B[a]P/CB and B[a]P exposures may indicate
35      that other adducts besides the BPDE-DNA adduct are important in potential carcinogenic effects
36      of B[a]P/CB exposures. Another factor is the possible influence of a chronic inflammatory
        2/1/98                                    10-18      DRAFT-DO NOT CITE OR QUOTE

-------
  1      response, cell injury, or cell proliferation, all of which accompany long-term exposures'to inhaled
  2      insoluble particles (Morrow, 1986).  Such responses are generally greater after prolonged
  3      exposure than in the current 12-week exposure. These responses might be factors in progression
  4      to tumors in long-term inhalation exposures of rodents, when large lung burdens of particles
  5      accumulate (Morrow, 1986), and in the increased incidence in tumors when B[a]P is merely
  6      mixed with Fe2O3 particles versus adsorbed onto the particle (Saffiotti etal., 1965).
  7             Studies have also been conducted to evaluate DMA adduct formation in the lungs of
  8      animals exposed both to DE and to particles that are nearly devoid of organics (e.g., CB) or that
  9      completely lack organic fractions (e.g., TiO2).
10             DNA adduct formation in the lungs of animals subjected to long-term exposure to whole
11      DE has been described by  Wong et al. (1986).  Using tissues from animals of the Mauderly et al.
12      (1987) study, these investigators reported an increase in DNA adduct formation in male and
13      female F344 rats exposed to whole DE (7.1  mg of particles/m3) for 7 h/day, 5 days/week  for up to
14      30 mo.  32P postlabeling was applied to DNA that was extracted from six control and six exhaust-
15      exposed rats (males and females). Characterization of the adducts and identification of the
16      exhaust components responsible for their formation were not within the scope of the study. The
17      lungs of exhaust-exposed rats were darkly pigmented and contained diesel particle-laden
18      macrophages. Aggregates of these macrophages were frequently associated with alveolar wall
19      fibrosis, bronchiolar metaplasia and, occasionally, squamous metaplasia. Lungs from control rats
           [
20      were not darkly pigmented and had relatively unaltered airways and structures.  Autoradiographic
21      analysis revealed elevated  levels of DNA adducts in the  exhaust-exposed rats. The authors
22      indicated that quantitative  and qualitative data regarding DNA adducts resulting from diesel
23      exhaust exposure may be useful for extrapolation to potential effects in humans.
24            A study by Bond etal. (1989) addressed several key topics regarding the role of DNA
25      adducts in the pulmonary carcinogeniciry of DE.  Using  groups of rats exposed to whole DE at
26      particle concentrations of 0,0.35,3.5, 7.0, or 10.0 mg/m3 for 12 weeks, the relationship between
27      DNA adduct levels and exposure concentration was examined. The data for the exposure levels
28      employed indicated that DNA adduct formation (about 14 adducts per 109 bases) was similar
29      across all exposure concentrations and was approximately twice that of the sham-exposed group.
30      The fact that DNA adduct  formation was independent of exposure concentration may be
31      explained, in part, by previously reported data (Bond and Mauderly,  1984) showing that
32      metabolism of organics associated with diesel exhaust by the isolated perfused rat lung could be
33      saturated at high concentrations, thereby limiting the production of metabolites required for the
34      formation of DNA adducts.
35            The time course for DNA adduct formation was also examined by Bond et al. (1989).
36      Over a 12-week period of exposure to diesel exhaust (7 mg/m3 DPM), lung DNA adducts. were
        2/1/98  .                                 10-19       DRAFT-DO NOT CITE OR QUOTE

-------
 1      found to accumulate slowly. The highest adduct level was reached at 12 weeks, followed by a
 2   .   decline-to control level by 4 weeks postexposure. Throughout the exposure period, lung DNA
 3      adducts remained constant and at a lower level in sham-exposed rats. The investigators
 4      suggested that the rapid repair of adducts relative to their formation might result in a steady-state
 5      level of DNA adducts during long-term exposure.
 6            A dosimetry study examined the distribution of DNA adducts in the respiratory tract to
 7      determine if increased DNA adduct formation occurred in regions of the lung where diesel
 8      exhaust-induced tumors are formed (Bond et ai., 1988).  For this study, rats were exposed for 12
 9      weeks to DE at a particle concentration of 10 mg/m3. DNA adduct levels were highest in
10      peripheral tissue, which is the same region in which tumors occurred in rats in long-term
11      exposure studies (Mauderly et al., 1987). Although these findings suggest that DNA adduct
12      formation and tumor formation are related, the data do not prove the association.
13            The previous studies provided data regarding the role of DNA adducts in the pulmonary
14      carcinogenesis of DE, but were not designed to provide insight into possible target cells. An
15    .  additional molecular dosimetry study by Bond et al. (1990a) addressed this topic and also
16      compared the effects of DPM with CB particles that were virtually free of the adsorbed organics
17      found on DPM. In this study, rats were exposed to whole DE (6.2 mg/m3), CB particles (6.2
18      mg/m3 DPM), or clean air 16 h/day, 5 days/week for 12 weeks. Relative to clean air controls, a
19      significant increase in the total DNA adduct level in Type II cells was noted  for rats exposed to
20      DE and CB. The exposure to CB and DE resulted in an approximate fourfold increase in adduct
21      level compared with controls. However, the investigators noted that there was a large region of
22      unresolved adducts in the chromatograms from DE-exposed rats and that the total adduct level in
23      these animals may be underestimated.  Whether the small amount (=0.04%) of extractable
24      organics on the CB particles was responsible for the observed DNA adduct formation or the
25      adducts were the result of inflammatory responses to the particles was not determined. This
26      study does, however, demonstrate that Type II cells are possible targets for diesel exhaust
27      exposure.
28            The report by Bond et al. (1989) provided additional information to suggest a possible
29      relationship between DNA adducts in the lung and DE-induced pulmonary carcinogenesis.  For
30      this study, rats, mice, hamsters, and monkeys were exposed to DE at a particle concentration of
31      8.1 mg/m3 for 12 weeks.  Following this exposure, the levels of lung DNA adducts in rats, a
32      species susceptible to  DE-induced carcinogenesis, were shown to be 60% greater than for sham-
33      exposed controls. However, lung DNA adduct levels in mice and hamsters (species that have
34      been shown to be resistant to exhaust-induced carcinogenesis) were very similar to those of
35    '  respective controls. Also of interest was the finding that DNA adduct levels in the lungs of
36      diesel exhaust-exposed monkeys were 80% greater than those of sham-exposed controls.
        2/1/98               -                    10-20       DRAFT-DO NOT CITE OR QUOTE

-------
  1             In study by Wolff et al. (1990) described earlier in this chapter, levels of DNA adducts as
  2      determined by "2P postlabeling were significantly higher for DE-exposed rats than for CB-
  3      exposed rats after 12 weeks of exposure to DPM concentrations of 10 mg/m3. The results of this
  4      study suggested that DNA adduct levels are influenced by the organic content of the
  5      carbonaceous particles and that the organic constituents may initiate carcinogenesis. It was
  6      hypothesized that the continued inflammatory and proliferative responses may then promote cell
  7      transformations.  Although the DNA adduct formation may have been the result of very small
  8      amounts of organics desorbed from the carbon particles, it is also possible that these adducts are
  9      the result of oxygen radicals or other reactive agents released from neutrophils and macrophages.
10      More recent reports have affirmed the latter contention, Randerath et al. (1992) and Williams et
11      al. (1992) provided data showing DNA adduct levels to be increased following exposures to  CB
12      and that no organic fraction was involved.
13             Gallagher et al. (1994) noted that DNA adduct-like compounds were formed in rat lungs
14      following exposures to DE, Ti02,,or CB, but that the total adduct levels were not significantly
15      elevated by DE exposure.  Exposure to DE resulted in DNA adducts, possibly from nitro-PAHs
16      associated  with the organic fraction of the DPM, but the overall significance of this mechanism
17      in the carcinogenic response is uncertain, -
18         •   The uncertainty of the role of a genotoxic mechanism was further shown by S wafford et
19      al. (1995),  who detected no differences in mutational patterns in CB-induced or diesel exhaust-
20      induced pulmonary carcinomas. It was, however, noted that inactivation of thep5J may have a
21      role in neoplastic responses with a squamous-cell carcinoma component.
22             Based on the assumption that DNA adduct formation is a critical step in the initiation of
23      carcinogenesis (Harris, 1985), increased residence time of PAHs in the lung would increase the
24      opportunity for metabolism and subsequent adduct formation. Some involvement of the organic
25      components is suggested by data showing the formation of DNA adducts in exposed animals and
26      by the known carcinogenic and mutagenic potential of many of the compounds in diesel exhaust.
27      Several studies affirm the bioavailability from inhaled diesel exhaust particles of compounds
28      such  as B[a]P and 1-NP, which are known to be carcinogenic or mutagenic. Furthermore, the
29      fact that xenobiotics may undergo biotransformation to reactive intermediates following their
30      entry into the body via inhalation of DPM has been demonstrated for B[a]P and various
31      nitroarenes. However, results from the metabolism/disposition studies using carbon particles to
32      which organics have been experimentally adsorbed must be interpreted with caution. The
33      concentration of organics on these particles is probably much greater than the monomolecular or
34      bimolecular layer on actual DPM and, therefore, might facilitate desorption of the organics from
35      these experimentally prepared particles.

        2/1/98                                   10-21       DRAFT-DO NOT CITE OR QUOTE

-------
 1            The role of DNA adducts in DE-induced pulmonary carcinogenesis requires further
 2      investigation (Bond, 1993).  Because DNA adduct formation is observed after exposure to CB
 3      particles, formation of these adduets by particle-adsorbed organics does not appear to be a
 4      definitive mechanism for explaining the carcinogenic response observed in rats. However, a
 5      possible (albeit minor) contributory role for particle-adsorbed organics cannot be totally
 6      dismissed.  Higher concentrations of CB than DPM were required to induce adduct formation,
 7      and Gallagher et al. (1994) reported a nuclease-sensitive adduct in rats exposed to DPM but not
 8      CB. It was hypothesized that this adduct may have resulted from exposure to nitro-PAHs.
 9            Several investigators have reported on DNA adducts  in humans exposed to diesel
10      exhaust. In studying biomarkers of exposure, distinct adduct patterns were found among garage
11      workers occupationally exposed to diesel exhaust when compared to nonexposed controls
12      (Nielsen and Autrup, 1994). Furthermore, the findings were concordant with the adduct patterns
13      observed in groups exposed to low concentrations of PAHs from combustion processes.
14      Hemminki et al, (1994) also reported elevated levels of DNA adducts in lymphocytes from
15      garage workers with known diesel exhaust exposure compared to nonexposed mechanics.  The
16      adduct levels (up to 3.63 adducts/108 nucleotides) were significantly greater in diesel exhaust-
17      exposed groups when compared to nonexposed groups.  Houetal. (1995) found elevated adduct
18      levels in bus maintenance workers exposed to diesel exhaust. Although no difference in mutant
19      frequency was observed between the groups, the adduct levels were significantly different (3.2
20      vs. 2.3 x 10"8). Nielsen et al. (1996) measured three biomarkers in DE-exposed bus garage
21      workers: lymphocyte DNA adducts, hydroxyethylvaline adducts in hemoglobin, and  1-
22      hydroxypyrene in urine. Significantly increased levels were  reported for all three.  Qu et al.
23      (1996) detected increased adduct levels, as well as increases  in some individual adducts, in the
24      blood of underground coal miners exposed to DE.
25
26      10.5. SUMMARY OF METABOLISM AND MECHANISM OF ACTION OF
27           CARCINOGENIC COMPONENTS OF DIESEL EXHAUST
28            Recent studies have shown tumor rates resulting from exposures to nearly organic-free
29      CB particles to be similar to those observed for DE exposures, thus providing strong evidence for
30      an epigenetic mechanism underlying DE-induced pulmonary carcinogenesis in rats at high doses.
31      A nongenotoxic mechanism is also supported by the fact that carbon particles per se cause
32      inflammatory responses and increased epithelial cell proliferation and that AM function may be
33      compromised under conditions of particle overload. A mechanism was proposed in which .
34      particle-overloaded phagocytic cells secreted a variety of inflammatory mediators, stimulating
35      cell proliferation and increasing the likelihood that any oxidant-induced or spontaneously
36      occurring genetic damage would become fixed and clonally expanded, resulting in an increased
        2/1/98                                   10-22       DRAFT-DO NOT CITE OR QUOTE

-------
 1      probability of neoplastic change.  The development of lung tumors in rats following chronic
 2      exposures to diesel exhaust was detected under conditions in which AM-mediated particle
 3      clearance from the lung provides support for this hypothesis. It should be noted, however, that
 4      the bioassays lacked sensitivity to detect tumors at nonoverload conditions.
 5            It is generally accepted that one of the underlying mechanisms of careinogenesis involves
 6      the formation of covalent adducts with DNA, resulting in the alteration of cellular genetic
 7      information.  Several reports have provided data indicating that such adducts are formed in
 8      animals after long-term exposure to DE. The premise that DNA adduct formation plays a role in
 9      DE-induced careinogenesis is  substantiated by several findings, including an increase in DNA
10      adducts in the same pulmonary regions where tumors occur and higher DNA adduct levels in
11      species known to be susceptible to DE-induced tumors.  However, the lack of an exposure
12      response for DNA adduct formation, as demonstrated by the molecular dosirnetry studies
13      reported by Bond et al. (1990b), suggests the involvement of additional mechanisms.
14            While particle overload mechanisms are likely to predominate at high exposure
15      concentrations, organics are likely to play a greater role for any effects that may occur at non-
16      overload exposure conditions.  Although the lung's AMs, which phagocytize deposited DPM,
17      may participate in the gradual  in situ extraction and metabolism of procarcinogens associated
18      with the diesel particles, it is uncertain if the mutagenic organics are or can be eluted in sufficient
.19      amounts to be relevant to a carcinogenic response in the rat model. The much slower particle
20      clearance rates in humans, however, allow for greater efficiency of extraction. While both DE
21      and CB have been reported to  increase DNA adducts to a similar degree, DE is effective in non-
22      particle-overload conditions, suggesting again that the organics may play a role in tumor
23      induction. The participation of organics is supported by reports of increased DNA adducts in
24      humans occupationally exposed to diesel exhaust (see  Section 10.4).
25            Low-dose effects of DPM must also be considered, including the findings of Riebe-Imre
26      et al. (1994), who showed that CB may induce chromosomal damage and cytoskeletal alterations
27      in vitro at doses that are not cytotoxic. The  role of reactive oxygen species must also be
28      considered relative to low-dose DE exposure. The normal tumoricidal activities of the AMs may
29      be compromised upon interaction with excessive numbers of diesel particles, and diesel particle-
30      macrophage interactions could lead to the generation of reactive oxygen species that have been
31      shown to be at least mutagenic. The in vitro formation of active oxygen radicals by DE exposure
32      reported by Sagai et al. (1993) indicated that reactive oxygen species could be formed without
33      bioactivation, and may thus occur even under nonoverload conditions. The in vivo induction of
34      oxygen radical-induced DNA damage (Nagashima et al., 1995) provided additional support for
35      low-dose particle effects.

        2/1/98              •                     10-23       DRAFT-DO NOT CITE OR QUOTE

-------
  1              Caution must be exercised in extrapolating observations made in animal models to

  2     ,  humans when assessing the potential for DE-induced pulmonary carcinogenesis.  The

  3       carcinogenic response and the formation of DNA adducts in rats-exposed to diesel exhaust and

  4       other particles at high exposure concentrations may be species-specific and not particle-specific.

  5       However, DNA adduct data and the possible involvement of reactive oxygen species may be

  6       relevant in long-term occupational exposure of humans to low concentrations.

  7

  8       10.6. REFERENCES
  9
10       Adamson, IYR; Bowden, DH, (1978) Adaptive response of the pulmonary macrophagic system to carbon, II.
11       Morphologic studies. Lab Invest 38:430-438.
12
13       Allen, RC; Loose, LD. (1976) Phagocytic activation of a luminol-dependent chemiluminescence  in rabbit alveolar
14       and peritoneal macrophages. Biochem Biophys Commun 69:245-252.
15
16       Ball, LM; King, LC. (1985) Metabolism, mutagenicity, and activation of 1-nitropyrene in vivo and in vitro. Environ
17       Int 11:355-362.                                                       '
18    .;
19       Beall, GD; Repine, JE; Hoidal, JR; Raps, FL. (1977) Chemiluminscence by human alveolar macrophages:
20       stimulation with heat killed bacteria or phorbol myristate acetate. Infect Immunol 17:117-120.
21
22       Bitterman, PB; Aselberg, S; Crystal, RG. (1983) Mechanism of pulmonary fibrosis: spontaneous  release of the
23       alveolar macrophage-derived growth factor in interstitial lung disorders. J Clin Invest 72:1801-1813.
24
25       Bohning, DE; Atkins, HL; Cohn, SH. (1982) Long-term particle clearance in man: normal and impaired. Ann Occup
26       Hyg 26:259-271.
27
28       Boldt, J; Man, MC;  Wang, YC; Smith, BA; Beland, FA; Maher, VM; McCormick, JJ. (1991) Kinds of mutations
29       found when a shuttle vector containing adducts of 1,6-dinitropyrene replicates in human cells. Carcinogenesis
30       12:119-126.                                                                       .
31
32       Bond, JA. (1993) DNA adducts. In: U.S. EPA research needs for risk assessment of inhaled paniculate matter.
33       Washington, DC: Office of Health and Environmental Assessment; EPA/600/R-93/104.
34
35       Bond, JA; Mauderly, JL. (1984) Metabolism and macromolecular covalent binding of [14C]-1-nitropyrene in isolated
36       perfused and ventilated rat lungs. Cancer Res 44:3924-3929.
37
38       Bond, JA; Mitchell, CE; Li, AP. (1983) Metabolism and macromolecular covalent binding of benzo[a]pyrene in
39       cultured Fischer-344 rat lung type II epithelial cells. Biochem Pharmacol 32:3771-3776.
40
41       Bond, JA; Butler, MM; Medinsky, MA; Muggenburg, BA; McClellan, RO. (1984) Dog pulmonary macrophage
42       metabolism of free and particle-associated [14C]benzo[a]pyrene. J Toxicol Environ Health 14:181-189.
43
44       Bond, JA; Sun, JD;  Mitchell, CE; Dutcher, JS; Wolff, RK; McClellan, RO. (1986) Deposition, metabolism and
45       excretion of '"C- 1-nitropyrene coated on diesel particles as influenced by exposure concentration. Toxicol Appl
46       Pharmacol 85:102-117.
47
48       Bond, JA; Wolff, RK; Harkema, JR; Mauderly, JL; Henderson, RF; Griffith, WC;  McClellan, RO. (1988)
49       Distribution of DNA adducts in the respiratory tract of rats exposed to diesel exhaust. Toxicol Appl Pharmacol
50       96:336-346.
51


         2/1/98                                       10-24        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Bond, JA; Wolff, RK; Harkema, JR; Mauderly, JL; Henderson, RF; Griffith, WC; McCIetlan, RO. (1988)
  2       Distribution of DNA adducts in the respiratory tract of rats exposed to diesel exhaust. Toxicol Appl Pharmacol
  3       96:336-346.
  4                                                                       -
  5 •   •  Bond, JA; Harkema, JR; Henderson, RF; Mauderly, JL; McClellan, RO; Wolff, RK. (1989) Molecular dosimetry of
  6       inhaled diesel exhaust. In: Mohr, U., ed. Assessment of inhalation hazards. New York: Springer-Verlag; pp. 315-
  7       324.
  8
  9       Bond, JA; Johnson, NF; Snipes, MB; Mauderly, JL. (1990a) DNA adduct formation in rat alveolar type II cells: cells
10       potentially at risk for inhaled diesel exhaust. Environ Molec Mutagen 16:64-69.
11
12       Bond, JA; Mauderly, JL; Wolff, RK. (1990b) Concentration- and time-dependent formation of DNA adducts in
13       lungs of rats exposed to diesel exhaust. Toxicology 60:127-135.
14
15       Bowden, DH; Adamson, IYR. (1978) Adaptive responses of the pulmonary macrophagic system to carbon. I. Kinetic
16       studies. Lab Invest 38:422-438.
17
18       Boyd, MR. (1984) Metabolic activation and lung toxicity: a basis for cell-selective pulmonary damage by foreign
19       chemicals. Environ Health Perspect 55:47-51.
20
21       Boyland, E. (1980) The history and future of chemical carcinogehesis. Br Med Bull 36:5-10.
22
23       Cantrell, ET; Tyrer, HW; Peirano, WB; Danner, RM. (1980) Benzo[a]pyrene metabolism in mice exposed to diesel
24       exhaust: II. Metabolism and excretion. In: Pepelko, WE; Danner, RM; Clarke, NA, eds. Health effects of diesel
25       engine emissions. Cincinnati, OH: Health Effects Research Laboratory, U.S. Environmental Protection Agency;
26       EPA-600/9-80iQ57a.   ,                                                       -                    '
27
28       Chen, K-C; Vostal, JJ. (1982) Aryl hydrocarbon hydroxylase induction  in rat pulmonary alveolar macrophages
29       (PAM) by diesel particulates or their extracts. Pharmacologist 24:151.
30
31       Cohen, SM; Ellwein, LB. (1991) Genetic error, cellproliferation and carcinogenesis. Cancer Res 51:6493-6505.
32                                                 .                                                        '
33       Crapo, JD; Young, SL; Fram, EK; Pmkerton, KE; Barry, BE; Crapo, RO. (1983) Morphometric characteristics of
34       cells in the alveolar region of mammalian lungs. Am Rev Respir Dis 128:542-546.
35
36       Denholm, EM; Phan, SH. (1989) The effects of bleomycin on alveolar macrophage growth factor secretion. Am J
37       Pathol 134:355-363.                                                   '
38       .'
39       Drath, DB; Karnovsky, ML. (1975) Superoxide production by phagocytic leukocytes. J Exp Med 141:257-262.
40
41       Driscoll, K. (1995) Role of inflammation in the development of rat lung tumors in response to chronic particle
42       exposure. In:  Proceedings of the MIT toxicology symposium: particle overload in the rat lung and lung cancer:
43       relevance for human risk assessment, Cambridge, MA: Massachusetts Institute of Technology; 29 pp.
44                                                                                                    .
45       Driscoll, KE; Carter, J; Oberdorster, G. (1995) Inflammation cytokine expression and mutagenesis in rat lung after
46       carbon black  inhalation. Toxicologist 15:46.
47
48       El-Bayoumy, K; Rivenson, A; Johnson, B; DiBello, J; Little, P; Hecht, SS. (1988) Comparative tumorigenicity of 1-
49       nitropyrene, 1-nitrosopyrene, and 1-aminopyrene administered by gavage to Sprague-Dawley rats. Cancer Res
50       48:4256-4260.
51                                                                                                  ' '
52       Evans, M; Shami, SG; Martinez, LA. (1986) Enhanced proliferation of pulmonary alveolar macrophages after
53       carbon instillation in mice depleted of blood monocytes by strontium-89. Lab Invest 54:154-159.
54                                                 .
         2/1/98                                         10-25        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Gallagher. J; Heinrich. U; George, M; Hendee, L; Phillips, DH; Lewtas, J. (1994) Formation of DNA adducts in rat
  2       lung following chronic inhalation of diesel emissions, carbon black, and titanium dioxide particles. Carcinoaenesis
  3       15:1291-1299.                                     ,                              •
  4
  5       Gerde. P; Medinsky, MA; Bond, JA. (1991) Particle-associated polycyclic aromatic hydrocarbons—a reappraisal of
  6       their possible role in pulmonary carcinogenesis. Toxicol Appl Pharmacol 188:1-13.
  7
  8       Glatt, HR; Oesch, F. (1976) Phenolic benzo[a]pyrene metabolites are mutagenic. Mutat Res 36:379-384.
  9
 10       Goldstein, IM; Roos, D; Kaplan, HB; Weissman, G. (1975) Complement and immunoglobulins stimulate superoxide
 11       production by human leukocytes independently of phagocytosis. J Clin Invest 56:1155-1163.
 12
 13       Harris, CC. (1985) Future directions in the use of DNA adducts as internal dosimeters for monitoring human
 14       exposure to environmental mutagens and carcinogens. Environ Health Perspect 62:185-191.
 15
 16       Hatch, GE; Gardner, DE; Menzel, DB. (1980) Stimulation of oxidant production in alveolar macrophages by
 17       pollutant and latex particles. Environ Res 32:121-136.
 18
 19       Heinrich, U. (1990) Presentation at workshop on health effects of diesel emissions. U.S.  EPA, Research Triangle
 20       Park, July 1990.
 21
 22       Heinrich, U; Muhle, H; Takenaka, S; Ernst, H; Fuhst, R; Mohr, U; Pott, F; Stober, W. (1986) Chronic effects on the
 23       respiratory tract of hamsters, mice and rats after long-term inhalation of high concentrations of filtered and unfiltered
 24       diesel engine emissions. J Appl Toxicol 6:383-395.
 25
 26       Heinrich, U; Dungworth, DL; Pott, F; Peters, L; Dasenbrock, C; Levsen, K; Koch, W; Creutzenberg, 0; Schulte, A.
 27       (1994) The carcinogenic effects of carbon black particles and tar-pitch  condensation aerosol after inhalation
 28       exposure of rats. Ann Occup Hyg48:suppl. 4, pp. 351-356.
 29
 30       Heinrich, U; Fuhst, R; Rittinghausen, S; Creutzenberg, O; Bellmann, B; Koch, W; Levsen, K. (1995) Chronic
. 31       inhalation exposure of wistar rats and two different strains of mice to diesel engine exhaust, carbon black, and
 32       titanium dioxide. Inhal Toxicol 7:533-556.
 33
 34       Hemminki,  K; Soderling, J; Ericson, P; Norbeck, HE; Segerbaack, D. (1994) DNA adducts among personnel
 35       servicing and loading diesel vehicles. Carcinogenesis 15:767-769.
 36
 37       Hengst, JCD; Kan-Mitchell, J; Kempf, RA; Strumpf, U; Sharma, OP; Hibbs, JB; Chapman, HA, Jr; Weinberg, JB.
 38       (1978) The macrophage as an antineoplastic surveillance cell: biological perspectives. J Reticulendothel Soc 24:549-
 39       570.
 40
 41       Hirafuji, M; Sakakibara, M; Endo, T; Murakami, S; Mori, Y; Sagai, M; Minami, M. (1995) Biological effects of
 42       diesel exhaust particles (DEP) on tissues and cells isolated from respiratory  tracts of guinea pigs. Res Com Mol
 43       Pathol Pharamacol-90:221-233.
 44'   .
 45       Hirose, M; Lee, M-S; Wang, CY; King, CM. (1984) Induction of rat mammary gland tumors by 1-nitropyrene,  a
 46       recently recognized environmental mutagen. Cancer Res 44:1158-1162.
 47
 48       Hoidal, JR; Repine, JE; Beall, GD; Rasp, FL; White, JG. (1978) The effect of phorbol myristate acetate on
 49       metabolism and ultrastructure of human alveolar macrophages. Am J Pathol 91:469-482.
 50
 51       Hou, S-M; Lambert, B; Hemminki, K. (1995) Relationship between hprt mutant frequency, aromatic DNA adducts
 52       and genotypes for GSTM1 and NAT2 in bus maintenance workers. Carcinogenesis 16:1913-1917.
 53
 54       Howard, PC; Gerrard, JA;  Milo, GE; Fu, PP; Beland, FA; Kadlubar,  FF. (1983) Transformation of normal human
 55       skin fibrob-lasts by 1-nitropyrene and 6-nitrobenzo[a]pyrene. Carcinogenesis 4:353-355.


          2/1/98   .                                      10-26        DRAFT-DO NOT CITE OR QUOTE

-------
   1       Howard, PC; Gerrard. JA; Milo, GE: Fu. PP; Beland, FA; Kadlubar, FF, (1983) Transformation of normal human
   2       skin fibrobiasts by 1-nitropyrene and 6-nitrobenzo[a]pyrene. Carcinogenesis 4:353-355.
   3-    •
   4       Howard, AJ; Mitchell, CE; Dutcher, JS; Henderson, TR; McClellan, RO, (1986) Binding of nirropyrenes and
   5   '    benzo[a]pyrene to mouse lung deoxyribonucleic acid after pretreatment with inducing agents. Biochem Pharmacoi
   6       35:2129-2134.
-   7
   8       Ichinose, T; Yamanushi, T; Seto, H; Sagai, M. (1997) Oxygen radicals in lung carcinogenesis accompanying
   9       phagocytosis of diesel exhaust particles, IntJ Oncol 11:571-575
 10
 11       International Agency for Research on Cancer. (1989) Diesel and gasoline engine exhaust and some nitroarenes.
 12       IARC monographs on the evaluation of carcinogenic risks to humans, vol. 46.  Lyon, France:  IARC; pp. 189-373.
 13
 14       Johnson, C; Finkelstein, J; Gelein, R; Baggs, R;  Mercer, P; Corson, N; Ngyen, K; Oberddrster, G. (1993) Early
 15       alterations in the mRNA abundance of IL-1 a, IL-1 P, iNOS, MIP-2, TGFp, and VEGF associated with ultrafine
 16       particle exposure. Fundam Appl Toxicol 50:138 (abstr).
 17
 18       Jordana, M; Richards, C; Irving, LB; Gauldie, J. (1988) Spontaneous in vitro release of alveolar macrdphage
 19       cytokines after the intratracheal instillation of bleomycin in rats: characterization and kinetic studies. Am Rev Respir
 20       Dis 137:1135-1140.     .
 21
 22       Kan-Mitchell, K; Hengst, JCD; Kempf, RA; Rothbart, RK; Simons, SM; Brooker, AS; Kortes, VL; Mitchell, MS.
 23       (1985) Cytotoxic activity of human pulmonary alveolar macrophages. Cancer Res 45:453-458.
 24
 25       Kawabata, Y; Iwai, K; Udagawa, T; Tukagoshi,  K; Higuchi, K. (1986) Effects of diesel soot on unscheduled DNA
 26       synthesis of tracheal epithelium and lung tumor  formation. In: Ishinishi, N; Koizumi, A; McClellan, R; Stober, W,
 27       eds. Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam: Elsevier; pp. 213-222.
 28
 29       Kawabata. Y; Udagawa, T; Iwai, K, (1993) One year exposure to diesel engine exhust causes lung tumors. Toxic and
 30       carcinogenic effects of solid particles on the respiratory tract. Fourth International Inhalation Symposium, Hannover;
 31       pp.61.
 32
 33       Kensler, TW; Egner, PA; Moore, KG; Taffe, BG; Twerdok, LE; Thrush, MA. (1987) Role of inflammatory cells in
 34       the metabolic activity of polycyclic aromatic hydrocarbons in mouse skin. Toxicol Appl Pharm 90:337-346.
 35
 36       King, CM. (1988) Metabolism and biological effects of nitropyrene and related compounds. Health Effects Research
 37       Institute, Research Report No. 16; pp. 1-31.
 38
 39       King, LC; Jackson, M; Ball, LM; Lewtas, J. (1983) Binding of l-nitro['4C]pyrene to DNA and protein in cultured
 40       lung macrophages and respiratory tissues. Cancer Lett 19:241-246.
 41     :           •.
 42       Klebanoff, SJ; Clark, RA. (1978) The metabolic burst. In: The neutrophil: function and clinical disorders.
 43       Amsterdam: North-Holland; pp. 283-368.
 44                                 ••
 45       Kumagai, Y; Arimoto, T; Shinyashiki, M; Shimojo, N; Nakai, Y; Yoshikawa, T; Sagai, M.  (1997) Generation of
 46       reactive oxygen species during interaction of diesel exhaust particle components with NADPH cytochrome P450
 47       reductase and involvement of the bioactivation in the DNA damage. Free Radical Biol Med 22:479-487.
 48
 49       Lee, KP; Trochimowicz, HJ; Reinhardt, CF. (1985) Pulmonary response of rats to titanium dioxide TiO2 by
 50       inhalation for two years. Toxicol Appl Pharmacoi 79:179-192.
 51
 52       Lee, KP; Henry, NW III; Trochimowicz, HJ; Reinhardt, CF. (1986) Pulmonary response to impaired lung clearance
 53       in rats" following excessive TiO, dust deposition. Environ Res 41:144-167.
 54                   '.•".''
          2/1/98                                         10-27       DRAFT-DO NOT CITE OR QUOTE

-------
 1       Lee. K.P; Ulrich, CE; Geil, RG; Trochimowicz, HJ. (1988) Effects of inhaled chromium dioxide dust on rats exposed
 2       for two years. Fundam Appl Toxicol 10:125-145.
 3
.4       Lehnert, BE; Valdez, YE; Bomalaski,,SH. (1988) Analyses of particles in the lung free cell, tracheobronchial lymph
 5       nodal, and pleural space compartments following their deposition in the lung as related to lung clearance
 6       mechanisms. Ann Occup Hyg 32:125-140.
 7
 8       Lehnert, BE; Ortiz, JB; London, JE; Cline, AF. (1989) Migratory behavior of alveolar macrophages during clearance
 9       of light to heavy lung burdens of particles. 1989 Annual Meeting of the American Thoracic Society, Cincinnati, OH,
10       May 14-17.        ,
11
12       Leung, H-W; Henderson,  RF; Bond, JA; Mauderly, JL; McClellan, RO. (1988) Studies on the ability of rat lung and
13       liver microsomes to facilitate transfer and metabolism of benzo[a]pyrene from diesel particles. Toxicology 51:1-9.
14
15       Lloyd, JW. (1971) Long-term mortality of steelworkers V. Respiratory cancer in coke plant workers. J Occup Med
16       13:53-68.
17
18       Lowrie, DB; Aber, VR. (1977) Superoxide production by rabbit alveolar macrophages. Life Sci 21:1575-1584.
19
20       Manabe, Y;  Kinouchi, T; Ohnishi, Y. (1985) Identification and quantitation of highly mutagenic nitroacetoxyarenes
21       and nitrohydroxypyrenes in diesel-exhaust particles. Mutat Res 158:3-18 (as cited in IAR.C, 1989).
22
23       Mauderly, JL; Jones, RK; Griffith, WC; Henderson, RF; McClellan, RO. (1987) Diesel exhaust is a pulmonary
24       carcinogen in rats exposed chronically by inhalation. Fundam Appl Toxicol 9:208-221.
25
26       Mauderly, JL; Snipes, MB; Barr, EB; el al. (1991) Influence of particle-associated organic compounds on
27       carcinogenicity of diesel exhaust. Eighth Health Effects Institute AnnualConference, April 21-24, Colorado Springs,
28       CO.
29
30       McLemore,  TL; Martin, RR;  Wray, NP; Cantrell, ET; Busbee, DL. (1981) Reassessment of the relationship between
31       aryl hydrocarbon hydroxylase and lung cancer. Cancer 48:1438-1443.
32
33       Medinsky, MA; Shelton, H; Bond, JA; McClellan, RO. (1985) Biliary excretion and enterohepatic circulation of 1-
34       nitropyrene  metabolites in Fischer-344 rats. Biochem Pharmacol 34:2325-2330.
35
36       Miles, PR; Lee, P; Thrush, MA; Van Dyke, K. (1977) Chemiluminescence associated with phagocytosis of foreign
37       particles in rabbit alveolar macrophages. Life Sci 20:165-170.
38
39       Mitchell, CE. (1982) Distribution and retention of benzo[a]pyrene in rats after inhalation. Toxicol Lett  11:35-42.
40
41       Morrow, PE. (1986) The setting of particulate exposure levels for chronic inhalation toxicity studies. J Am Coll
42       Toxicol 5:533-544.
43
44       Morse, MA; Carlson, GP. (1985) Distribution and macromolecular binding of benzo[a]pyrene in Sencar and
45       BALB/c mice following topical and oral administration. J Toxicol Environ Health 16:263-276.
46
47       Mumford, JL; Chapman, RS; Harris, DB. (1989) Indoor air exposure to coal and wood combustion emissions
48       associated with a high lung cancer rate in Xuan Wei, China Environ Int 15:315-320.
49
50       Nagashima, M; Kasai, H; Yokota, J; Nagamachi, Y; Ichinose, T; Sagai, M. (1995) Formation of an oxidative DNA
51       damage, 8-hydroxydeoxyguanosine, in mouse lung DNA after intratracheal instillation of diesel exhaust particles and
52       effectsof high-dietary fat and beta-carotene on this process. Carcinogenesis 16:1441-1445.
53
54       Nesnow, S;  Triplett, LL; Slaga, TJ. (1984) Tumor initiating activities of 1-nitropyrene and its nitrated products in
55       Sencar mice. Cancer Lett 23:1-8.


         2/1/98                                         10-28        DRAFT-DO NOT CITE OR QUOTE

-------
   1       Nielsen, PS: Andreassen, A: Farmer, PB,; 0vreb0, S; Autrup, H. (1996) Biomonttoring of diesel-exhaust exposed
   2       workers. DNA and hemoglobin adducts and urinary 1-hydroxyproline as markers of exposure. Toxicol Lett 86'?7-
   3       37,                                                   '
   4
   5       Nielsen, PS; Autrup, H. (1994) Diesel exhuast-related DNA adducts in garage workers. Clin Chem 40:14^6-1458
   6
   7       Nikula, KJ; Snipes, MB; Barr, EB; Griffith, WC; Henderson, RF; Mauderly, JL, (1995) Comparative pulmonary
   8       toxicities and carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl
   9       Toxicoi 25:80-94.
 10
 11       Oberdorster, G; Yu, CP. (1990) The carcinogenic potential of inhaled diesel exhaust: a particle effect? J Aerosol Sci
 12       2l(suppl. 1):S397-S401.
 13
 14       OberdOrster, G; Baggs, R; Gelein, R; Corson, N; Mercer, P; Nguyen, K; Pepelko, W. (1995) Pulmonary effects of
 15       inhaled carbon black in rats. Toxicologist 15:46.
 16
 17       Opresko, DM; Holleman, JW; Ross, RH; Carroll, JW. (1984) Problem definition study on emission by-product
 18       hazards from diesel engines for confined space army workplaces. ORNL Report No. 6017, Oak Ridge, TN: Oak
 19       Ridge National Laboratory.
 20
 21       Pepelko, WE; Peirano, WB. (1983) Health effects of exposure to diesel engine emissions. J Am Coll Toxicol 2:253-
 22       306.
 23
 24       Phillips, BJ; James, TEB; Anderson, D. (1984) Genetic damage in CHO cells exposed to enzymatically generated
 25       active oxygen species. Mutat Res  126:265-271.
 26
 27       Pitts, JN, Jr; Lokensgard, DM; Harger, W; Fisher, TS; Mejia, V; Schuler, JJ; Scorziell, GM; Katzenstein, YA.
 28       (1982) Mutagens in diesel exhaust particulates, identification and direct activities of 6-nitrobenzo[a]pyrene, 9-
 29       nitroanthracene, 1-nitropyrene, and 5H-phenanthra(4,5-bcd)pyran-5-one. Mutat Res 103:241-249.
 30.                                   .  .
. 31       Plopper, CO; Hyde, DM; Weir, AJ. (1983) Centriacinar alteration in lungs of cats chronically exposed to diesel
 32       exhaust. Lab Invest 49:391-399.
 33
 34       Pott, F; Stdber, W. (1983) Carcinogenicity of airborne combustion products observed in subcutaneous tissue and
 35       lungs of laboratory rodents. Environ Health Perspect 47:293-303.
 36
 37       Pott, F; Dungworth, DL; Heinrich, U; Muhle, H; Kamino, K; Germann, P-G; Roller, M; Rippe, RM; Mohr, U.
 38       (1994) Lung tumours in rats after intratracheal instillation of dusts. Ann Occup Hyg 38:357-363, Suppl. 1.
 39
 40       Qu, SX; Leigh, J; Koelmeyer, H; Stacey, NH. (1996) Detection of DNA adducts in coal miners exposed to diesel
 41       emissions in Australia. Fundam Appl Toxicol 50:237 (abstr).
 42
 43       Randerath, K; Randerath, E; Agarwal, HP; Gupta, RC; Schurdak, ME; Reddy, MV. (1985) Postlabeling methods for
 44    .   carcinogen-DNA adduct analysis.  Environ Health Perspect 62:57-65.
 45                                                                      -               •
 46       Randerath, K; Randerath, E; Danna, TF; Putnam, KL. (1992) Diesel exhaust-induced DNA damage as determined by
 47       "P postlabeling analysis. 9th Health Effects Institute Annual Conference, Monterey, CA.
 48
 49       Reddy, MV; Randerath, K. (1986) Nuclease PI-mediated enhancement of sensitivity of phosphorus-32-postlabeling
 50       test for structurally diverse DNA adducts. Carcinogenesis (London) 7:1543-1551.
 .51                           -                                            .         •
 52       Riebe-Imre, M; Aufderheide, M; Gartner-Hilbsch, S; Peraud, A; Straub, M. (1994) Cytotoxic and genotoxic effects
 53       of insoluble poarticles in vitro. Toxic Carcinog Effects Solid Part Resp Tract, pp. 519-523.
 54                          .                            .
          2/1/98  .                                       10-29        DRAFT-DO NOT CITE OR QUOTE

-------
 1       Rister, M: Baehner, RL, (1977) Effect of hyperoxia on superoxide anion and hydrogen peroxide production of
 2       polymorphonuclear leukocytes and alveolar macrophages. Br J Haematol 36:241-248.
 3
 4       Root, RK; Metcaif, JA. (1977) H2O2 release from human granulocytes during phagocytosis, J Clin Invest 60:1266-
 5       1279.
 6                                                                                           .
 7       Rosenkranz, HS. (1995) Genotoxicity of diesel emissions. Part II: The possible role of dinitropyrenes in lung cancer.
 8       In: Diesel exhaust: a critical analysis of emissions, exposure, and health effects. Cambridge, MA: Health Effects
 9       Institute; pp. 245-250.
10
11       Rosenkranz, HS. (1996) Mutagenic nitroarenes, diesel emissions, particle-induced mutations and cancer: an essay on
12       cancer causation by a moving target. Mutat Res 367:65-72.
13
14       Saffiotti, U; Cefin, F; Kolb, LH; Shubik, P. (1965) Experimental studies of the conditions of exposure to carcinogens
15       for lung cancer induction. J Air Pollut Cont Assoc 15:23-25.
16
17       Sagai, M; Saito, H; Ichinose, T; Kodama, M; Mori, Y. (1993) Biological effects of diesel exhaust particles. I. In
18       vitro production of superoxide and in vivo toxicity in mouse. Free Radical Biol Med 14:37-47.
19
20       Sandusky, CB; Cowden, MW; Schwartz, SL. (1977) Effect of particle size on regurgitive exocytosis by rabbit
21       alveolar macrophages. In: Pulmonary macrophages and epithelial cells, CONF-760927, National Technical
22       Information Service, U.S. Dept. of Commerce, pp. 85-105.
23
24       Schuetzle, D; Riley, TL; Prater, TJ; Harvey, TM; Hunt, DF. (1982) Analysis of nitrated polycyclic aromatic
25       hydrocarbons in diesel partieulates. Anal Chem 54:265-271.
26
27       Sims, P; Grover, PL; Swaisland, A; Pal, K; Hewer, A. (1974) Metabolic activation  of benzo[a]pyrene proceeds by a
28       diol-epoxide. Nature (London) 252:326-328.
29                                                •           •  •
30       Slaga, TJ; Viage, A; Berry, DL; Bracken, W. (1976) Skin tumor initiating activity of benzofojpyrene, 4,5-7,8- and
31       •7,8-diol-9,10-epoxidesand7,8-diol.CancerLett2:115-122.
32
33       Sone, S. (1986) Role of alveolar macrophages in pulmonary neoplasias. Biochim Biophys Acta 823:227-245.
34
35       Sone, S; Brennan,  LM; Creasia, DA. (1983) In vivo and in vitro NO2 exposures enhance phagocytic and tumoricidal
36       activities of rat alveolar macrophages. J Toxicol Environ Health 11:151 -163.
37
38       Stenback, F; Rowland, J; Sellakumar, A. (1976) Carcinogenicity of benzo[a]pyrene and dusts in the hamster lung
39       (instilled intratracheally with titanium oxide, aluminum oxide, carbon and ferric oxide). Oncology 33:29-34.
40
41       Stowers, SJ; Anderson, MW. (1985) Formation and persistence of benzo[a]pyrene  metabolite-DNA adducts.
42       Environ Health Perspect 62:31 -39.
43                  .    •   '
44       Strom, KA. (1984) Response of pulmonary cellular defenses to the inhalation of high concentrations of diesel
45       exhaust. J Toxicol Environ Health 13:919-944.
46
47       Sun, JTD; Wolff, RK; Kanapilly, GM. (1982) Deposition, retention, and biological fate of inhaled benzo[o]pyrene
48       adsorbed onto ultrafine particles and as pure aerosol.  Toxicol Appl Pharmacol 65:231-244.
49
50       Sun, JD; Wolff, RK; Kanapilly, GM; McClellan, RO. (1984) Lung retention and metabolic fate of inhaled
51       benzo[a]pyrene associated with diesel exhaust particles. Toxicol Appl Pharmacol 73:48-59.
52
53       Sun, JD; Wolff, RK.; Maio, SM.; Barr, EB. (1988) The influence of adsorption to carbon black particles on the
54       retention and metabolic activation of benzojajpyrene in rat lungs following  inhalation exposure or intratracheal
55       instillation. Inhalation Toxicol 1:79-94.


         2/1/98                                         10-30        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Swafford, DS; Nikula, KJ: Mitchell, CE; Belinsky, SA, (1995) Low frequency of alterations inp53, K-ros, and
  2       mdm2 in rat lung neoplasms induced by diesel exhaust or carbon black. Carcinogenesis 16:1215-1221.
  3
  4       Thyssen, J: Althoff, J; Kimmerle, G; Mohr, U. (1981) Inhalation studies with benzo[a]pyrene in Syrian golden
  5       hamsters, J NatI Cancer Inst. 66:575-577.
  6                                                                                         .
  7       Tyrer, HW; Cantrell,  ET; Horres, R; Lee, IP; Peirano, WB; Danner, RM. (1981) Benzofajpyrene metabolism in
  8       mice exposed to diesel exhaust: I. Uptake and distribution.  Environ Int 5:307-311.
  9
10       Urban, JL; Shepard, HM; Rothstein, JL; Sugarman, BJ. (1986) Tumor necrosis factor: a potent molecule for tumor
11       cell killing by activated macrophages. Proc Natl Acad Sci USA 83:5233-5237.
12
13       Vostal, J. (1986) Factors limiting the evidence for chemical carcinogenicity of diesel emissions in long-term
14       inhalation experiments. In: Ishinishi, N; Koizumi, A; McClellan, R; StOber, W, eds. Carcinogenic and mutagenic
15       effects of diesel engine exhaust. Amsterdam: Elsevier; pp. 381-396,
16
17       Weinstein, IB; Troll, W. (1977) National Cancer Institute workshop on tumor promotion and cofactors in
18       Carcinogenesis. Cancer Res 37:3461-3463.
19
20       Weiss, SJ; Rustagei, PK; LoBuglio, AF. (1978) Human granulocyte generation of hydroxyl radical, J Exp Med
21       147:316-323.
22
23       Weitzman, SA; Stossel, TP. (1981) Mutation caused by human phagocytosis. Science 212:546-547,
24
25       White, HJ; Garg, BD. (1981) Early pulmonary response of the rat lung to inhalation of high concentration of diesel
26       particles. J Appl Toxicol 1:104-110.
27
28       Williams, PL; Randerath, K; Mauderly, JL. (1992) Statistical analysis of correlated measurements of DNA adduct
29       formation in lungs of F344 rats exposed to diesel exhaust or carbon black. 9th Health Effects Institute Annual
30       Conference, Monterey, CA.
31
32       Wislocki, PG; Bagan, ES; Lu, AYR; etal. (1986) Tumorigenkity of nitrated derivatives of pyrene,
33       benz[a]anthracene, chrysene, and benzo[a]pyrene in the newborn mouse assay. Carcinogenesis (London) 7:1317-
34       1322.
35
36       Wolff, RK; Bond, JA; Sun, JD; Henderson, RF; Harkema, JR; Griffith, WC; Mauderly, JL; McClellan, RO. (1989)
37       Effects of adsorption of benzo[a]pyrene onto carbon black  particles on levels of DNA adducts in lungs of rats
38       exposed by inhalation. Toxicol Appl Pharmacol 97:289-299.
39
40       Wolff, RK; Bond, JA; Henderson, RF; Harkema, JR; Mauderly, JL. (1990) Pulmonary inflammation and DNA
41       adducts in rats inhaling diesel exhaust or carbon black. Inhal Toxicol 2:241-254.                             :
42                                                          ;
43   ,    Wong, D; Mitchell, CE; Wolff, RK; Mauderly, JL; Jeffrey,  AM, (1986)  Identification of DNA damage as a result of
44       exposure of rats to diesel engine exhaust. Carcinogenesis 7:1595-1597.
45
46       Wood, AW; Levin, W; Lu, AYH; Yagi, H; Hernandez, O; Jerina, DM; Conney, AH. (1976) Metabolism of
47       benzo[a]pyrene and benzo[a]pyrene derivatives to mutagenic products by highly purified hepatic microsomal
48       enzymes. JBiolChem 251:4882-4890.
49
         2/1/98                                         10-31        DRAFT-DO NOT CITE OR QUOTE

-------

-------
                 11.  QUALITATIVE AND QUANTITATIVE EVALUATION OF THE
                      CARCINOGENICITY OF DIESEL ENGINE EMISSIONS

  1      11.1. INTRODUCTION
  2            Concern about the carcinogenic risk of exposure to diesel engine emissions was
  3      stimulated in the late 1970s by a report indicating that diesel particle extracts are mutagenic
  4      (Huisingh et al., 1978), by the knowledge that diesel exhaust (DE) contains known carcinogens,
  5      and by the projected increase in the use of diesel engines in passenger vehicles. This concern
  6      culminated in a U.S. Environmental Protection Agency (EPA)-sponsored quantitative cancer risk
  7      estimate for diesel engine emissions (Albert et-al., 1983). Its estimate was based on a
  8      "comparative potency" approach because of a lack of either chronic animal cancer bioassays or
  9      definitive epidemiologie data. Since 1983, several chronic animal inhalation studies and
10      epidemiologic investigations designed to assess the eareinogenicity of diesel engine emissions
11      have been completed. These studies are summarized in Chapters 7 and 8.  Data relating to
12      mechanisms of DE-induced eareinogenicity are summarized in Chapter 10. Because of the
13      increase in the availability of such data and because of the need to provide an up-to-date
14      evaluation of the hazards of DE inhalation for U.S. EPA's Office of Mobile Sources, a
15      qualitative as well as a quantitative assessment of the cancer risk from exposure to DE was
16      undertaken.  These assessments are presented in this chapter.
17
18      11.2. QUALITATIVE EVALUATION OF THE CARCINOGENIC RISK OF
19           EXPOSURE TO DIESEL EXHAUST
20      11.2.1.  Hazard Evaluation Based on EPA 1986 Guidelines for Carcinogen Risk Assessment
21            Lung cancer incidence has been studied in human populations exposed  to diesel engine
22      exhaust. An increased incidence of lung cancer was observed in 6 mortality studies (Howe et al.,
23      1983; Wong et al., 1985; Boffetta and Stellman, 1988; Garshick et al., 1988; Gustavsson et al.,
24      1990; Ahlman et al, 1991) and in 10 case-control studies (Williams et al., 1977; Hall and
25      Wynder, 1984; Damber and Larsson, 1987; Garshick et al., 1987; Benhamou et al., 1988; Hayes
26      et al., 1989; Gustavsson et al., 1990; Steenland et al., 1990; Burns and Swanson, 1991; Emmelin
27      et al., 1993). A dose-response trend was observed in three of the cohort studies (Howe etaL,
28      1983; Wong et al., 1985; Boffetta and Stellman, 1988) and in three of the case-control studies
29      (Garshick et al., 1987; Steenland et al., 1990; Emmelin et al., 1993). An increased incidence of
30      lung cancer was not observed in some other studies (Waller, 1981; Rushton et al., 1983; Wong et
31      al., 1985; Edling et al., 1987; Lerchen et al., 1987), but each had several methodologic
32      limitations, such as small sample sizes, short follow-up, and lack of adjustment for confounding
33      factors.  The studies reporting increased incidences also had some major limitations even though
       .2/1/98                                  11-1       DRAFT-DO NOT  CITE OR QUOTE

-------
 1      some of them, especially those reported by Garshick and co-workers (1987, 1988), were able to
 2      eliminate most of the confounding variables.
 3            An association between DE exposure and lung cancer is therefore suggested by the results
 4      of a number of epidemiologic studies.  Because of the uncertainties created by limited exposure
 5      -data and the possibility of exposure to other agents, the evidence for carcinogenicity of diesel
 6      engine emissions in humans is considered to be limited under the EPA Cancer Assessment
 7      Guidelines (U.S. EPA, 1986).
 8            In animal experiments, inhalation of whole DE resulted in the induction of lung tumors in
 9      F344 rats (Ishinishi et al., 1986; Iwai et al., 1986; Mauderly et al., 1987; Brightwell et al., 1989;
10      Nikula et al., 1995), in Wistar rats (Heinrich et al., 1986a; Heinrich et al., 1995), in Sencar mice
11      (Pepelko and Peirano, 1983), in NMRI mice (Heinrich et al., 1986a; Stober, 1986), and in C57BL
12      mice (Takemoto et al., 1986).  It should be noted, however, that lung tumor incidences did not
13      exceed those of historical controls in the 1986 NMRI mouse study, and in a later study reported
14      by Heinrich et al. (1995), using both NMRI and C57BL mice, significant increases in lung
15      tumors were not seen. Lung tumors also were induced by implantation of DE condensate in
16      Osborne-Mendel rats (Grimmer et al., 1987) and by intratracheal instillation of diesel particulate
17      matter in Wistar rats (Pott and Roller, 1994).  Skin painting of diesel particle extracts induced
18      dermal tumors  in strain A mice (Kotin et al., 1955) and in Sencar mice following promotion with
19      tetradecanoylphorbol-13-acetate (Nesnow et al., 1982). Extensive studies with Salmonella
20      mutagenesis assays have, in most cases, demonstrated direct-acting mutagenic activity in both
21      particle extracts and gaseous fractions of DE. Positive results also have been reported for gene
22      mutations and chromosome effects in mammalian cell systems after exposure to diesel particle
23      extracts.
24            Based on the induction of lung tumors via inhalation in at least two strains of rats and two
25      strains of mice and by lung implantation of DE condensate, subcutaneous tumors following
26      injection of exhaust particle organic extracts, and skin tumors following dermal application of
27      exhaust particle organic extracts and supported by positive mutagenicity results, the evidence for
28      carcinogenicity of DE in animals is considered to be sufficient under U.S. EPA's Cancer
29      Assessment Guidelines (U.S. EPA, 1986).
30            The International Agency for Research on Cancer (IARC) (1989)  also evaluated the
31      evidence for carcinogenicity of DE and concluded that the evidence for carcinogenicity in
32      humans is limited. This conclusion was based primarily on cohort studies of railroad workers in
33      the United States (Garshick et al.,  1988) and Canada (Howe et al., 1983) and two case-control
34      studies (Howe  et al., 1983;  Garshick et al., 1988) using individuals drawn from the same group
35      of workers. Three further studies of cohorts with less certain exposure to diesel engine exhaust
36      also were considered: two  studies of London bus  company employees (Raffle, 1957; Rushton et
        2/1/98                                   11-2       DRAFT-DO NOT CITE OR QUOTE

-------
  1      al, 1983) and one of Swedish dock workers (Edling et al, 1987). IARC concluded that the
  2      evidence for carcinogenicity of whole engine exhaust in experimental animals was adequate,
  3      The conclusions were based on positive tumorigenic effects in two different strains of rats in five
  4      of six experiments (Karagianes et al.. 1981; Heinrich et al,, 1986b; Ishinishi et al., 1986; Iwai et
  5      al., 1986; Mauderly et al., 1987) and on positive effects in mice in two studies (Pepelko and
  6      Peirano, 1983; Stober, 1986).
  7             The IARC conclusions are thus in general agreement with those of EPA. Both agencies
  8      concluded that human evidence is limited.  IARC considered the animal evidence to be adequate
  9      for whole DE as well as for extracts of exhaust particles but inadequate for the gaseous phase.
10      Although EPA has not specifically evaluated the gaseous phase and did not consider whole
11      exhaust and particle extracts separately, both agencies agree that whole diesel engine exhaust is
12      carcinogenic in animals.
13             On the basis of limited evidence for carcinogenicity in humans, diesel engine emissions
14      are considered to best fit into the cancer weight-of-evidence Category Bl, according to 1986
15      EPA Cancer Assessment Guidelines.  This  classification is supported by sufficient evidence in
16      animals and positive results in mutagenicity studies and is consistent with the presence of known
17      carcinogens on diesel particles. Chemicals classified in Category B1 are considered to be.
18      probable human carcinogens. IARC (1989) also considers DE to be probably carcinogenic in
19      humans and has therefore classified it as Category 2A.
20
21      11.2.2. Hazard Evaluation Based on EPA Proposed Guidelines for Carcinogen Risk
22             Assessment
23             Under EPA's Proposed Guidelines for Carcinogen Risk Assessment (U.S. EPA, 1996),
24      DE is  considered to be a likely human carcinogen by the inhalation route of exposure. Its
25      carcinogenic potential is indicated by limited evidence for carcinogenicity in  occupationally
26      exposed humans, positive carcinogenicity in laboratory animal studies, positive evidence for
27      mutagenicity, and the presence of known human carcinogens in the mixture.  In comparison with
28      other agents designated as likely carcinogens, the overall weight of evidence for this agent places
29      it at the upper end of this grouping. This is because of positive data on tumor responses in both
30      hurnans and animals.
31             Concerning humans, increases in relative risks for  lung cancer have been reported in a
32      variety of populations occupationally exposed to DE. These increases have been small, generally
33      less than 2. While the inability to eliminate all confounding variables increases uncertainty, most
34      of the  variables have been eliminated in at least one large study.  Moreover, the reporting of
35      increased relative risk in a number of studies with differing confounding factors increases the
36      likelihood that a true response has occurred.
        2/1/98              •                    11-3       DRAFT-DO NOT CITE OR QUOTE

-------
 1          " Concerning animals, clear evidence for induction of lung tumors in rats was reported in
 2      several well-designed and well-conducted experiments. Results in mice were equivocal with the
 3      only definitive response occurring in a sensitive strain of mice exposed from conception. Syrian
 4      hamsters did not show a positive response but are known to be resistant to induction of deep lung
 5      tumors. Experiments with cats, monkeys, and Chinese hamsters, while negative for tumor
 6      induction, were of insufficient exposure duration and/or exposure level.
 7            In addition,-DE contains a variety of compounds that are considered to be known or likely
 8      human carcinogens.  These include several polycyclic aromatic compounds such as
 9      benzo[a]pyrene and nitroaromatics such a dintropyrene, 1,3-butadiene, and benzene. This
10      evidence is limited because of the low concentration of most of these compounds and because of
11      their uncertain bioavailability.  Extracts of diesel exhaust particulate matter (DPM) are also
12      genotoxic in most tests.
13            Biological information on DE is contradictory in terms of how to quantitate potential
14      cancer risks. Among animal species, only rats have shown unequivocal tumor responses, raising
15      the possibility that this response may be unique to rats. Tumor induction also was noted only at
16      exposure concentrations resulting in lung particle overload and accompanying pathological
17      effects. Because of limited sensitivity of the studies, however, the possibility of tumor induction
18      at exposure levels below those responsible for particle overload still exists. Mechanistic studies
19      support the concept that high-dose induction of lung cancer is a result of lung particle overload
20      resulting in release of inflammatory cytokines, proteolytic enzymes, and reactive oxygen species,
21      resulting in genotoxic effects with increased likelihood of fixing harmful mutations because of
22      rapid cell turnover. On one hand, while epidemiologic data are limited because of small
23      increases in relative risk ratios and an inability to eliminate all confounding factors, the data
24      nevertheless suggest that humans are likely to respond with lung cancer under nonparticle-
25      overload conditions. Low-dose effects are more likely due to elution of organics from the diesel
26      particle, although direct effects of particles cannot be ruled out.
27                                             .
28      11.3.  QUANTITATIVE ESTIMATION OF THE CARCINOGENIC RISK OF
29            EXPOSURE TO DIESEL EXHAUST
30      11.3.1. Review of Previous Quantitative Risk Estimates
31            Early attempts to quantitatively assess the carcinogenicity of diesel  engine emissions were
32      hindered by a lack of both positive epidemiologic studies and. long-term animal studies. One
33      means of overcoming these obstacles was the use of the "comparative potency" method (Albert
34      et al.,  1983). A second involved estimating risk based on equivocal epidemiologic evidence
35      (Harris, 1983).


        2/1/98                                   11-4       DRAFT-DO NOT CITE OR QUOTE

-------
  1            In the comparative potency method, a combustion or pyrolysis product was selected that
  2      had a previously determined cancer potency estimate based on epidemiologic data.  The ratios of
  3      the potency of this agent (e.g., coke oven emissions) to diesel exhaust particulate matter (DPM)
  4      extract in a variety of in vivo and in vitro tests are then multiplied by the epidemiology-based
  5      potency estimate for coke oven emissions and averaged.  If epidemiology-based estimates from
  6      more than one pollutant are used, the derived potencies are generally averaged to obtain an
  7      overall mean.
  8            The comparative potency estimate of Albert et al. (1983) is probably the best known.
  9      Their results were obtained using epidemiology-based unit cancer risk estimates for coke oven
10      emissions, cigarette smoke condensate, and roofing tar. Samples of particulate matter were
11      collected from three light-duty engines (a Nissan 220 C, an Oldsmobile 350, and a Volkswagen
12      turbocharged Rabbit), all ran on a highway fuel economy test cycle, and from a heavy-duty
13      engine (Caterpillar 3304), run under steady-state, low-load conditions. The particulate matter
14   .   was extracted with dichloromethane and tested in a variety of assays. Dose-dependent increases
15      in response were obtained for the four assays  listed below:
16            »   Ames Salmonella typhimurium (TA98) reverse mutation,
17            •   Gene mutation in L5178Y mouse lymphoma cells,
18            »   Sencar mouse skin tumor initiation test, and
19            •   Viral enhancement of chemical transformation in Syrian hamster embryo cells.
20            Only the first three assays were used to develop comparative potency estimates because
21      of variability of responses in the enhancement of the viral transformation assay. The in vitro
22      studies were carried out both in the presence and absence of metabolic activators. The potency,
23      defined as the slope of the dose-response curve, was measured for each sample in each short-term
24      assay.
25            The skin tumor initiation test was positive for all the. engines tested except the Caterpillar
26      engine.  Only  the Nissan engine, however, gave strong dose-response data. Because skin tumor
27      initiation was  considered to be the most biologically relevant test, it was used to derive potency
28      estimates for the Nissan engine. An estimate for the Nissan engine was then derived by
29      multiplying the epidemiology-based potency estimates for each of the three agents (coke oven
30      emissions, roofing tar, and cigarette smoke condensate) by the ratios of their potencies in the skin
31      tumor initiation test to that of the Nissan diesel engine. According to this method, three 95%
32      upper-bound estimates of lifetime  cancer risk per microgram per cubic meter of extractable
33      organic matter were derived for the Nissan diesel, based on potency comparisons with each of the
34      three agents. These values are: coke oven emissions, 2.6 * 10'4; roofing tar, 5.2 * 10"4; and
35      cigarette smoke condensate, 5.4 *  10"4. The average of the three equals 4.4 * 10"4.

        2/1/98                                    11-5       DRAFT-DO NOT CITE OR QUOTE

-------
 1            The potency of the other diesel emission samples was not estimated directly because of
 2      the weak response in the skin tumor initiation test.  Instead, their potency relative to the Nissan
 3      engine was estimated as the arithmetic mean of their potency relative to the Nissan in the
 4      Salmonella assay in strain TA98, the sister chromatid exchange assay in Chinese hamster ovary
 5      cells, and the mutation assay in mouse lymphoma cells. The estimated lifetime cancer risk per
 6      microgram per cubic meter of extractable organic matter for extracts from these engines are as
 7      follows: Volkswagen, 1.3 x 10'4; Oldsrhobile 1.2 x 10'4; and Caterpillar, 6.6 x 10'6.
 8            To convert these values to a lifetime risk per microgram per cubic meter of total DPM,
 9      the results were multiplied by the fraction of extractable organic matter in the particles. This
10      conversion was based on the assumption that the carcinogenic effects were caused solely by the
11      organic fraction.  These fractions were as follows:  Nissan, 0.08; Volkswagen, 0.18; Oldsmobile,
12      0.17; and Caterpillar, 0.27. After this adjustment, the resulting estimated potencies per
13      microgram per cubic meter of inhaled DPM varied from 3.5 x  10"5 for the Nissan to 1.8 x 10"6for
14      the Caterpillar.
15            Harris (1983) developed comparative potency estimates for the same four engines used by
16      Albert et al. (1983) but used only two epidemiology-based potency estimates: those for coke
17      oven emissions and for roofing tar. He employed preliminary data from three of the same assays
18      used by Albert et al. (1983)—the Sencar mouse skin tumor initiation assay, enhancement of viral
19      transformation in Syrian hamster embryo cells,  and the L5178 mouse lymphoma test. The mouse
20      lymphoma test was used both with and without metabolic activation, whereas the Salmonella
21      assay was not used.
22            The diesel cancer potency estimates by Harris (1983) were then derived by multiplying
23      the epidemiology-based cancer potency estimates for both coke oven emissions and roofing tar
24      by the ratio of their potencies compared with DE particles in each of the four bioassays. For
25      example, the epidemiology-based relative risk of exposure to 1 ug/m3 of coke oven emissions
26      was estimated to equal 4.4 x 10"4.  In the skin tumor initiation test, 2.1 papillomas per mouse
27      were reported for the coke oven sample, compared with 0.53 for the Nissan engine extract. The
28      benzene-extractable fraction was assumed to equal 0.06 (slightly less than that in the Albert et
29      al., 1983, studies). The  diesel potency estimate using this comparison is then equal to (0.53/2.1)
30      x 0.06 x 4.4 x 10"4/ug/m3, or 6.6 x 10~6/ug/m3 DPM. A total of eight comparisons were made for
31      each engine, four bioassays times two epidemiology-based potency estimates.
32            The Harris (1983)  estimates are not comparable to those of Albert et al. (1983)  without
33      adjustment. The unit risk  estimates of Albert and co-workers are based on absolute risk during
34      lifetime exposure, whereas Harris reported his values in terms of relative risk per year of
35      exposure.  To adjust this to lifetime risk for continuous exposure, it is necessary to multiply

        2/1/98                                   11-6       DRAFT-DO NOT CITE OR QUOTE

-------
  1      Harris' values by a factor of 2.7 = (70 * 0.039), where 70 reflects the lifetime exposure (70 years)
  2      and 0.039 is the lifetime lung cancer mortality rate in the U.S. population.
  3             The range of potencies varied from 0.2 * 10"5 to 0.6 * 10"5 for the Nissan sample,
  4      0.1  x 1Q-5 to 2.4 x 10'5 for the Oldsmobile 350, 0.2 x 10'5to 27.8 x 1Q-5 for the Volkswagen
  5      Rabbit, and 0.1 x 1 (r5 to 2.5 x 10-5/ug/m3 DPM for the Caterpillar sample.  Harris (1983) derived
  6      an overall mean relative risk value of 3.5 x 10"5/ug/m3 for the three light-duty engines with a 95%
  7      upper confidence limit of 2.5 x IQ"4, Individual mean values for each engine were not reported.
  8      After multiplying by 2.7 to convert to  a unit risk, the upper-bound estimate of potency from the
  9      three light-duty engines was equal to 6.8 x 104/ug/m3 DPM.  McClellan (1986), Cuddihy et al.
10      (1981,1984), and Cuddihy and McClellan (1983) estimated a risk of about 7.0 x 10'5/ug/m3
11      DPM using a comparative potency method similar to those reported in the preceding paragraph.
12      The database was similar to that used by Albert et al. (1983) and Harris (1983). This estimate
13      agrees quite well those reported by Albert et al. (1983). Although the Harris (1983) estimate is
14      somewhat greater, it should be remembered that it was based on preliminary data.
15             With the availability of chronic cancer bioassays, more recent assessments were based on
16      lung tumor induction in rats. Albert and Chen (1986) reported a risk estimate based on the
17      chronic rat bioassay conducted by Mauderly et al. (1987). Using a multistage model and
18      assuming equivalent deposition efficiency in humans and rats, they derived a 95% upper
19      confidence limit of 1.6 x 10"5 for lifetime risk of exposure to 1 ug/m3. Pott and Heinrich (1987)
20      used a linear extrapolation, including data reported by Brightwell et al. (1989), Heinrich et al.
21      (1986a), and Mauderly et al. (1987). They reported risk estimates of 6 x 1Q'5 to 12 x  10"5/ug/m3.
22      More recently, Smith and Stayner (1990), using time-to-tumor models based on the data of
23      Mauderly et al. (1987),  derived 95% upper confidence limits ranging from 1.5 x  10"5 to 3 x 10"5/
24      Pg/m3.  Pepelko and Chen (1993) developed unit risk estimates based on the data of Brightwell et
25      al. (1989), Ishinishi  et al. (1986), and Mauderly et al. (1987) using a detailed dosimetry model to
26      extrapolate dose to humans and a linearized multistage model. Taking the geometric mean of
27      individual estimates from the three bioassays, they derived unit risk estimates of 1.4 x 10"5/ug/m3
28      when dose was based on carbon particulate matter per unit lung surface area rather than whole
29      DPM and 1.2 x 10~4/ng/m3 when based on lung burden per unit body weight. Hattis and Silver
30      (1994) derived a maximum likelihood estimate for occupational exposure of 5.2 x 10"5/|ig/m3
31      based on lung burden and bioassay data reported by Mauderly et al. (1987) and use of a five-
32      stage Armitage-Doll low-dose extrapolation model.
33            Lung  cancer risk estimates based on epidemiologic data were derived by two
34      investigators. Harris (1983) assessed the risk of exposure to diesel engine emissions using data
35      from the London Transport Worker Study reported by Waller (1981). Five groups of employees
36      from the London Transport Authority (LTA) were used: bus garage engineers, bus drivers, bus
        2/1/98                                    11-7       DRAFT-DO NOT CITE OR QUOTE

-------
 1      conductors, engineers in central works, and motormen and guards. The first group was
 2      considered to have received the highest exposure; the next two, intermediate; and the last two,
 3      none.  When cancer death rates for the high-exposure group were compared with those of
 4      London males, there was no increase in the observed-to-expected (O/E) ratios.  The author, in
 5      fact, considered the results to be negative. However, because the low rate of lung cancer in all
 6      the LTA exposure groups may have been the result of a "healthy worker" effect, Harris (1983)
 7      compared the exposed groups with internal controls. He merged the three exposed groups and
 8      compared them with the two groups considered to be unexposed.  An adjustment was made for
 9    .  the estimated greater exposure levels of garage engineers compared with bus drivers and
10      conductors. Using this method, the relative risk of the exposed groups was greater than  1 but
11      was statistically significant only for garage engineers exposed from 1950 to 1960.  In this case,
12      the O/E ratio was 29% greater than the presumed unexposed controls.
13             Harris (1983) identified a variety of uncertainties relative to potency assessment based on
14      this study. These included:
15             •   Small unobserved differences in smoking incidences among groups, which could  have
16                a significant effect on lung cancer rates;
17             •   Uncertainty about the magnitude of exposure in the exposed groups;
18             •   Uncertainty regarding the extent of change in exposure conditions over time;
19             •   Random effects arising from the stochastic nature of the cancer incidence; and
20             •   Uncertainty in the mathematical specification of the model.
21             Taking the uncertainties into account, he derived a maximum likelihood relative  risk
22      estimate of 1.23 x 10'4 with a 95% upper confidence limit of 5 x 10'4/|ig/m3 DPM per year.
23      These estimates are equal to 5  x 10"4 and 2 x 10"3, respectively, when converted to an absolute
24      risk for lifetime exposure to 1 |ig/m3 particulate matter. It should be noted that, because of the
25      high degree of uncertainty, the 95% lower confidence limit would predict no risk.
26          .   More recently, McClellan et al. (1989) reported risk estimates based on the Garshick  et al.
27      (1987) case-control study in which lung cancer in railroad workers was evaluated. Using a
28      logistic regression, the expected relative risk of lung cancer death was estimated to rise 0.016 per
29      year of exposure to DE.  Adjustments were made to convert to continuous exposure (168 vs.  40
30      hours) for 70 years. Because exposure levels could not be defined exactly, two sets of
31      calculations were made, assuming inhaled DPM concentrations of either 500 or 125 |ig/m3 DPM.
32      Using a 95% upper confidence limit, the  number of excess cancer deaths per year in the  United
33      States was estimated to range from 1,900 to 7,400. These values could then be converted to  a
34      lifetime 95% upper confidence limit of the risk of exposure to 1 |ig/m3 DE, by dividing the
35      estimated excess number of annual cancer deaths in the United States by the total population and

        2/1/98               •                    11-8       DRAFT-DO NOT CITE OR QUOTE

-------
1      multiplying by 70, the estimated mean lifespan. Using this approach, unit risks of 0.6 x  10"3 to 2
2      x 10° were derived.  Even using the 95% lower confidence limits, an excess of 100 to 400 deaths
3      is predicted, unlike the Harris (1983) study in which no excess deaths could be predicted based
4      on the lower confidence limit. The estimates discussed in this section are listed in Table 11-1.
5             Each of the methods used to assess cancer risk has differing strengths and weaknesses,
6      which will be discussed later in this chapter.                          .    •
             Table 11-1.  Estimated 95% upper confidence limits of the lifetime risk of
             cancer from inhalation of 1 jig/m3 diesel participate matter (DPM)
               Method
Potency
                                              Comments
Reference
Comparative potency       3.5 x IO"5
Comparative potency       2,6 x 10"5
Comparative potency       7.0 x 10"5
Comparative potency       6.8 x IO"4
Multistage model           1.6 x 10"5
Straight-line
extrapolation               6-12 x 10"5
Time-to-tumor model       2-3 x IO"5
Logistic regression           8 x 10'5
Multistage model           1,4 x 1Q'5
Armitage-Do 11 model       5.2 x 10"5
Epidemiologic analysis      1.4 x IO"3
                                                 Nissan engine         Albert et al., 1983
                                                 Average of 3 engines   Albert et al., 1983
                                                 Light-duty engines     Cuddihy etal, 1984
                                                 Average of 3 engines   Harris, 1983
                                                 Lung cancer ratsa       Albert and Chen, 1986
                                                 Lung cancer ratsb
                                                 Lung cancer rats"
                                                 Lung cancer ratsc
                                                 Lung cancer ratsd
                                                 Lung cancer ratsa*e
                                                 London transport
                                                 study
                                    Pott and Heinrich, 1987
                                    Smith and Stayner, 1990
                                    McClellan et al, 1989
                                    Pepelko and Chen, 1993
                                    Hattis and Silver, 1994
                                    Harris, 1983
        Epidemiologic analysis     .6-2 x 10~3     Railroad workers       McClellan etal., 1989
      "Useddata from studies by Mauderly et al, 1987.
      bUsed data from studies by Brightwell et al., 1989; Heinrich et al., 1986a; and Mauderly et al., 1987.
      cUsed data from studies by Brightwell et al., 1989; Ishinishi et al., 1986; Iwai et al., 1986; and Mauderly et a!., 1987.
      dUsed data from studies by Brightwell et al., 1989; Ishinishi et al., 1986; and Mauderly et al., 1987.
      .'Maximum likelihood estimate based on 53 years of exposure, 8 hours/day, 240 days/year.
      2/1/98
              11-9
                                                    DRAFT-DO NOT CITE OR QUOTE

-------
 1      11.3.2. Current Approaches to Quantitation of Human Cancer Risk From Exposure to
 2             Diesel Exhaust
 3      11.3.2.1.  Dose-Response Estimates Using Human Epidemiologic Data
 4            In an attempt to quantitatively estimate risk using humans, a detailed analysis of the
 5      Garshick et al. (1988) study of railroad workers was carried out by ICF Clements (Crump et al.,
 6      1991). This study was selected because the cohort was quite large, significant increases in
 7      relative risk ratios were reported, exposure data collected during the later years of exposure were
 8      available, and some estimates of early exposures were available (Woskie et al., 1988a,b).  The
 9      investigators were also able to eliminate many of the confounding factors present in earlier
.10      studies.  Garshick et al. (1988) analyzed information obtained from the Railroad Retirement
11      Board on 55,407 white males who began railroad employment between 1939 and 1949, who
12      were between the ages of 40 and 64 in 1959, and who in 1959 worked at 1  of the 39 jobs selected
13      to represent a range of potential DE exposure.  Two analyses that indicated effects of exposure to
14      DE on lung cancer risk in this cohort were reported:  (1) A relative risk for lung cancer of 1.45
15      (95% confidence interval [CI] = 1.11, 1.89) was observed for DE-exposed  workers who were 40
16      to 44 years of age in 1945 and who consequently had the longest potential  exposure to DE;
17      relative risk was progressively lower among DE-exposed workers who were older in 1959 and
18      who had potentially shorter exposures to DE; and (2) the relative risk of lung cancer increased
19      monotonically with increasing duration of work in 1959 or later in a job involving DE exposure
20      (disregarding exposures in the current year and in the most recent 4 years); this risk was 1.72
21      (95% CI = 1.29, 2.23) in the group with the longest exposure (15 to 17 years).
22             The Garshick et al. (1988) database was used by two groups of investigators for the
23      purpose of quantitating cancer risk from exposure to diesel engine emissions (Crump et al., 1991,
24      unpublished; Dawson, California EPA, personal communication to William Pepelko, U.S. EPA).
25      The results of these analyses were not in agreement. While Crump and co-workers failed to
26      derive a positive dose-response relationship, Dawson did report increasing incidence of lung
27      cancer with exposure level. These differences have not been resolved.  Until they are resolved,
28      utilization of the Garshick et al. (1988) study to quantitate cancer risk from DE is not anticipated.
29      It should be noted, however, that despite the decision to not use this study for quantitative
30      analysis at this time, it still provides the strongest qualitative evidence for a causal association
31      between DE exposure and lung cancer.
32             As described previously, McClellan et  al. (1989) estimated annual lung cancer mortality
33      in the U.S. population, per  |ig/m3 DPM, using  the Garshick et al. (1987) case-control study,
34      which minimizes the issue of the dose response by assuming a single exposure level for all the
35      cases. Mean exposure levels of 125  or 500 u£/m3 DPM were used for deriving risk estimates
36      because exposure data, while limited, indicated that  actual exposures were unlikely to fall outside
        2/1/98                                   11-10       DRAFT-DO NOT CITE OR QUOTE

-------
  1      this range (Woskie et al., 1988a,b). The unit risk estimate of 2 x 10"V((ig/m3), assuming a mean
  2      exposure concentration of 125 ug/m3, is then quite conservative because it is based not only on
  3  ,    the upper-bound estimate of lung cancer mortality, but also on the lower estimate of mean
  4      exposure. The risk estimates  of either 0.6 or 3 x 10~3/(ug/m3), depending on the level of exposure
  5      assumed, will be approximately halved  if maximum likelihood estimates of mortality are used.
  6
  7      11.3.2.2. Dose-Response Estimates Based on Animal Bioassay Data
  8             Several cancer potency estimates for DE have been developed by U.S. EPA and others.
  9      Most of them used somewhat similar approaches, and as a result most of the estimates did not
10      differ greatly. The approaches used here are an extension of these efforts.  To develop unit risk
11      estimates, it was necessary to address several issues, including (a) determination of the critical
12      target site for DE; (b) determination of the fraction of exhaust responsible for tumor induction;
13      (c) use  of existing methods or development of new dosimetric methods for accurately
14      extrapolating dose from experimental animals chronically exposed to high concentrations of
15      exhaust to humans exposed at ambient concentrations; and (d) selection of the most suitable low-
16      dose risk extrapolation model.                                               .
17             The critical target organ was considered to be the lung. Although Iwai et al. (1986)
18      reported the induction of malignant lymphomas in the spleen as well as lung tumors in rats
19      following DE exposure, the lung was the only target site in other experimental studies with this
20      species. Potential carcinogenic agents present in DE may be adsorbed from the lungs, enter the
21      bloodstream^ and be transported systemically.  Although increases in DNA adduct levels in
22      peripheral lymphocytes have been reported in workers occupationally exposed to DE (Hemminki
23      et al., 1994), there is insufficient evidence for induction of cancer in humans at systemic
24      locations. Particle-adsorbed organics also may reach systemic targets via the gastrointestinal
25      tract. Particles deposited in the conducting airways are cleared to the mouth quite rapidly and
26      swallowed. A considerable volume of particles are also likely to be ingested as a result of
27      grooming during whole animal exposures (Wolff et al., 1982), resulting in the possible uptake of
28      carcinogens by the gastrointestinal tract Because half-times for elution of organics from the
29      particles are considerably longer than passage through the gastrointestinal tract, however, the
30      fraction adsorbed is expected to be small.
31        •     The site of action in the lungs is  assumed to be the epithelial lining  of the alveoli and
32      small airways. According to Mauderly et al. (1987) and others, inflammation and tumors appear
33      to arise from this tissue. Although a connection between interstitial events and lung tumors has
34      been suggested for particles (i.e., fibrosis as a precondition for lung tumors [Kuschner, 1984]),
35      data are unavailable to support this view with respect to DE-induced tumors.

        2/1/98                                   11-11       DRAFT-DO NOT CITE OR QUOTE

-------
 1            Accurate extrapolation of dose from experimental studies using animals exposed at high
 2      concentrations of exhaust to humans exposed to ambient concentrations requires a variety of
 3      adjustments. These include adjustments for species differences in deposition efficiency and
 4      respiratory exchange rates. An important factor is the rate of particle clearance from the deep
 5      lung. Normal clearance half-times from the alveolar region are several times longer in humans
 6      than in rats (Chan et al., 1981; Bohning et al., 1982; Griffis et al, 1983). This may result in an
 7      underestimate of lung burden when extrapolating to humans. On the other hand, the high
 8      exposure concentrations used in some of the animal studies resulted in greatly slowed or even
 9      complete cessation of macrophage-based clearance (Chan et al., 1984).
10            To more accurately extrapolate dose from experimental studies to humans, detailed
11      dosimetry modeling was used to account for these factors.  One of these models (Yu et al., 1991)
12      is listed in Appendix C. A second approach reported by Hattis and Silver (1994) was also
13      employed for comparison.  Hattis  and Silver determined area under the curve/time, using
14      observed lung burden data reported by Mauderly et al. (1987). Human lung burden in the Hattis
15      and Silver approach was calculated using a one-compartment model suggested by Smith (1992),
16      assuming that at low doses the lung burden is at steady state. Dose for both approaches was
17      estimated in terms of either particle concentration per unit of lung surface area or lung
18      burden/body surface area.
19            The use of whole particle lung burden as the dosimeter is based on the assumption that
20      both the diesel particle as well as its surface-adsorbed constituents are responsible  for the
21      carcinogenic effects of DE. A particle-based assessment was considered to be reasonable for two
22      principal reasons.  First, exposure to the vapor phase alone did not result in detectable tumor
23      induction in rats (Ishinishi et al., 1986; Iwai et al., 1986; Stdber, 1986; Brightwell et al., 1989;
24      Heinrieh et al., 1995).  Second, while tumor induction at high doses is considered to be primarily
25      a particle effect (Heinrieh, 1990; Heinrieh et al.,  1995; Nikula et al., 1995), the particle-
26      associated organics are likely to play a greater role at lower doses.  Particle concentration is also
27      the most practical measure of dose because of the large number of surface-associated compounds
28      and because their concentrations are seldom reported.
29            As reviewed in Chapter 7, several bioassay studies showed positive lung tumor responses
30      in rats (Ishinishi et al., 1986; Iwai et al., 1986; Stober, 1986; Mauderly et al., 1987; Brightwell et
31      al.,  1989; Heinrieh et al., 1995; Nikula et al., 1995). Three of these studies (Tables 11-2 through
32      11 -4) were used for unit risk calculations because they were designed using multiple exposure
33      groups and thus are more appropriate for risk calculations. Use of these studies also allows
34      comparison with earlier risk estimates.  The time-to-event (i.e., death with or without tumors)
35      data are available for the Mauderly et al. (1987) study. These time-to-event data are used in all


        2/1/98                                    11-12       DRAFT-DO NOT CITE OR QUOTE

-------
        Table 11-2. Incidence of lung tumors in Fischer 344 rats (males and females
        combined) exposed to heavy-duty engine exhaust
Exposure
concentration
(mg/m3)
0
0.35
3.50
7.08
Dose
Weekly exposure
(mg/m3 x h)a
0
12
122
248
metric
Lung particle
burden (mg/cm2
lung surface)b
0
6.4 x 1Q-5
2.8 x IQ-3
6,0 x lO'3
Lung tumor
incidence1"
2/230
3/223
6/222
18/227
  aExposures were 7 hours/day, 5 days/week for 30 months. Numbers include those dying or moribund prior
  to 30 months,
  ""Calculated using mathematical models in Appendix C.
  cDoes not include squamous cysts.

  Source: Mauderly etal, 1987.
        Table 11-3.  Incidence of lung tumors in Fischer 344 rats (males and females
        combined) exposed to heavy-duty engine exhaust
Exposure
concentration
(mg/m3)
0
0.46
0.96
1.84
3,72
Dose
Weekly exposure
(mg/m3 x h)«
0
44
92
177
357
metric
Lung particle
burden (mg/cm2
lung surface)1"
0
2.5 x ID'4
2.0 x lO'3
4.2 x IQ-3
8.8 x IQ-3
Lung tumor
incidencec
1/123
1/123
0/125
4/123
8/124.
 "Exposures were 16 hours/day, 6 days/week for 30 months. Numbers include those dying or sacrificed moribund
. before 30 months.
 "Calculated using mathematical models in Appendix C.
 "All tumors reported were carcinomas.

 Source:  Ishinishi etal., 1986.


 2/1/98                •                     11-13      DRAFT-DO NOT CITE OR QUOTE

-------
              Table 11-4.  Incidence of lung tumors in Fischer 344 rats (males and females
              combined) exposed to diesel engine exhaust
Dose metric
Exposure
concentration
(mg/m3)
0
0.7
2.2
6.6
Weekly exposure
(mg/m3 x h)a
0
56
176
248
Lung particle
burden (mg/cm2
lung surface)1"
0
3.5 X 1CT4
4.2 x 10~3
1.3 x ia2
Lung tumor
incidence1
4/250
1/112
14/112
55/111
        'Exposures were 16 hours/day, 5 days/week.
        'Calculated using mathematical models listed in Appendix C.
        'The number of animals sacrificed at 6 and 12 months is excluded from the denominators.
        Source: Brightwelietal., 1989.
 1      the risk calculations based on the Mauderly et al. (1987) data. Use of a time-to-tumor model,
" 2      when data are available, allows more accurate modeling of dose response.
 3            Risk estimates were derived using two models.  The first one is the linearized low-dose
 4      extrapolation model. The linearized multistage (LMS) model is adopted by U.S. EPA as a
 5      default procedure to provide an upper bound estimate of risk when data useful to incorporate
 6      plausible mechanisms are not available. This model was selected as one of the approaches
 7      because, although mechanisms of carcinogenesis have been proposed (see Chapter 10), they
 8      remain largely unproven.
 9            The LMS model has the mathematical form P = 1 - exp(-Z), where Z is either Z = q^ +
10      q,xd + ... + qmxdm, a polynomial of concentration d, or Z = (Q0 + Q,xd + ... + Qmxdm) xtk, a
11      polynomial of concentration d multiplied by a time factor tk when time-to-event data are used.
12      The range of extrapolation is about three orders of magnitude in the present study. Because the
13      extra risk (P - P0)/(l - P0) is dominated by the linear term q,xd at low concentration, the 95%
14      upper  bound of qi is used to represent unit risk when d is expressed in ng/m3.  P0 denotes the
15      lifetime cancer risk at concentration 0. The resulting unit risk estimates are listed in Table 11-5.
        2/1/98
11-14
DRAFT-DO NOT CITE OR QUOTE

-------
               Table 11-5. Unit risk estimates per )*g/m3 of diesel exhaust based on the
               Yu et al, (1991) dosimetry model

                                       Lung burden/lung surface area       Lung burden/body weight'7'
         Mauderly et al., 1987                       3.4 x lO"5                      9.9 * 10'5
         Ishinishi et al., 1986                        1.6 x 1Q-5                      4.6 x 10'5
         BrightwelletaL, 1989   '                   7.1 * lO'5      x               2.1 x 10'5
         Geometric mean of three studies               3.4 x 1Q-5                .      9.9 x 10'5
 1             The potency estimates derived using the LMS model are an improvement over most of
 2      the earlier estimates because of improved methods for extrapolating dose from experimental
 3      animals to humans. These assessments as well as earlier ones, however, are based on the premise
 4      that carcinogenic effects are a function of the whole particle concentration.  Moreover, risk
 5      estimates are calculated using the default assumption that responses vary linearly with particle
 6      concentration even at ambient exposure concentrations. As noted in Chapter 10, however, while
 7      particle effects may be dominant at large lung burdens, other components such as polycyclic
 8      aromatic compounds, nitroaromatics, and surface-associated reactive oxygen species are likely to
 9      play a more important role as exposure concentrations approach those found in the ambient air.
10             In an attempt to account for the effects of particles as well as particle-associated
11      components of DE, a biologically based two-stage model with piecewise constant parameters
12      was developed.  This model assumes that both the carbon and the organic fraction have initiating
13      properties and that the carbon fraction also has effects  on the proliferation, conversion and
14      progression steps of carcinogenesis. In addition, it accounts for the tumor initialing and
15      promoting effects at high exposure concentrations. A description of this model has been
16      published by Chen and Oberdorster( 1996).  (See Appendix B for details.) The resultant two-
17      stage model is then used to predict tumor responses under various exposure scenarios and to
18      investigate the impacts of different biological assumptions on risk projection.
19             The  parameters for the two-stage model were statistically estimated initially, using tumor
20      response data from animals exposed to high concentrations of DE and assuming particles
21      continue to exert effects at low concentrations (i.e., nonthreshold for particle effects).  When the
22      two-stage biologically based dose-response model (BBDr) was compared with results derived
23      with the LMS model, using malignant tumor data from the Mauderly et al. (1987) study, the
24      results were nearly identical (Table 11-6). The maximum likelihood estimate (MLE) of cancer
25      risk for the BBDr model at an exposure concentration of 1 ug/m3 was calculated to be 8.2 x 10~6

        2/1/98                                   11-15       DRAFT-DO NOT CITE OR QUOTE

-------
              Table 11-6. Relative importance of particles and organics at different exposure
              concentrations assuming either threshold or nonthreshold effects of particles
                                       Excess risk in humans (MLE)
Concentration
0.1
1.0
100 -.
, i.ooo
Threshold effect of
particles
1.8 x lO;7
1.8 x 10'6
1.2 x ID'2
1.9 * lO'3
Nonthreshold effect of
particles
8.1 x IQ-7
8.2 x lO'6
5.6 x 1Q-4
2.6 x 1C)-2
Ratio"
4.5
'4.6
4.9
14.1
         "Ratio =ratio of column 3 and column 2.
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
with an upper 95% confidence limit of 1.7 x 10"5. The 95% upper confidence limits using the
LMS model were also 1.7 x IQ"5 when rounded to two significant numbers.
       Table 11-6 demonstrates that risk is much smaller if the assumption is made that particles
do not exert effects below a certain concentration. As can be seen in Table 11-6, risk is
decreased 4.6-fold at an exposure concentration of 1 |ig/m3 if a threshold is assumed for particles.
Importantly, not only is risk decreased if a threshold is assumed for particle effects, but while
particles play a major role in tumor induction at high exposure concentrations, organics and other
components of DE play an increasingly dominant role as concentrations become lower.
       An alternative approach that avoids the controversies related to lung overload with
concomitant pathology involves calculation of an upper-bound risk based only on animals
exposed to low diesel particle concentrations (e.g., below 0.5 mg/m3). This modeling is designed
to cap a theoretical upper-bound estimate on a nonobservable rat resonse to organics, while
assuming organics and possibly free radicals are driving carcinogenicity at low exposure
concentrations. As a general matter, the newly proposed Carcinogenicity Risk Assessment
Guidelines suggest identifying a point of departure on the dose-response curve, perhaps an ED10
or some other EDX, and then extrapolating risk from that point of departure. The low-dose
groups from the Mauderly et al. (1987) and Ishinishi et al. (1986) studies are the most suitable for
this purpose.  Both include exposure groups at less than 0.5 mg/m3, and have the same duration
follow-up (30 mo). By combining studies, statistical power is increased, resulting in a smaller
upper-bound estimate of risk than by either study alone. The resultant upper-bound estimate is
equal to 1.9 x  I0'54ig/m3.
       While this approach has the advantage of avoiding controversies related to lung particle
overload, it also has disadvantages.  First, because lung tumors were not significantly increased
        2/1/98
                                         11-16
DRAFT-DO NOT CITE OR QUOTE

-------
  1      at the doses used, the estimated risk is considered unlikely to exceed the upper bound, but may be
  2      as low as zero. Second, the upper-bound estimate of risk is greater than the unit risk estimates
  3      derived using data from animals exposed to all concentrations of DE.
  4
  5      11.3.2.3. Dose-Response Estimation Using Benzo(a)pyrene as a Dosimeter
  6            Pike and Henderson (1981) found good agreement when relating the concentration of
  7      benzo(a)pyrene (B[a]P) to lung cancer risk in smokers, British gas workers, U.S. coke oven
  8      workers, and U.S. hot pitch workers and when comparing residents of rural and urban locations.
  9      They concluded that while B[a]P is unlikely to be the only carcinogen present and perhaps not
10      even the most important one present in combustion emissions, nevertheless it serves as a
11      reasonably accurate dosimeter. Based on an estimated cancer risk of 1/1,500 per ng/m3B[a]P
12      and a reported B[a]P concentration of 3.9 ng/ug DPM in exhaust from a Volkswagen engine
13      (Heinrich et al.,  1995),  a maximum likelihood estimate of cancer risk from lifetime exposure to
14      1 ng/m3  can be calculated to be 3 x 1Q"6. The 95% upper bound was not derived, but is estimated
15      to be near 1 x 10'5.
16
17      11.3.3.  Strengths and Weaknesses of Differing Approaches for Quantitating Cancer Risk
18      11.3.3.1. Dose-Response Estimates Based on Human Epidemiotogic Data
19            A major advantage in the use of human data is the elimination of uncertainty due to
20      possible differences in sensitivity to cancer induction by DE among species. As will be
21      discussed later, there is some evidence that rats may respond differently from humans to DE.
22      Second,  epidemiology studies are based on occupational exposures, which generally .occur at
23      concentrations insufficient to result in lung particle overload.  Thus, lung cancer in the human
24      studies is likely to be induced by non-particle-overload mechanisms under either occupational or
25      ambient  exposure levels. Cancer induction in  the rat bioassays, on the other hand, was observed
26      only under particle-overload conditions, with concomitant pathology. Uncertainty in
27      extrapolating from occupational studies is therefore decreased, not only because low-dose
28      extrapolation occurs over a smaller range, but because mechanisms of cancer induction are less
29      likely to vary within this range with accompanying changes in the dose-response curve.
30            There is considerable evidence for nonoverload mechanisms of cancer induction by
31      products of fossil fuel combustion. Mumford et al. (1989) reported greatly increased lung cancer
32      mortality in Chinese communes burning so-called smoky coal containing high concentrations of
33      PAHs. Demonstration of the carcinogenicity of coke oven emissions in humans (Lloyd, 1971)
34      also provided evidence  for a role by organics because coke oven PM contains a high
35      concentration of PAHs but lacks an insoluble carbon core.  Increased levels of aromatic DNA
36      adducts were reported in bus maintenance and terminal workers by Hemminki et al. (1994) and
        2/1/98                                   11-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1      in garage workers and mechanics exposed to diesel exhaust (Nielsen and Autrup, 1994). Studies
 2      by Sagai et al. (1993) have indicated that DPM could produce superoxide and hydroxyl radicals
 3      in vitro, without any biologically activating systems. On the basis of these findings, they
 4      suggested that most DE toxicity in lungs is due to active oxygen radicals. In a more recent study,
 5      these investigators reported that instillation of only 0.1 mg of DPM into mouse lungs resulted in
                                V
 6      the production of 8-hydroxyguanosine in lung cell DNA. The critical lesion may thus be induced
 7      by oxygen free-radicals generated from DPM (Nagashima et al., 1995).
 8             An uncertainty associated with most of the diesel epidemiology studies was the inability
 9      to completely eliminate all confounding factors, resulting in possible errors in estimating relative
10      risk ratios. Small errors in adjustment for smoking, for example, can result in considerable error
11      because smoking has a much larger effect on relative cancer risk than is likely for DE. The
12      likelihood of significant confounding errors in the Garshick et al. (1987 and 1988) studies is
13      decreased by the considerable effort exerted to eliminate or reduce such factors, especially
14      smoking. Moreover, a meta-analysis by the California EPA (Cal-EPA, 1997) and more recently
15      one by Bhatia et al. (1998) using a number of diesel epidemiology studies resulted in relative risk
16      ratios quite similar to the one reported by Garshick et al. (1987).  Although exposure levels are
17      likely to have differed somewhat among studies, the agreement still suggests that a relative risk
18      near 1.4 is reasonable.
19             The greatest uncertainty in estimating DE-induced cancer risk from epidemiology studies
20      is determination of exposure levels.  Even though DPM concentrations were often measured near
21      the end of the studies, historic exposure data are generally lacking.  Such information is critical,
22      since there is indirect evidence, based on other pollutant measurements such as nitrogen oxides,
23      that exposure levels have decreased  considerably in recent years, especially in the railroad
24      industry (Woskie et al., 1988b).  In the only historic study found in which DPM were measured,
25      Heino et al. (1978) reported average concentrations of 2 mg/m3 in Finnish roundhouses.  Woskie
26      et al. (1988a), by contrast, reported a mean of 134  ug/m3 for roundhouse workers near the end of
27      tha Garshick et al. (1987, 1988) studies. While the relationship between DPM concentrations in
28      Finnish and U.S. railroad roundhouses during the 1970s is uncertain, it does point to the
29      likelihood that exposure levels have decreased over time.
30             Despite the uncertainties  discussed above, the epidemiology data-based risk assessments
31      are useful for bounding risk. The control of confounding factors in the Garshick et al. (1987)
32      study and the general agreement of relative risk estimates with those of other studies renders it
33      unlikely that the true risk is greatly different from the reported one. The mean DPM
34      concentration for all the occupational groups was somewhat less than 125 ug/m3. Historic
35      exposures were undoubtedly greater, but unlikely to approach 2 mg/m3 because the cases in this
36      study included not only roundhouse workers but other occupational groups as well, who were
        2/1/98                •                   11-18       DRAFT-DO NOT CITE OR QUOTE

-------
  1      likely exposed at lower concentrations. Based on the evidence available, there is considerable
  2      confidence that mean exposures are in the range of 125-500 ^ig/m3,
  3
  4      11,3.3.2. Dose-Response Estimate-Based Laboratory Animal Bioassays
  5             Cancer risk assessment from exposure to DE, based on available animal bioassays, has a
  6      variety of strengths.  The studies selected are adequately designed, eliminating confounding
  7      factors often present in epidemiology studies.  Exposure duration and exposure levels were
  8      precisely controlled and monitored. The presence or absence of tumors were verified by
  9      pathological evaluation. Although animal-to-human extrapolation of dose is required, the
10      development of dosimetry models has eliminated much of the uncertainty in this area.
11             Nevertheless, two important uncertainties remain: the adequacy of the rat as a model for
12      evaluating human risk of cancer from exposure to DE and the shape of the dose-response curve.
13      It is believed by some that the rat may be unique in its response to particulate matter, and
14      therefore its use for assessing human lung cancer risk is questionable (Mauderly, 1994; Watson
15      and Valberg, 1995).  As noted in Chapter 7, the rat is the only species that has unequivocally
16      been shown to develop lung cancer in response to DE exposure.  On the other hand, only  two
17      other species, Syrian hamsters and mice, have been tested adequately. Although responses in the
18      hamster have been negative, this species is known to be resistant to induction of lung cancer
19      (Heinrich, 1994). Mice also fail to respond in standard bioassays of DE (Heinrich et al,, 1995;
20      Mauderly etal,  1996). Positive results, however, have been reported in mice exposed from birth
21      or conception (Pepelko and Peirano, 1983; Takemoto et al., 1986), a condition likely to increase
22      sensitivity.  In the case of the study reported by Pepelko and Peirano, Sencar mice were used, a
23      strain sensitive to chemical induction of skin tumors. Since the alveolar lining of the lung is  also
24      epithelial tissue, it is possible these mice are more sensitive to the organic constituents of DPM.
25      Ichinose et al. (1997a,b) reported increased lung tumors in mice following intratracheal
26      instillation of DPM.  Positive results in mice may have therefore been due to use of a sensitive
27      strain, early exposure to increased sensitivity, or very high local concentrations due to the
28      instillation procedure.
29             It has also been argued that humans are resistant to particle-induced lung cancer; although
30      coal miners develop pneumoconiosis, lung cancer seldom occurs. Rats, on the other hand, were
31      reported to develop lung cancer in response to coal dust (Martin et al., 1977). This study,
32      however, was poorly described, and the number of animals exposed was small (4/36 developed
3.3      lung cancer). Moreover, exposure levels were very high and lung burdens were greater than
34      those generally encountered in coal miners (Mauderly, 1994). Finally, although lung cancer has
35      not been reported in most epidemiology studies of coal miners, Zhong and Dehong (1995)

        2/1/98  .                                 11-19       DRAFT-DO NOT CITE OR QUOTE

-------
 1      reported that Chinese workers suffering coal miners' pneumoconiosis did have an increased risk
 2      of lung cancer.
 3             The evidence for the uniqueness of the rat carcinogenic response to DE is therefore still
 4      equivocal. Nevertheless, the evidence to date does raise questions concerning the suitability of
 5      the rat for assessing human risk from exposure to DE. This concern is heightened by the fact that
 6      particle deposition patterns are different in the rat and human. Because of the absence of
 7      respiratory bronchioles in the rat, a greater fraction of inhaled particles deposit in the alveolar
 8      regions; in primates, deposition occurs to a large extent at the bifurcation of the small bronchi.
 9      Differing deposition patterns may result in different pathologic responses, as reported by Nikula
10      et al. (1997) for rats and monkeys.
11             The other important uncertainty relating  to use of rat bioassay data concerns low-dose
12      extrapolation.  Significant lung tumor increases  in experimental animals have generally been
13      obtained only at concentrations resulting in lung particle overload with concomitant pathological
14   .   effects.  A hypothesis first proposed by Vostal etal. (1986) is based on the belief that diesel
15      particles themselves induce lung cancer, through a secondary effect. Driscoll (1995) suggested
16      that the secondary effect was due to release of various inflammatory mediators by particle
17      overloaded phagocytic cells.  The resultant inflammatory response, with accompanying cell
18      division, can increase the  likelihood that any oxidant-induced or spontaneously occurring genetic
19      damage becomes fixed in a dividing cell and is clonally expanded. In support of this hypothesis
20      are the observations that lung cancer can be induced by inhalation of so-called "inert" particles,
21      such as titanium dioxide (Lee et al., 1986) and coal dust (Martin et al., 1977), or by intratracheal
22      instillation of activated carbon (Kawabata et al., 1986).  Inhalation of carbon black that was
23      virtually devoid of organics and is similar to the carbon core of diesel exhaust was also reported
24      to induce lung cancer in rats (Heinrich, 1990; Heinrich et al., 1995; Nikula et al., 1995). Finally,
25      Kawabata et al. (1993) induced lung tumors in rats through intratracheal instillation with DPM
26      from which the organic components had been extracted.
27             Although considerable evidence for the particle-overload hypothesis exists, occupational
28      exposures in which lung cancer induction has been reported in humans are generally considered
29      insufficient to induce lung particle overload. This suggests that other mechanisms are also likely
30      to be operating. Experimental evidence also provides support for low-dose mechanisms. Riebe-
31      Imre et al. (1994) reported that carbon black is taken up by lung epithelial cells in vitro, inducing
32      chromosomal damage and disruption of the cytoskeleton (lesions that closely resemble those in
33      tumor cells) at concentrations that did not induce measurable toxicity. As noted in the previous
34      section, diesel particles are capable of inducing toxicity and possible carcinogencity at low doses
35      through production of oxygen free radicals'. Finally, Dasenbrock etal. (1996) reported that
36      extraction of the organic fraction from diesel particles decreased their carcinogenic potency, even
        2/1/98                                   11-20       DRAFT-DO NOT CITE OR QUOTE

-------
  1      though specific particle surface area, a factor related to cancer potency, is increased following
  2      extraction,
  3             In summary, the use of rat data for quantitating human lung cancer risk from exposure to
  4      DE results in considerable uncertainty.  There is some doubt concerning the validity of the rat
  5      model for this purpose, although proof is still lacking. The greatest uncertainty concerns low-
  6      dose extrapolation of lung cancer risk. Information to date indicates that DE induces lung cancer
  7      by more than one means, including particle overload-induced pathology, carcinogenic organics
  8      present on the particle, and oxygen free radicals. While all these methods are likely to function
  9      at high doses, particle effects are likely to be  absent or minimal at low exposures, resulting in a
10      high degree of uncertainty regarding the shape of the dose-response curve.
11             Despite these uncertainties, rat data are still useful for comparison with other estimates,
12      especially those derived using human epidemiology data. The greater risk estimates derived using
13      epidemiology data, for example, suggest that even though rats are the only laboratory species in
14      which lung cancer can be regularly induced by DE and other fine particulate matter, they may be
15      less sensitive than humans to honoverload mechanisms induced by DE at low concentrations.
16      Unfortunately, limitations on animal numbers do not allow testing this possibility.
17                                            -
18      11.3.3.3.  Dose-Response Estimates Based on the Biomarker Approach
19             This approach provides reasonably good estimates of lung cancer risk in spite of the fact
20      that B(a)P may constitute a relatively small fraction of the carcinogens present in combustion and
21      pyrolysis .products of coke ovens, hot pitch gas productions, etc. Risk estimates were also based
22      on well-documented lung cancer rates in the occupationally exposed groups.  On the other hand,
23      while predictions are good for the pollutants tested, the particles present, unlike diesel particles,
24      generally lack an insoluble carbon core. As noted below, adsorption to an insoluble particle core
25      may influence risk estimates. Estimates of cancer risk will also vary based on B(a)P
26      concentration .on the particle. The variability in B(a)P concentration among different DE sources
27      and its effect on cancer potency have not been evaluated.
28                                                                  .
29      11.3.3.4.  Dose-Response Estimates Based on the Comparative Potency Approach
30             In this method, the potency of diesel DPM extract is compared with other combustion or
31      pyrolysis products, coke oven emissions, roofing tar, and cigarette smoke condensate for which
32      epidemiology-based unit risk estimates have been developed. Comparisons are made using
33      short-term tests such as skin painting, mutation, and mammalian cell transformation. The ratio
34      of the potency of DPM extracts to each of these agents is then multiplied by their unit risk
35      estimates to obtain the unit risk for DE.  Because no new studies have been carried out since the

        2/1/98                                    11-21      DRAFT-DO NOT CITE OR QUOTE

-------
 1      early 1980s, the values considered are those published by Albert et al. (1983) and listed in Table
 2      11-1,
 3            This test is based on the belief that cancer induction at low doses is due to the organic
 4      fraction present on diesel particles. A major strength of this approach is avoidance of lung
 5      particle overload effects. The tests also have shown that the organic fraction of DE can induce
 6      cancer. - A degree of uncertainty is due to the possibility that the presence of particles may alter
 7      the carcinogenicity of the organics. For example, particle-associated organics may deposit in
 8      different regions of the lungs than free organics do.  Association with particles may increase
 9      residence time, thus increasing potency. On the other hand, a fraction of the organics may not be
10      bioavailable because of tight adherence to the particle. Another uncertainty involves the
11      assumption that cancer potency in short-term tests is the same as that for lung cancer induction.
12      This is somewhat mitigated by the fact that the assessment is based to a large degree on skin
13      painting studies. Skin and lung surfaces are both epithelial tissue, increasing the likelihood that
14      responses will be similar. Nesnow et al. (1995) have, in fact, showed that for many poly cyclic
15      aromatic compounds relative cancer potency by skin painting and lung adenoma induction in
16      strain A mice was  quite similar.
17
18      11.4. SUMMARY AND RECOMMENDATIONS
19            Based on limited evidence from human epidemiologic studies and adequate evidence
20      from animal experiments, DE is considered to be a probable human carcinogen.  The evidence
21      places DE in the upper range of agents classified as probable human carcinogens because of
22      positive findings in both animal experiments and human data as well as the presence of known
23      carcinogens in the mixture. The inability to completely eliminate confounding factors in most of
24      the epidemiology studies, the lack of clearcut carcinogenic responses in more than one animal
25      species, and the low concentration of organics precluded classifying DE as a known human
26      carcinogen.
27            Cancer potency has been estimated using human epidemiology data, chronic animal
28      cancer bioassays, a comparative potency approach, and use of B[a]P as a dosimeter. Each of
29      these methods has important uncertainties that preclude recommendation of a point estimate of
30      risk. Uncertainties relating to use of epidemiology studies include limited historical exposure
31      data, inability to eliminate all possible confounding factors, and small increases in relative risk
32      that can be more easily influenced by confounding factors. Uncertainties in use of chronic
33      animal bioassays stem from questions regarding  relative sensitivities of rats and humans to DE
34      and shape of the low-dose response curve.  In the comparative potency method, the assumption
35      that relative potency in short-term tests accurately reflects relative cancer potency of different
36      combustion emissions requires confirmation.  While use of B[a]P as a dosimeter has been shown
        2/1/98                                    11-22       DRAFT-DO NOT CITE OR QUOTE

-------
  1      to reflect the potency of several combustion emissions reasonably well, DE differs from these
  2      emissions because of the presence of an insoluble carbon core.
  3             Despite these uncertainties, they are considered, collectively, adequate to place reasonable
  4      bounds on risk.  An upper-bound lung cancer risk from lifetime exposure to DE at a paniculate
  5      matter concentration of 1 fig/m3 of 2 x 10"3, based on the estimate derived by McClellan et al.
  6      (1989), using the case-control study of Garshick et al. (1987) and assuming a mean exposure
  7      concentration of 125 u-g/m3 is recommended.  It is considered to be a reasonable upper bound
  8      because relative risk ratios were either less or only slightly greater in other epidemiology studies
  Q      with DE,  Furthermore, exposures were unlikely average less than 125  ug/m3 paniculate matter
10      assumed.  A lower bound of 1 * 10"s/(ug/m3) is recommended. MLEs slightly less than this value
11      were obtained from animal bioassay data by Chen and Oberdorster (1996) using a biologically
12      based dose-response model and assuming nonthreshold effects for particles and by estimates
13      based on use of B[a]P as a dosimeter according to the Pike and Henderson (1981) report and
14      B [a]P content of DE reported by Heinrich et al. (1995). Most comparative potency estimates
15      were only slightly greater than 1  * ] 0"Vu.g/m3.  The lower bound is considered to be reasonable
16      for at least two reasons. First, the latter two approaches are based on effects of the organic
17      constituents. Second, while they are likely to be the primary cause of lung cancer at low doses,
18      other factors such as particle effects and reactive oxygen species may play some role. Thus, risk
19      is unlikely to be significantly lower. This conclusion is supported by animal data-based
20      calculations of Chen and Oberdorster (1996) using a biologically based, low-dose extrapolation
21      model.
22             Finally, the 95% lower-bound estimate of risk, assuming exposures of 500 (ig/m3, is 5  x
23      10~5/u.g/m3. The range of human data-based risk estimates therefore encompasses a large portion
24      of the recommended bounds of 1 to 200 x 10"5/jig/m3.
25             ^     .                    •                                                 .
26      11.5. 'REFERENCES '
27        '    ..    .             .
28      Ahlman, K; Koskela, RS; Kuilla, P; Koponen, M; Annanmaki, M. (1991) Mortality among sulfide ore miners. Am  J
29      IndMed 19:603-617.
30
31      Albert, RE; Chen, C. (1986) U.S. EPA diesel studies on inhalation hazards. In: Ishinishi, N; Koizumi, A; McClellan,
32      R; Stober, W, eds. Carcinogenicity and mutagenicity of diesel engine exhaust. Amsterdam: Elsevier, pp. 411-419.
33
34      Albert, RE; Lewtas, J; Nesnow, S; Thorslund, TW; Anderson, E. (1983) Comparative potency method for cancer
35      risk assessment: application to diesel particulate emissions. Risk Anal 3:101-117.
36 '    •  •           :            '
37      Benhamou, S; Benhamou, E; Flamant, R. (1988) Occupational risk factors of lung cancer in a French case-control
38      study. BrJIndMed 45:231-233.
39    .
40      Bhatia, R; Lopipero, P; Smith, AH. (1998) Diesel exhaust exposure and lung cancer.  Epidemiology 9:84-91.
41

        2/1/98                '                    H-23       DRAFT-DO NOT CITE OR QUOTE

-------
  1       Boffetta, P; Stellman, SD. (1988) Association between diesel exhaust exposure and multiple myeloma: an example
  2       of confounding. Prev Med 17:236-237.
  3
  4       Bohning, DE; Atkins, HL; Cohn, SH. (1982) Long-term particle clearance in man: normal and impaired. Ann Occup
  5       Hyg 26:259-271.
  6
  7       Brightwell, J; Fouillet, X; Cassano-Zoppi, AL; Bernstein, D; Crawley, F; Duchosal, F; Gatz, R; Perczel, S; Pfeifer,
  8       H. (1989) Tumours of the respiratory tract in rats and hamsters following chronic inhalation of engine exhaust
  9       emissions.  J Appl Toxicol 9:23-31.
10
11       Burns, PB; Swanson, GM. (1991) The occupational cancer incidence surveillance study (OCISS): risk of lung cancer
12       by usual occupation and industry in the Detroit metropolitan area. Am J Ind Med 19:655-671.
13
14       Chan, TL;  Lee, PS; Hering, WE. (1981) Deposition and clearance of inhaled diesel exhaust particles on the
15       respiratory tract of Fischer 344 rats. J Appl Toxicol 1:77-82.
16
17       Chan, TL;  Lee, PS; Hering, WE. (1984) Pulmonary retention of inhaled diesel particles after prolonged exposure to
18       diesel exhaust. Fundarh Appl Toxicol 4:624-631.
19
20       Chen, CW; Oberdorster, G. (1996) Selection of models for assessing dose-response relationships for particle-
21       induced lung cancer. Inhal Toxicol 8(suppl):259-278.
22
23       Crump, KS; Lambert, T; Chen, C. (1991) Assessment of risk from exposure to diesel engine emissions. Clement
24       International Corporation. Prepared for U.S. EPA under contract no. 68-02-4601; 56 pp.
25
26       Cuddihy, RG; McClellan, RO. (1983) Evaluating lung cancer risks from exposures to diesel engine exhaust. Risk
27       Anal 3:119-124.
28
29       Cuddihy, RG; Griffith, WC; Clark, CR; McClellan, RO. (1981) Potential health and environmental effects of light-
30       duty diesel vehicles. Report No.  LMF-89. Albuquerque, NM: Inhalation Toxicology Research Institute.
.31
32       Cuddihy, RG; Griffith, WC; McClellan, RO. (1984) Health risks from light-duty diesel vehicles. Environ Sci
33       Technol 18:14-2 la.
34
35       Damber, LA; Larsson, LG. (1987) Occupation and male lung cancer: a case-control study in northern Sweden. Br J
36       Ind Med 44:446-453.
37
38       Dasenbrock,  C; Peters, L; Creutzenberg, O; Heinrich, U. (1996) The carcinogenic potency of carbon particles with
39       and without PAH after repeated intratracheal administration in the rat. Toxicol Lett 88:15-21.
40                                            .
41       Driscoll, K. (1995) Role of inflammation in the development of rat lung tumors in response to chronic particle
42       exposure. Inhal Toxicol 8(Suppl): 139-153.
43
44     . Edling, O; Anjou, CG; Axelson, O; Kling, H. (1987) Mortality among personnel exposed to diesel exhaust. Int Arch
45       Occup Environ Health 59:559-565.
46
47       Emmelin, A; Nystrom, L; Wall, S. (1993) Diesel exhaust exposure and smoking: a case-referent study of lung cancer
48       among  Swedish dock workers. Epidemiology 3:237-244.
49
50       Garshick, E;  Schenker, MB; Munoz, A; Segal, M; Smith, TJ; Woskie, SR; Hammond, SK; Speizer, FE. (1987) A
5.1       case-control study of lung cancer and diesel exhaust exposure in railroad workers. Am Rev Respir Dis  135:1242-
52       1248.
53
         2/1/98   .                                      11-24        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Garshick, E; Schenker, MB; Munoz, A; Segal M; Smith, TJ; Woskie, SR; Hammond, SK; Speizer, FE, (1988) A
  2       retrospective cohort study of lung cancer and diesel exhaust exposure in railroad workers. Am Rev Respir Dis
  3       137:820-825.
.  4
  5  •     Griffis, LC; Wolff, RK; Henderson, RF; Griffith, WC; Mokler, BV; McClellan, RO. (1983) Clearance of diesel soot
  6       particles  from rat lung after subchronic diesel exhaust exposure. Fundam Appl Toxicol 3:99-103.
  7
  8       Grimmer, G; Brune, H; Deutsch-Wenzel, R; Dettbarn, G; Jacobs, J; Navjack, KW; Mohr, U; Ernst, H, (1987)
  9       Contribution of poiycyclic aromatic hydrocarbons and nitro-derivatives to the carcinogenic impact of diesel exhaust
 10  .     condensate evaluated by implantation into the lungs of rats. Cancer Lett 37:173-180.
 11
 12       Gustavsson, P; Plato,  N; Lidstrom, EB; Hogstedt, C. (1990) Lung cancer and exposure to diesel exhaust among bus
 13       garage workers. ScandJ Work Environ Health 16:348-354.
 14
 15       Hall, NE; Wynder, E. (1984) Diesel exhaust exposure and lung cancer. Environ Res 34:77-86.
 16
 17       Harris, JE. (1983) Diesel emissions and lung cancer. Risk Anal 3:83-100.
 18
 19       Hattis, D; Silver, K, (1994) Use of mechanistic data in occupational health risk assessment: The example of diesel
 20       particulates. In: Smith, MC; Christiani, DC; Kelsey, KT eds. Chemical risk assessment and occupational health.
 21       Auburn House: Westport,  CT, pp. 167-178.
 22  •
 23       Hayes, RB; Thomas T; Silverman, DT; Vineis, P; Blot, WJ; Mason, TJ; Pickle, LW; Correa, P; Fontham, ETH;
 24       Schoenberg, JB. (1989) Lung cancer in motor exhaust occupations. Am J Ind Med 16:685-695.
 25                       -•
 26       Heinrich, U. (1990) Carcinogeniciry of carbon black "Printex 90" in female Wistar rats. Presented at: U.S. EPA
 27       Diesel Emissions Peer Review  Workshop; July 18-19; Chapel Hill, NC.
 28
 29       Heinrich, U. (1994) Carcinogenic effects of solid particles. In: Mohr, U; Dungworth, DL; Mauderly, JL;
 30       Oberdorster, G, eds. Toxic and carcinogenic effects of solid particles in the respiratory tract. Washington, DC: ILSI,
 31       pp. 57-73.                   '                                                      '
 32                                      '             •
 33       Heinrich, U; Muhle, H; Takenaka, S; Ernst, H; Fohst,  R; Mohr, U; Pott, F; Stober, W.  (1986a) Chronic effects on the
 34       respiratory tract of hamsters, mice and rats after long-term inhalation of high concentration of filtered and unfiltered
 35       diesel engine emissions. J  Appl Physiol 6:383-395.
 36
 37       Heinrich, U; Pott, F; Rittinghausen, S. (1986b) Comparison of chronic inhalation effects in rodents after long-term
 38       exposure to either coal oven flue gas mixed with pyrolized pitch or diesel engine exhaust. In: Ishinishi, N; Koizumi,
 39       A; McClellan, RO; Stober, W, eds. Carcinogenic and mutagenic effects-of diesel engine exhaust. Amsterdam:
 40       Elsevier,  pp. 441-457.
 41
 42       Heinrich, U; Fuhst, R; Rittinghausen, S; Creutzenberg, O; Bellmann, B; Koch, W; Levsen,  K. (1995) Chronic
 43       inhalation exposure of Wistar rats and two different strains of mice to diesel engine exhaust, carbon black, and
 44       titanium dioxide. Inhal Toxicol 7:553-556.
 45                       .             •                    '                  •'....
 46       Hemminki, K;  SOderling, J; Ericson, P; Norbeck, HE; Segerback, D. (1994) DNA adducts among personnel
 47       servicing and loading diesel vehicles. Carcinogenesis  15:767-769.
 48
 49       Howe, GR; Fraser, D; Lindsay, J; Presnal, B; Yu, SZ.  (1983) Cancer mortality (1965-77) in relation to diesel fume
 50       and coal  exposure in a cohort of retired railway workers. J Natl Cancer Inst 70:1015-1019.
 51                                                                                                     '
         2/1/98                                         11-25        DRAFT-DO NOT CITE OR QUOTE

-------
 1       Huisingh, J; Bradow, R; Jungers, R; Claxton, L; Zweidinger, R; Tejada, S; Bumgarner, J; Duffield, F; Waters, M;
 2       Simmon, V; Hare, C; Rodriguez, C; Snow, L. (1978) Application of bioassay to the characterization of diesel
 3       particle emissions. In: Waters, MD; Nesnow, S; Huisingh, JL; Sandhu, SS; Claxton, L, eds. Application of short-
 4       term bioassays in the fractionation and analysis of complex environmental mixtures. New York: Plenum Press, pp.
 5       383-418.
 6
 7       Ichinose, T; Yajima, Y; Nagashima, M; Takenoshita, S; Nagamachi, Y; Sagai, M. (1997a) Lung carcinogenesis and
 8       formation of 8-hydroxyguanosine in mice by diesel exhaust particles. Carcinogenesis  18:185-192.
 9
10       Ichinose, T; Yamanushi, T; Seto, H; Sagai, M. (1997b) Oxygen radicals in lung carcinogenesis accompanying
11       phagocytosis of diesel exhaust particles. Int J Oncol 11:571-575.
12
13       International Agency for Research on Cancer. (1989) IARG monographs on the evaluation of carcinogenic risks to
14       humans, vol. 46. Diesel and gasoline engine exhausts and some nitroarenes. Lyon, France: IARC, pp. 41-185.
15
16       Ishinishi, N; Kuwabara, N; Nagase, S; Suzuki, T; Ishiwata, S; Kohno, T. (1986) Long-term inhalation studies on
17       effects of exhaust from heavy and light duty diesel engines on F344 rats. In: Ishinishi, N; Koizumi, A; McClellan,
18       RO; Stober, W, eds. Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam: Elsevier, pp. 329-
19       348.
20
21       Iwai, K; Udagawa, T; Yamaguchi, M; Yamada, H. (1986) Long-term inhalation studies of diesel exhaust on F344
22       SPF rats. Incidence of lung cancer and lymphoma. In: Ishinishi, N; Koizumi, A; McClellan, RO; Stober, W, eds.  .
23       Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam:  Elsevier, pp. 349-360.
24
25       Karagianes, RF; Palmer, RF;  Busch, RH. (1981) Effects of inhaled diesel emissions and coal dust in rats. Am Ind
26       HygAssocJ 42:382-391.                        •
27
28       Kawabata, Y; Iwai, K; Udagawa, T; Tukagoshi, K; Higuchi, K. (1986) Effects of diesel soot on unscheduled DNA
29       synthesis of tracheal epithelium and lung tumor formation. In: Ishinishi, N; Koizumi, A; McClellan, R; Stober, W,
30       eds. Carcinogenic and mutagenic effects of diesel engine exhaust. Amsterdam: Elsevier, pp. 213-222.
31
32       Kawabata, Y; Udagawa, T; Iwai, K. (1993) One year exposure to diesel engine exhaust causes lung tumors. Toxic
33       and carcinogenic effects of solid particles in the respiratory tract. Fourth International Inhalation Symposi,
34       Hannover, 61 pp.
35
36       Kotin, P; Falk, H; Thomas, M. (1955) Aromatic hydrocarbons. III. Presence in the paniculate phase of diesel engine
37       exhausts and the carcinogenicity of exhaust extracts. Arch Ind Health  11:113-120.
38
39       Kuschner, M. (1984) The  causes of cancer. Am Rev Respir Dis 98:573-590.
40                     ..
41       Lee, KP; Henry, NW, III;  Trochimowicz, HJ; Reinhardt, CF. (1986) Pulmonary response to impaired lung clearance
42       following excessive Ti02  dust deposition. Environ Res 41:144-167.
43
44       Lerchen, ML; Wiggins, CL; Samet, JM. (1987) Lung cancer and occupation in New Mexico. J Natl Cancer Inst
45       79:639-645.
46
47       Lloyd, JW. (1971) Long-term mortality study of steelworkers V. Respiratory cancer in coke plant workers. J Occcup
48       Med 13:53-68.
49
50       Martin, JC; Daniel, H; LeBouffant, L. (1977) Short- and long-term experimental study of the toxicity of coal mine
51       dust and some of its components. In: Walton, WH, ed. Inhaled particles IV. Vol 1. Oxford, Pergamon: pp. 361-370.
52
53       Mauderly, JL. (1994) Contribution of inhalation bioassays to the assessment of human health risks from solid
54       airborne particles. In: Mohr, U; Dungworth, DL; Mauderly, JL; Oberdorster, G, eds. Carcinogenic effects of solid
55       particles in the respiratory tract. Washington, DC: ILSI, pp. 355-365.


         2/1/98                                          11-26        DRAFT-DO NOT CITE OR QUOTE

-------
  1       Mauderly, JL; Jones, RK; Griffith, WC; Henderson, RF; McClellan, RO. (1987) Diesel exhaust is a pulmonary
  2       carcinogen in rats exposed chronically by inhalation. Fundam Appl Toxicol 9:208-221.
  3       ,
  4       Mauderly, JL; Banas, DA; Griffith, WC; Harm, FF; Henderson, RF; McClellan, RO. (1996) Diesel exhaust is not a
  5       pulmonary carcinogen under conditions carcinogenic to F344 rats. Fundam Appl Toxicol 30:233-242.
  6
•  7       McClellan, RO. (1986) Health effects of diesel exhaust: a case study in risk assessment. Am Ind Hyg Assoc J 47:1-
  8       13.              .
  9
10       McClellan, RO; Cuddihy, RG; Griffith, WC; Mauderly, JL. (1989) Integrating diverse data sets to assess the risks of
11       airborne pollutants. In: Mohr, U, ed. Assessment of inhalation hazards. New York: Springer Verlag, pp. 3-22.
12
13       Mumford, JL; Chapman, RS; Harris, DB. (1989) Indoor air exposure to coal and wood combustion emissions
14       associated with a high lung cancer rate in Xuan Wei, China. Environ Int 15:315-320.
15
16       Nagashima, M; Kasai, H; Yokata, J; Nagamachi, Y; Ichinose, T; Sagai, M. (1995) Formation of an oxidative DNA
17       damage, 8-hydroxyguanosine,  in mouse lung DNA after intratracheal instillation of diesel exhaust particles and
18       effects of high dietary fat and beta carotene on this process. Carcinogenesis 6:1441-1445.
19
20       Nesnow, S; Evans, C; Stead, A; Creason, J; Slaga, TJ; Triplett, LL. (1982) Skin carcinogenesis studies of emission
21     •  extracts. In:  Lewtas, J, ed. Toxicological effects of emissions from diesel engines. Amsterdam: Elsevier, pp, 295-
22       320.
23
24       Nesnow, S; Ross, JA; Stoner, GD; Mass, MJ. (1995) Mechanistic linkage between DNA adducts, mutations in
25       oncogenes and tumorigenesis of carcinogenic environmental polycyclic aromatic hydrocarbons in strain A/J mice.
26       Toxicology  105:403-413.
27
28       Nielsen, PS; Autrup, H. (1994) Diesel exhaust-related DNA adducts in garage workers. Clin Chem 40:1456-1458.
29                                                          .
30       Nikula, KJ; Snipes, MB;  Barr,  EB; Griffith, WC; Henderson, RF; Mauderly, JL. (1995) Comparative pulmonary
31       toxicities and carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl
32       Toxicol 25:80-94.                                 '
33       .'
34       Nikula, KJ; Avila, KJ; Griffith, WC; Mauderly, JL. (1997) Lung tissue responses and sites of particle retention differ
35       between rats and Cynomolgous monkeys exposed chronically to diesel exhaust and coal dust. Fundam Appl Toxicol
36 •      37:37-53.
37
38       Pepelko, WE; Chen, C. (1993) Quantitative assessment of cancer risk from exposure to diesel engine emissions. Reg
39       Toxicol Pharmacol 17:52-65.
40                                 '
41       -Pepelko, WE; Peirano, WB. (1983) Health effects  of exposure to diesel engine emissions. J Am Coll Toxicol 2:253-
42       306.
43
44       Pike, ME; Henderson, BE. (1981) Epidemiology of polycyclic aromatic hydrocarbons: quantifying the cancer risk
45       from cigarette smoking and air pollution. In: Gellbond, HV; Tso, POP, eds. Polycyclic hydrocarbons and cancer.
46       New York: Academic Press, pp. 317-334.
47
48    -   Pott, F; Heinrich, U, (1987) Dieselmotorabgas und Lungenk auf die Gefahrdung des Menschen. In: Umwelthygiene,
49       vol. 19. Med. Institut. f. Umwelthygiene, Annual Report 1986/87. Dusseldorf, F.R.G., pp. 130-167.
50
51       Pott, F; Roller, M. (1994) Relevance of nonphysiologic exposure routes for carcinogenicity studies of solid particles.
52       In: Mohr, U; Dungworth, DL; Mauderly, JL; Oberddrster, G, eds. Toxic and carcinogenic effects of solid particles in
53       the respiratory tract. Washington, DC: ILSI, pp. 109-125.
54
55       Raffle, PAB. (1957) The health of the worker. BrJ Ind Med 14:73-80,


         2/1/98                                         11-27        DRAFT-DO NOT CITE OR QUOTE

-------
 1       Riebe-Imre, M; Auferheide, M; Gartner-Hiibsch, S; Peraud, A; Straub, M. (1994) Cytotoxic and genotoxic effects of
 2       insoluble particles in vitro. In: Mohr, U; Dungworth, DL; Mauderly, JL; Oberdorster, G, eds. Toxic and carcinogenic
 3       effects of solid particles in the respiratory tract. Washington, DC: ILSI, pp. 519-523.
 4
 5       Rushton, L; Alderson, MR; Nagarajah, CR. (1983) Epidemiological survey of maintenance workers in London
 6       Transport executive bus garages at Chiswick Works. Br J Ind Med 40:340-345.
 7
 8       Sagai, M; Saito, H; Ichinose, T; Kodama, M; Mori, Y. (1993) Biological effects of diesel exhaust particles. I. In
 9       vitro production of superoxide and in vivo toxicity in mouse. Free Radic Biol Med 14:37-47.
10
11       Smith, TJ. (1992) Occupational exposure and dose over time: limitations of cumulative exposure. Am J Ind Med
12       21:35-51.
13
14       Smith, RA; Stayner, L. (1990) An exploratory assessment of the risk of lung cancer associated with exposure to
15       diesel exhaust based on a study with rats. Final report. Division of Standards Development and Technology Transfer;
16       Cincinnati, OH: NIOSH.
17
18       Steenland NK; Silverman, DT; Homung, RW. (1990) Case-control  study of lung cancer and truck driving in the
19       Teamsters Union. Am J Public Health 80:670-674.
20
21       Stober, W.  (1986) Experimental induction of tumors in hamsters, mice and rats after long-term inhalation of filtered
22       and unfiltered diesel engine exhaust In: Ishinishi, N; Koizumi, A; McClellan, RO; Stober, W, eds. Carcinogenic and
23       mutagenic effects of diesel engine exhaust Amsterdam: Elsevier, pp. 421-439.
24
25       Takemoto,  K; Yoshimura, H; Katayama, H. (1986) Effects of chronic inhalation exposure to diesel exhaust on the
26       development of lung tumors in di-isopropyl-nitropyrene-treated F344 rats  and newborn C57B1 and ICR mice. In:
27       Ishinishi, N; Koizumi, A; McClellan, RO; Stober, W, eds. Carcinogenic and mutagenic effects of diesel engine
28       exhaust. Amsterdam: Elsevier, pp. 311-327.
29
30       U.S. Environmental Protection Agency. (1986, Sept. 24) Guidelines for carcinogenic risk assessment. Federal
31       Register 51(185):33992-43003.
32
33       U.S. Environmental Protection Agency, (1996, April 23) Proposed  guidelines for carcinogen risk assessment.
34       Federal Register 61 (79): 17960-18011.
35
36       Waller, RE. (1981) Trends in lung cancer in London in relation to exposure to diesel fumes. Environ Int 5:479-483.
37
38       Watson, AY; Valberg, PA. (1996) Particle-induced lung tumors in rats: evidence for species specificity in
39       mechanisms. Inhal Toxicol 8(Suppl):227-257.
40
41       Williams, RR; Stegens, NL; Goldsmith, JR. (1977) Associations of cancer site and type with occupation and industry
42       from the Third National Cancer Survey Interview. J Natl Cancer Inst 59:1147-1185.
43
44       Wolff, RK; Griffis, LC; Hobbs, CH; McClellan, RO. (1982) Deposition and retention of 0.1 urn 67Ga2O3 aggregate
45       aerosols in rats  following whole body exposures. Fundam Appl Toxicol 2:195-200.
46
47       Wong, O; Morgan, RW;  Kheifets, L; Larson,  SR; Whorton, MD. (1985) Mortality among members of a heavy
48       construction equipment operators' union with potential exposure to diesel  exhaust emissions. Br J Ind Med 42:435-
49       448.
50
51       Woskie, SR; Smith, TJ; Hammond, SK; Schenker, MB; Garshick, E; Speizer, FE. (1988a) Estimation of the diesel
52       exhaust exposures of railroad workers: I. Current exposures. Am J Ind Med 13:381-394.
53
54     '  Woskie, SR; Smith, TJ; Hammond, SK; Schenker, MB; Garshick, E; Speizer, FE. (1988b) Estimation of the diesel
55       exhaust exposures of railroad workers: II. National and historical exposures. Am J Ind Med 13: 395-404.


         2/1/98                                         11-28        DRAFT-DO NOT CITE OR QUOTE

-------
1      Yu, CP; Yoon, KJ; Chen, YK, (1991) Retention modeling of diesel exhaust particles in rats and humans. J Aerosol
2      Med 4:79-115,
3
4      Zhong, Y; Dehong, L. (1995) Potential years of lifework and tenure lost when silicosis is compared with
5      pneumoconiosis, Scand I Work Environ Health 21(Suppl 2):91-94.
       2/1/98  .                                   11-29      DRAFT-DO NOT CITE OR QUOTE

-------
           12. HEALTH RISK CHARACTERIZATION FOR DIESEL ENGINE EMISSIONS


  1     12.1.  INTRODUCTION
  2     12.1.1. Scope
  3           Earlier chapters focused on specific risk assessment topics and developed key findings for
  4     these topics or provided an overview of relevant background information. This chapter will
  5     integrate and summarize the key findings about the health hazards and risk potential for humans
  6     exposed to diesel exhaust (DE). This chapter will first integrate the key information and pertinent
  7     uncertainties for two of the three basic components of risk assessment; hazard identification and
  8     dose-response assessment. Exposure assessment, the other basic component, is not in the scope of
  9     this report, though an exposure perspective is included as background information. The final
10     section will characterize the hazards and risk in a plain-language summary form.
11           For introductory purposes, a quick overview of findings in the key assessment areas will
12     help put the remainder of this chapter into perspective.
13
14           •    The DE particle and its coating of organics, as well as the accompanying gases and
15                semivolatiles, have biochemical and toxicological properties that raise a suspicion
16                about adverse health effects given sufficient exposure.  Because DE is found only as a
17                mixture, the choice of dosimeter for measuring exposure is an important issue.  For
18.               DE, ng/m3  of diesel particulate matter is used.
19           «    Carcinogenicity: Epidemiologic data are strongly suggestive of a carcinogenic hazard
20                to the lung  under occupational exposure conditions.  Some rat and mouse studies
21                show a similar effect at high test exposures.  Mode-of-action information poses the
22                challenge of sorting among high-dose particle effects and possible low-dose effects
23                from  mutagenic/genotoxic organics.
24           •    Noncancer  toxicity: For chronic exposure, there is limited human and much animal
25                evidence for adverse respiratory effects, such as airway restriction, other measures of
26                reduced pulmonary function, and immunologic allergenic reactions. Acute exposure
27                in humans elicits various reactions from some individuals, ranging from annoying or
28                temporarily debilitating symptoms reflecting tissue irritation, up to permanent harm to
29                the respiratory system from very high acute exposure episodes.
30           •    Ambient exposures to DE vary widely depending on whether an occupational element
31                is involved, the setting is urban or rural, or near to or distant from a source of DE
32                emissions.  General average ambient exposures run in the 0.6 - 3.2 ug/m3 (of
33                particulate) range, though some areas can be in the 4-22 ug/m3 range at certain
34                periods during the year.
35
36     12.1.2. What Is Diesel Eihaust in a Risk Assessment Context?
37           .As reviewed in more detail in Chapter 2 and other chapters, a health risk characterization
38     for DE is more complicated than for most environmental pollutants because DE is a complex
39     mixture. The mixture consists of particulate matter made up of an elemental carbon core with

        2/1/98                                   12-1       DRAFT-DO NOT CITE OR QUOTE

-------
 1     hundreds of compounds adsorbed to the surface (particle-phase emissions) and a volatile fraction
 2     that is also made up of many organic and inorganic compounds (gaseous-phase emissions). The
 3     particles, the particle coating of adsorbed compounds, and the volatile elements each have known
 4     properties from which hazards can be inferred, in addition to the aggregate hazard potential posed
 5     by the whole mixture.
 6            The exhaust particles are formed through the condensation of even smaller particles in the
 7     engine and engine exhaust pathway. They average about 0.2 urn in diameter and have a very large
 8     surface area (50-200 m2/gm).  The main constituent is carbon, which accounts for approximately
 9     80% of total particle mass.  Approximately 70% of the total carbon occurs in the form of
10     elemental carbon, the remainder being various organic compounds, called organic carbon. The
11     particle constituents coating the particle surface include inorganics and hundreds of hydrocarbons.
12     At least 16 hydrocarbons adsorbed onto the particles have been classified as having a carcinogenic
13     potential for humans. Many of the compounds emitted as gases are potentially carcinogenic or
14     otherwise toxic at some dose.  These include benzene, 1,3-butadiene, various aldehydes, ethylene
15     dibromide, nitroaromatics, oxides of nitrogen, and sulfur compounds. Additionally, there is
16     evidence that reactive oxygen or hydroxal species (free radicals) may be formed on the particle
17     surface that could cause or exacerbate damage to biological cells. Inorganic compounds are also
18     present, including nitrates and compounds of sulfur.
19            The quantitative physical-chemical composition of DE exhaust is variable and depends on
20     numerous factors, including operating conditions, heavy-duty versus light-duty engines, engine
21     design, engine age, fuel technology, and exhaust control technology. Heavy-duty and off-road
22     diesel engines have the largest U.S. particulate emissions. The human and animal health studies
23     are pegged to specific engine exhaust generated at some time in the past, so the question of how
24     relevant those exposures are to current conditions is a valid inquiry. There is no single answer to
25     this question.  However, health studies focus on particle mass as a surrogate for the DE mixture,
26     so as the mass changes so may the applicability of the assessment findings.
27            Once diesel emissions are released in the air, they are subject to dispersal, dilution, and
28     chemical and physical transformations. Newly emitted exhaust is "fresh" and has free radicals
29     that pose some extra hazard, the magnitude of which is not discernable. DE that is more than a
30     day old is "aged," largely because of atmospheric alterations, and is thought to have fewer free
31     radicals.  The atmospheric alterations produce secondary pollutants that also have hazardous
32     properties or potential toxicity. The formation of the secondary pollutants will vary depending on
33     atmospheric conditions. A comprehensive assessment of the health risks posed by DE would also
34     consider the risks posed by the atmospheric reaction products, a task that is  not addressed in this
        2/1/98   .                                  12-2       DRAFT-DO NOT CITE OR QUOTE

-------
 1     assessment. Table 12-1 lists many of the particle-phase and gaseous-phase emissions in DE, as
 2     well as the atmospheric reaction products associated with each of these emissions.
 3
 4     12,2. HAZARD ASSESSMENT    .   •                            -       .      '
 5            Hazard assessment reviews what is known about the ability of DE to cause adverse effects
 6     (i.e., toxicity) in humans and laboratory animals and characterizes the likelihood that these effects
 7     are, in fact, human hazards. It also discusses the biological mechanisms that may be causing the
 8     toxicity and comments on the reliability and uncertainties of key studies.
 9
10     12.2.1. Hazard Assessment for Health Effects Other Than Cancer
11            As reviewed in Chapter 5, exposure to diesel exhaust has been shown to induce a number
12     of effects in humans and in experimental animals. As exposure progresses from episodic to more
13     frequent, from shorter and longer duration, or from small to large concentrations, the evidence
14     shows that symptoms progress from being annoying to being more temporarily disabling and
15     become more severe, with increasing likelihood of permanent damage at high enough or long
16     enough exposure. Although this section sets the stage for drawing conclusions about the potential
17     for hazard, quantitative evaluations to estimate acceptable exposure levels are discussed in
18     Chapter 6, the Inhalation Reference Concentration section.
19
20     12.2.1.1.  Effects From Acute-Exposure
21            The most readily identified acute noncancer health effect of DE on humans is its ability to
22     elicit complaints of eye, throat, and bronchial irritation as well as physiological symptoms such as
23     headache, lightheadedness, nausea, vomiting, and numbness and tingling of the extremities. Such
24     symptoms have been reported by individuals exposed to DE on busy city streets or in bus stations.
25     Recent human and animal studies also show that acute DE exposure episodes play a role in the
26     development of allergic disease (immunological allergic reactions), resulting in prolonged
27     hypersensitivity  to DE and perhaps other ambient contaminants. We do not know what DE.
28     concentrations, per se, induce allergic responses, though particle concentrations of 106 per cm3
29     induced symptoms of eye and nasal irritation and airway resistance in one study.
30            Acute animal studies have also shown that the gaseous components of DE elicit toxic
31     responses. These include NO2 (lung damage) and aliphatic aldehydes (irritation). Other animal
32     evidence shows that acute exposure to high enough  concentrations of whole DE does cause lung
33     damage. In laboratory animals acutely exposed to high concentrations of whole DE, "pulmonary
34     edema (an excessive accumulation of fluid) often occurs during the first few days of exposure.
35     After several days, aggregations of particle-laden macrophages have been observed in the


        2/1/98  • ,                               12-3       DRAFT-DO NOT CITE OR QUOTE

-------
       Table 12-1. Particle-phase and gaseous-phase emissions from diesel exhaust
       and their atmospheric reaction products
             Emission component
          Atmospheric reaction products
 Particle-phase emissions
 Elemental carbon
 Inorganic sulfate
 Hydrocarbons (C14-C35)

 PAHs (£4 rings) (e.g., pyrene, benzo[a]pyrene)
 Nitro-PAHs (>3 rings) (e.g., nitropyrenes)

 Gaseous-phase emissions
 Carbon dioxide
 Carbon monoxide
 Oxides of nitrogen
 Sulfur dioxide
 Hydrocarbons
  Alkanes(
-------
  1     alveolar regions of the lungs. At the same time, a proliferation of cuboidal type II alveolar cells is
  2     observed, replacing thinner type I cells and thickening the alveolar walls.  Because of the key role
  3     alveoli play in the exchange of gases, these changes may inhibit the efficiency of pulmonary
  4     function. Human studies are inadequate to evaluate potential toxic effects resulting from acute
  5     exposures.
  6            The observed effects from single high-dose or episodic exposures in test animals are
  7     consistent with effects seen with lower and more long-term exposures. This suggests that total
  8     accumulated dose may be one basis for characterizing DE hazards, but this could be too simplistic
  9     for all aspects of DE toxicity because dose-rate aspects cannot be ruled out.
10
11     12.2.1.2. Effects From Short-Term and Chronic Exposure
12            Based on suggestive evidence from human occupational studies, combined with multiple
13     controlled laboratory animal studies in several species,  a high level of confidence exists that
14     chronic exposure to DE constitutes a noncancer respiratory hazard for humans. As DE exposure  .
15     levels and duration increase, the onset of respiratory symptoms in humans is observable on a
16     limited basis, whereas in animal studies the onset of symptoms is more clear and replicable.
17     Current data do not support confident identification of health hazards other than for the respiratory
18     system. Chapter 5 discusses this topic in more depth.
19            Several studies of workers occupationally exposed to DE on a short-term basis have
20     monitored pulmonary function at the beginning and end of work shifts to see if this marker of
21     respiratory distress has been unpaired by exposures. Short-term symptoms are seen (e.g., bus
22     garage workers experienced burning and watering of the eyes, coughing, labored breathing,
23     chest tightness, and  wheezing), but no reduction in pulmonary function.  A study of stevedores
24     showed an adverse effect  on pulmonary function, but normal function returned a few days after
25     DE exposure stopped.  It  was noted in one study that occupationally exposed smokers appeared
26     to demonstrate larger work-shift respiratory function decrements and increased incidence of
27     respiratory symptoms.  Other occupational studies did not find statistically significant effects
28     from short-term exposure, though these studies have limitations that reduce the ability to detect
29     responses.
30            Noncancer effects  of chronic DE  exposure have also been evaluated in epidemiologic
31     studies of occupationally exposed workers (metal and nonmetal miners, railroad yard workers,
32     stevedores, and bus garage mechanics).  Some of the data indicate an absence of increased
33     respiratory disease associated with exposure. In a few studies, though, a higher prevalence of
34     respiratory symptoms, primarily cough, phlegm, or chronic bronchitis, was observed among the
35     exposed, but usually without significant  changes in pulmonary function. However, two studies,

        2/1/98                                     12-5       DRAFT-DO NOT CITE OR QUOTE

-------
 1     one of stevedores and one of coal miners, detected statistically significant decrements in baseline
 2     pulmonary function consistent with obstructive airway disease, thus providing some suggestion
 3     that impairment of pulmonary function among occupational populations may be occurring.
 4     Recent investigations have also indicated that human exposure to DE may result in the
 5     development of immunologic-driven allergic hyper sensitivity; this would be a new health concern
 6     for DE exposure.
 7            Overall, these results are suggestive of adverse chronic effects for humans, but database
 8     limitations preclude drawing more definitive conclusions.  The epidemiplogic investigations
 9     suffer from methodological limitations that confound the observations and limit the assignment of
10     these observations to DE exposure. The limitations include (1) incomplete information on the
11     extent of exposure to DE, necessitating in some studies estimations of exposures from job titles
12     and resultant misclassification; (2) the presence of confounding variables such as smoking or
13     occupational exposures to other toxic substances (e.g., mine dusts); (3) the short duration and low
14     intensity of exposure; and (4) unlike in animal experiments, pathological evaluation of human
15     lung tissue is seldom available for confirmatory analysis.
16            The suggestive human experience, however, is reinforced by  a considerable body of
17  .   animal evidence that clearly correlates DE exposure with pulmonary injury.  The combined
18     human and animal data are sufficient to infer that this  hazard likely exists for humans. Short-term
19     animal exposures to DE containing high concentrations of particulate matter (PM) resulted in
20     histological and cytological changes in the lungs, but only minimal effects on pulmonary function.
21     A number of long-term laboratory studies with  rats, mice, Chinese hamsters, Syrian golden
22     hamsters, cats, and Cynomolgus monkeys found varying degrees of adverse lung pathology.
23     Exposures for several months or longer to levels markedly above environmental ambient
24     concentrations resulted in accumulation of particles in the animal lungs and an impaired ability to
25     clear particulate matter from the lungs. Histological studies also showed a variety of changes in
26     respiratory tract tissue, including focal thickening of the alveolar walls, replacement of Type I
27     alveolar cells by type II cells, and fibrosis. Because these effects were seen in a wide range of
28     animal species, there is a compelling basis to believe that humans would also be at hazard under
29     some condition of exposure.
30            Respirable particles in general have been implicated as etiologic factors in various types of
31     chronic human lung diseases (U.S. EPA, 1996), Ambient particulate matter (PM) is associated
32     with increased morbidity and mortality, aggravation of respiratory and cardiovascular disease,
33     changes in lung function and increased respiratory symptoms, changes to lung tissues and
34     structure, and altered respiratory defense mechanisms. The effects vary as one considers "fine"
35     particles (e.g., PM 2 5) compared with "coarse" particles (e.g., PM,0).  The vast majority of DE

        2/1/98                                    12-6        DRAFT-DO NOT CITE OR QUOTE

-------
  1     particles are fine and ultrafme in size and thus contribute to ambient levels of PM2 5. In addition,
  2     DE contributes significantly to total ambient PM. For instance, Yoshizumi (1989) indicated that
  3     for Tokyo, the yearly mean concentration of DE particles was about 40% of the total particle
  4     concentration.
  5            Epidemiologic studies of the effects of DE on organ systems other than the pulmonary
  6     system are scant. Animal studies have suggested that liver and kidney changes may be occurring
  7     at high concentrations, along with some indication of neurotoxic effects.  Impaired growth rates
  8     have also been observed in animals chronically exposed to DE. However, since these effects are
  9     only seen at relatively high exposure levels, this does not imply a hazard for humans at low
10     ambient exposures.                                                          ,
11
12     12.2.2. Toxicity Mode of Action
13            An understanding of the mode of action(s) (MO A) for toxicity allows one to make
14     informed choices about how to translate observed toxicity data into specific risk assessment
15     guidance that protects human health. For example, MOA information may help answer several
16     questions:  (1) Are all humans equally susceptible or just some population subgroups?  (2) Are
11     animal responses predictive of potential human responses because the animal MOA is thought to
18     operate in  humans?  (3) Are high-dose effects extrapolatable to low ambient levels of exposure,
19     and what does the shape of the low dose-response curve look like? and (4) a number of other
20     specific qualitative and quantitative matters. In the absence of convincing or reasonably clear
21     MOA information, scientific inference or default assumptions based on science policy are used in
22     order to facilitate risk assessment conclusions, if a hazard potential is suspected.
23            Chapters 7, 9, and  10 contain more in-depth review of the mode-of-action topic.
24            With DE being a mixture and having several distinct components, the topic of MOA(s) is
25     complex.  For the carbon core particle component of DE, the pathogenic sequence following the
26     inhalation  of the diesel particle begins when alveolar macrophages (AMs), "scavenger" cells that
27     defend the lungs from invading foreign matter, ingest diesel particles. When AMs ingest particles
28     in large numbers they are activated and release chemical signals that attract neutrophils (a
29     component of white  blood cells) and additional AMs. As the lung burden of diesel particulate
30     matter increases, particle-laden AMs aggregate in alveoli adjacent to terminal bronchioles. The
31     overloaded neutrophils and AMs produce and release compounds that mediate inflammation. The
32     particle-laden macrophages also become less mobile, thus decreasing their ability to clear particles
33     from the lung. The latter series of events may result in inflammation of the lung, with
34     replacement of very  thin type I alveolar cells with more cuboidal type II cells, slowing exchange
35    ' of oxygen  and carbon dioxide. Continued exposure may result in further consequences, such as
36     lung fibrosis and/or emphysema. This mode of action has a dose threshold because at some point
        2/1/98                •                     12-7       DRAFT-DO NOT CITE OR QUOTE

-------
 1     the normal detoxification mechanisms become overloaded and consequential toxicity ensues.
 2     This MOA may also contribute to a carcinogenic response as discussed below. At least in the rat,
• 3  ,   the particle-driven toxicity and carcinogenicity can be viewed as being driven by the elemental
 4     carbon particle or a combination of the physical particle and the accompanying organics,
 5     depending on how several types of studies are interpreted. Intratracheal instillation of unaltered
 6     DE particles gave a stronger carcinogenic response than particles stripped of the organics,
 7     suggesting that particle-associated compounds play a role in tumor induction. Inhaled pure carbon
 8     particles (printex), however, were as effective in the induction of lung tumors as DE. Printex has
 9     a larger surface area per unit mass than diesel particulate matter. An initial notion that organics
10     play an insignificant role is not necessarily indicated, however,, because the very large surface area
11     of the printex, a factor known to be related to cancer potency, may well make up for its lack of
12     surface-adhering organics. In fact, a recent study has shown that stripping the organics from the
13     DE particulates reduced their carcinogenic potency.  Several animal studies also show that filtered
14  .   DE (e.g., the gases without the coated particles) is ineffective in producing lung tumors in rats.
15            As for the chemical compounds that coat the particles, mutagenicity and genotoxicity
16     assays reveal activity for the intact coated  particles as well as organics extracted from the
17     particles. Some of the gaseous fractions of the DE mixture are also mutagenic. The correlation of
18     mutagencity-genotoxicity with carcinogencity isn't rigid, but the presence of this activity gives
19     rise to support for a carcinogenic hazard, as well as suggesting some possible  modes of action for
20     cancer.
21            The operable mode of action for, toxicity of the DE mixture taken as a whole is not
22     definitively known.  There are likely to be several modes operating, given the many DE
23     constituents.  A simplistic combination of modes to consider might have the following elements:
24     (1) DE organics at low or high doses can initiate DNA damage which, if particles were not
25     present, would have a probability to advance to cancer as a function of increasing total dose. This
26     would also be considered a likely nonthreshold process because of the mutagenic properties of
27     many of the organics on the particles and in the gases, and by EPA preference would be modeled
28     for extrapolation purposes with a linear model; (2) with particles also present that overload
29     normal lung detoxification at higher exposures, the particles continue to deliver organics to the
30     deep lung, resulting in an increased residence time for the organics. The particles also add a
31     second and seemingly dominant MOA (at least in the rat model) that induces a particle-specific
32     proliferation of cells (because of the inflammatory consequences of the particles), thus
33     accelerating growth of any DNA-damaged cells; (3) the presence of particles, even at nonoverload
34     levels, may also influence carcinogenicity by contributing to additional DNA damage via the free
        2/1/98                                    12-8       DRAFT-DO NOT CITE OR QUOTE

-------
  1     radicals present on the particle surface, which are potent oxidizing agents for bio-organic
  2     substrates.
  3            The observation that some of the DE-induced rat tumors are different pathologically from
  4     what one would expect in humans and that some are similar may also support the thought that
  5     multiple MO As may be present.
  6            Given the information at hand, dual modes of action are more likely than a single mode.
  7     One would be driven by particles and one by the organic/inorganic components, with a shifting
  8     influence as the exposure/dose varies from high to low. The role of contributory toxicity from
 '9     other constituents, such as N02 or oxygen free radicals, among others, is uncertain.
10            It would be ideal if the MOA were informative about whether all humans were equally
11     susceptible or whether some population subgroups were more or less susceptible. The human
12     evidence, i.e., occupational and population-derived studies, does not provide any particular insight
13     about variations in susceptibility, nor do the animal studies. Neither female versus male nor adult
14     versus children's susceptibility differences are specifically indicated: the former may have been
15     discernable given the nature of the studies, but there is no study basis to address possible
16     children's risk issues.  Later, in this  section, a qualitative discussion about susceptibility is
17     included that acknowledges that toxicological wisdom suggests that background respiratory
18     system conditions could make some in the population more susceptible to chronic effects of DE
19     exposure.  Similarly, infants and children could have greater  susceptibility simply because their
20     developing organs and defense systems may be less effective at dealing with insults from DE
21     exposure.  These suggestions are not unique to DE exposure, though.
22
23     12.2.3.  Carcinogenic Hazard Assessment
24            For inhalation exposure, both human studies and animal bioassays are available to assess
25     the chronic exposure carcinogenic potential for DE. In fact, both the human and the animal
26     studies provide evidence that exposure to DE has potential to be carcinogenic to humans under
27     some condition of exposure. Chapters 7 and 8 review these data in detail. A finding about the
28     hazard potential does not specify the magnitude of the impact, information on which is discussed
29     in Chapter 11, the carcinogenicity dose-response evaluation section.
30
31     12.2.3,1. Human Evidence                                     ,
32            A total of 26 key epidemiologic studies have been evaluated to examine the association
33     between exposure to DE and increased cancer response.  The positive human evidence consists of
34     observed increases in lung cancer mortality .in a number of occupational exposure studies and
35     some suggestion of other possible cancer sites. Cohort, case control, and population-based studies
                                                                       *
        2/1/98                                    12-9       DRAFT-DO NOT CITE OR QUOTE

-------
 1     are available. Exposure is most often defined indirectly by occupation or job title in the industry-
 2     related studies and is self-reported in population-based studies.  The lack of direct exposure
 3     measurements, a condition common to retrospective epidemiology studies, is an overall limitation
 4     in the database.  An excess risk (e.g. elevated standardized mortality ratios, relative risks, or odds
 5     ratios > 1.0) for lung cancer was observed in 5 of 9 cohort studies and 8 of 10 case-control studies.
 6     Of these studies, three cohort and three case-control studies observed a dose-response relationship
 7     by using duration of employment as a surrogate for dose.
 8            The most convincing evidence that exposure to DE can induce lung cancer in humans
 9     comes from case-control and cohort studies among U.S. railroad workers and truck drivers. The
10     study of railroad workers, a well-conducted and well-analyzed study, is evaluated and published
11     as both a cohort and case-control study with varied controlling for confounders. The case-control
12     study is the best for control of confounders, especially the question of smoking and its possible
13     role as a confounder for the reported lung cancer increase. Statistically significant higher risks of
14     41% to 43% for lung cancer were found in the case-control study for 20 or more years of
15     exposure, and these risks were not confounded by smoking or asbestos exposure, adjustments for
16     which were rigorously accounted for in the study methodology. In the retrospective cohort study
17     of these same railroad workers, the risks varied from 20% to 72% higher than the general
18     population, all statistically significant depending on the duration of exposure.  Although
19     adjustments for possible asbestos exposure were accounted for, there was no adjustment for the
20     possible role of smoking.  However, recognition was given to the rigorous smoking adjustments
21     in the case-control study, which showed no effect on risks. Though the overall risks were
22     increased in the railroad worker cohort study, the identification of a dose-response relationship is
23     a subject of debate.  A case-control  study of truck drivers showed statistically significant
24     increased risks of 80% to 240%, depending on data stratification and duration of exposure after
25     adjustment for smoking.
26            There is a notable consistency in finding elevated, although not always statistically
27     significant, increases in lung cancer among workers exposed to DE in several industries. There
28     are industry-specific findings of elevated lung cancer risk from truck drivers, professional drivers,
29     and railroad workers, with some of the studies having adjusted for smoking. When the possible
30     role of smoking as a confounder was accounted for, the increased risks prevailed.
31             A very recent meta-analysis (Bhatia et al., 1998) shows the consistency of elevated risks
32     in the epidemiology database and lends clear support to a causal association between increased
33     risks for lung cancer and exposure to DE.  Using 29 epidemiology studies, 23 of which met
34     inclusion criteria, statistically significant relative risks (RR) for all studies were 1.33 (95% CI =
35     1.24-1.44). A subanalysis of case-control studies showed a RR of 1.33 (95% CI = 1.18-1.51); for

        2/1/98                                    12-10      DRAFT-DO NOT CITE OR QUOTE

-------
  1     studies that specifically controlled for smoking, the results were nearly identical.  These findings
  2     play an important role in analytically showing the trend of the evidence across much of the
  3     epidemiology database. The quantitative findings of aggregate relative risks may also be useful
  4     for dose-response analysis, if individual study RRs are not thought to be suitable for some reason.
  5            The relative risks from all studies with elevated risks, statistically significant or not, are
  6     on the low side, generally < 2.  Several of the best individual studies have RRs in the 1.4-1.7
  7     range. Low relative risks are harder to interpret as being definitive because of the possibility that
  8     unresolvable uncertainties could be responsible for or could be influencing the elevated risks.
  9     When only one or two positive studies with low RRs exist, there is some uncertainty about
10     whether the inference of an effect is the proper interpretation, compared to a case in which several
11     studies have low RRs, where the uncertainty is less and the inference that a real effect  is being
12     seen is more confident. With several low-RR diesel studies available and the meta-analysis clearly
13     demonstrating the increased RR pattern, the low-RR situation would not discount the inference
14     about an effect from exposure.
15            In most risk assessment situations positive epidemiologic findings are weighed within a
16     framework of "causality" criteria. The causality framework, with its related biological factors and
17     physical factors, helps to rationalize the increased epidemiologic responses within a broader
18     context of plausibility.  When these criteria are applied to the positive DE studies, all of them
19     apply well. Other assessors using similar criteria may come to a different conclusion because
20     there is no rigid attainment measure within the causality criteria.
21            With all evidence and analysis taken into account, it is concluded that the human evidence
22     is "highly suggestive" of an association between DE exposure and lung cancer in retrospective
23     occupational settings. This is just short of a more definitive finding of a "known"  human
24     carcinogen, primarily because of deficiencies in exposure information.
25            Human study evidence for other forms  of cancer is inconclusive.
26
27     12.2.3.2. Animal Evidence
28            Animal  studies show that DE is carcinogenic in test animals, which in the simplest sense
29     raises a concern that DE has a carcinogenic potential in humans. The direct inhalation exposure
30 '    rat studies are best for direct observation  of inhalation responses.  Chronic-exposure animal
31     studies conducted prior to the 1980s did not use inhalation as the route of exposure; instead,
32  •   exposure was artificially produced by lung implantation, intratracheal installation, subcutaneous
33     and intraperitoneal injection, or dermal application. Of note here is that organic extracts from the
34     particles, as well as the whole particles (with the absorbed organics), are carcinogenic in many of
35     the older artificial exposure route assays.

        2/1/98                                    12-11       DRAFT-DO NOT CITE OR QUOTE

-------
 1            When focused on the newer inhalation studies and recognizing the reproducibility of
 2     results, one concludes that if exposures are adequate, inhalation of diesel exhaust will induce lung
 3     cancer in rats and, under some conditions, in mice, albeit at higher concentrations than in rats.
 4     Generally, rats showed significant increases in lung tumors beginning at exposures of 2,200 ug/m3
 5     or higher (the human equivalent concentration [HEC] for 2,200 ug/m3 is about 700-900 ug/m3).
 6     These exposures are higher than those thought to be present in the human occupational studies
 7     (i.e., 125 up to 500 ug/m3), which in turn are higher than the ambient exposures of interest for
 8     humans (e.g., 0.6-3.6 ug/m3). Along with the increased tumor incidence in the rat studies, there
 9     was evidence of particle-induced inflammation and particle clearance overload in the lung.
10     Exposures to rats below 2,200 ug/m3 did not elicit an observable lung cancer response, while the
11     accompanying evidence of inflammation trailed off more gradually as dose was lowered below
12     2,200; some inflammation was  still seen at 1,000 ug/m3.  Inhalation of diesel exhaust has also
13     induced significant increases in lung tumors in a few, but not all, strains of female mice; exposure
14     concentrations of 6,000 ug/m3 or greater were needed to see a significant response. Although
15     responses were not detected in rats or mice at lower exposure concentrations of 350-2,200 ug/m3,
16     these studies with nominally 50 animals probably lack the sensitivity to reveal a threshold or a
17     response to a less potent mode-of-action-driven carcinogenic process. Attempts to produce
18     positive responses by inhalation exposure in Syrian golden hamsters, cats, or monkeys were
19     unsuccessful. The negative  results in cats and monkeys may be explained by an inadequate
20     exposure duration (only 2 years) in these longer-lived species, whereas hamsters are known to be
21     less sensitive to lung tumor  induction compared to rats and mice.
22            There is convincing  evidence, based on numerous published studies, that DE constituents
23     are also mutagenic or, in a broader sense, genotoxic and/or carcinogenic. This supports the
24     concern for a carcinogenic hazard for humans as well as suggesting possible mode-of-action and
25     related approaches for low-dose risk estimation. The whole particle, particle extracts, and gaseous
26     portions of diesel exhaust have been shown to cause changes in genetic  material, which is not
27     surprising because each component contains one or more mutagenic constituents. Human studies
28    • show increased formation of DNA adducts with DE exposure, and cultured human cells show an
29     increased occurrence of sister chromatid exchange.  Extensive studies with Salmonella have
30     unequivocally demonstrated mutagenic activity  in both particulate and gaseous fractions of DE.
31     The induction of gene mutations has been reported in several in vitro mammalian cell lines after
32     exposure to extracts of diesel particles. Particles have also induced structural chromosome
33     aberrations and sister chromatid exchanges in mammalian cells. The question of germ-cell
34     interaction, however, and the potential for human germ-cell mutagenic risk from exposure to
35     diesel emissions remain unanswered.

        2/1/98  -  .                               12-12       DRAFT-DO NOT CITE OR QUOTE

-------
  1     12.2.4.  Weight-of-Evidence Summary
  2            A conclusion about the likelihood of a human carcinogenic hazard involves a weighing of
  3     the evidence from the human and animal studies, mode-of-action information, and evidence from
  4     ancillary studies. Scientific uncertainty or debates about some aspects of the evidence as well as
  5     different guidance about how to weigh evidence can easily lead to modestly different conclusions.
  6     EPA uses Carcinogen Risk Assessment Guidelines to frame its approach to weighing evidence
  7     and judging the likelihood of hazard. Overall, EPA considers that the human evidence for DE
  8     gives a clear signal about the likely presence of a human hazard.  Subtle, deficiencies in the human
  9     data influenced EPA in holding back a definitive call for a "known" human carcinogen based on
10     the human data alone.  Other conclusions that give a high likelihood to DE being a human
11     carcinogen are also defensible. Animal evidence clearly shows that at high experimental
12     exposures, DE produces  lung cancer. Numerous DE particle extracts have been shown to pose
13     carcinogenic hazards to animals and thus probably humans., even though some artificial exposure
14     routes are used and carcinogenic responses are not necessarily in the lung. When mode-of-action
15     information is considered, valid questions are raised about whether the lifetime rat bioassay, per
16     se, is a good model for defining the hazard for humans at low exposures. Still, many of the
17     individual components of the DE mixture are known or thought by the health science community
18     to potentially pose a carcinogenic hazard. Supporting evidence for mutagenicity-genotoxicity in
19     human cell cultures and in various other in vitro test systems for the whole DE mixture, the
20     particle  coatings, and the DE gases also support a likely mammalian-based carcinogenic hazard
21     potential. For low levels of exposure, the hazard is hypothesized to be related to the organics
22     present, while at high doses the presence of particles seems to exert a definite influence.
23     Everything considered, EPA believes that DE is  "very or highly likely" to be carcinogenic in
24     humans by the inhalation route. This discussion and finding is consistent with the provisions of
25     EPA's Proposed Cancer  Risk Assessment Guidelines (U.S. EPA, 1996).
26           DE is considered to be a "probable" human carcinogen by inhalation exposure and to best
27     fit into cancer weight-of-evidence category B1 according to EPA's 1986 Cancer Risk Assessment
28    . Guidelines (U.S. EPA, 1986). This conclusion evolves from positive yet "limited" evidence in
29     the human studies, a "sufficient" level of evidence  in bioassays, and consideration of the
30     supporting information from mutagenicity and genotoxicity data.
31           In comparison with other agents classified into this category, the confidence in the DE
32     weight-of-evidence call,  on a relative scale of high, medium, or low, is on the high side. The
33     conclusion that DE is a "probable" or a "very likely" rather than  a "known" human carcinogen is
34     due to several subtle limitations of the human studies, including the difficulty of reconstructing
35     reliable  estimates of exposures many years in the past, the lack of systematic quantitative records

        2/1/98  .                                 12-13      DRAFT-DO NOT CITE OR QUOTE

-------
 1     of ambient air quality, and the inability to completely eliminate confounding factors such as
 2     exposure to other pollutants, especially tobacco smoke, and the small increases in relative risk that
 3     make possible confounding more critical. The weight of the evidence for a human health hazard
 4     is nevertheless only just shy of being sufficient to view DE as a "known" human hazard, and for
 5     this reason "very likely" may be a better descriptor than "probable."
 6           Once the hazard potential for humans is identified, information about the possible impact
 7     of the hazard (i.e., measures of risk and risk estimation, etc.) is treated as a related but separate
 8     dose-response issue. A complete conclusion about carcinogenicity thus, has two components.
 9
10     12.3. DOSE-RESPONSE ASSESSMENT
11           Dose-response assessment focuses on the relationship of exposure/dose to a biological
12     response and how the response might change with dose; it also investigates the translation of this
13     relationship to human low-exposure circumstances.  The response(s) are the ones previously
14     identified as important in the hazard assessment, and the low-dose aspects are approached by
15     extrapolation, if appropriate, from an observable response range to lower exposure/dose levels,
16     such as ambient levels  of interest. Key dose-response assessment choices are influenced by
17     definitive knowledge and informed reasoning about the mode of action.  In the absence of such
18     information, assumptions (i.e., defaults) are used, some of which may be conservative. Chapter
19     11 contains a more detailed review of dose-response issues,
20
21     12.3.1. Considerations in Modeling the Dose-Response Relationship
22           In order to know the extent to which a substance poses a hazard to human health, it is
23     necessary to estimate the magnitude or incidence of adverse effects induced in humans by a
24     particular concentration or dose of a substance. Since intentional experimentation with suspected
25     toxicants on humans is not acceptable, animal  studies are usually a major source of information
26     about possible human hazards, yet the animals may have a variety of differences from humans
27     that could affect the translation and extrapolation of the response. Animal studies usually involve
28     exposure to high concentrations of toxicants, so that dose-response for low exposures must be
29     inferred via informed assumptions and extrapolation. Even studies of humans are often at higher
30     exposures than most individuals would experience at ambient conditions, so that conclusions
31     about the effects of lower levels of exposure once again require assumptions and extrapolation.
32            Dose-response  analysis of the positive  human data may be approached in any of several
33     ways, depending on what the data allow. One almost always has to struggle with exposure
34     uncertainties when using retrospective human  studies, as well as with questions of confounding
35     exposure from other carcinogens, such as from smoking, when lung cancer is involved.

        2/1/98                                    12-14      DRAFT-DO NOT CITE OR QUOTE

-------
  1     Statistically significant mortality increases for lung cancer, in this case in cohort or case-control
  2     studies, may be useful analytical starting points.  Adjustments in exposure dosimetry for humans
  3     wouldn't be needed, as is the case with animal studies.
  4            In the DE dose-response analysis based on animal studies, several issues about exposure-
  5     dose to the lung can be reasonably accounted for because of available dosimetry research and
  6     modeling. The issues include species differences in particle deposition efficiency, particle
  7     clearance rates, lung surface area, respiratory rates, and transport of particles to lung-associated
  8     lymph nodes. Another set of issues involves the suitability of the rat model for assessing human
  9     risk from the standpoint of biological sensitivity as well as exposure-related MO A even after dose
10     adjustment to equivalent concentrations. A final series of issues involves the choice of dose-
11     response extrapolation model to be used (i.e., threshold or nonthreshold, linear or nonlinear, and
12     biologically based or curve-fitting models).
13            While whole DE has multiple components, the various gaseous/semivolatile components
14     of DE are in rough proportion to the particle concentration. Particle concentrations have been
15     adopted by researchers  to document whole DE mixture exposure and hence are used in this
16     assessment as a dosimeter in evaluating carcinogenic and other adverse effects. For DE, the
17     critical site of action is  in the lower respiratory tract, particularly the alveolar region of the lung.
18     The exposure route of concern for humans is inhalation, which is also the method of exposure
19     used in some of the available experimental data, though important information from animal or
20     other in vitro assays is gleaned from studies where the exposure is not related to inhalation per  se
21     but to artificial experimental exposures.
22
23     12.3.2. Dose-Response Assessment for Health  Effects Other Than Cancer
24            A considerable body of evidence clearly shows a major noncancerous health hazard may
25     be presented to the respiratory system following inhalation of DE. Based on pulmonary function
26     and histopathological and histochemical effects, a rough estimate can be made concerning what
27     chronic dose/exposure rates of DE (measured in terms of the concentration of diesel particulate
28     matter) can be observed to cause an adverse effect and which exposures do not; this then is a
29     starting point for establishing adequate margins of safety. A reliable experimental database and
30     established EPA dose-response evaluation methods have been used to derive an inhalation
31     reference concentration (RfC) for chronic exposure to DE as a guide in determining a level of
32     long-term exposure that is thought to have an acceptable margin of safety from hazard for chronic
33     exposure adverse effects other than cancer.
        2/1/98                                    12-15       DRAFT-DO NOT CITE OR QUOTE

-------
 1     12.3.2.1. Derivation of an Inhalation Reference Concentration
 2            The derivation of an inhalation reference concentration (RfC) for DE is a dose-response
 3     approach used by EPA for noncarcinogenic, threshold chronic effects. An RfC is defined as an
 4     estimate of a continuous inhalation exposure to the human population, including sensitive
 5     subgroups, with uncertainty spanning perhaps an order of magnitude, that is likely to be without
 6     appreciable risks of deleterious noncancer effects during a lifetime.  The RfC approach is based
 7     on the assumption that a threshold exists for the human population below which no effect will
 8     occur.
 9            As an alternative to using a no-observed-adverse-effect-level (NOAEL) in the RfC
10     methodology as a dose-response marker, this assessment examined a benchmark
11     dose/concentration (BMC) analysis approach, a newer approach that EPA has used to derive the
12     RfCs for carbon disulfide, chlorodifluoromethane, and several other chemicals (U.S. EPA, 1995).
13     The BMC refines the ascertainment of the NOAEL.
14            The database EPA chose to work from  for RfC derivation consisted of 10 long-term
15     (greater than 1 year) studies of inhalation of diesel engine emissions in laboratory rats.  The
16     available human studies, as discussed earlier, were qualitatively suggestive of adverse effects but
17     were inadequate for RfC consideration. The selected rat studies were conducted by the Inhalation
18     Toxicology Research Institute (ITRI) and the Japanese Health Effects Research Program (HERP).
19     These studies were selected because each identified respiratory effects after chronic exposure and
20     provided good information about pulmonary histopathology. Further, the selected studies
21     spanned a wide range of exposures, from 350 to 7,000 jag/m3 with three exposures in the 350-960
22     |ag/m3 range.  Human equivalent concentrations were calculated using a dosimetry model
23     developed by Yu et al. (1991) that accounted for species differences in respiratory exchange rates,
24     particle deposition efficiency, differences in particle clearance rates at high and low doses, and
25     transport of particles to lymph nodes.
26            The adopted RfC comes from the HERP study, which showed a NOAEL of 460 |ig/m3
27     (human equivalent concentration, HEC, = 155 |ag/m3). While particle overload is thought to be
28     still present at 1,000 jag/m3 to some degree, at 460 jag/m3 the overload is thought to be much less,
29     if not minimal.  Consistent with standard RfC practice for a good chronic animal study, two types
30     of uncertainty factors were used to further lower the NOAEL-HEC to a value having a sufficient
31     margin of safety for humans. An uncertainty factor of 3 out of 10 was used to account for
32     interspecies sensitivity; that is, humans could be somewhat more sensitive than rats. Out of a
33     possible factor of 10, credit is given to the dosimetry adjustment procedures used.  We would also
34     'note that some researchers believe rats could be more biologically sensitive than humans to DE
35     particles, but we do not know whether rats are more or less sensitive to the organics. A second

        2/1/98               '                    12-16       DRAFT-DO NOT CITE OR QUOTE

-------
 1     uncertainty factor of 10 is used to account for sensitive members within the human population,
 2     this being standard practice unless mechanistic or other data suggest otherwise. The resulting total
 3     uncertainty factor is 30. With 155 ug/m3 divided by 30, the resulting RfC is;
 4
 5                          RfC = 5 p,g/m3 of dieselparticulate matter (DPM).
 6
 7     A BMC analysis also supports the use of an RfC of 5 micrograms per cubic meter of air. The
 8     derived RfC is considered reliable because of the high quality of the animal studies. Additional
 9     support for this RfC level is provided by a retrospective analysis of an earlier monkey study in
10     which monkeys were exposed to DE at a concentration of 2,000 ug/m3 for 2 years. Two years
11     would be considered subchronic in this longer lived species. On the basis of minimal effects on
12     the lungs of the monkeys, it could be argued that the exposure level is either a NOAEL or a
13     marginal LOAEL (lowest observed adverse effect level).  After adjustment to the equivalent
14     human concentration and using appropriate uncertainty factors, an RfC value similar to that
15     derived from the rat can be  shown.
16            It is interesting to note that EPA's recently adopted particulate matter (PM2 5) standard,
17     with an adequate margin of safety, is 15 jJg/m3, as a 3-year average, based on human studies.  The
18     noncancer effects from DE  are qualitatively similar to some of those for PM2 5, though particle
19     differences exist, as do the presence of absorbed organics and gases. A comparative discussion of
20     the PM, 5 standard and the DE RfC is not currently pursued in this assessment.
21                  '        ..'                          -              '       '
22     12.3,3. Dose-Response Assessment for Carcinogenic Effects
23            Evidence shows that DE is likely to induce lung cancer by the inhalation route of
24     exposure, though the mode  of action is imperfectly understood. Using a range  of studies, some
25     assumptions, extrapolation models, and animal-human dosimetry factors appropriate for DE, it is
26     conceptually possible to estimate human cancer potency as a function of lifetime exposure
27     assuming that DE is a human carcinogen.
28
29     12.3.3.1.  High- to Low-Dose Extrapolation andMode-of-Action Considerations
30            Because the biological mechanisms that result in cancer may be different for DE at low
31     and high doses, a good understanding of the MO A would be needed in order to generate DE-
32     specific, biologically based probability models for extrapolating from high to low doses.
33     Currently, however, there are significant gaps in understanding of the mode of action(s) by  which
34    • DE produces cancer, and  thus the development of a rigorous DE-specific model is not pursuable.
35     Given this imperfect situation, the risk assessor falls back on making choices about modeling,

        2/1/98               '                    12-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1     some of which can be informed choices while others are more public-health-driven policy
 2     choices.
 3  .          To recount the mode-of-action discussion earlier, at low to high doses the DE cancer
 4     hazard is believed to have at least two modes. One involves mutagenic/genotoxic mechanisms
 5     associated with the organic substances adsorbed to DE particles as well as those in the gases.  The
 6     organics include substances that are known to be mutagenic and/or carcinogenic in animal
 7     bioassays. Superoxide or hydroxal free radicals on the surface of fresh DE particles may also
 8     contribute to this mode of action. The mutagenic organics are present in approximate proportion
 9     to the particulate concentration, which likely varies somewhat by diesel engine circumstances.
10     These substances could initiate cells or be complete carcinogens.  A second mode involves
11     nongenotoxic particle-specific mechanisms associated with lung particle overload, which is likely
12     to dominate at high doses, according to rat studies, by exacerbating the promotion of initiated
13     cells. The influence of the particles at low, nonoverload exposures may be limited to delivering
14     organics, but this is an unknown.
15            If we favor the organics (gases or absorbed organics on the particles) as a likely etiologic
16     agent for a suspected low-exposure carcinogenic hazard, the issue of bioavailabiliry of the
17     organics is an important consideration. There would be little doubt about the bioavailabiliry of
18     the gaseous-phase compounds in the DE mixture;  whether they present a large enough dose to
19     make a carcinogenic response observable is a different question. The bioavailabiliry of the
20     organics coating the particles is a more complex topic. Rat studies using radiolabeled BaP and
21     nitropyrene-coated carbon particles have  shown that lung retention time of the organic is
22     significantly increased compared to a nonparticle instillation and that the gradual elution of BaP,
23     for example, was faster than the lung clearance of the carrier particle itself.  An understanding of
24     the role extracellular fluids play in extracting the organics from the particles is unclear.
25     Extraction of the organics by alveolar and other cell types is theorized but not well understood.
26     At least in the rat and mouse models, a carcinogenic response was not seen below the whole-DE
27     exposure levels that were also associated with particle overload, implying that a dose adequate to
28     cause tumors was not present. On the other hand, one recent bioassay (Dasenbrock,  1996) has
29     shown by intratracheal instillation that rat tumor yield is decreased (from 17% to 4%) when diesel
30     particulates are stripped of their organics. Overall, limited data make it plausible to assume that
31     there is a gradual release of organics from the particles, with resulting exposure to the respiratory
32     epithelium.
33            Ideally, the dose-response curve that would be used to extrapolate from high to low
34     exposures would reflect changes in the MOAs, a result of the possible transition from primarily
35     nongenotoxic modes of action at high exposures to primarily genotoxic ones at lower exposures.

        2/1/98                                    12-18       DRAFT-DO NOT CITE OR QUOTE

-------
  1     However, data needed to construct this unifying dose-response curve are lacking. While the
  2     exposure levels at which primarily nongenotoxic MOAs may predominate can be inferred from
  3  .   lung pathology in rats (e.g., Chen and Oberdosterf 1996] say the transition range probably lies in .
  4     the > 100 to 1,000 |ig/m3 exposure range for rats), the change in slope of the dose-response
  5     relationship can only be crudely approximated by the use of a biologically based model, some of
  6     whose parameters are estimated, or by estimating low dose risk from nonsignificant responses.  In
  7     general, experimental studies inherently lack sensitivity for estimating dose-response in the low-
  8     dose range, except for extremely potent carcinogens, and DE seems notto be that potent.  Use of
  9     human data to estimate low dose risks is also pursued, but this is not as robust as one would like
10     because of the underlying exposure uncertainties and low relative risks limit, which for some
11     assessors limits the confidence in the human-study-based risk estimates.
12            High- to low-dose extrapolations from various types of animal studies have been pursued
13     using a variety of modeling concepts, all of which have low-dose linearity because  of the
14     inference of mutagenicity/genotoxicity.  A variety of approaches is useful when information is
15     inadequate to make a clear or reasonable call about the mode of action or when definitive  insight
16     into dose-response aspects is lacking.  The margin of exposure (MOE) is also investigated for
17     additional insight about the magnitude of extrapolation.
18
19     12.3.3.2, Estimating the Cancer Risk of Diesel Exhaust
20            In attempting to determine the cancer risk (i.e., potency) of DE, several kinds of studies
21     and dose-response approaches were considered, including relative risks from epidemiologic
22     studies, animal bioassay responses, a comparative potency approach, and the use of a biomarker,
23     B[a]P, as a dosimeter.
24
25     12.3.3.2.1. Assessing risk using epidemiologic studies. As an ideal, epidemiology data are
26     preferred for risk estimation if the available data are rigorous enough. One quantitative approach
27     is to identify and use the dose-response relationship from an epidemiology study. This has been a
28     debatable issue with the best of the epidemiologic studies, the cohort study of railroad workers by
29     Garshick et al. (1988). This was a well-designed study with one of the largest cohorts.  Because
30     of the ongoing debate regarding the dose-response, EPA sees no benefit to generating more
31     analysis of the issue until the cohort study can be updated or consensus can be reached about how
32     to best use the current railroad worker cohort data. Garshick et al. (1987) also published a  nested
33     case-control study of the railroad workers, which showed increased relative risks of lung cancer
34     from DE exposure. These can be used to back into a risk derivation by using a proportional
35     population risk approach and overlaying this with separate assumptions about average exposure.

        2/1/98                                    12-19      DRAFT-DO NOT CITE OR  QUOTE

-------
 1     For convenience EPA has started with a published proportional population risk analysis
 2     (McCellan, 1989) based on the Garshick case-control study relative risks. Additionally, EPA
 3     selected a reasonable exposure estimate of 125 ug/m3 (and also included 500 ug/m3 for
 4     comparison), corrected the numerical risk estimates for occupational versus ambient exposure,
 5     and thus back-calculated equivalent estimates of unit cancer risk. The resulting risk estimates
 6     have 95% upper and lower bounds as well as MLE values.
 7            The adoption of a particular exposure value for risk derivation from the railroad worker
 8     study is a critical choice since risk magnitude is directly proportional to .the exposure. The true
 9     exposure of the railroad workers beginning in the late 1940s-1980s period is an unknown, though
10     Woskie (1988a, b) in conjunction with the railroad worker epidemiology study, evaluated current
11     levels of exposure for the railroad worker job categories.  Woskie sought to evaluate the current
12     exposures and comment on historical exposures for railroad workers by job category and did so
13     by collecting limited personal monitoring data for the job categories as well as employing
14     modifying factors to account for a number of influencing  factors. The job category exposure
15     estimates for the late 1980s showed that geometric mean exposures might range across all job
16     categories from about 17 u/m3 to 134 ug/m3 at a 95% confidence level, these included an
17     adjustment for cigarette smoke. It was also suggested that exposures were approximately constant
18     over the 1950s to 1980s given the circumstances of railroads that were sampled in the study.
19     National projections were slightly higher at 31-35 ug/m3  up to 125-157 ug/m3.  Woskie mentions
20     that anecdotal reports suggest that exposures were likely higher in the early period versus the later
21     years due to, for example, minimal ventilation, but the magnitude of difference could be
22     ascertained.
23            From these crude approximations, one has to make a choice as to how appropriate overall
24     estimates of 125 or 500 ug/m3 are. Woskie's work would suggest that 125 is a reasonable choice
25     and that 500 seems too high. Woskie's estimates were for worker shifts, while EPA's interest
26     would be for a 24-hour exposure which could be a lower value.  The diesel  locomotive engines
27     went through two generations, the first generation lasting through the  1950s, with a second
28     generation starting to appear in the 1960s. Woskie's data was mostly derived from first
29     generation engines that were still in use in the 1980s and which probably were higher emitters of
30     DPM.  The selection of 500 ug/m3 has no particular support, though it can't be ruled out. It does
31     show the sensitivity of the risk calculations to a fourfold higher (125 vs 500) exposure estimate.
32            An exposure estimate of 125 ug/m3 is in general agreement with measurements made
33     during the period of the underlying railroad worker study. As there is indirect evidence, but few
34     actual measurements, indicating that historical occupational exposures were higher, the use of 125
35     ug/m3 is a reasonable conservative choice, though not without uncertainty.  The selection of a best

        2/1/98                                   12-20       DRAFT-DO NOT CITE OR QUOTE

-------
  1     exposure value is an unresolvable question, though reason would suggest that the true exposures
  2     are in this range.
  3            For each exposure, three estimates of risk can be provided:  an upper end, an MLE
  4     (maximum likelihood estimate), and a lower end.  The case- control, proportional risk study-based
  5     estimates define a range as follows:
  6
  7     Exposure = 125 ug/m3:       Upper end risk       200 x 10"5 per ug/m3
  8                                MLE is              100 x IQ-5
  9                                Lower end risk        10 x 10"5
10     Exposure = 500 ug/m3:       Upper end risk        50 x 10'5
11                                MLE is               30xlO-5
12                                Lower end risk         3 x \Q-5
13            	.	.	.	.	  .
14            Upper bound from biomarker (BaP)      ~ 1 x  1Q'5 per ug/m3 (as discussed in next
15                                                 section, MLE = 0.26 xlQ-5)
16                                                                -
17     EPA is aware that the authors of the proportional risk analysis may have some reservations about
18     their conclusions using the case-control study, given the dose-response debate about the larger
19     cohort study, but EPA risk assessors believe the merit of using the case study relative risks is not
20   '  diminished by the cohort study debate.
21           A few comments about the human databased estimates are important to the question: How
22     good are they? The average occupational exposure of 125 ug/m3 is unlikely to be large enough to
23     induce lung particle overload in humans (a human level of 500 ug/m3 is likely to be in the
24     overload range, extrapolating from rat studies).  When using an average estimate for exposure, we
25     must remember that about half the time the exposures are higher, and while noting that
26     occupational exposure occurs about 40 h per week, the long-term occupational exposure is less
27     than lifetime, and thus mean particle concentrations adjusted to continuous exposure are unlikely
28     to exceed 100 ug/m3 and may be notably lower. Because cancer induction at nonparticle-
29     overload conditions is reasoned to be influenced by the mutagenically active organic fraction and
30     perhaps by the oxygen radicals present in fresh exhaust, the use of linearized low-dose
31     extrapolation is a supportable modeling approach. Individual smoking data were obtained in the
32     Garshick et al. (1987) case-control study, as was information about possible asbestos exposure,
33     thus eliminating a major potential source of bias. Also, relative risk ratios from other diesel
34     epidemiologic studies did not differ greatly from the 1.41 (CI = 1.06,1.88) reported in the
35     Garshick et al. (1987) study, increasing confidence in the Garshick-based relative risk findings.

        2/1/98                                   12-21       DRAFT-DO NOT CITE OR QUOTE

-------
 1     For example, in a recent meta-analysis of 34 studies by the State of California (Cal-EPA, 1997a)
 2     relative risk ratios ranging from 1.12 to 1.43 were derived, depending on the model used and
 3     whether or not they were corrected for smoking.
 4           Exposure estimation is a notable uncertainty in the risk estimates derived from the railroad
 5     worker case control study.  However, according to Woskie et al. (1988), the 125 ug/m3 estimate
 6     was probably reasonable as an average exposure near the end of the study period. The EPA risk
 7     estimate of 200 * 10"5 per ug/m3  defines the high end of a range of plausible upper-bound
 8     estimates for DE cancer risk, while 3 xlO"5 ug/m3 defines the lowest end of the human data range.
 9     The MLEs of the two exposure scenarios define a tighter range of 30-100 x. 10"5 per ug/m3.
10
11     12.3.3.2.2. Assessing risk using a biomarker.  A second approach using human data is the use of
12     a biomarker as a dosimeter.  Pike and Henderson (1981) related the concentration of
13     benzo[a]pyrene (B[a]P) to smokers, British gas workers, U.S.  coke oven workers, U.S. hot pitch
14     workers, and residents of rural and urban locations and found good agreement in predicting lung
15     cancer risk. They concluded that while B[a]P is not the only carcinogen present in DE,  and
16     perhaps not even the most important, it is a reasonably accurate dosimeter for assessing risk from
17     combustion or pyrolysis of petroleum products or tobacco and could therefore be appropriately
18     used for DE risk assessment. Based on Pike and Henderson's estimated lung cancer risk of
19     1/1,500 per ng/m3 B[a]P and a reported B[a]P concentration of 3.9 ng per ug of diesel particulate
20     matter in exhaust from a Volkswagen engine (Heinrich et al., 1995), a maximum likelihood
21     estimate of lung cancer risk of 2.6 x 10"6 per ug/m3 of diesel particulate matter can be derived.
22     The 95% upper bound of this value, while not calculated by the authors, is near 1 x 10"5 per
23     ug/m3.
24           A strength of the biomarker approach is its moderately good accuracy in estimating cancer
25     risk from exposure to a number of combustion/pyrolysis pollutants using B[a]P as a dosimeter.
26     However, while most of the combustion products assessed in the study contain organics similar to
27     DE, unlike DE they have little or no insoluble particulate matter.  Although this approach, like the
28     comparative potency method, fails to account for the carcinogenic effect of the particle itself or
29     for modifications in potency because of the association of the  organic component with particles,
30     the human exposures of interest presumably don't have particle-driven effects either. Of course,
31     the B[a]P concentration might also vary in diesel samples because of different types and sizes of
32     engines being run under varying  conditions and using different fuels.
33
34     12.3.3.2.3. Assessing risk using animal studies. Previous EPA attempts to quantitatively
35     estimate cancer risk based upon chronic rat bioassays used some form of a linearized model to

        2/1/98                                    12-22      DRAFT-DO NOT CITE OR QUOTE

-------
  1     extrapolate risk to low doses.  This approach was based upon the assumption that cancer response
  2     is a direct function of exposure concentration at all exposure levels. As discussed previously, this
  3     assumption is no longer supported by the available data because of the apparent difference in
  4     mode of action at high versus low exposures.  At high concentrations, rat lung cancer is believed
  5     to be induced by an interaction of particle overload with associated pathology, carcinogenic
  6     organics, and possibly oxygen free radicals. At lower doses, only the latter two are likely to have
  7     a major impact. This is likely to result in a dose-response curve that is nonlinear at least in the
  8     high-dose region.  Unfortunately, the animal bioassays lack sensitivity, in particular, the number
  9     of animals per group are too small to directly measure low-dose responses or show clearly where
10     the mode of action transition occurs on the dose-response curve. Clearly, at exposures above
11     2,200 u-g/m3 the dose response would be nonlinear. At some exposure level below 2,200, the
12     particle driven mode of action is diminished. For environmental levels of exposure, the risk
13     estimation objective would be to partition the dose-response curve and develop risk estimates
14     from the low-exposure portion.
15            Several quantitative modeling approaches were used with animal data to gain some insight
16     about low exposure only risks—none were completely satisfactory. The simplest approach is to
17     derive a 95% upper-bound estimate of risk using the highest concentrations of DE not inducing
18     pathologic responses and lung cancer. This conceptually places an upper bound on risk using
19     elevated but not statistically significant responses. The low-dose groups from the Mauderly and
20     Ishinishi studies (i.e., concentrations < 500 ^ig/m3) are suitable for such an analysis. Combining
21     these studies increases the statistical power somewhat, though the variability statistics of the
22     minimal response would tend to increase the risk estimate.  The resulting risk estimate using a
23     linear multistage extrapolation model (LMS) is 19 x 10"5 per jo.g/m3. This is several fold higher
24     than LMS estimates using all (including the high-dose responses) exposure groups (about 3.4 x
25     10"5) but approaches the higher risks suggested by the human data.    .
26            An alternative approach (mentioned in EPA's  Proposed Guidelines for Carcinogen Risk
27     Assesmenf) involves identifying a point of departure on the dose-response curve, perhaps ED]0 or
28     ED01, the concentrations inducing carcinogenic responses in 10% or 1% of the animals, and then
29     extrapolating risk from that point as long as extrapolation was justifiable. This was done for the
30     rat data, both EDW and ED01 risks were estimated using an LMS model.  The answer was nearly
31     identical to the 3.4 x 10~5 obtained from using the entire dose-response data. The linear model
32     seems insensitive for this particular data set.
33            A third approach published by Chen and Oberdorster (1996) involved development of a
34     biologically based dose-response (BBDR) extrapolation model. They applied it to the rat dose-
35     response data to consider the risk sensitivity to a threshold-particle driven MOA compared with a

        2/1/98                                    12-23       DRAFT-DO NOT CITE OR QUOTE

-------
 1     nonthreshold-mutagen driven MOA. If particle effects are assumed at all concentrations, i.e., no
 2     threshold for particle effects, then the model predicts risk virtually identical to those predicted by
 3     the linearized multistage model. If a threshold for particle-induced cancer effects is assumed at
 4     1,000 ug/m3, the BBDR risk estimate is about fivefold lower at a concentration of 1 ug/m3. This
 5     difference is consistent with the hypothesis that organics and oxygen radicals play a more
 6     important role as concentration becomes lower. The results, however, only suggest the influence
 7     of mutagenic-nonthreshold versus mutagenic-threshold modes of action, because data are
 8     insufficient to validate the model.
 9            There is also some debate, and hence uncertainty, regarding the biological adequacy of the
10     rat as a model for evaluating any human risk.  This begs the question whether rats may be
11     uniquely sensitive to particle-induced cancer, perhaps because of their different respiratory tract
12     anatomy. Only two other rodent species have been adequately tested, hamsters and mice.  The
13     response in mice was equivocal, and negative in hamsters. Hamsters, which do not respond to DE
14     exposure, are resistant to induction of lung cancer from DE.  Since there is evidence for human
15     carcinogenicity of DE in epidemiologic studies, one could argue that the rat may be more
16     qualitatively similar to humans in response to this agent than are the other laboratory species.
17     Furthermore, rats have been shown to respond similarly to humans when exposed to cigarette
18     smoke {Finch et al., 1995). Although the rat could be more sensitive from a particle standpoint,
19     we wouldn't necessarily say the same in regard to the organic components, which are playing a
20     role as well. Thus, while the use of rat data does involve uncertainty, it is not justifiable to fully
21     discount the rat as a plausible model for establishing a range of possible human risk from
22     exposure to DE.
23            A strength of rat-based risk estimates is that the studies are of good quality and provide
24     tumor responses that are in essential agreement with one another.  Another strength is the ability
25     to use a sophisticated dosimetry model to derive equivalent target tissue concentrations between
26     rats and humans. However, with all statistically significant rat responses occurring at exposures
27     where particle overload is also occurring, the use of rat data to define possible risk in any manner
28     has distinct shortcomings and resulting uncertainties.  For reference purposes, an LMS-derived
29     averaged risk estimate of 3.4 x  1Q"5 per jig/m3 was calculated from the high-dose rat responses of
30     Mauderly, Ishinishi, and Brightwell, Use of measured rather than modeled lung particle burdens
31     from the Mauderly et al. (1987) study did not significantly affect results. As a matter of curiosity,
32     if dose equivalence is based on lung burden per unit body weight374, a more traditional means of
33     species extrapolation for organics, LMS risk estimates would increase about threefold.
34            As it turns out, all of the rat-based risk estimates are near the low end of estimated risks.
35

        2/1/98                                     12-24       DRAFT-DO NOT CITE OR QUOTE

-------
  1     12.3.3.2.4. Assessing risk using a comparative potency method.  A third approach, developed
  2     before either animal bioassay or human epidemiologic data became available, is the comparative
  3     potency method of Albert et al. (1983). In this method, exhaust particle extracts obtained from
  4     three light-duty engines, manufactured by Nissan, General Motors, and Volkswagen, and a
  5     Caterpillar heavy-duty engine were evaluated for potency in a variety of short-term tests.  The
  6     ratios of potencies of these extracts were then multiplied by the unit risk estimates of related
  7     combustion or pyrolysis products, such as roofing tar, cigarette smoke condensate, and coke oven
  8     emissions, for which unit risk estimates based on human data had already been derived. The
  9     potency ratio of DE to each of these was multiplied by their unit risk estimate.  Using this method
10     Albert et al, (1983) derived a unit risk estimate for DE averaging about 3 x 10"5 per ug/m3.
11            The comparative potency method was developed more than 15 years ago because bioassay
12     data were  lacking at the time and because it was believed that cancer induction was a function of
13     the organic components. Conceptually, this approach can still be of use in rationalizing lower
14     limits on risk.  However, one must accept the assumption that relative potency in short-term tests
15     will be similar to relative potency for lung cancer induction.  Any reluctance may be balanced by
16     considerable reliance on dermal exposure in some of the short-term studies. Since both skin and
17     lungs are epithelial tissues, confidence that relative potencies are similar increases.  Another
18     possible weakness in the comparative approach is a failure to account for potential differences in
19     the relative bioavailability of the organic fraction during exposure to whole exhaust versus
20     exposure to particle extracts, or the possibility that association of organics with particles may alter
21     their effectiveness. It is not known whether these issues raise significant concerns.
22
23                         Tabular summary of animal-based risk estimates
24                 19 x  10"5 per (ig/m3           Upper limit from low-dose rat studies
25                3.4 *  10"5per ug/m3           Upper bound from high-dose rat responses
26                 3 x  10~5per ug/m3           Upper bound from BaP  comparative potency
27                                             analysis
28                  1  x  10"5 per ug/m3           Linear extrapolation using BBDR model
29
30     12.3.3.3. Bounds for Cancer Risk and Margin of Exposure
31            Each of the dose-response and  risk derivation approaches, using human or animal data, has
32     known uncertainties  that seemingly preclude  selecting one as the "most scientifically valid" or
33     best estimate. Taken collectively, the approaches provide a numerical basis for defining a range
34     of plausible upper-bound risks, upper-bound meaning that conservative assumptions (e.g., linear
35     low-dose extrapolation) have been used and this is not likely to result in an underestimate of risk.

        2/1/98                                    12-25      DRAFT-DO NOT CITE OR QUOTE

-------
 1           For the reasons cited earlier, the risk estimate of 2 * 10~3 per ug/m3 from human studies is
 2     the highest estimate and thus defines the upper-end value for a range of estimates. Three different
 3     animal-based approaches that assumed that lung cancer at low doses is a function of the organic
 4     components (the comparative potency method, the biomarker method, and low-dose estimation of
 5     risk from nonsignificant rat response data) resulted in upper-bound risk estimates from near 1  *
 6     10'5 up to 19 x 10"5 per ug/m3. A BaP human biomarker-based estimate of 0.26 x 10'5 as an MLE
 7     also exists, with an upper bound of the MLE being about 1 x 10~5. Since no threshold is assumed
 8     for the organic components of DE, if particles do contribute to risk at low exposures, the upper
 9     bound may not be as conservative as would normally be the case, i.e., normally we would say the
10     true risk, which is not definable, could be between the upper bound and be as low as zero.  The 1
11     x 10~5 per ug/m3 risk estimate from various animal-based approaches seems to be a floor of all
12     estimates and therefore,  establishes the low end for a range of risks.
13           The upper-end risk estimate of 2 x 10° per ug/m3 from the epidemiologic study is based
14     on an assumed long-term average exposure of 125 ug/m3. If a higher exposure were chosen the
15     risk estimate would decrease by a proportional amount.  Risk estimates based upon animal
16     bioassays, comparative potency, and the biomarker approach are as much as 200 times lower.
17     The range of possible human risks encompasses mixed MOA assumptions and does not fully
18     discount any experimental animal model system. Given the presence of known uncertainties with
19     the adequacy of the various animal test systems as a predictor of human hazard, and with the
20     uncertainties for the railroad worker exposures, neither the human nor animal test system provides
21     a scientifically compelling basis for selection of a single best risk estimate. Therefore, a range of
22     plausible risk estimates is recommended to characterize the possible public health impact of
23     exposure. Public health policy preferences might be appropriate, such as recognizing that the
24     human data-based estimates at least avoid uncertainties of species extrapolation and biological
25     relevance associated with the animal estimates, and thus a preference for the range of human data-
26     based risk estimates would be reasonable. An argument also could be made that the human
27     estimates derived from 125 ug/m3  exposure have more support than 500 ug/m3 because Woskie's
28     assessment was more consistent with a level of 125.  Some may also favor the MLE risk estimates
29     from exposure assumptions of 125 and 500 as a preferable selection of risks.
30           Upper bound risks  ranging from  1 x 1Q-5 to 200 x 1Q'5 (i.e., 2 x 1Q-3 ) per ug/m3 are
31     recommended for bounding the risk of human lung cancer induced by exposure to DE.  The risk
32     estimates evolving directly from human data run from 3 to 200 x 10'5 per ug/m3, with a subrange
33     of maximum likelihood  estimates spanning 30-100 x 1Q'5 per ug/m3. Animal-based estimates
34     range from 1 to 19 x 1Q:5 per ug/m3.
        2/1/98                •                   12-26       DRAFT-DO NOT CITE OR QUOTE

-------
  1            It may be insightful to consider the margin of exposure (MOE) for these risk estimates.
  2     An MOE compares ambient exposure of interest to LOAELs (adjusted to human equivalent
  3     LOAELs) from the human and from the animal studies. The resulting margin shows the
  4     extrapolation range from, e.g., an observable cancerous effect level to an ambient human exposure
  5     of interest. The margin value illustratates the range of extrapolation that has occurred when risk
  6     models have been employed to predict dose-response below the observable data.  The greater the
  7     magnitude of extrapolation, i.e., the margin, the more general uncertainty there is. If 2 u-g/m3 is
  8     an ambient human DE exposure of interest, if an estimate of 125 pg/m3 is selected from many
  9     possible choices for the human study of railroad workers, and 3.5 mg/m3 is a LOAEL rat exposure
10     from the Mauderly rat study (human equivalent concentration: 0.36 mg/m3 = 360 jag/m3 after
11     adjusting to continuous exposure conditions and across species), all of the human equivalent
12     exposure information needed for an MOE comparison is available. What is seen is that the range
13     of extrapolation from the human studies is about 1/3 of that for the animal studies, thus suggesting
14     a reduced margin for uncertainty compared to animal estimates. The ambient levels of interest are
15     nevertheless 10-60 times lower than the lower of the exposure scenarios associated with the
16     human studies.
17            The MOE comparison is displayed graphically in Figure  12-1. This figure may provide a
18     clearer picture of the relationship between exposure and various risk estimation recommendations
19     coming from this assessment.
20                                                                  -
21     12.3.4. Susceptible Subgroups
22            The hazards previously identified, i.e., acute symptoms including exacerbation of asthma,
23     and chronic effects such as reduced pulmonary functions and other respiratory weaknesses, are
24     assumed to be possible consequences in individuals of average health and in their adult years.
25     Individuals with preexisting lung burdens of particulates may have may have less of a margin of
26     safety from DE hazard consequences, though this cannot be quantified.  In reality, DE exposure is
27     probably additive to many other minute or larger exposures to mutagenic organies and paniculate
28     matter, but the magnitude of this additivity has not been estimated in this assessment.  For
29     example, adults who predispose their lungs to increased particle retention (e.g., smoking or high
30     paniculate burdens from nondiesel sources), have existing respiratory or lung inflammation or
31     repeated respiratory infections, or have chronic bronchitis, asthma, or fibrosis (e.g., silica
32     exposure) would have a much lower margin of safety and thus would be at greater hazard from
33     DE exposure. It hasn't been shown per se in DE studies, but infants and children may have a
34    -greater susceptibility to the acute/chronic toxicity of DE for the conventional public health reason
        2/1/98               '                     12-27       DRAFT-DO NOT CITE OR QUOTE

-------
WBrf (DOS)!'


>w/6rt (09£)
                               ,iu/6rl (921.) Apnis idg - ajnsodxg
                                                       ueujn|-|
                                                                     (g) juaiqinv ueiun|-|
                                                                Cuu/6r1 (z) juaiqujv ueuunn
                                                                                       o
                                                                                     I— O
                                                                                       o
                                                   o
                                                  • o
                                                                                           O)
                                                                                           3.
                                                                                          LU
                                                                                          W
                                                                                          O
                                                                                          Q
                                                             W
                                                             o
                                                             S
                                                             •o
                                                             •SS
                                                                                                 Jg
                                                                                                 K
                                                                                                 u
                                                                                                 a
                                                                                                 O

                                                                                                 o
                                                                                                 a:
                                                                                                 •c
                                                                                                 a,
                                                                                                 S
                                                                                                 o

                            I
                            CO
       I
      t
                                                                                       O
2/1/98
        12-28
DRAFT-DO NOT CITE OR QUOTE

-------
  1     that their developing pulmonary and immunologic systems may be more susceptible than an
  2     average adult's.
  3
  4     12.3.5.  Other Comprehensive Diesel Exhaust Health Assessments in the United States
  5            In 1997 Cal-EPA (California Environmental Protection Agency) released a draft hazard
  6     and risk assessment for DE emissions. Cal-EPA concluded that a reasonable, very likely
  7     explanation for the increased risks of lung cancer seen in experimental animal and epidemiologic
  8     studies is a bona fide causal association between DE exposure and lung cancer. Because of
  9     evidence for both carcinogenic and noncarcinogenic toxic effects, they also proposed to list DE as
10     a toxic air contaminant in California.  Their conclusion for the carcinogenic hazard is similar but
11     not identical to EPA's, (i.e., DE is highly likely to be carcinogenic in humans).  Cal-EPA also
12     provided a range for cancer risk per unit of lifetime exposure that is virtually identical to EPA's
13     risk range. The Cal-EPA risk range included a subrange of estimates from the animal studies and
14     a subrange of estimates from the human studies, as has EPA in this assessment. EPA indicates
15     that DE is "highly likely" to be a human carcinogen, whereas Cal-EPA is slightly more certain
16     about the likelihood. Given the health data uncertainties, data gaps, and the discretionary choices
17     that are made in risk assessment, the differences in the EPA and Cal-EPA assessment findings are
18     not significant. Cal-EPA's recommendations for noncancer respiratory hazards are identical to
19     EPA's because they adopted the EPA RfC of 5 ug/m3. As of February 1998, the Cal-EPA
20     assessment was still in draft.
21
22     12.4. EXPOSURE PERSPECTIVE
23            Diesel emissions are complex mixtures containing thousands of organic and inorganic
24     constituents in both gas and partieulate phases with differing chemical reactivities. After entering
25     the atmosphere, they are transported and transformed according to their distinctive characteristics,
26     undergoing physical and chemical changes that may form secondary pollutants more harmful than
27     their predecessors.  Thus, a knowledge of diesel emissions at or near their sources is not sufficient
28     to fully assess their impact on human health and welfare. However, data on how DE contributes to
29     exposure levels for these secondary pollutants are currently lacking.
30            Determining the amount of DE present in the ambient air is also complicated by the
31     difficulty of distinguishing organic compounds and particles that originate in diesel engines from
32     those that originate in gasoline engines or come from other sources.  This source speciation is not
33     well sorted out in the ambient characterization of DE.
34           Nonoccupational exposure to DE is worldwide in urban areas, with lesser exposure in rural
35     areas. Certain working populations are also exposed to higher levels of DE than the rest of the

       2/1/98                                   12-29       DRAFT-DO NOT CITE OR QUOTE

-------
 1     population. The level of exposure will differ within geographic areas based on the number and
 2     types of diesel engines in the area, as well as atmospheric patterns of dispersal and the location of
 3     the individual relative to the emission sources.
 4            While a detailed exposure assessment for DE has not been conducted as part of this study,
 5     the following exposure data are provided to give some context for the hazard assessment and
 6     dose-response analysis.
 7            Estimates of annual average concentrations of particulate matter in the ambient air,
 8     published in Chapter 9 of EPA's Motor Vehicle-Related Air Toxics Study (U.S. EPA, 1993), may
 9     be used to generate a crude estimate of the concentration of particulates from diesel exhaust in the
10     ambient air.  The total concentration of particulates from DE in ambient air in urban areas for
11     1995 was estimated as 2 ug/m3, the concentration in rural areas as 0.6 ug/m3, and the nationwide
12     average concentration as 1.1 ug/m3.
13            In an alternative estimation, using ambient monitoring data, total suspended particulate
14     matter (TSP) for  1990 was determined to equal 48 ug/m3. With approximately 5% of total
15     particulate matter associated with DE, multiplication of the total by the fraction contributed by
16     diesel exhaust and adjusting for time spent indoors results in an integrated estimate of 1.5 ug/m3,
17     according to EPA's hazardous air pollution exposure model (HAP-EM,  1988).
18            Exposure estimates for more highly polluted locations are somewhat greater.  Estimated
19     mean concentrations of DPM for Los Angeles  were reported to be 2.7 ug/m3 (Sienicki and Mago,
20     1992). McGlellan (1986) estimated concentrations on urban freeways and street canyons to be  as
21     great as 15 ug/m3.
22            Recent Cal-EPA (1996) studies show winter period estimates in three California locations
23     for diesel PM10 of 4 to 22 ug/m3. A broader Cal-EPA analysis shows average ambient outdoor
24     diesel PM10 to range from 0.2 to 3.6  ug/m3 across 14 California air basins with a population-
25     weighted average of 3.2 ug/m3. Concentrations in occupational settings may be higher.
26            Recent studies, including a study of the Baltimore Harbor Tunnel (conducted by the
27     Desert Research Laboratory for the American Petroleum Institute) and an ORD measurement
28     study of tailpipe emissions from a moving heavy duty  diesel truck, have  confirmed that dioxins
29     are formed and emitted from heavy-duty diesel trucks. ORD's draft dioxin source emission
30     inventory, being developed as part of the ongoing Dioxin Reassessment effort and scheduled to
31     undergo peer review in April 1998 estimates that 60 g  TEQ are emitted from U.S. trucks. When
32     this estimate is compared with total estimated U.S. emissions of 3,000 g TEQ, it appears that
33     diesel trucks are not a major dioxin source.  However, it is unknown whether such emissions
34     could have significant local impacts, since current information does not permit us to rule out the
35     possibility of exposures of interest. The types of exposures that have been of particular concern

        2/1/98                                    12-30       DRAFT-DO NOT CITE OR QUOTE

-------
 1     from stationary dioxin sources, e.g., incinerators, have been noninhalation exposures associated
 2     with ingestion of certain foods, e.g., beef and vegetables, contaminated by the deposition of stack-
 3     emitted dioxin. There is a potential for deposition of stack emissions from trucks to soil and
 4     water adjacent to some highways resulting in similar impacts. We recommend that appropriate
 5     data be collected.
 6           The changing composition of DE (i.e., older engines vs. newer technology ones, heavy-
 7     duty vs. light-duty, and engines ran under varied operating conditions) gives rise to questions
 8     about how the health data and the risk assessment findings in this report, which are based on pre-
 9     1998 engines, can be applied to present-day engine exhaust emissions and resulting ambient
10     exposures.  This is a complex question that is not rigorously addressed in this assessment. It is
11     clear that newer technology engines will have somewhat different emission composition (i.e.,
12     perhaps reduced NOX with increased fine particles), not to mention emission controls, which
13     would reduce certain exhaust components, presumably larger particles.  Since particle mass is the
14     surrogate dosimeter used to correlate toxicity with exposure and public health impact, the
15     implication is that we have a basis for scaling to account for exhaust changes. This relationship
16     may be too simplistic, and thus further investigations of current-day emissions, and how they
17     average out across the fleet or stationary engine use, may be warranted.
18
19     12.5. DIESEL EXHAUST HEALTH RISKS-A PLAIN-LANGUAGE OVERVIEW OF
20          KEY INFORMATION
21          This section reviews key information about diesel emission hazards and risks by posing and
22     answering simple questions. It is mostly duplicative of information found earlier in this chapter,
23     though some added explanation has been added for background purposes.
24
25     What is diesel exhaust (DE)? What happens when it enters the environment?
26          Diesel engines are very durable and have performance characteristics that make them a
27     desirable  alternative  in certain uses. The diesel engine, whether it be in an automobile, truck, off-
28     road equipment, locomotive., or ship, produces exhaust from the combustion of diesel fuel. The
29     exhaust is a mixture  of organics and inorganic constituents (i.e., products of incomplete
30     combustion).  The exhaust can be invisible or be seen as a gray or black  smoke. When visible,
31     what is seen is the particle fraction of the exhaust mixture, usually from  an engine that is not
32     required to control its emission or one that is not well maintained. In the simplest terms, DE is a
33     mixture of carbon core particles that have a coating of various inorganic/organic compounds, as
34     well as gases and semivolatiles. The identifiable organic and inorganic compounds number in the
35     hundreds. The particles have a spongelike structure and a very large surface area per gram, which

        2/1/98                                   12-31       DRAFT-DO NOT CITE OR QUOTE

-------
 1     make them an excellent carrier for adsorbed inorganic/organic compounds. The amount of
 2     specific chemical compounds present and the size of the particles depends on the engine design,
 3     load, operating speed, fuel consumption, and whether or not the engine is in a well-maintained
 4     state. The diameter of diesel particles is very small, typically 0.1 to 0.25 urn, with more than 75%
 5     of the particles smaller than 1 um. Light-duty diesel engines emit 50-80 times and heavy-duty
 6     engines 100-200 times more particulate matter than catalytically equipped gasoline engines. The
 7     heavy-duty and off-road diesel engines, as a group, account for most of the DE particulate
 8     emissions discharged into ambient air.
 9          When the exhaust first escapes to the ambient air it is called "fresh"  exhaust, which also
10     connotes some increased chemical and biological reactivity properties, compared to "aged"
11     exhaust, which after a day or so has diminished reactivity. Once in the ambient air, some of the
12     organic, inorganic compounds and oxygen free radicals associated with the fresh exhaust begin to
13     transform into other chemical compounds because of their exposure to sunlight or other
14     atmospheric elements.  The exhaust mixture also becomes dispersed and thus diluted in the
15     ambient air. The topography and/or climatic conditions in a particular area may promote
16     dispersal, or in some cases hinder dispersal to the point of causing a slow accumulation of DE
17     components, e.g., because of ground-level air stagnation. The particle fraction of DE contributes
18     to the background ambient particulate matter (PM) in the air, while the various gaseous
19     inorganic/organic components add to other background loadings in the ambient air.
20          The measurement of whole DE is complex because it is a mixture of particles and gases.
21     The approach adopted by researchers has been to use particle mass as a surrogate for the whole
22     mixture. The mass measurement is in units of weight (ug, microgram) per volume of air (m3,
23     cubic meter). This emphasizes the particle fraction of the whole DE mixture together with its
24     adsorbed organic/inorganic components, rather than the gaseous organic/inorganic constituents.
25     The latter are, however, in relative proportion to the particle mass present.
26
27     How are individuals exposed to DE? How does it enter and leave the human body? Is there
28    .a test to determine whether exposure has occurred?
29          Individuals may be exposed to DE when they are in an area where diesel engines are in use
30     and the exhaust mixture is breathable.  Diesel engines are nearly everywhere, so it becomes a
31     matter  of relative frequency of contact (i.e., exposure), duration of exposure, and concentration of
32     the exhaust mixture. Some occupational settings may be prone to more frequent and higher
33     exposures, such as in engine maintenance shops, heavy equipment operations, mining, or bus
34     terminal operations, to suggest a few. A nonoccupational setting that may have a higher than
35     average ambient exposure could be, for example, among  those who spend a notable part of their

        2/1/98  .                                12-32       DRAFT-DO NOT CITE OR QUOTE

-------
  1     day in the vicinity of diesel roadway traffic, such as in or around highways or urban street
  2     canyons. For some DE emissions, one can see smoke or soot, indicating that some relatively large
  3     particles are present. Emissions with the more typical  small-diameter particle are virtually
  4     invisible. The odor threshold, at least according to one study, is about 200 ug/m3 of particulate, or
  5     greater.
  6           This assessment does not determine exposure levels across the population, but some
  7     exposure estimates from various sources are rioted to provide a frame of reference. For example,
  8     several evaluations show average rural and urban ambient levels  of DE to be in the range of 0.6-
  9     3.2 fig/m3. Recent California-EPA studies show winter-period estimates in three California
 10     locations for diesel PMIO (particulate matter < 10 um) of 4 to 22 ug/m3.  A broader Cal-EPA
 11     analysis shows average ambient outdoor diesel PM10 to range from 0.2 to 3.6 ug/m3 across 14
 12     California air basins with a population-weighted average of 3.2 ug/m3. Concentrations in
 13     occupational settings may be higher.
 14           DE exhaust most easily enters the body by breathing, though in some occupations portions
 15     of the exhaust may cling to skin or hair and thereafter possibly be ingested as a consequence of
 16     hand-to-mouth activity. By far, the major exposure pathway is from breathing.
 17           Once inhaled into the nasal passage, some DE mixture  components could be deposited or
.18     absorbed along the upper nasal and respiratory tract, but most of the mixture travels into the lungs
 19     where the particles and gases are deposited on lung tissue. The inhale-exhale pattern of breathing
 20     results in some exhalation of the particles and gases; the remaining particles stay in the lung until
 21     the body's natural defense mechanisms mobilize to clear them out. The remaining gases and
 22     organics/inorganics coated on the particles are eventually absorbed into the lung  tissue, then into
 23     the bloodstream, and thereafter begin a process of normal detoxification by various body organs
 24     followed by removal from the body via urine and feces.
 25          There is no single medical test to determine if a DE exposure has occurred. Many
 26     symptoms of episodic DE exposure are similar to symptoms caused by other agents or, in some
 27     cases, the onset of a common cold.  Invasive sampling  of particle deposits in the  upper respiratory
 28     tract or lung could be done, yet such particles may not be readily distinguishable from particulate
 29     matter from other sources.
 30
 31     How does exposure affect human health and how certain are we about these effects?
 32          One way to consider the possible harmful effects of DE is to consider acute exposure (i.e.,
 33     episodic/infrequent contact) versus chronic exposure (i.e., fairly continuous over long periods of
 34     time, such as years). As the exposure frequency and/or duration of the contact increase, acute
 35     exposure and its effects give way to chronic exposure and its consequences. Most health studies

        2/1/98        .                           12-33       DRAFT-DO NOT CITE OR QUOTE

-------
 1     are designed to evaluate either acute or chronic effects. Another aspect of the exposure event is to
 2     realize that in a general sense total cumulative exposure and the rate at which the exposure is
 3     received in some manner influences the nature and/or the extent of a harmful effect, this being a
 4     traditional toxicological concept not unique to DE. This relationship is not always definable in a
 5     rigorous way, but is useful as a concept in relating exposure features to toxic consequences.
 6     Many DE components have a potential of being harmful because they are toxic at some exposure
 7     in their own right.  A DE component may start out being  toxic, may start out being nontoxic but
 8     be changed into a toxic substance by the body's defense mechanisms, or have no toxicity at all in
 9     any phase. The question is, do the concentrations in the DE mixture, when taken as a total, cause
10     harmful effects?
11           The pure  carbon core DE particle, the organics coating the particle, and the gas/vapor phase
12     components of the mixture all have health study evidence that shows toxicity, and thus potential
13     to be hazardous under some regime of exposure.  Taken individually, both the particles and some
14     of the chemical compounds can be irritants and cause inflammation in the respiratory system, or
15     in larger amounts cause more permanent harmful effects. For example, among the many
16     hydrocarbons found in DE, 19 of them are believed or known to pose a human carcinogenicity .
17     hazard, with the magnitude of the risk thought to be proportional to total exposure over a lifetime.
18     It is not clear precisely what components of the DE mixture are key to causing the acute effects or
19     the more permanent chronic effects and, in fact, most of the components may play a role, which
20     may change as the human exposure changes from low to  higher levels.
21           Effects of acute exposure on humans have not been systematically and comprehensively
22     studied, but there are some symptoms seen, depending on the person and the concentration of DE.
23     The symptoms are a biological response to irritation of human tissues. The symptoms range from
24     no effect, to annoying and quickly passing effects, to effects that cause temporary impairment, up
25     to and including symptoms that may indicate permanent impairment, e.g., from a very high one-
26     time exposure. Examples at the lower levels include: headache, runny eyes and  nose, or nausea,
27     up to and including restricted breathing due to respiratory resistance (asthmalike  response).
28     Immunologic allergic reactions resulting in long-term hypersensitivity to DE and perhaps other
29     ambient agents may also be possible, according to some very recent human and animal research.
30     Animal studies have confirmed the irritational aspects of contact with DE and further suggest that
31     high acute exposures may cause lung damage. The supposition is that different people have
32     different levels of tolerance or susceptibility and that the  higher the exposure, the more likely
33     people are, in general, to experience an unpleasant symptom or perhaps have a more permanent
34     adverse effect such as hypersensitivity or lung damage.
        2/1/98                                    12-34       DRAFT-DO NOT CITE OR QUOTE

-------
  1           As the exposure instances increase, changing from episodic to more continuous and
  2     increasing from weeks to months to years, it is clear that too much exposure increases the
  3     likelihood of noncancer respiratory system damage or the risk of lung cancer, and thus we say DE
  4     at some level of chronic exposure poses a respiratory  hazard for humans.  Both human and animal
  5     studies provide evidence of this. But the human evidence specifically for diesel exhaust is not as
  6     clear as that from the well-controlled studies in test animals. As the total exposure over a lifetime
  7     increases, basic respiratory functions can' be impaired, and there is a probability (i.e., risk) that
  8     lung tumors may appear later in life. Part of the permanent harm may be caused by the particle
  9     portion of the exhaust, part may be caused by the other inorganic/organic constituents on the
10     particle  or in the DE gases, or all of these may be interacting to influence the adverse outcome.
11     With animal studies being conducted at high test exposures, and with the occupational human
12     studies being somewhat lower in exposure but greater than ambient levels of DE, it  is necessary to
13     rationalize whether DE also poses a low-exposure hazard, whereas it is clear that at  high
14     exposure/doses it does. Plausible explanations supported by observations about adverse effects at.
15     higher doses are not matched with equal information about how effects may develop at low doses,
16     and thus there is uncertainty about how to best estimate the low-dose hazards or risk. These
17     questions are a current pursuit of health researchers. EPA also takes the position that chronic DE
18     exposure, at high or low concentrations, is very likely to increase the hazard and risk of an
19     adverse  consequence.
20           Specific individuals inherently have varied susceptibility to these adverse outcomes,
21     depending on whether they already have a weakened or compromised respiratory system, perhaps
22     by smoking or having allergic or asthmatic symptoms. Although it is not demonstrated, one could
23     hypothesize that episodic or frequent exposure of young children to DE could disproportionally
24     increase their lifetime  respiratory hazard or lifetime cancer risk, because damage early in life
25     could increase their susceptibility, in addition to their having a longer period to accumulate
26     exposure. Current data do not permit any more definitive explanations, nor is there  a confident
27     identification of additional human health hazards beyond those to the respiratory system.
28
29     What recommendations exist to protect human health?
30           This health assessment identifies the likely human health hazards and uses additional risk
31     assessment tools to assist decision-makers in understanding what is important for protection of
32     public health. These tools are in the form of concentration-based exposure-response relationships
33  *   based on the identified hazard likelihood (i.e., respiratory damage or lung cancer). These
34     relationships facilitate the rough estimation of acceptable/unacceptable exposures or health
        2/1/98                                    12-35       DRAFT-DO NOT CITE OR QUOTE

-------
 1     impacts.  Use of these measures requires caution and recognition of the biological and risk
 2     assessment uncertainties that are present.
 3           For the acute effects of DE exposure, there is no specific recommendation from this
 4     assessment for a concentration not to be exceeded in order to avoid acute effects, because of an
 5     absence of sufficient data.  Clearly, if acute symptoms are noted one would want to remove
 6     oneself from the locale as soon as practicable, if for no other reason than personal comfort.  As the
 7     level of exposure increases, the acute symptoms can become more annoying and be indicative of
 8     temporary impairment. With the inherent variability of susceptibility and sensitivity in the human
 9     population, it is not possible to judge the outcome of a specific exposure incident. The same
10     statements could generally be made about exposure to gasoline exhaust and many other agents as
11     well.
12           For the chronic effect hazards, EPA believes that for many people, keeping long-term
13     exposures at or below 5 ug/m3 of diesel particulate matter provides an adequate margin of safety
14     for noncancer respiratory hazards. This level also includes a 10-fold margin to account for
15     variability in the human population. This is not an absolute demarcation of acceptable versus
16     unacceptable exposure, since an order-of-magnitude range of uncertainty is thought appropriate
17     for this recommendation.  For practical purposes, the belief is that as the long-term average
18     exposure concentration exceeds this value, the likelihood of respiratory distress increases.  The 5
19     ug/m3 value conies from test animals who experienced respiratory distress at higher experimental
20     exposures to which a margin of safety (e.g., uncertainty factor) and animal-to-human equivalence
21     factors have been used to derive the 5 ug/m3 level for humans. It was necessary to depend on
22     animal studies because the human database was not robust enough.  While children should not, a
23     priori, be assumed to be protected by adult recommendations, at this time there is no separate
24     recommendation for children, except for the general wisdom to minimize exposure as much as
25     possible.
26           For carcinogenic hazard and risk of cancer over a lifetime, EPA is recommending that
27     exposure be viewed as likely to pose a risk at low levels, as well as high levels, and is offering a
28     crude range of cancer risks per unit of lifetime exposure in order to gauge the public health
29     acceptability of exposures. The risk values provide an upper bound to the possible human risk,
30     rather than a true estimate; the true estimate is undefmable and could be much lower.  A range of
31     risk estimators was provided because the available cancer data had too many uncertainties to
32     justify the selection of one scientifically best estimate. The risk range is thought to bracket the
33     upper limits of possible risk, and these values would not likely underestimate the true risk.
34           Assuming that DE is a cancer hazard for humans, EPA believes that the cancer risks for DE
35     are not likely higher than  1 * 10"5 to 200 * 10'5 per ug/m3 of diesel particulate. These values

        2/1/98               .                    12-36       DRAFT-DO NOT CITE OR QUOTE

-------
  1     evolve from consideration of human occupational exposure-responses, several types of high-dose
  2     animal studies, and extrapolation to provide an appropriate estimate for low exposures,
  3     Extrapolation below the range of observation has uncertainty but is necessary from a public health
  4     perspective, because low-exposure cancer effects are often not detectable, yet low exposure
  5     hazards are expected for DE given our knowledge about the DE mixture components.
  6     Numerically, the risk estimates are the same as saying the probability of cancer could be, and is
  7     not likely higher than, 1/100,000 to 200/100,000 per microgram of diesel particulate matter per
  8     cubic meter of air (|ig/m3, an average particle concentration over a lifetime).
  9           With exposure information, a risk assessor can make crude estimates of the highest possible
10     impacts on a population, if such analysis is desirable.  Under this type of evaluation, for an
11     individual the development of cancer is a matter of chance. A person either gets cancer or doesn't
12     by the end of their lifetime, but in the interim they have a probability (i.e., a risk) of getting
13     cancer.  Nationally, the lifetime risk of being diagnosed with any type of cancer is about 1 in 4
14     (1/4), and as a cause of death is about 1 in 5 (1/5).  Cancer of the lung runs about 1 in 12 for males
15     and 1 in 19 for females.
16           As an example of how the crude risk estimations can be informative, at a hypothetical
17     human average lifetime DE exposure of 2 u*g/m3, there would be little likelihood of noncancer
18     respiratory hazard because this is less than 5 u-g/m3, whereas the cancer risks are not likely higher
19     than 2 7100,000 .(1 in 50,000) up to about 400/100,000 (1 in 250). At a concentration of 5 ug/m3,
20     which is still presumed protective for respiratory effects, there remains an upper-limit cancer risk
21     of 5/100,000 (1 in 20,000) up to 1,000/100,000 (1/100).  It should be noted  that as these all of
22     these risks are upper bound, the true risk is unlikely to be greater and may well be less.
23
24     Are there other  important considerations of DE exposure?
25           Particulates (i.e, particulate matter, PM) are a prominent constituent in DE.  Breathing
26     nonspecific PM is a public health concern in its own right, as evidenced by EPA's 1997 Ambient
27     Air Quality standards for PM. When present, DE plays a role in contributing to ambient PM,
28     especially PM 2 5 (PM less than 2.5 urn in diameter). Diesel particulates are small; more than  75%
29     of them can be less than 1 Jim, which means that EPA's new PM25 standard provides another
30     health-based reference point.  DE particulates are potentially a more toxic fraction in a PM  2.5
31     mixture because the smaller DE particles (<1 um) can be deposited deeper in the lung. Because
32     of their small size they also have a large surface area per unit mass and carry a coating of organic
33     compounds with  them.  Though diesel particulates are associated with a carcinogenic hazard,  this
34    • is not indicated, per se, for ambient PM exposure.
        2/1/98                                    12-37      DRAFT-DO NOT CITE OR QUOTE

-------
  1            Older diesel engines emit higher levels of nitrogen oxides (NOJ than do gasoline engines,

  2      and NOX is also an ambient urban contaminant that EPA seeks to reduce because of its influence

  3      on ozone formation, formation of nitrate PM, acid rain, and the eutrophication of coastal waters.

  4      Some new engine design is focused on reducing the NOx.

  5            Those individuals who already carry a significant burden of particles in their lungs or have

  6      weakened respiratory systems (e.g., from allergies, asthma, or other respiratory system

  7      inflammation) could be at higher hazard/risk. These special population subgroups are difficult to

  8      enumerate, but they do exist.  These same individuals might also be more sensitive to a number of

  9      insults, such as general PM or gasoline engine exhaust or smog.

10

11      12.6. REFERENCES

12
13      Albert, RE; Lewtas, J; Nesnow, S; Thorslund, TW. (1983) Comparative potency method for cancer risk assessment:
14      application to diesel particulate emissions. Risk Anal 3:101-117.
15
16      Bhatia, R; Lopipero, P; Smith, AH. (1988) Diesel exhaust exposure and lung cancer. Epidemiology 9:84-91.
17
18      Brightwell, J; Fouillet, X; Cassano-Zoppi, AL; Bernstein, D; Crawley, F; Duchosal, F; Gatz,  R; Perczel, S; Pfeiffer,
19      H. (1989) Tumors of the respiratory tract in rats and hamsters following chronic inhalation of engine exhaust
20      emissions. J Appl Toxicol 9:23-31.
21
22      California Environmental Protection Agency (Cal-EPA). (1996) Health risk assessment for diesel exhaust, public and
23      scientific panel review draft; Office of Environmental Health Hazard Assessment, California Environmental
24      Protection Agency.
25
26      California Environmental Protection Agency (Cal-EPA). (1997a) Health risk assessment for diesel exhaust, public
27      and scientific panel review draft; Office of Environmental Health Hazard Assessment,  California Environmental
28      Protection Agency; March 1997.
29
30      California Environmental Protection Agency (Cal-EPA). (1997b) Proposed identification of diesel exhaust as a toxic
31      air contaminant. Part A exposure assessment, public comment and SRP version; California Environmental Protection
32      Agency; May 1997.
33
34      Chen, CW; Oberdorster, G. (1996) Selection of models for assessing dose-response relationships for particle-induced
35      lung cancer. Inhal Toxicol 8(supp):259-278.
36
37      Dasenbrock, C; Peters, L; Creutzenberg, O; Heinrich, U. (1996) The carcinogenic potency of carbon particles with
38      and without PAH after repeated intratracheal administration in the rat.- Toxicol Lett 88:15-21.
39
40      Finch, GL; Nikula, KJ; Barr, EB; Bechtold, WE; Chen, BT; Griffith, WC; Hobbs, CH. (1995) Lung tumor synergism
41      between 239PuO2and cigarette smoke inhaled by F344 rats. Toxicologist 15:47 (abstr).
42                                                              •             •                         .
43      Garshick, E; Schenker, MB; Munoz, A;  Segal, M; Smith, TJ; Woskie, SR; Hammond,  SK; Speizer, FE. (1987) A
44      case-control study of lung cancer and diesel exhaust exposure in railroad workers. Am Rev Resp Dis 135:1242-1248.
45
46      Garshick, E; Schenker, MB; Munoz, A;  Segal, M; Smith, TJ; Woskie, SR; Hammond,  SK; Speizer, FE. (1988) A.
47      retrospective cohort study of lung cancer and diesel exhaust exposure in railroad workers. Am Rev Resp Dis
48      137:820-825.

         2/1/9.8                                       12-38        DRAFT-DO NOT CITE OR QUOTE

-------
  1     Health Effects Institute. (1995) Diesel exhaust: a critical analysis of emissions, exposure, and health effects.
  2     Cambridge, MA: Health Effects Institute.
  3
  4     Heinrich, U; Fuhst, R; Rittinghausen, S; Creutzenberg, O; Bellmann, B; Koch, W; Levsen, K. (1995) Chronic
  5     inhalation exposure of Wistar rats and two different strains of mice to diesel engine exhaust, carbon black and
  6     titanium dioxide. Inhal Toxicol 7:553-556.
  7
  8     Ishinishi, N; Kuwabara, N; Nagase, S; Suzuki, T; Ishiwata, S; Kohno, T. (1986) Long-term inhalation studies on
  9     effects of exhaust from heavy and light duty diesel engines on F344 rats. In: Carcinogenic and mutagenic effects of
10     diesel engine exhaust: proceedings of the international satellite symposium on toxicological effects of emissions
11     from diesel engines. Ishinishi, N; Koizumi, A; McClellan, RO; Stober, W, eds. July 1986, Tsubuka Science City,
12     Japan. Amsterdam, Holland: Elsevier Science Publishers B.V., pp. 329-348. (Developments in toxicology and
13     environmental science: v.  13.)
14
15     Mauderly, JL; Jones, RK; Griffith, WC; Henderson, RF; McClellan, RO. (1987) Diesel exhaust is a pulmonary
16     carcinogen in rats exposed chronically by inhalation. Fundam Appl Toxicol 9:208-221.
17
18     McClellan, RO. (1986) Health effects of diesel exhaust: a case study in risk assessment. Am Ind Hyg Assoc J 47:1-
19     13.                                     •        •
20
21     McClellan, RO; Cuddihy, RG; Griffith, WC; Mauderly, JL. (1989) Integrating diverse data sets to assess the risks of
22     airborne pollutants. In: Assessment of inhalation hazards. Mohr, U, ed. New York: Springer Verlag, pp. 3-22.
23
24     Pike, ME; Henderson, BE. (1981) Epidemiology of polycyclic aromatic hydrocarbons: quantifying the cancer risk
25     from cigarette smoking and air pollution. In: Gellbond, HV; Tso, POP, eds. Polycyclic hydrocarbons and cancer.
26     New York: Academic Press, pp. 317-334.
27            •
28     Sienicki, EJ; Mago, RS. (1992) Re-evaluation of diesel particulate matter inventories. In: Toxic air pollutants from
29     mobile sources:  proceedings of an international specialty conference. Pittsburgh, PA: Air and Waste Management
30     Association, pp. 151-164.
31                                                                 <
32     U.S. Environmental Protection Agency. (1986, Sept. 24) Guidelines for carcinogenic risk assessment. Federal
33     Register 51(185):33992-43003.
34
35     U.S. Environmental Protection Agency. (1993) Motor vehicle-related air toxic study. Ann Arbor, MI: Office of
36     Mobile Sources. EPA/420/R-93/005. Available from National Technical Information Service, Springfield, VA,
37     PB93-182590/XAB.
38
39     U.S. Environmental Protection Agency. (1995) Integrated Risk Information System (IRIS).  Online. Cincinnati, OH:
40     National Center for Environmental Assessment,
41
42     U.S. Environmental Protection Agency. (1996) Proposed guidelines for carcinogen risk assessment. EPA/600/P-
43     92/003C,
44
45     Woskie, SR; Smith, TJ; Hammond, SK; Schenker, MB; Garshick, E; Speizer, FE. (1988). Estimation of Diesel
46     Exhaust Exposures of Railroad Workers: I. Current exposures. Am J Ind Med 13:381-394.
47
48     Yoshizumi, K; Watanabe, T; Ishii, K; Yoshioka, H; Asakino, K; Tamura, T; Kan, M; Yoshino, N. (1989).
49     Assessment of the sources contributing to suspended particulate matter in Tokyo by receptor model. Ann Rep Tokyo
50     Metrop Res Inst Environ Prot 3-10.
51
52     Yu, CP; Yoon, KJ; Chen, YK. (1991) Retention Modeling of Diesel Exhaust Particles in Rats and Humans. J Aerosol
53     Med 4:79-115.
         2/1/98                                         12-39        DRAFT-DO NOT CITE OR QUOTE

-------
                      Appendix A

         Experimental Protocol and Composition
                of Exposure Atmospheres
2/1/98                      A-l    DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF
EXPOSURE ATMOSPHERES"
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/m3)
Particle size
CO, (%)
CO (ppm)
NO, (ppm)
NO (ppm)
SO, (ppm)
SO/2 (Atg/m3)
Ozone (ppm)
Aliphatic aldehydes
(ppm)
Formaldehyde (ppm)
Acrolein (ppm)
NH/
THC (ppm)
PAHs
Benzo(a)pyrene
Benzo(e)pyrene
Benzo(a)anthracene
Benzo(k)fluoranthene
Fluoranthene
Pyrene
Phenanthrene
Chrysene
Perylene
Indeno(l,2,3-Cd)
fluoranthene
Indeno(l,2,3-Cd)
pyrene
Benzo(ghi)perylene
U.S. Environmental Protection Agency
Bhatnageretal., 1980; Campbell et al., 1980, 1981;
Hyde et al., 1985; Moorman et al., 1985; Pepelko et
al., 1980b, 1981; Pepelko, 1982b; Pepelko and
Peirano, 1983; Plopper et al., 1983
Nissan CN 6-33., 3.24 L, 6-cylinder
Federal short cycle
No. 2 diesel
0.15%
8 h/d, 7 d/week, 124 weeks
Control
0.00

0.04 + 0.002
2.20 + 0.50
0.03 + 0.03
0.05 + 0.04
0.03 ± 0.02
-

0.00
0.00
0.00
-
2.82 ± 0.50












Exhaust -
weeks 1-61
6.34 + 0.81
Exhaust -
weeks 62-124
11.70 ±0.99
90% < 1 jtm by mass;
50% <, 0.3 urn by mass
0.30 ± 0.04
20.17 ± 3.01
2.68 ±0.80
11.64 + 2.34-
2.12 + 0.58
-

0.177 +0.043
0.106 + 0.029
0.025 + 0.003
-
7.93 + 1.42
0.52 ± 0.04
33.30 ± 2.94
4.37 ±1.19
19.39 + 3.80
5.03 ± 1.03
-

0.338 ± 0.057
0.251 ± 0.059
0.034 ± 0.009
-
11.02 + 1.04
15.9 /ig/g extract
28.6 ftg/g extract
53.8 jug/g extract
77. 8 inglg extract (k+b)
155.8 /zg/g extract
198 jtg/g extract
145.2 fjig/g extract
71.6 /ig/g extract
3.5 /tg/g extract
10.9 jtg/g extract
14.8 pg/g extract
21.1 fig/g extract
Laurie et al., 1980; Laurie and
Boyes, 1980, 1981
3.24 L, 6 cylinder
Federal short cycle
No. 2 diesel
0.15%
8 h/d, 7 d/week, 16 weeks
Control
0.01

0.05 + 0.00"
1.86 + 0.06"
0.03 ± 0.00b
0.08 + 0.01b
0.46 ± 0.02"






3.22 ± 0.08b












Exhaust
5.97 ± 0.17b

0.28 + 0.01"
19.20 ±0.35b
2.51 ±0.10b
11.14 ±0.43"
1.82 + 0.07"






7.29 ± 0.11"












a All + are S.D., unless specified otherwise.
h Standard error of mean values.
2/1/98
A-2
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF
EXPOSURE ATMOSPHERES"
Facility /Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/m3)
Particle size
CO, (%)
CO (ppm)
NO2 (ppm) '
NO (ppm)
SO, (ppm)
SO/2 (Mg/m3)
Ozone (ppm)
Aliphatic aldehydes
(ppm)
Formaldehyde (ppm)
Acfolein (ppm)
NH^
THC (ppm)
PAHs
Benzo(a)pyrene
Nitropyrene
U.S. Environmental Protection Agency
Wiesteretal., 1980
Nissan CN6-33, 3.24 L, 6 cylinder
California cycle, modified
No. 2 diesel
0.15%
20 h/d, 7 d/week, 4 weeks
Control
0.00

0.04
2.0
0.07
O.ll
0.0
0.00
0.0 -




0.00


Exhaust
6,32 ±1.31
Exhaust - irradiated
6.83 ± 1.44
0. 1-1.0 ^tm
0.261 + 0.01
17.4 ± 2,5
2.3 ± 0.4
5.9 ±0.6
2.1 ±0.8
0.57 ±0.12
0.0




31.6 + 2.3


0.25 ± 0.03
16.7 ±4.0
2.9 + 0.7
5.0 + 1.2
1.9 ±0.8
0.57 ±0.13
<0.01




26.1 ± 1.6


Pepelkoetal., 1980a
3.24 L, 6 cylinder
California cycle, modified
No. 2 diesel

20 h/d, 7 d/week, 4 weeks
Exhaust
6.40 ± 0.36"

0.26 ± 0.008b
14.61 + 0.90"
2. 13 ±0.09" '
6.13 ±0.18"
2.10 ±0.21"
0.577 ± 0.019"





31.56 ± 1.25"


* All ± are S.D., unless specified otherwise.
b Standard error of mean values.
2/1/98
A-3
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE
ATMOSPHERES"
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/m3)
Particle size (^m)
MMD" (GSD)C
C02(%)
CO (ppm)
NO, (ppm)
NO (ppm)
NO, (ppm)
SO, (ppm)
SO/2 (/ig/m3)
0, (%)
Ozone (ppm)
Aliphatic aldehydes
Formaldehyde (ppm)
Acrolein (ppm)
NH/
Hydrocarbons (ppm)
PAHs
Benzo(a)pyrene
Nitropyrene
U.S. Environmental Protection Agency
Pepelko, 1982a
Nissan, 6 cylinder, 3.24 L
California cycle, modified
No. 2 diesel

20 h/d, 7 d/week, 4 weeks
Control


















Exhaust
6.40 ± 0.36

0.247 ±
0.003
16.9 ± 1.1
2.49 ± 0.18
5.71 ± 0.21

2.10 ±0.21
577 ± 19






31.6 + 3.8


Exhaust -
irradiated
6.75 ± 0.39

0.244 ±
0.007 '
16.1 ± 1.3
2.76 ± 0.15
4.53 ± 0.15

1.86 ± 0.21
569 ± 19






26.1 ± 3.4


Leeetal., 1978, 1980
3.24 L, 6 cylinder
California cycle, modified
No. 2 diesel

20 h/d, 9 weeks
Control


0.040
2.0
0.07
0.11


0.0






2.0


Exhaust
6.32

0.252
15.7
2.19
5.85

2.13
0.57






15.6


Exhaust -
irradiated
6.83

0.255
15.4
2.73
4.94

1.91
0.57

<0.01




15.0


   a All ± are standard errors of weekly means.
   b Mass median diameter.
   c Geometric standard deviation.
2/1/98
A-4
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE ATMOSPHERES3
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fue! sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/nf)
Respirable particles1'
(mg/m3)
Particle size (tan)
MMD" (GSD)"
CO, (%)
CO (ppm)
NO, (ppm)
NO (ppm)
SO, (ppm)
SO;2 (ftg/m3)
Aliphat. aldehydes (ppm)
Formaldehyde (ppm)
Acetaldehyde (ppm)
Acrolein (ppm)
NH, (ppm) '
NH4+ (ppm) •
THC (ppm)
PAH (wB/m3)
Benzo(a)pyrene
Benzo(a)anthracene
Benzo(k)fluoranthene
Fluoranthene
Pyrene '
National Institute for Occupational Safety and Health '
Castranovaetai., 1985; FedanetaL, 1985; Hahonetal., 1985; Lewis etal, 1986, 1989;
Mentnech et aL, 1984; Vallyathan et al,, 1986
Caterpillar 3304, 7 L, 4-cycIe
8-mode mining cycle, 60% idling
No. 2 diesel
0.34%
7 h/d, 5 d/week, 104 weeks .
" Control



0.08 ± 0,02
2,2 ± 0.9
0.06 ± 0.04
0.08 ± 0.14


0.02 ± 0.01
0.0076 ± 0.0035
0.0015 ± 0.0035
0.0030 + 0.0033
6.52 ± 0.28

4,1 ± 1.9





Exhaust

1.95 ± 0,25
0.23 (± 2.5)e
0.36 (±2.0)'
0.20 ± 0.06
11.5 ± 3.1
1.5 ± 0.5
8.7 + 3.6
0.81 ± 0,38
29.0 ± 24,9
0.12 ± 0.06
0.0383 ± 0.0230
0.0387 ± 0.0153
0.0602 ± 0.0245
0.64 + 0.71
0,027 ± 0.0307
7,5 ± 2.2 (cold)
13.5 ± 6.8
19.6 ± 9.9
5.6 ±2.3
139.3 ±98,1
123.4 + 72.2
Coal dust
4.98 ± 0.82 .
2.00 ± 0,41

0.09 ± 0,05
2.2 ± 0.9
0.06 ± 0.05
0,08 ± 0.29
0.01 ± 0.07
16.8 ± 17,9
0.02 ± 0.01
0.0074 ± 0.0041
0.0009 ± 0.0025
0.0062 ± 0.0047
0.57 ± 0.52
0.0065 ± 0.0143


3.2 ± 2.2

26.5 ± 11.5'
32.3 ± 15.1
Exhaust + coal dust
3.23 ± 0.60
2.02 ± 0.30

0.20 ± 0.07
10.9 ± 2.8
1.6 ±-0.5
8.3 ± 3.2
0,61 ± 0.29
42.3 ± 33.8
0.12 ±0.05
0.0374 ± 0.0266
0.0377 ± 0.014
0.0578 ± 0.0205
0.48 ± 0.55
0.0165 ± 0.0233
7.4 ± 2.0 (cold)
10.2 ± 6.5
11.2 ± 5.2
3.6 ± 2.4
67.5 ± 52.4
60.0 ± 36.6
    1 All ± are S.D., unless specified otherwise.
    c Mass median diameter.
    d Geometric standard deviation,
    c Electrical aerosol size analyzer.
    'Scanning electron microscope.
2/1/98
A-5
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE ATMOSPHERES0
Facility /Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulta
Exposure regime
Exposure conditions
Particle cone, (mg/m1)
Respirable particles1'
(mg/m3)
Particle size (^m)
MMDC (GSD)4
CO, (%)
CO (ppm)
NO2 (ppm)
NO (ppm)
SO, .(ppm)
SO4'2 (rt/m3)
Aliphatic aldehydes
. Formaldehyde (ppm)
Acetaidehyde (ppm)
Acrolein (ppm)
NH, (ppm)
NH4* (ppm)
THC (ppm)
PAH ffjg/m5)
Benzo(a)pyrene
Nitropyrene
National Institute for Occupational Safety and Health
Green et al., 1983; Rabovsky et al,, 1986
Caterpillar, 7 L, 4 cylinder, 4-cycle
8-mode mining cycle, 60% idling
No. 2 diesel
< 0,5% '
7 h/d, 5 d/week, 12 mo.
Control



0.08
2.3
0.04
0.07
0,01





0.63




' Exhaust
2
2.01

0.21
12.7
1.6
9.7
0.83





1.13




Coal dust
5
1.97

0.09
2.4
0.04
0.08






0.83




Exhaust +
coal dust
3
2.08

0.20
11.1
1.3
1.3
0.56





0.54




Rabovsky etal., 1984
Caterpillar, 7 L, 4 cylinder, 4-cycle
8-mode mining cycle, 60% idling
No. 2 diesel

7 h/d, 5 d/week, 24 mo.
Control



0.07 ±
0.02
2.0 ±
0.9
0.06 ±
0.04
0.08 ±
0.13






0.5 ±
0.6




Exhaust

1.9 ± 0.3

0.16 +
0.04
10.5 ±
2.3
1.5 ± 0.5
7.8 ± 3.1
0.6 + 0.4





0.6 ± 0.8




Coal dust

2.1 ± 0.4

0.08 ±
0.04
2.1 ± 0,8
0.07 ±
0.05
0.08 ±
0.28
0.003 ±
0.05





0.6 ± 0.7




Exhaust +
coal dust

2.0 + 0.3

0.17 ±0.06
10.3 + 2.0
1.5 ± 0.05
7.6 + 2.8
0.5 ± 0.3





0.4 ± 0.3




a All ± are S.D. unless specified otherwise.
" < Turn.
* Mass median diameter
d Geometric standard deviation.
2/1/98
A-6
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE
ATMOSPHERES"
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/m3)
Particle size (jan)
MMD" (GSD)
CO, (%)
CO (mg/m3)
NO, (ppm)
NO (ppm)
NO, (mg/m3)
Sulfur (mg/m3)
SO, (ppm)
Aliphatic aldehydes
Formaldehyde (ppm)
Acrolein (pprn)
NH4+
THC (ppm)
PAHs
Benzo(a)pyrene
Nitropyrene
General Motors Research Lab
Barnhart et al., 1981, 1982; Chaudhari et al., 1980,
1981; Chaudhari and Dutta, 1982; Chen and Vostal,
1981; Dziedzic, 1981; Eskelson et al., 1981; Penney et
al., 1981: Misiorowski et al., 1980, 1981; Navarro et al.,
1981; Schneider and Felt, 1981; Schrecketal., 1980,
1981; Strom, 1984; Vostal et al., 1981; Wallace et al.,
1987; White and Garg, 1981
1978 350D Oldsmobile, 5.7 L, 4-cycle
1350 rpm, 96 N-m
Amoco type 2D
0.27%
20 h/d, S'/a d/week, 104 weeks
Control
0.007 ±
0.009


2.2 ±0.6


0.05









Exhaust
. 0.258 ±
0.087
Exhaust
0.796 ±
0.228
Exhaust
1.533 ± '
0.346
0.19

3.4 ± 0.8


2.1 ±0.6










5.3 ±0.9


5.0 + 1.2










7.9 ± 2.1


9.2 ± 1.6









Gross, 1981
5.7 L
1350 rpm, 96 N-m
Amoco type 2D
• 0.27%
20 h/d, 51/2 d/week, 87 weeks
Control
0.007
± 0.009.


1.9


<0.04









Exhaust
1.533
± 0.346
0.2

7
0.5
6.7
7.2
1.4








 All + are S.D., unless specified otherwise.
 Mass median diameter.
2/1/98
A-7
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE
ATMOSPHERES"
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/m3)
Particle size (/im)
MMD" (GSD)C
C02 (%)
CO (ppm)
NO, (ppm)
NO (ppm)
SO, (ppm)
SO/2 Gig/in3)
Aliphatic aldehydes
(ppm)
Formaldehyde (ppm)
Acrolein (ppm)
Ammonia (ppm)'
Hydrocarbons (ppm)
PAHs
Benzo(a)pyrene
Nitropyrene
Inhalation Toxicology Research Institute
Bice et al., 1985; Cheng et al., 1984; Henderson et al., 1983, 1985, 1988; Mauderly et al.,
1983, 1984, 1987a, b, 1988; McCIellan et al., 1986; Wolf et al., 1987
1980 Oldsmobile V8, 5.7 L
Federal Test Procedure, urban driving cycle
Phillips No. 2 diesel
0.34%
7 h/d, 5 d/week, 130 weeks . .
Control
0.013 ± 0.006

0.2005 ±
0.0390
1.0 ±0.7
0
0





1.1 ±3.0
2.6 ± 0.6


Exhaust
0.353 ± 0.071
0.183 ± 0.04(4.8 ±
0.28)"
0.262 ± 0.06 (4.2 ±
0.24)e
0.2284 ± 0.0371
2.9 ± 1.0
0.05 ± 0.09
0.7 ±0.3





1.4 ± 1.3
3.8 + 0.9


Exhaust
3.469 ± 0.447
0.184 ±0.02(5.3 ±
0.64)"
0.249 ± 0.03 (4.5 ±
0.54)e
0.4355 ± 0.0590
16.5 ±7.1
0.34 ± 0.22
5.7 + 1.5





0.9 ± 0.9
8.7 ± 5.2


Exhaust
7.082 ± 0.808
0.213 ± 0.06 (4.7 ±
0.94)d 0.234 ± 0.06
(4.4 + 0.88)e
0.6643 ± 0.1320
29.7 ± 12.9
0.68 ± 0.48
10.0 + 2.6





0.7 ±0.6
13.4 + 8.3


a All ± are S.D. unless specified otherwise; data for particles through 30 mo.; data for gases from 35th week through 30 mo.
b Mass median diameter.
c Geometric standard deviation.
11 Lovelace multiple jet impactor, mass median aerodynamic diameter.
" Impactor/parallel flow diffusion battery, mass median diameter.
2/1/98
A-8
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE ATMOSPHERES"
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuei sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/m3)
Particle size Om)
MMD" (GSD)11
CO, (%)
CO (ppm)
NO, (ppm)
NO (ppm)
NO, (ppm)
SO, (ppm)
SO/2 (fig/m3)
0, (%)
Ozone (ppb)
Aliphatic aldehydes
Formaldehyde (ppm)
Acrolein (ppm)
Ammonia
Hydrocarbons (ppm)
HTHC (ppm)
PAHs
Benzo(a)pyrene
Nitropyrene
Inhalation Toxicology Research Institute
Inhalation Toxicology Research Institute
- Annual Report, 1980
1980 GM, 5.7 L
California 7-mode urban cycle
Phillips No. 2 diesel

7 h/d, 5 d/week, 12 weeks
Control
0,039
± 0.020


1.1 ±
0.6










2.8 ± -
0.7




Exhaust
0.230
± 0.073


1-5 ±
0.6










3.2 ±-
0.8




Exhaust
1.030
±0.340


3.7 ±
1.1










2.9 ±
0.9




Exhaust
4.260
± 1.110

0.2080 ±
0.04
11.5 ± 2.6
0.4 ±0.4
0.80 ±
0.25




14.6 ± 3.1



2.5 ± 0.7
4.0 ± 0.8



Mauderly et al.,
1981b
1980 GM, 5.7 L
California 7-mode urban cycle
Phillips No. 2 diesel

7 h/d, 5 d/week, 19 weeks
Control
0.050
± 0.024


















Exhaust
0.210
± 0.070


















Exhaust
1.020
± 0.350


















Exhaust
4,380
+ 1.160















•


* All ± are S.D. unless specified otherwise.
h Concentrations of gaseous components reported to be proportional to these in 12-week study.
c Mass median diameter.
d Geometric standard deviation.
2/1/98
A-9
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE ATMOSPHERES"
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure
conditions
Particle cone.
(mg/m3)
Particle size (fim)
MMD" (GSD)°
CO, (%)
CO (ppm)
NO, (ppm)
NO (ppm)
NO, (ppm)
SO, (ppm)
SO;2 (Mg/m3)
0, (%)
Aliphatic
aldehydes
Formaldehyde
(ppm)
Acrolein (ppm)
NH/
LTHC (pprn)
HTHC (ppm)
PAHs (ne/ma)
Benzo(a)pyrene
BenzQ(k)fluoranthene
Benzo(ghi)perylene
1-Nitropyrene
Japan Automobile Research Institute Inc. (Health Effects Research Program - HERP)
HERP 1988; Ishinishi et al. , 1986; Ishinishi et al. . 1989
Light duty, 1 .8 L, 4-cylinder, swirl chamber
1700 rpm, eddy current dynamometer
Nippon Oil Co JIS No. 1 or 2 diesel
0.41%
16 h/d, 6 d/week, 30 mo.
Control
0.003

0.026
0.80
0.011
0.033
0.044
0.06
0.41
20.8

0.002


2.17
2.20




Exhaust
0.11

0.050
1.23
0.08
1.16
1.24
0.38
18.8 . .
20.8

0.01


2.27
2.44




Exhaust
0.41

0.105
2.12'
0.26
3,81
4.06
1.06
62.4
20.7

0.03


2.51
2.93




Exhaust
1.08
0.19
(2.37-
2.71)
0.219
3.96
0.70
9.44
10.14
2.42 • '
151
20.5

0.07


2.87
3.82




Exhaust
2.32
0.21-0.22
(2.23-2.93)
0.418
7.10
1.41
18.93
20.34
4.70
315
20.3

0.13


3.57
5.49
5.3 + 10.6
5.4 ± 7.7
2.7 ± 3.9
46.6 ±
44.0
Heavy duty, 11 L, 6-cylinder, direct injection
1200 rpm, eddy current dynamometer
Nippon Oil Co JIS No. 1 or 2 diesel
0.41%
16 h/d, 6 d/week, 30 mo.
Control
0.002

0.035
0.63
0.021
0.042
0.061
0.06
0.49
20.8

0.003


3.50
2.43




Exhaust
0.46

0.084
2.65
0.46 5
5,71
6.17
0.98
62.9
20.8

0.05


4.27
4.63




Exhaust
0.96

0.140
4.85
1.02
12.11
13.13
1.79
111
20.7

0.11


5.16
7.15
-



Exhaust
1.84
0.20-0.23
(2.73-3.07)
0.215
7.75
1.68
19.99
21.67
2.82
198
20.6

0.18


5.90
9.94




Exhaust
3.72
0.25-0.28
(2.75-3.18)
0.360
12.91
3.00
34.45
37.45
4.57
361
20.4

0.29


7.62
15.65
7.5 ± 3.2
6.0 ± 3.0
8.9 ±2.5
43.4 ±
9.8
"All ± are S.D., unless specified otherwise.
h Mass median diameter.
c Geometric standard deviation.
   2/1/98
A-10
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE
ATMOSPHERES'
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/m3)
Particle size (^rn)
MMD" (GSDf
CO, (%)
CO (ppm)
N02 (ppm)
NO (ppm)
NOX (ppm)
SO, (ppm)
SO/2 fe/m3)
O, (%) .
Aliphatic aldehydes
Formaldehyde (ppm)
Acrolein (ppm)
NH/
LTHC (ppm)
HTHC (ppm)
PAHs
Benzo(a)pvrene
Nitropyrene
Japan Automobile Research Institute Inc. (Health Effects Research Program - HERP)
HERP, 1988; Ishinisbi et al., 1986; Ishinishi et al., 1989
Heavy duty, 1 1 L, 6-cylinder, direct injection
1200 rpm, eddy current dynamometer
Nippon Oil Co JIS No. 1 or 2
•0.41%
16h/d, 6d/week, 30 mo.
Control
0.004

0.068
0.06
0.024
0.040
0.062
0.03
0.35
20.8

0.003


3.62
2.38


Exhaust, filtered
0.005

0.083
2.54
0.42
5.16
5.58
0.96
1,43
20.7

0.04


4.43
3.74


Exhaust
0,39

0.084
2.50
0.44
5,37
5.81
0.98
57.7
20.7

0.04


4.41
4.53


Exhaust, filtered
0,019

0.391
13,00
3.96
32.81
36.76
4.50
1.61
20.4

0.24


7.79
12.68


Exhaust
2.99
0.31-0.35
(2.58-2.83)
0,412
12.90
4.95
31.50
36.45
4.03
358
20.3

0.20


7.68
13.79


* Ail + are S.D. unless specified otherwise,
,b Mass median diameter.
c Geometric standard deviation.
2/1/98
A-ll
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE ATMOSPHERES'
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/nf)
Particle size (pm) MMD"
CO, (%)
CO (ppm)
NO, (ppm)
NO (ppm)
NOX (ppm)
SO, (ppm)
SO/2 (Mg/m3)
O2'(vol%)
Aliphatic aldehydes
Formaldehyde (ppm)
Acrolein (ppm)
NH4*
THC (ppm)
CH4 (ppm)
PAHs fue/e oarO:
Benzo(a)pyrene
Benzo(e)pyrene
Benz(a)anthracene
Fluoranthene
Pyrene
Benzo(a)fluoranthene
Benzo(b)fluoranthene
Benzo(ghi)perylene
Chrysene
Fraunhofer Institut fur Toxikologie und Aerosolforschung
Heinrich et al. 1982
2.4 L
Constant load of 16kW, 2400 rpm
European reference fuel
0.36%
7-8 h/d, 5 d/week, 104 weeks
Exhaust
3.9 ± 0.5
0.1
0.54 ± 0.15
18.5 ± 4.9
1.2 ± 1.7
16.5 + 5.8
18.6 ± 5.8
3.1 + 1.8

19.5 ± 0.6




9.3 ± 4.6
3.0 + 2.2
7.0
14.1
9.8
134.6
65.8
5.4
5.3
21.4
25.7
Exhaust, filtered


0.52 ± 0.13
18.0 ± 4.4
1.0 ± 1.5
17.2 ± 5.9
19.2 ± 6.1
2.8 ± 1.7

20.0 ± 0.7




7.9 ± 3.3
2.6 ± 1.8









Heinrich etal., 1986; Stober, 1986
1.6L
FTP (1972)
European reference fuel
0.36%
19 h/d, 6 d/week, 120-140 weeks
Control


0.10 ± 0.01
0.16 ± 0.27
_
. . .
.
.






3.5 ± 0.29
2.3 ± 0.17






•


Exhaust
4.24 ± 1.42
0.35 + 0.10
0.38 ± 0.05
12.5 ±2. 18
1.5 ±0.33
10.0 ± 2.09
11. 4 ±2.09
1.12 ± 0.89






5.5 ± 0.69
2.6 ± 0.19
3 (13 ng/m3)
- (21 ng/m3)



- (51 ng/m3)


Exhaust, filtered


0.35 ± 0.05
11.1 + 1.92
1.2 ± 0.26
8.7 ± 1.84
9.9 ± 1,80
1.02 ± 0.62






5.2 ± 0.65
2.4 ± 0.20



-





a All ± are S.D. unless specified otherwise.
h Mass median diameter.
2/1/98
A-12
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE ATMOSPHERES'
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/rf)
Particle size (p,aif
CO, (%)
CO (ppm)
NO, (ppm)
NO (ppm)
NO, (ppm)
SO, (ppm)
SO;2 (ttg/m3)
O, (vol%)
Aliphatic aldehydes
Formaldehyde (ppm)
Acrolein (ppm)
NH/
THC (ppm)
CH4 (ppm)
PAHs
Benzo(a)pyrene
Nitropyrene
Fraunhofer Institut fur Toxikologie und Aerosolforschung
Heinrichetal., 1979; Meissetal., 1981 . .'
2.4 L .
Constant load of 16 kW, 2400 rpm
European reference fuel
0.36%
7-8 h/d, 5 d/week, 5 mo.
Control


0.1
<1



<1






6



Exhaust
4
0.1
0.5
11
0.6
25
26
3'






8
5


Exhaust,
filtered


0.5
11
0.5
22
23
4






8
5


Exhaust •
11
0.1
0.9
25
1.5
43
45
8






11
5


Exhaust,
filtered


0.95
27
1.3
43 .
44
8






12
5 •


Exhaust
17
0.1
1.4
42
2.6
75
78
13






13
5


Exhaust,
filtered


1.6
45
2.7 ' .
68
71
12 '






13
5


" Values estimated from graphically depicted data.
b Aerodynamic diameter of the modal peak of the particle mass distribution.
2/1/98
A-13
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE
ATMOSPHERES'
Facility/Sponsor
Reference
Engine _type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/m3)
Particle size (pm)
C02 (%)
CO (ppm)
NO2 (ppm)
NO (ppm)
NO, (ppm)
SO, (ppm)
SO/2 (Mg/m3)
0, (%)
Aliphatic aldehydes
Formaldehyde (ppm)
Acrolein (ppm)
NH4+
Hydrocarbons (ppm)
PAHs
Benzo(a)pyrene
Nitropyrene
Southwest Research Institute
Kaplan etal., 1983; White eta!,, 1983
5.7 L
Steady state, 1347 rpm, equivalent to constant 40 mph
Emissions 2D
0.23-0.24%
.20 h/d, 7 d/week, 65 weeks
Control
0.01 ± 0.009

0.0649 ±
6.0020
5,81 ± 0.2

0
0.05 ± 0.0







3.43 ± 0.2


Exhaust
0.242 ±
0.049
88-93% <
1.0
79-85% <
0.5
0.0781 ±
0.0028
6.39 ± 0.3

0.56
0.65 ± 0.1







3.76 + 0.3


Exhaust
0.735 ± 0.084
88-94% <1,0 '
76-84% <0.5
0.1026 ±
0.0043
7.43 ± 0.3

1.69
1.85 ±0.2







4.31 ± 0.3


Exhaust
1.500 ±
0.136
91-94% <1.0
81-85% <0.5
0.1355 ±
0.0062
9.40 ± 0.5

3.42
3.73 ± 0.4







4.99 ± 0.3


Kaplan et al.,
1982
5.7 L
Steady state,
40 mph


20 h/d, 7
d/week,
12-13 weeks
Exhaust
1.500
















   " All ± are S.D. unless specified otherwise.
2/1/98
A-14     DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE ATMOSPHERES'
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/nf)
Particle size (pm)
MMD" (GSD)C
C02 (%)
CO (ppm)
NO, (ppm)
NO (ppm)
NO, (ppm)
SO2 (ppm)
SO,,4 (Mg/m3)
0, (%)
Aliphatic aldehydes
Formaldehyde (ppm)
Acrolein (ppm) ,
NH/
Hydrocarbons (ppm)
PAHs
Benzo(a)pyrene
Nitropyrene
Battelle-Geneva Research Center
Brightwell et al., 1986; Bernstein et al., 1984
1.5L
FTP - 1972


16 h/d, 5 d/week, 104 weeks
Control



1 ±3


0.1 ±
0.1










Exhaust
0,7
















Exhaust
2.2
















Exhaust
6.6

0.46 ±
0.03'
32 ± 11

5 J ± 2.0*
8± 1







18.9 + 4.1"


Exhaust,
filtered


0.47 ±
0.03*
. 32 ± 11

6.0 ± 2.0"
8 ±2







18:8 ± 4.1"


Japan Anti-Tuberculosis Association
Iwaietal,, 1986
2.37L
Steady state, 1000 rpm


8 h/d, 7 d/week, 96 weeks
Control

















Exhaust,
filtered



7,0 + I A*
1.8 ± 1.8*

30.9 ± 10.9*
13.1 ± 3.6*









Exhaust
4.9 ± 1.6


7.0 ± 1.4"
1.8 ± 1.8"

30.9 ±
10,9"
13.1 ± 3.6"









* All ± are S.D. unless specified otherwise.
h Mass median diameter.
c Geometric standard deviation.
d Samples from dilution tunnel, exposure chamber reported to have approximately the same concentrations.
" Data from first year of study (Bernstein et al., 1984).
2/1/98
A-15
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE
ATMOSPHERES3
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/m3)
Resp. particles (mg/m3)
Particle size (/xm)
MMD" (GSD)C
CO, (%)
CO (ppm)
NO, (ppm)
NO (ppm)
NO, (ppm)
SO, (ppm)
SO/2 (Aig/m3)
0, (%)
Aliphatic aldehydes (ppm)
Formaldehyde (ppm)
Acrolein (ppm)
Ammonia (ppm)
Hydrocarbons (ppm)
PAHs
Benzo(a)pyrene
Nitropyrene
Battelle, Pacific Northwest Laboratory
Karagianesetal., 1981
43 bhp, 3 cylinder
Simulated mining cycle
Equivalent to VV-F-800 A grade DF-2
_
6 h/d, 5 d/week, 87 weeks
Control
_

















Exhaust
8.3 +2.0
95% respirable
0.71 (2.3)

50 ± 3
4-6


<1


<1


26-40



Exhaust + coal
dust
13.5 ± 4.0




































   a All ± are S.D. unless specified otherwise.
   h Mass median diameter.
   c Geometric standard deviation.
2/1/98
A-16
DRAFT-DO NOT CITE OR QUOTE

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE ATMOSPHERES"
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone, (mg/rrf)
Particle size (ftm)
CO, (%)
CO (ppm)
NO, (ppm)
NO (ppm)
NOX (ppm)
SO, (ppm)
SO;2 (pig/m3)
0, (%)
Aliphatic aldehydes
Formaldehyde (ppm)
Acetaldehyde
Acrolein (ppm)
NH/
Hydrocarbons (ppm)
Benzene (ppm)
Toluene (ppm)
PAHs (^e/m3):
Benzo(a)pyrene
Nitropyrene
University of Pittsburgh
Battigelli, 1965
7 hp, four cycle, single cylinder



15-60 min
Dilution A


0.1
<20
1.3


0.2

20.5
<1.0
<0.1

<0.05

<2.0




Dilution B


0.9 .
30
2.8,-


0,5

20.0

-------
APPENDIX A. EXPERIMENTAL PROTOCOL AND COMPOSITION OF EXPOSURE ATMOSPHERES'
Facility/Sponsor
Reference
Engine type
Operating mode
Fuel type
Fuel sulfur
Exposure regime
Exposure conditions
Particle cone.
(mg/m3)
Particle size (#tm)
CO, (%)
CO (ppm)
N02 (ppm)
NO (ppm)
NO, (ppm)
SO, (ppm)
H,S04 (ppm)
Oxidants
(ppm as O3)
Aliphatic aldehydes
Formaldehyde
(ppm)
Acrolein (ppm)
NH4+
Hydrocarbons
(ppm as CH4)
PAHs
Benzo(a)pyrene .
Nitropyrene
U.S. Environmental Protection Agency
Gillespie, 1980; Hyde et al., 1980; Malanchuk, 1980; Orthoefer, 1980; Stara et al., 1980
Automobile gasoline engine
Urban cycle


16 h/d, 7 d/week, 68 mo.
Control



4.9
0.04
0.04

0.03
-
0.02




2.7


Non-irradiated
gasoline exhaust
(R)



97.5 ± 10.0
0.05 ± 0.02
1.45 ± 0.42

-
-
-




27.5 ± 4.4


Irradiated
gasoline
exhaust (I)



94.5 ± 19.6
0.94 ± 0.36
0.19 ± 0.29

-
-
0.20 ± 0.09




23.9 ±6.1


SO2 +
H2SO4



-
-
-

0.42 ±
0.22
0.02 ±
0.01
-




-


R +
SO2 +
H2SO4



98.4 ±
13.8
0.05 ±
0.03
1.51 ±
0.44

0.48 ±
0.23
0.02 ±
0.01
-




27.4 +
4.3


I +
S02 +
H2S04



-
0.89 ±
0.36
0.19 ±
0.29

0.42 ±
0.21
0.03 ±
0.01
0.20 ±
0.08




23.9 ±
6.0


Nitrogen
oxides



-
0.64 ±
0.12
0.25 ±
0.06

-
-
-




-


Nitrogen
oxides



-
' 0.15 ±
0.33
1.67 ±
0.21

-
-
-




-


'All ± are S.D. unless specified otherwise.
2/1/98
A-18
DRAFT-DO NOT CITE OR QUOTE

-------
                       Appendix B

           Alternative Model for Diesel Cancer
                     Risk Assessment
2/1/98                       B-l     DRAFT-DO NOT CITE OR QUOTE

-------
 1      B.I.  INTRODUCTION
 2             As previously discussed in Chapter 11, the most appropriate method to assess cancer risk
 3      of diesel exhaust is to take into account effects of both particles (carbon core) and organics
 4      because evidence exists that both agents are involved in carcinogenic process. The reasons for
 5      this conclusion are based on the following observations:  (1) organics include a variety of
 6      polycyclic aromatic hydrocarbons (PAHs) and nitroaromatics, many of which are known to be
 7      carcinogenic; (2) the results of recent studies on inert particles and carbon black in rats strongly
 8      support the hypothesis that the carbon core of the diesel particle may be the primary component
 9      responsible for the induction of lung cancer; (3) PAHs are unlikely responsible for all observed
10      tumors because they account for less than 0.1  ng/mg particulate matter (Tong and Karasek,
11      1984); and (4) the observation of disproportionate high tumor incidence in high exposure animals
12      coincides with disproportionate increase of cumulative lung burden of diesel particle as exposure
13      concentration increases,
14             A workshop on Research Needs for Risk Assessment  of Inhaled Particulate Matter was
15      organized and sponsored by the U.S. Environmental Protection Agency (EPA) in March, 1992.
16      The purpose of the workshop was to determine the extent of information that can be used for
17      quantitative risk assessment and to discuss  mechanisms of particle-induced lung tumors to serve
18      as a guidance for future research needs. Two major, among several other, conclusions that are
19      relevant to quantitative risk assessment were reached by the Workshop:
20
21             (1)  • particle overloading of the lung tissue may induce both initiation (by PAH specific
22                  adducts and adducts through oxygen radicals) and cell proliferation steps in tumor
23                  formation, and
24
25             (2)   more research is needed to improve the risk assessment of particle-induced lung
26                  cancers.
27
28             Although there are not enough data available to construct a biologically based
29      dose-response model, it is desirable to investigate implications of the hypothetical mechanisms
30     • proposed by the workshop. The purpose of the alternative modeling presented in this report is to
31      do just that.  Briefly, the biological issues and their implications to quantitative risk assessment
32      that we would like to consider are the following.
33
34      1. Particles deposited in lung are phagocytized by alveolar macrophages.  Because the
35         phagocytizing macrophages in animals from high-dose group may be more likely to be
36         activated to release mediators including reactive oxygen species, cytokines, and growth
37         factors, it is of interest to determine whether or not the available tumor response data are
        2/1/98   •  ,                               B-2       DRAFT-DO NOT CITE OR QUOTE

-------
  1          consistent with the hypothesis that the particle burden affects both initiation and proliferation
  2         in carcinogenic process.
  3
  4     2,  Organic materials can also induce specific adducts which may contribute to cell initiation.
  5         However, given its low content, the contribution of organics to tumor induction may be very
  6         small. Can a dose-response model that is consistent to the proposed biological concept be
  7         constructed with both organics and particles as dose metrics?
  8
  9     3.  If a model that has the above biological interpretation and is consistent with the bioassay data
10         can be constructed, what would be its implications on quantitative risk assessment of diesel
11          exhaust emissions, and how would its results compare with those predicted by the linearized
12         multistage (LMS) model?
13
14     B.2. PRELIMINARY CONSIDERATIONS
15            In order to evaluate the impact of various biological assumptions on diesel risk
16     assessment, it is necessary to construct a mathematical dose-response model that takes into
17     account the biological mechanisms discussed in the EPA workshop. Because an issue of
18     significant importance in diesel risk assessment is  the effect of lung overloading on tumor
19     induction, the model should possess the fol/lowing properties.
20
21      1.  It should depend on both types of dose metrics: organics, and carbon core.  It should allow
22         one to account for the contribution of organics and carbon core individually and/or jointly to
23         tumor induction/formation.
24
25     2.  It should allow for the possibility that model parameters can change with time because of the
26         increasing lung burden during exposure.
27
28     3.  The cell proliferation and tumor induction/formation should be stochastic in nature; it is not
29         realistic to assume deterministic clonal growth. For instance, it should not be required to
30         assume that all cells divide at the same age.
31
32            To accomplish these goals, we assume that a normal cell can be initiated by both organics
33     and carbon core.  Denote the initiation rate by u^, which is a function of background and diesel-
34     induced rates (as specified below). Because an initiated cell (I-cell) eventually either goes into
35     cell death, or enters the cell cycle (including cells in quiescence, G0), it is reasonable to assume
36     that the cell lifetime for an I-cell follows certain probability distribution. Under this model, a cell
37     in G0 phase is equivalent to the case where it has a very long lifetime with certain probability
38     (i.e., in the right-hand tail of the cell lifetime distribution). At the end of an I-cell's lifetime, it
39     either dies (death) with probability P, divides into two daughter  cells (birth) with probability a,
40     or divides into one I-cell and one  malignant cell (second transition) with probability [i2; a + P +
41      [i2 = 1 •  Instead of assuming that a single malignant cell is equivalent to a tumor as in the MVK

        2/1/98    -  ,                               B-3       DRAFT-DO NOT CITE OR QUOTE

-------
 1      model proposed by Mooigavarkar and colleagues (1979, 1981), we assume that a malignant cell
 2      has.a certain probability to become a tumor; this probability is assumed dose-dependent, thus
 3      allowing for an evaluation of dose effect on tumor progression. It should be noted that the
 4      proposed model does not exclude the possibility that it may take more than one step (for a normal
 5      cell) to become "initiated." The rate of initiation used in the model should be viewed as a net
 6      rate which represents several genetic alterations and repairs.
 7
 8      B.3.  MATHEMATICAL MODEL AND PARAMETERS ESTIMATION
 9            We shall proceed to construct a dose-response function P(t;d,D), probability of cancer by
10      time (age) t, which depends on both organic, d, and particle (carbon core), D, and incorporates
11      the biological concept discussed previously.  Because the model parameters that are not directly
12      observed in laboratory can only be statistically estimated from high concentration cancer
13      bioassay data, the model constructed should not be considered a valid model of diesel-induced
14      carcinogenesis; uncertainty about the low-dose extrapolation still remains. Some discussions
15      about the need for further laboratory measurements will be given later.
16            The model with the desirable features discussed above falls into one of several classes of
17      stochastic models that have been developed by EPA's Office of Health and Environmental
18      Assessment (OHEA): namely, the stochastic model which was originally proposed by Chen and
19      Farland (1991) and extended into one with time varying parameters by Tan and Chen (1992).
20      This model will be used as the basis for constructing a biologically based dose-response model.
21      A brief mathematical description is presented in Appendix B-2.
22            The time to event data from Mauderly et al. (1987) are used to estimate model
23      parameters. The data from Mauderly et al. are useful because they contain information on natural
24      mortality and serial sacrifice of animals with or without (malignant) tumors, valuable information
25      for estimating tumor latency.  To utilize the information from serial sacrifice in Mauderly et al.
26      an (E-M) algorithm is derived (see Appendix B-l) and used to calculate maximum likelihood
27      estimates of parameters.
28                                                                              _          .
29      B.3.1. Model Parameters and Notations
30            The following parameters are incorporated in the dose-response model, which includes
31      initiation rate (u^), proliferation rate (ya), conversion rate (7^2)'an<^ probability of progression
32      (q). The death rate for the initiated cells is implicitly defined by y(l -  ^2 ~ °0- These
33      parameters are all dose dependent.
34
35      D:   Dose of carbon core, mg/cm of lung epithelial surface; D varies over time
36

        2/1/98                                    B-4       DRAFT-DO NOT CITE OR QUOTE

-------
  1      d:   Dose of organics, mg/cm of lung epithelial surface
  2
  3     (i j:  Dose-related initiation rate (per cell per day) that depends on u-0 (background rate), d, and
  4  •        D by Uj =-|j.0(l + ad + bD); a and b are paramaters to be estimated.
  5             .   •
  6     u.2:  Probability of producing a malignant cell at the end of an initiated cell (I-cell) lifetime
  7
  8     a:   The probability that an I-cell divides into two daughter cells at the end of its lifetime
  9
10     q:   Probability that a single malignant cell will develop into a malignant tumor
11
12     Y:   1/Y is the mean I-cell lifetime in days; a cell lifetime ends if it either goes into mitosis, or
13          cell death.  Note  that if one assumes that the probability for a cell to get into mitosis is
14          about the same as cell death then the mean cell lifetime can be conveniently interpreted as
1 5          time to mitosis (i.e., cell turnover time); thus, shorter cell lifetime implies more frequent
16          cell division. Note that the time to mitosis is a random variable here, not a fixed constant
17          as in the assumption made in the Greenfield et al. (1984)  model that has been used
18          extensively by Cohen and Ellwein( 1988) to analyze experimental bladder cancer.
19
20     N:   Number of (normal) target cells
21
22     B.3.2.  Practical Considerations
23            By statistical theory alone the E-M algorithm developed in this report provides an elegant
24     procedure which can be used to test hypotheses whether a particular parameter is influenced by
2 5     organics and carbon core individually or both together. For instance, one could postulate that the
26     parameter y (reciprocal of which represents mean cell lifetime) is given by y(d5Dj) = y0 + ytld +
27     y 12Di5  and then proceed to test a null hypothesis that y 11 = 0, no effect of organics on cell
28     lifetime.  This temptation, however, must be resisted because there would be too many
29     parameters that must be estimated if such statistical tests are to be performed. Therefore, rather
30     than performing such a statistical exercise, we proceed with a biologically plausible assumption
31      that parameters q and y depend only on lung burden of carbon core, C.
32            The duration of the Mauderly et al. study was about 940 days. To construct a dose-
33     response model with time-dependent lung burden, the time interval (0,940] is divided into
34     five subintervals; each subinterval spans 6 mo except for the last subinterval, which spans from
35     730 (2  years) to 940 days. Corresponding to an ambient air concentration of diesel emissions in
36     mg/m  , the deposition-retention model developed by Yu et al. is used to calculate dosimetric (d,
37     Dj), 1=1,2,..., 5, where organics dose, d, is not changing with time because it reaches steady
38     state quickly after exposure begins and Dj is the lung burden of carbon core during the ith
39     subinterval.                            ,
40            The assumptions about dose-parameters relationship are given below.

        2/1/98                                     B-5       DRAFT-DO NOT CITE OR QUOTE

-------
 1      1.  The initiation rate associated with a lung burden {d, Di; I = 1, 2,..., 5} is given by
 2         u,(d,D;) = u0(l + a * d + b * D;), for 1=1,2, ..., 5. This is the only parameter that is
 3         assumed to depend on both d and D.
 4
 5      2.  Probability of tumor formation from a malignant cell is assumed to be dependent on lung
 6         burden D  by q(D;) = q0 + qjD;, 1 = 1,2,..., 5. To simplify calculation, the possibility that q is
 7         also dependent on organics d is not considered.
 8
 9      3.  The cell lifetime parameter y is assumed related nonlinearly to lung burden D by Y(Dj) = Yo
10         + YiLog(l+Dj),I=l,2,...,5.
11
12             To reduce the number of parameters that must be estimated from the Mauderly data, some
1 3      of the background parameters (u0, q0, and YQ) for the dose-response model are estimated from
14      the National Toxicology Program (NTP) historical control rate on Fischer-344 rats (reconstructed
1 5      from Portier et al., 1986). Giving these background parameters, the dose-related parameters are
16      then estimated by the E-M algorithm, which is described in Appendix B-2.  Using tumor
1 7      response data from Mauderly et al. (1987) and the corresponding dosimetric in Table B-l, the
1 8      resultant parameter estimates for the model are given in Table B-2.  To have some appreciation
1 9      about the implication of the Mauderly et al. .(1987) study, the estimated initiation and
20      proliferation (for I-cells) rates for the study are given in Table B-3.  Although these values may
21      not represent  reality (because they are not actual laboratory measurements), they could be used as
22      a guidance for future research planning.  For instance, Table B-3 (along with a discussion about
23      Table Br7) suggests that a slight increase of proliferation rate could cause a drastic increase on
24      tumor incidence, but only if the initiation rate is high enough. This  conclusion seems to suggest
25      that although the promotion effect of growth factors is important for tumor induction, the
2 6      initiation effect of carbon core and/or organics is also essential.
27
               Table B-l. Dosimetric (rag/cm  lung surface) use in modeling3
Exposure group d
0.35
3.5
7.08
2.5E-6
3.6E-5
7.3E-5
D,
6.23E-5
7.54E-4
1.98E-3
D,
8.75E-5
2.40E-3
5.49E-3
D,
8.97E-5
3.91E-3
8.56E-3
D,
9.02E-5
5.25E-3
1.12E-2
D,.
9.02E-5
6.29E-3
1.44E-2
        ad is organics; D,, 1=1,2,..., 5, are average lung burden of carbon core over five time intervals. These values are
        calculated by Yu et al. retention model in Appendix C.
        2/1/98                                    B-6      DRAFT-DO NOT CITE OR QUOTE

-------
       Table B-2. Maximum likelihood estimates for model parameters
 Parameter
             Estimate
 a
 b
 Yo
 Yi
 a
 N  (given)
             1.033E-7
             1.103E+4
             3.214E+2
             7.907E-7
             1.035E-1
             5.332E-2'
             1.662E-2
               2.647
             5.443E-1
             8.80E+7
Background parameters n0, q0, and YO are estimated separately from NTP historical control data.
 The number of target cells N is assumed to be 10-fold of Type II cells in mice, which is given in Kauffinan (1974).
 It is not essential for N to be given accurately because Nu0 appears as a single term in the model; the estimated u0
 will compensate for the under- or over-estimation of N.
       Table B-3. Relative magnitude of initiation  _J	!___  and proliferation
       (r[D]/T0) potential for the exposed versus control groups in Mauderly et al.
              (1987) study



Initiation



Proliferation


Time Interval
1
2
3
4
5
1
- 2
3
4
5

Low
1.048
1.056
1.056
1.056
1.056
1.004
1.006
1.006
1.006
1.006
Exposed Groups
Mid
1.639
2.168
2.654
3.084
3.419
1.052
1.166
1.270
1.362
1.424 .

High
2.442
3.570
4.556
5.405
6.433
1.137
1.379
1.590
1.770
1.989
2/1/98
B-7
DRAFT-DO NOT CITE OR QUOTE

-------
  1    • B.4. RESULTS             '  .  .
  2          The dose-response model predicts only probability of tumor occurrence (i.e., tumor
  3     incidence) by time t due to an exposure scenario. Because the probability of tumor occurrence is
  4     not directly observable (note that when an animal dies with a tumor, it only tells us that a tumor
  5     occurred before that time), the model can not be evaluated against the observed data. (Although
  6     it is possible to use the dose-response model, with additional assumptions, to calculate  tumor
  7     mortality, we prefer to evaluate the dose-response model alone because it will be used to predict
  8     cancer risk). To evaluate the reasonableness of the model constructed, it is possible only to
  9     compare the observed tumor mortality  rate (after adjusting for the competing risk) and  the
 10     predicted tumor incidence rate, with the understanding that observed values are expected to be
 11     smaller (this is particularly true at an early stage of tumor development when a tumor is small)
 12     than the predicted tumor incidence.  Table B-4 appears to bear this out; all of the predicted tumor
 13     incidences are either greater than the observed tumor mortality rates, or within the confidence
 14     bounds calculated from the observed tumor mortality. The observed tumor mortality rate is
 15     calculated by life-table approach. The  probability of tumor mortality can only be calculated up to
 16     about 900 days because after 900 days  tumors are no longer discovered by natural mortality only;
 17     in fact, the majority of tumors are discovered by sacrifice.
 18                                      .
 19     B.5. RISK PREDICTIONS UNDER VARIOUS EXPOSURE SCENARIOS
.20          For comparison, excess lifetime risks (see Tables B-5 and B-6) due to various exposure
 21     scenarios are calculated by the alternative model and the linearized multistage (LMS) model.
 22     Both point (maximum likelihood estimate) and 95% upper bound estimates are provided for the
 23     alternative model, whereas only upper  bound estimate is provided for  the LMS model because its
 24     linear component (which is notoriously unstable) is estimated to be 0. The 95% upper  bound for
 25     the alternative model is calculated by the same approach as for the LMS model;  (i.e., by
 26     increasing parameters a and b until the log-likelihood exceeds a critical value). To extrapolate
 27     from animal-based risk estimates to human; two assumptions are made:  (1) lung burden in terms
                 2
 28     of ug/cm of lung surface is equally potent between  animals and humans, and (2) 6 mo of animal
 29     life is equivalent to 18 years of human life. The latter assumption is necessary because life-span
 30     must be divided into five subintervals to account for different parameter values.
 31          Table B-5 compares predicted risks for humans due to continuous exposure (24 h/day)
 32     calculated by alternative and LMS models. It is interesting to see that risk calculations under
 33     various exposure concentrations are very similar using the two different models. Table B-6 gives
 34     excess risks due to exposure to 2.6  ^g/m  of diesel emissions, 16 h/day, 7 days/week; and 15
         2/1/98  .                                 B-8      DRAFT-DO NOT CITE OR QUOTE

-------
             Table B-4. Comparison of observed tumor mortality rate and predicted
             probability of cancer occurrence by time (T) when a (malignant) tumor
             was observed in rats
Exposure Time tumor
(mg/m ) observed (days)
Control

0.35

3.50

7.08













538
551
710
863
891
895
646
672
701
729
742
798
810
839
840
847
856
859
883
895
Observed tumor mortality Predicted tumor rate
rate3 (95% C. I.) by time (t)
0.0051
0.010
0.007
0.025
0.016
0.036
0.006
0.013
0.021
0.027
0.039
0.052
0.066
0.081
0.096
0.112
0.129
0.146
0.168
0.191
(0,0.015)
(0, 0.028)
(0, 0.022) •
(0, 0.063)
(0,0.126)
(0,0.052)
(0,0.019)
(0, 0.037)
(0, 0.041)
(0,0.059)
(0.004, 0.075)
(0.009, 0.095)
(0.016,0.115)
(0.023,0.138)
(0.032,0.161)
(0.041,0.183)
(0.052, 0.207)
(0.063, 0.229)
(0.077, 0.259)
(0.091, 0.291)
0.002
0.003
0.005
0.008
0.039
0.040
0.039
0.046
0.054
0.064
0.069
0.097
0.104
0.123
0.123
0.128
0.135
0.137
0.156
0.166
      Calculated by the life table procedure. Note that observed values are mortality, which are expected to be smaller
       than the (predicted tumor) incidence. This expectation is particularly true at early stage of tumor development
       when a tumor was small.

1     l^g/rn3, 8 h/day, 5 days/week. The concentration 2.6 ug/m  was reported by EPA's Office of
2     Mobile Sources to be the annual mean exposure of the U.S. population to diesel particulate
3     matter in 1986 and is only slightly higher than the most recent estimate of 2.03 ug/m in an EPA
4     draft document (Motor Vehicle-Related Air Toxic Study, April, 1993); the concentration 15
5     ug/m3 was reported to be the particulate exposure for workers on urban freeways in an EPA
6     report by Carey (Air Toxics Emissions from Motor Vehicles, 1987, EPA-AA-TSS-PA-86-5).

      2/1/98                                  B-9      DRAFT-DO NOT CITE OR QUOTE

-------
             Table B-5,  Comparison of excess risk for humans due to continuous
             exposure of various concentrations of diesel exhaust emissions under
             two different models

Exposure concentration (ug/m )
0.01
0.1 .
1.0 (unit risk)
100
1,000
Alternative
MLE
7.68E-8
8.12E-7
8.16E-6
5.58E-4
2.60E-2
model8
95% u.b.
1.35E-7
1.41E-6
1.65E-5
9.63E-4
4.22E-2
LMS modelb
1.71E-7
1.72E-6
1.74E-5
1.74E-4
3.33E-2
       "MLE:  maximum likelihood estimate; 95% u.b.: 95% upper bound estimate.  These are calculated using the
       alternative dose-response model
       bLMS:  calculated by linearized multistage model (slope = 9.04 per mg/cm2 of lung surface), using carbon core as
       dosimetric.  Only malignant tumors are used in the calculations.
             Table B-6. Excess lifetime risk for humans due to exposure to
             diesel exhaust emissions, under various exposure scenarios
Exposure pattern
Alternative model3
MLE 95%u.b.
LMSb
        2.6 ug/m , 16 h/day, 7 days/week
        Normal person

        2.6 (ig/m , 16 h/day, 7 days/week
        20 pack-year smoker
        15 p-g/m  , 8  h/day, 5 days/week
   1.41E-5
   2.32E-5
   3.12E-5
        2.44E-5
        3.61E-5
        5.17E-5
3.00E-5
5.38E-5
6.18E-5
       "MLE: maximum likelihood estimate; 95% u.b.; 95% upper bound estimate. These are calculated using the
       alternative dose-response model.
       bLMS; calculated by linearized multistage model, using carbon core as dosimetric. Only malignant tumors are used
        in the calculations.
1      For the general population exposed to an ambient air concentration of 2.6 ug/m , the risk to
2      normal (i.e., persons with normal respiratory functions) and smokers of 20 pack-years (as defined
3      by Bohning et al., 1982) are provided.  According to Bohning et aL, the retention half-life for
4      insoluble particle increases from 296 days for persons with normal respiratory function to 519
       2/1/98
B-10
DRAFT-DO NOT CITE OR QUOTE

-------
  1      days for persons with a smoking history of 20 pack-years.  This information is used to reduce the
  2     alveolar clearance rate in the dosimetric calculations using the same retention model that is also
  3     used to calculate dosimetric in Table B-l.
  4          The excess lifetime risks in Tables B-5 and B-6 are calculated by actuarial life-table
  5     approach using the survival probability of the NTP control animals provided in Portier et al.
  6     (1986).  Conceptually, this approach can be viewed as a weighted average of the probability of
  7     cancer occurrence over entire lifetime, weighted by survival probability.  This approach is more
  8     appropriate than the one used previously in the draft report in which probability of cancer
  9     occurrence at a preselected time (730 days) is used to represent the lifetime risk; it is more
10     appropriate because tumors here occur very late in life.
11
12     B.6. IMPLICATIONS OF THE ALTERNATIVE MODEL
13          Before proceeding to discuss implications of the alternative model on risk assessment, it
14 ,    should be noted that the parameters used in the model  are estimated on the basis of high exposure
15     concentration cancer bioassay data, not on the basis of data from laboratory measurements (e.g.,
16     mitotic rates for cells from normal and preneoplastic lesions measured over time), which usually
17     can be obtained over lower range of exposure concentrations. Therefore,  uncertainty associated
18     with low-dose extrapolation still remains.  For this reason, we will refrain from using the model
19     to evaluate low-dose risk estimations, but  rather to evaluate the relative contribution of each
20     biological component (e.g., initiation by organic and carbon core, individually or jointly) in the
21      model to cancer induction.
22          On the basis of the constructed alternative dose-response model, some specific inferences
23     could be made from Table B-7 by changing parameter values of the original model. Table B-7
24     provides a comparison of risks calculated  with changed parameters, assuming that animals are
25     exposed to 7.08 mg/m3 of diesel exhaust emissions, 7 h/day, 5 days/week (which is identical to
26     the exposure pattern of the highest exposed group in Mauderly et al., 1987).
27
28           The following observations can be made from Table B-7.
29
30     1. When there is no diesel induced  initiation (Case 2), the risk is 32% of the original model (i.e.,
31         the model without changing parameters), in contrast to 42% when exposure concentration is
32        reduced from 7.08 to 1 mg/m3 (not shown here). Therefore, the rple of diesel-induced
33        initiation in cancer induction increases with increasing exposure concentrations. This
34        conclusion is intuitively obvious because spontaneous induction of initiated cells play a
35        bigger role in cancer induction when concentration is lower. A practical implication of this
        2/1/98                                   B-ll      DRAFT-DO NOT CITE OR QUOTE

-------
               Table B-7.  Effect of changing dose-dependent initiation and promotion
               parameters (animals are assumed to be exposed to 7.08 mg/m3 of diesel
               emissions in air, 7 h/day, 5 days/week for life [i.e., the highest exposure group
               in Mauderly et al., 1987])
Case number
1

2

3
4
5

6
7
8
Parameters changed
None
(original model) :
a = 0
b = 0
a = 0
b = 0
a = 2,813aa
b = 0
Yl = 1.378Yla
Yi = 1.756Yla
Yi=0
Risk at 938 days
0.2067

0.0663

0.1616
0.1165
0.2007

0.3513
0.4537
0.0319
Ratio to original model
1.00

0.321
'
0.782
0.564
0.971

1.700
2.195
0.154
         a = 2.813a implies that a is increased by 2.813 times of its original value.
 1         observation is that reduction of non-diesel-induced initiation (e.g., by smoking) could have
 2         greater proportion of cancer risk reduction when diesel concentration is low than when the
 3         concentration is high.
 4
 5      2.  Cases 3 and 4 indicate that initiation by either carbon core, or organics contributes
 6         significantly to tumor incidence.
 7
 8      3.  Case 5, along with observation Number 2 above, suggests that although diesel-indueed I-cells
 9         play an important role in cancer induction, the role of initiation, however, could be assumed
10         by either organics or carbon core alone by increasing their respective proportions. One
11         implication is that, although existence of I-cells are important for tumor induction, these
12         I-cells could be induced by any agent that initiates (e.g., smoking).
13
14      4.  Cases 6 to 8 suggest that a small change of proliferation parameter y could have a
1 5         disproportionate change of cancer risk. Because this parameter is assumed a function of
16         carbon core dose, lung burden overloading has a significant effect on cancer incidence. In the
17         absence of better information, it is assumed in this report that carbon core continues to have
18         effect at low doses.
19
20             These four observations together suggest that while effect of growth factors (which may
21      increase value of Y) by particle overloading is important, the initiation effect of carbon core
22      and/or organics is also essential. Although this conclusion is only tentative because the model

        2/1/98                                    B-12      DRAFT-DO NOT CITE OR QUOTE

-------
  1      parameters are estimated on the basis of high concentration bioassay data, they do suggest the
  2      importance of studying the role of carbon core and organics on initiation and promotion at low-
  3      versus high-exposure concentrations. Does the relative initiation potential between organics and
  4      carbon core differ at high and low concentrations? Along with the results in Table B-6, it also
  5      suggests that a subcohort of workers who were smokers and exposed to high concentrations of
  6      diesel exhaust for a long duration would be expected to have higher lung cancer mortality.
  7          .   It is interesting to observe from Table B-8 that, under the same exposure conditions, risk
  8      is greater when exposure begins later in life. This model-based conclusion is due to the fact that
  9      older animals have more spontaneously (including non-diesel) induced initiated cells that have
10      potential to be proliferated, converted to malignant cell, and then progressing to cancer. (Note
11      that the above observation would not contradict any observation that might show that younger
1 2      animals are more sensitive to diesel exposure than older animals if a treatment induces more
13      initiated cells in the younger animals).  Assuming that the above model-generated hypothesis is
14      realistic, an important implication is that initiation by nondiesel agents should be considered
15      when assessing risk to humans due to exposure to diesel emissions.
16
17      B.7.  CONCLUSIONS AND SUMMARY
18      1.  The risk predictions by both alternative and LMS models are comparable over a range of
19         exposure concentrations that is of practical interest. However, this conclusion is valid only
20         under the assumption that the  effect of carbon core on each biological component (e.g.,
21         initiation) in the model continues to exist at low doses (see further discussions about
               Table B-8. Excess cancer risk to rats 200 and 300 days after termination
               of 6-mo exposure to D = 7.30E-5 mg/cm  of organics, and D = 1.89E-3
               mg/cm of carbon core3
                                                  Days After Exposure Terminated
Exposure Period (mo)
>6
6 to 12
12 to -18
200
1.48E-3
4.01E-3
8.32E-3
300
1.96E-3
5.25E-3
1.05E-2
            lung burden is assumed to be zero over unexposed periods. It may not be a realistic assumption because the
         lung burden is expected to linger over the following periods after exposure terminated; however, the assumed
         exposure condition serves the purpose better here.
        2/1/98  .                                 B-13       DRAFT-DO NOT CITE OR QUOTE

-------
 1         uncertainties below). Based on the Mauderiy et al. (1987) study, the risks associated with
 2         continuous exposure to 1 jig/m of diesel emissions calculated by two different models are
 3         summarized below:
 4
 5
Alternative Model
Lung Tumor Data Used
Malignant tumors
Total tumors
MLE
8.16E-6
N.A..
95% u.b.
1.65E-5
N.A.
LMS Model
95% u.b.
1.74E-5
3.44E-5
(taken from Chapter 11)
 1      2,  The model suggests that populations with higher expected background rate (e.g., smokers)
 2         may be subjected to higher lung cancer risk than the populations with lower background rate.
 3         It is noted that U.S. females have about the same background lung cancer rates as the
 4         Fischer-344 rats (about 1 to 2%), whereas U.S. males have a background rate of 6%.
 5         However, because most of lung cancers are smokers, the risk to nonsmokers (males or
 6         females) should be about the same.  The use of the unit risk estimate provided in Chapter 11
 7         may somewhat underestimate risk to smokers (or other respiratory-impaired persons) unless
 8         adjustment on lung burden is made. Table B-6 provides an example of such adjustment.
 9
10      B.8. DISCUSSIONS ABOUT UNCERTAINTIES OF RISK ESTIMATES
11            Although, it is interesting to note that risk estimates by the LMS model are comparable to
12      those calculated by the alternative model, there are uncertainties about low-dose extrapolation by
13      the alternative (as well as by the LMS) model: first, the model parameters are estimated
14      statistically, not measured in the laboratory; and second, the model parameters are estimated on
1 5      the basis of high-exposure data, the relationship between a parameter and exposure below the
16      exposure range remains unknown, and the dose-parameter relationship used in the model may
17      not be adequate for low-dose extrapolation. For instance, it is assumed that initiation rate is
18      linearly related to doses of carbon core. Such an assumption needs be evaluated. The risk at low
19      doses would be overestimated in this report if the relationship between initiation rate and carbon
20      core is sublinear (concave upward). The sublinear assumption would be reasonable if there is no
21      effect of initiation by carbon core dose (D) 'at low concentration.  On the other hand, the risk
22      would be underestimated if the relationship is supralinear. Therefore, it is important to evaluate

        2/1/98                                   B-14       DRAFT-DO NOT CITE OR QUOTE

-------
  1      how increase of diesel-exposure concentration affects initiation rate over low-exposure
  2     concentrations. Similarly, it is important to know the relationship between dose of carbon core
  3     and cell (I-cells in particular) proliferation at low concentration.
  4            Another aspect of uncertainty is the use of lung burdens (organics and carbon core)
  5     calculated by mathematical model, rather than actually measured. However, the impact of this
  6     uncertainty with regard to the conclusions reached in this report is not expected to be
  7     significantly altered even if the model-based dosimetrics are not accurate because the relative
  8     patterns of lung burdens between high- and low-exposure concentrations, and between animals
  9     and humans should be about the same. Although there is some observed total lung burden, these
10     data are not used because of the folio wing reasons.
11
12         1. The observed data are not separated by organics and carbon core.
13
14        2. There are no human data—these data are needed to predict risk in humans.
15
16        3, The observed data do not go beyond 730 days.
17
18        4. The desire is to be consistent with Chapter 11 so that results can be compared.
19
20     B.9. DISCUSSIONS ABOUT FUTURE RESEARCH NEEDS
21             The single most important use of a biologically based dose-response model in the cancer
22     risk assessment is to reduce uncertainty of low-dose extrapolation when the mechanism for tumor
23     response observed at high doses differs drastically from the low doses.  However, this report can
24     focus only on the use of the model to guide future research rather than to actually reduce
25      uncertainty of risk estimate because of our inability to obtain biological parameters in the model.
26     If a chemical is known to induce disproportionately larger cell proliferation (in normal, initiated,
27      and/or malignant cells) at high doses than at low doses, then a model that reflects this fact would
28      be useful.  With this in mind, our effort should be to identify "components" of carcinogenesis
29      (e.g., increase of mitotic rate) that are disproportionately more affected at high doses than at low
30     doses and to develop models that incorporate those high-dose effects.  For the diesel risk
31      assessment, the "components" that require further study include effects of organics and carbon
32      core, individually or jointly, on initiation, proliferation, conversion, and progression steps of
33      carcinogenesis. In order to use biologically based models of carcinogenesis  in risk assessments,
34     one needs to know the relationship between parameter values in a model and exposure (or dose).
35      Ideally, some of these parameters, if not all, should be measured directly in the laboratory, or
36      indirectly estimating from neoplastic and preneoplastic lesions (e.g., number of foci, adenomas,
37      and tumors in a lung).

        2/1/98                                    B-15      DRAFT-DO NOT CITE OR QUOTE

-------
 1             Cell proliferation is an increase in the cell population of different stages: normal,
 2     _ initiated, or malignant cells.. Enhanced cell proliferation of normal target cells may itself increase
 3      the frequency of mutations, either by inducing error in replication or by converting DMA adducts
 4      to mutations before DNA repair can occur.  The model implies that tumor incidence is linearly
 5      proportional to initiation rate.  On the other hand, enhanced cell proliferation of initiated cells
 6      could lead to more than linear increase of tumor incidence. Therefore, proliferation of I-cells has
 7      a greater impact on tumor incidence than proliferation of normal cells. However, this does not
 8      mean that initiation potential of compounds (organics or carbon core) is  not important. As
 9      discussed previously, it is important to determine the ability of these compounds to initiate at low
10      versus high doses; this has a significant implication for low-dose extrapolation. From the
11      viewpoint of mathematical modeling, cell proliferation is the result of a decrease of cell death
1 2      rate and/or an increase of mitotic rate, regardless of underlying biological mechanisms.
13      Therefore, it is logical to construct a model (as is done here) with a proliferation component
14      involving cell death and mitosis, and important to obtain data at the cellular level even if
1 5      biological mechanism at the molecular level is not yet known. If a more precise mechanism is
16      known and the quantitative data are available, then the proliferation component of the model can
17      be improved by incorporating the available biological information.  Most of the two-stage
18      models consider a single malignant cell to be equivalent to a tumor. If a compound is known to
19      affect the cell proliferation of tumor cell population, a model that incorporates tumor progression
20      should be used. For the diesel modeling, we assume that particles could enhance the
21      proliferation of malignant cells. This assumption needs to be verified. Another model-generated
22      hypothesis is that persons with higher number of initiated cells are subjected to higher lung
23      cancer risk when exposed to diesel emissions.  (A person could have a higher number of initiated
24      cells due to exposure to diesel and/or nondiesel agents, or simply by acquiring more
25      spontaneously induced initiated cells through aging).
26             In summary, information that is necessary to construct a biologically based dose-response
27      model includes (1) identifying roles that are played by organics and carbon core (individually or
28      jointly) with respect to initiation, proliferation, conversion, and progression, at low versus high
29      doses; (2) quantitative measurements of cellular dynamics (e.g., mitotic  rate) for cells at different
30      stages and exposure concentrations; and (3) relationship between parameters and exposure or
31      dose. Because many biological parameters are expected to be age-dependent, they should be
32      measured over different time points. Furthermore, frequency and size of preneoplastic foci,
33      nodules, and tumors could also provide useful information toward improving  risk assessment.
34      Some-of these data may be obtained by initiation-promotion type of study.
        2/1/98                                    B-16       DRAFT-DO NOT CITE OR QUOTE

-------
  1       B.10.  RFERENCES
  2
  3      Bohning D., Atkins, H., and Cohn, S. Long-term particle clearance in man: normal and impaired. Ann. Occup. Hyg.
  4      26:259-271.
  5        '              •
  6      Chen, C., and Farland, W. Incorporating Cell Proliferation in Quantitative Cancer Risk Assessment: Approach,
  7      Issues, and Uncertainties. In: Chemically Induced Cell Proliferation:. Implications for Risk Assessment, B.
  8      Butterworth, T. Slaga, W. Farland, and M. McClain, eds.; pp. 481-499 (Wiley-Liss, Inc., New York, 1991).
  9'
10      Cohen S. and Ellwein L. Cell growth dynamics in long-term bladder carcinogenesis, Toxicol. Letters, 43:151-173,
1 1       1988.
12
1 3      Dempster, A., Laird, N., and Rubin, D. Maximum likelihood from incomplete data via the EM algorithm. Roy.
14      Statist. Soc.,B 39:1-38, 1977.
15
1 6      Greenfield, R., Ellwein, L., and Cohen, S. A general probabilistic model of carcinogenic analysis experimental
1 7      bladder cancer, Carcinogenesis 5:437-445, 1984.
18
1 9      Kauffrnan S. Kinetics of alveolar epithelial hyperplasia in lungs of mice exposed to urethane I. quantitative analysis
20      of cell populations, Laboratory Investigations, 30:170-175, 1974.
21
22      S, Moolgavkar and D. Venzon, "Two-Event Models for Carcinogenesis:  Incidence Curve for Childhood and Adult
23      Tumors," Math. Biosciences 47, 55-77 (1979),
24
25      S. Moolgavkar and A. Knudson, "Mutation and Cancer: A Model for Human Carcinogenesis," Journal of the
26      National Cancer Institute 66,1037-1052 (1981).  •
27
28      Portier C, Hedges, J., and Hoel, D. Age-specific models of mortality and tumor onset for historical control animals in
29      the National Toxicology Program's carcinogenicity experiments. Cancer Research 46:4372-4378, 1986.
30
31       W. Tan and C. Chen, A nonhomogeneous stoehastic.model of carcinogenesis for assessing risk of environmental
32      agents. To appear in:  Mathematical Population Dynamics, Proceedings of the Third International Conference, Pau,
33      France, 1992.
34
         2/1/98                                        B-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1                                    APPENDIX B-l
 2                                 E-M ALGORITHM
 3
 4            The E-M algorithm is derived below. It will be used to calculate maximum likelihood
 5     estimate of parameters of the alternative model. Data used for the E-M algorithm is taken from
 6     Mauderly et al. (1987), which includes time when an animal died (natural mortality or sacrifice)
 7     with or without (malignant) tumors. The computer program for the E-M calculations was
 8     developed by Mr. Daliang Chang of the Computer Science Corporation under an EPA contract.
 9     The theory of E-M algorithm can be found in Dempster et al. ( 1 977).
1 0            Assume that the distinct times when animals died by either natural mortality or sacrifice
1 1     are tj < t2 < ...< tm.  The observations can be classified as follows;
12
1 3            aixW-     observed number of natural deaths without tumor at time tj in the treatment
14                      group x (There are four groups for diesel data [i.e., x = 1 , 2, 35 4.]),
15                  /
16            ^x®'     observed number of natural deaths with tumor at time tj in the treatment
17                      group x,
18
19            blx(I):     series sacrifice at time tj without tumor in the treatment group x,
20
2 1            b2x(I):     series sacrifice at time tj with tumors in the treatment group x.
22
23
24     Let Td represent the time an animal died and T the time a tumor developed.
25
26             tj, x} (conditional probability of death without tumor)
27
28            px(I|u) = Pr{Td = tj|Td > tj, Te^, \], x} (related to deaths with tumors)
29
30     Define,
                                                 n  - «xG)i.
                                              j=i
31
        2/1/98                                  B-l 8      DRAFT-DO NOT CITE OR QUOTE

-------
                                 Sx(t)  - Pr(T >  t xl  = exp[-fh(x)dx].
 1      The function Sx(t) is the probability of tumor free by time t.  The exact form of the hazard
 2      function h(x) and Sx(t) are given in the next section.
 3
 4      Let
 5
 6            a2x(I|u) = number of natural death at tj with tumor developed during (ty.j, tj, in the
 7            treatment group x, u < I,
 8
 9            b2x(I|u) = number of sacrifice at tj with tumor developed during (ty.j, tj, in the treatment
1 0            group x, u < I,
11
12      Then
13
14      Let
15

                             b r, *     A
                             p (i|u)  = --
16
17      and
18
                                  	     Av(u)S ft )A (t,,)Bfi|u)
                                                    W XK W
19
20      where
21
22
        2/1/98               '                   B-19       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17

18
19
20
21
22
23
24
25
Thus,
       Given a2x(x), {a2x(i|u), u = 1,.,,, i}, is an (i ~ l)-dimension multinomial with parameter
E[a2x(iu)la2x(i)]=a2x(i)Px(i|u).
Similarly, {b2x(i u), u = 1,..., i}, is an (i - l)-dimension multinormial with parameters (b2x(i),
Qx(i|u),u=l,...,i},and

                             E[b2x(iu)ib2x(i)] = b2x(i)Qx(iU).

It can be shown that the likelihood function is proportional to
where
                                             tii
                                          = E
Let
                                                             mxO)],'and
Let
                                  0  =
be a vector of parameters in function S;
                          Px(u)  =
                                              ), «x(2),  ..., ax(m)], and
                                              |u),  Px(2|u),  ...,  Px(m|u)]
be vectors of parameters related to conditional probabilities of death with and without tumors.
These parameters, along with those in ,0X will be estimated by the E-M algorithm described
below,   ,         -
        2/1/98
                                         B-20       DRAFT-DO NOT CITE OR QUOTE

-------
  1      TheM-step:
  2      Given initial values *a2x(iu) and b2x(iu), estimate
  3
                                1.  ^T(i)  = alx(i)/RJx(i) and
                                2.  p^iju) ^a2x(iju)/R2x(i|u), and
                                3,  obtain 6x by maximizing the log of L.

  4
  5      TheE-Stepi          '                     '       ^                         .   •    .
  6             Given the estimated values on ccx(i), px(i), and 0X from the M-step, compute Px(i|u) and
  7      Qx(i|u), and obtain estimates of 32X(i|u) and b2x(i|u) by
  8
                                       a2x(i|u)  = a2!£(i)Px(i|u), and
.9
1 0             With the estimated values of a2x(i|u) and b2x(i|u) available from the E-step, go back to the
1 1      M-step. Repeat the same process until estimates are stabilized.
        2/1/98                                    B-21       DRAFT-DO NOT CITE OR QUOTE

-------
  1                                      APPENDIX B-2

  2                          A  TUMOR GROWTH MODEL
  3
  4      -       The tumor growth model with piece-wise constant parameters is taken from Tan and
  5      Chen (1992), which is an extension of a stochastic model developed by Chen and Farland (1991).
  6      This model has a similar biological motivation as the two-stage model proposed by Greenfield
  7      et al. (1984), which has been used by Cohen and Ellwein (1988) to analyze bladder tumors.
  8      However, the two models differ from each other with respect to their mathematical formulations;
  9      the one adopted in this report is a stochastic model, whereas the other is a deterministic model
10      and does not allow for parameters estimation because the model does not have complete
11      mathematical expression.
1 2             Although its most general form will not be used here because of the lack of data, it is
1 3      worthwhile to note that the stochastic model by Chen and Farland (1991) has two desirable
1 4      features: (1) it allows for any cell growth distributions (e.g., Gompertz), rather than limiting only
15      to the exponential distribution as in other existing models; and (2) it incorporates the birth and
16      death of tumor cells, rather than assuming that a tumor is born once a single tumor cell occurs, an
17      assumption made by the MVK model (a model proposed by Mpolgavkar et al., 1979, 1981).
1 8      Therefore, if information on cell lifetime distribution, and the progression phase of tumor
1 9      development is available, a reasonably realistic model can be constructed.
20             For completeness of the report, a brief description of the model will be presented here.
21      The following notations are needed for the model:
22
23             N(t): number of normal (target) cells at time t,
24
25             Uj:   initiation rate, and
26
27             f(t):  the probability density function for the lifetime of an initiated cell (I-cell).
28
29             For an I-cell, at the end of its lifetime it either divides (mitosis) or dies (programmed or
30      nonprogrammed death). If it enters into mitosis, it either divides into two I-cells with probability
31      a, or divides into one I-cell and one malignant cell (M-cell) with probability ji2-  Note that, at the
32      end of a cell's lifetime, the probability for the cell to die is p = 1 - a - u2.  A similar setup (i.e.,
33      to allow for any cell lifetime distribution) can be made for an M-cell. However, we will confine
34      ourselves to a simpler version assuming that an M-cell lifetime follows an exponential

        2/1/98                                    B-22       DRAFT-DO NOT CITE OR QUOTE

-------
21
  1      distribution. Thus, we can simply assume that an M-cell follows a simple birth-death process; it
  2      can either divide into two M-cells with a rate am or die with a rate Pm.
  3             When parameters are constant over time (ages), the hazard function is given by

                          h(t) = (ijH2|N(s)m(t - s) ds
                                   o
  4      where
                                           (y2 -  y,)2exp[A(t)o(y2 - y,)]
                                m(t)  =	—;
                                       
-------
                                           '2J - l)exp[A.(t)o.(yy - y,p]}2

 2
                                            2
 3      where y y < y2j are two real roots of Ojy  - (aj + fy + (J^q^y + Pj = °; aj + Pj + P-2j = 1 » Ij = 1 ""
 4      pmj/amj,j = l,2,...,k.
 5             When exponential distribution (i.e., A:(t) = Yjt and q: = 1 are assumed, the model is
 6      equivalent to the MVK model with piece-wise constant parameters. A special case that may be
 7      more appropriate than the exponential distribution is when the Gompertz distribution is assumed
 8      (i.e., when Aj(t) = { 1 - exp[- Yjt]}/Yj)-
 9             For the diesel alternative model, the total time is divided into five (i.e., k = 5)
1 0      subintervals.  It is shown in Tan and Chen (1992) that, under the assumption of exponential cell
1 1      lifetime distribution, the tumor free distribution function, Sx(t), can be written as
12
                               S(t) -
                                           j=i
13
14      where S: = t: if j < k and Sj = t if j = k, and
15

         A (t   s)-= 2Nn  ^  — 1 - {-(s - t  )
          u  }    }       J  1J 2jW  + Z     J    J
                                                     Yr(Wj - Zj)

16
                                               YH(W!
                                               zi + (
                                           Cj + Zj + (Wj
                                          «?  -j.gr  + /«jy.
                                          w I   £i,   V^T    tJT/w-"-rV" T~"i-i-1 iV*!*1*!/ \
                                     -log[_J	!	^	L	   ;  J    J j),
                                          W  + 7  + IW  — 7 >
                                           I   £I   v"l     F
17     where,
18
19           Wj      = [(a + p + u2q)2 •
20           Zj       = a - P -  (J.2q, and
21            Ajj(s,t)   = Yj(t"- s) if both s and t are in the same closed subinterval [tj,l5 tj] and
        2/1/98                '                    B-24       DRAFT-DO NOT CITE OR QUOTE

-------
                                       r=i-H



 if seLj, teL. with t. < t.
                                          Yr(tr - v,
2/1/98                             B-25     DRAFT-DO NOT CITE OR QUOTE

-------
                      Appendix C




          Models for Calculating Lung Burdens
2/1/98                       C-l     DRAFT-DO NOT CITE OR QUOTE

-------
 1      C.I.  INTRODUCTION
 2            As discussed in Chapter 4 the lung burden of diesel exhaust particles (DEPs) during
 3      exposure is determined by both the amount and site of particle deposition in the lung and,
 4      subsequently, by rates of translocation and clearance from the deposition sites. Mathematical
 5      models have often been used to complement experimental studies in estimating the lung burdens
 6      of inhaled particles in different species under different exposure conditions. This section
 7      presents a mathematical model that simulates the deposition and clearance of DEPs in the lungs
 8      of rats and humans.
 9            Diesel particles are aggregates formed from primary spheres of 15 to 30 nm in diameter.
10      The aggregates are irregularly shaped and range in size from a few molecular diameters to tens of
11      microns.  The mass median aerodynamic diameter (MMAD) of the aggregates is approximately
12      0.2 nm. The primary sphere consists of a carbonaceous core (soot) on which numerous kinds of
13      organic compounds are adsorbed.  The organics normally account for 10 to 30% of the particle
14      mass. However, the exact size distribution of DEPs and the specific composition of the adsorbed
15      organics depend upon many factors, including engine  design, fuels used, engine operating
16      conditions, and the thermodynamic process of exhaust. The physical and chemical
17      characteristics of DEPs have been reviewed extensively by Amann and Siegla (1982) and
18      Schuetzle (1983).
19            Four mechanisms deposit diesel particles within the respiratory tract during exposure:
20      impaction, sedimentation, interception, and diffusion.  The contribution from each mechanism to
21      deposition, however, depends upon lung structure and size, the breathing condition of the
22      subject, and particle size distribution.  Under normal breathing conditions, diffusion is found to
23      be the most dominant mechanism. The other three mechanisms play only a minor role.
24            Once DEPs are deposited in the respiratory tract, both the carbonaceous cores and the
25      adsorbed  organics of the particles will be removed from the deposition sites as described hi
26      Chapter 4. There are two mechanisms which facilitate this removal: (a) mechanical clearance,
27      provided by mucocilliary transport in the ciliated conducting airways as well as macrophage
28      phagocytosis and migration in the nonciliated airways, and (b) clearance by dissolution.  Since
29      the carbonaceous soot of DEPs is insoluble, it is removed from the lung primarily by mechanical
30      clearance, whereas the adsorbed organics are removed principally by dissolution.
31
32      C.2. PARTICLE MODEL
33            To develop a mathematical model which simulates the deposition and clearance of DEPs
34      in the lung,  an appropriate particle model characterizing a diesel particle must first be introduced.
35      For the deposition study, we employed an'equivalent sphere model for the diesel particle
36      developed by Yu and Xu (1987) to simulate the dynamics and deposition of DEPs in the
        2/1/98                                   C-2        DRAFT-DO NOT CITE OR QUOTE

-------
  1      respiratory tract by various deposition mechanisms. For the clearance study, we assume that a
  2      diesel particle is composed of three different material components according to their
  3      characteristic clearance rates: (1) a carbonaceous core of approximately 80% of the particle
  4      mass, (2) absorbed organics of about 10% of particle mass, and that are slowly cleared from the
  5      lung (3) adsorbed organics quickly cleared from the lung accounting for the remaining 10% of
  6      particle mass. The presence of two discrete organic phases in the particle model is suggested by
  7      observations that the removal of particle-associated organics from the lung exhibits a biphasic
  8      clearance curve (Sun et al, 1984; Bond et al,, 1986) as discussed in Chapter 4. This  curve
  9      represents two major kinetic  clearance phenomena: a fast phase organic washout with a half-
10      time of a few hours and a slow phase with a half-time that is a few hundred times longer.  The
11      detailed components involved in each phase of the clearance are not known.  It is possible that
12      the fast phase consists of organics which are leached out primarily by diffusion mechanisms
13      while the slow phase might include any or all of the following components: (a) organics which
14      are "loosened" before they are released, (b) organics which have become intercalated in the
15      carbon core and where release is thus impeded, (c) organics which are associated  for longer
16      periods of time due to hydrophobic interaction with other organic phase materials, (d) organics
17      which have been ingested by macrophages and as a result effectively remain in the lung for a
18      longer period of time due to metabolism by the macrophage; metabolites formed may interact
19      with other cellular components, and (e) organics which have directly acted on cellular
20      components such as the formation of covalent bonds with DNA and other biological
21      macromoleeules to form adducts.
22            The above distinction of the organic components is largely mechanistic and it does not
23      specifically imply the actual component nature of the organics adsorbed on the carbonaceous
24      core; however, the distinction is made to account for the biphasic clearance of DEPs.  However,
25      this distinction is necessary in appreciating the dual phase nature of DEPs. For aerosols made of
26      pure organics such as benzo(a)pyrene (BaP) and nitropyrene (NP) in the same size range of
27      DEPs, Sun et al. (1984) and Bond et al, (1986) observed a nearly monophasic clearance curve.
28      this might be explained by the absence of intercalative phenomena (a) and of hydrophobic
29      interaction imposed by a heterogeneous mixture of organics (b).  The measurement of a pure
30      organic might also neglect that quantity which has become intracellular (c) or covalently bound
31      (d).
32
33      C.3. COMPARTMENT AL LUNG MODEL
34            To study the transport and removal of DEPs from the lungs, we used a compartmental
35      model consisting of four anatomical compartments: the nasopharyngeal or head (H),
36      tracheobronchial (T), alveolar (A), and lung associated lymph node (L) compartments as shown
        2/1/98                                    C-3       DRAFT-DO NOT CITE  OR QUOTE

-------
 1     in Figure C-l.  In addition, we used two outside compartments B and G representing,
 2     respectively, the blood and gastrointestinal (GI) tract. The alveolar compartment in the model is
 3     obviously the most important compartment for long-term retention studies.  However, for short-
 4     term consideration, retentions in other lung compartments may also be significant. The presence
 5     of these lung compartments and the two outside compartments in the model therefore provides a
 6     complete description of all clearance processes involved.
 7            In Figure C-1, r %  r$ and r .fare, respectively, the mass deposition rates of DEP material
 8     component i (i=l [core], 2 [slowly cleared organics], and 3 [rapidly cleared organics]) in the
 9     head, tracheobronchial and alveolar compartments; and /^, represents the transport rate of
10     material component i from any compartment X to any compartment Y. Let the mass fraction of
11     material component i of a diesel particle be //. Then
                              4° =f,rH  ,                                              (C-l)
12
                              r? = ft rT  ,                                              (C-2)
13
                              ^-f'trA  ,                                              (C-3)
14
15      where rH, rT, and rA are, respectively, the total mass deposition rates of DEPs in the H, T, and A
16      compartments, determined from the equations:
17
                         rH = c(TV)(RF)(DF)H  ,                                         (C-4)

18
                         rT = c(TV)(RF)(DF)T  ,                                         (C-5)

19
                    •     rA = c(TV)(RF)(DF)A  .                                         (C-6)

20
21            In equations C-4 to C-6, c is the mass concentration of DEPs in the air, TV is the tidal
22      volume, RF is the respiratory frequency, and (DF)H, (DF)T, and (DF)A are, respectively, the
23      deposition fractions of DEPs in the H, T, and A compartments over a respiratory cycle. The
24      values of (DF)H, (DF)T, and (DF)A which vary with the particle size, breathing conditions and
25      lung architecture were determined from our deposition model (Yu and Xu, 1987).
26
        2/1/98                                   C-4       DRAFT-DO NOT CITE OR QUOTE

-------
-


. (i)
_ ATB

B
(i)


(i)



(i)
rH
O . (i)
HHG
-
TT
o . 0)
T ATG
1
i
^AT fA


,**
i
L


,-- ~i
^.i I
*"i '
I
l ~
« G
i ^^
i i
i i
\ i_ « -—
)


L



      Figure C-l. Compartmental model of DEP retention.
1
2
3
4
5
      The differential equations for w§, the mass of material component i in compartment X, as
a function of exposure time t can be written as

Head(H)
                 dm.
                  dt

                         H
      2/1/98
                                     C-5
DRAFT-DO NOT CITE OR QUOTE

-------
 1     Tracheobronchial (T)

                     = r ^  +  A^Lm ^ - A.1'' »j M)  -  A.M'1 m V)                                  (C-8)
dmT     (i)  .  ,(/)   (I)   ,(/)  (I)
                dt
       Alveolar (A)
               dm C0
                  ^   _  (0  _ }(0   (0 _ i(0   (0 _  jO)  (0                                  (C Q\
               ————    r *     A ATftl A    /\*AffFl A     f* A Of** A   J                               ^V--~7/
                 dt
       Lymph Nodes (L)

 4     Equation C-9 may also be written as
                         dmf     (t
                          dt   ~TA
                                                                       (C-ll)
       where
 6     is the total clearance rate of material component i from the alveolar compartment. In equations
 7     C-7 to C-10, we have assumed vanishing material concentration in the blood compartment to
 8     calculate diffusion transport.
 9            The total mass of the particle-associated organies in compartment X is the sum of m ($
1 0     and m f^the total mass of DEPs in compartment X is equal to
                        m.x = mf  +  mf] + mf                                    ,   (C-13)

1 1     The lung burdens of diesel soot (core) and organies are defined, respectively, as
12     and
13     Because the clearance of diesel soot from compartment T is much faster than from compartment
14     A,m(^
-------
  1             Solution to equations C-7 to C-10 can be obtained once all the transport rates /l(f are
  2      known.  When /i§- are constant, which is the case of linear kinetics, equations C-7 to C-10 will
  3      have a solution that increases with time at the beginning of exposure but eventually saturates and
  4      reaches a steady-state value. This is the classical retention model developed by the International
  5      Commission of Radiological Protection (ICRP, 1979). However, as discussed in Chapter 4, data
  6      have shown that when rats are exposed to DEPs at high concentration for a prolonged period, the
  7      diesel soot accumulates in various peribronehial and subpleural regions in the lung and the long-
  8      termed clearance is impaired. This is the so-called overload effect, observed also for other
  9      insoluble particles.  The overload effect cannot be predicted  by the classical ICRP model.
10      Soderholm (1981) and Strom et al. (1987, 1988) have proposed a model to simulate this effect by
11      adding a separate sequestrum compartment in the alveolar region. In the present approach, a
12      single compartment for the alveolar region of the lung is used and the overload effect is
13      accounted for by a set of variable transport rates /ij| J$u and A® which are functions of mA.  The
14  .    transport rates /if and /|f in equations C-7 to C-10 can be determined directly from experimental
15      data on lung and lymph node burdens, and A$- and Afy from  equation C-12.
16
17      C.4.  SOLUTIONS TO KINETIC EQUATIONS
18            Equation C-l 1 is a nonlinear differential equation ofm(/ with known function of A$ For
19      diesel soot, this equation becomes
                         -       Ji      a  *•  A'  A
                        at

20      Because clearance of the particle-associated organics is much faster than diesel soot,
21      w^constitute only a very small fraction of the total particle mass (less than one percent) after a
22      long exposure and we may consider A^zs a function of m^alone. Equation C-17 is then reduced
23      to a differential equation with m^the only dependent variable.
24            The general solution to equation C-17 for constant rf%t any time, t, can be obtained by
25      the separation of variables to give
                             —	——  =  t  .                        •          .      (C-18)

26            If rj^is an arbitrary function of t, equation C-17 needs to be solved numerically such as by
27      a Runge-Kutta method.  Once m^is found, the other kinetic equations C-7 to C-10 for both diesel
28      soot and the particle-associated organics can be solved readily, since they are linear equations.
29      The solutions to these equations for constant r$, r$ and rf are given below:

        2/1/98                                    C-7        DRAFT-DO NOT CITE OR QUOTE

-------
 1     Head(H)
 2
 3
                        where   A   = AG + A                                         (C-20)
 4     Tracheobronchial (T)
       0  =  exp (-4° r )    . (  r« + A«  IH « ) exp (A«  t) dt + rr.                      (C.2l)
 5                       where  $  =  A.®  + A.®           -            .                 (C-22)

 6     Lymph Nodes (L)
             IH «  = exp(-A«  0 |Q' A.«« «exp(A.«) * + m<%)                            (C-23)

 7            In equations C- 1 9 to C-23 , m $0 represents the value of m $ at t = 0.
 8            In the sections to  follow, the methods of determining r$, r$ and r% or (DF)H, (DF)T,  and
 9     (DFA r(D$, r(D$ and /°f;as well as the values of /^ in the compartmental lung model are
10     presented.
11
12     C.5. DETERMINATION OF DEPOSITION FRACTIONS
1 3            The mathematical models for determining the deposition fractions of DEPs in various
14     regions of the respiratory tract have been developed by Yu and Xu (1986, 1987) and are adopted
15     in this report. Yu and Xu consider DEPs as a polydisperse aerosol with a specified mass median
1 6     aerodynamic diameter (MM AD) and geometrical standard deviation og. Each diesel particle is
1 7     represented by a cluster-shaped aggregate within a spherical envelope of diameter de. The
1 8     envelope diameter de is related to the aerodynamic diameter of the particle by the relation

                         d- - cj>-1/2(^)1/2)(f )1/2             '                          (C-24)
                         a         C  .    C
                           a          e     'o

1 9     where £ is the bulk density of the particle in g/cm3, £0 = 1 g/cm3;  is the packing density, which
20     is the ratio of the space actually occupied by primary particles in the envelope to the overall
21     envelope volume; and Cx is the  slip factor given by the expression:
                           A
               Cx  = 1  + 2—  [1.257 + 0.4 exp -(— r-*- )]           .                   (C-25)
                           dx                       A

22      in which A = 8 x 10'6cm3 is the mean free path of air molecules at standard conditions.  In the
23      diesel particle model of Yu and Xu (1986), C has a value of 1.5 g/cm3 and a  value of 0.3 is
        2/1/98                                    C-8       DRAFT-DO NOT CITE OR QUOTE

-------
 1      chosen based upon the best experimental estimates.  As a result, Equation C-24 gives de/da =
 2      1 ,35. In determining the deposition fraction of DEPs, de is used for diffusion and interception
 3      according to the particle model.
 4
 5      C.5.1.  Determination of (DF)H
 6            • Particle deposition in the naso- or oro-pharyngeal region is referred to as head or
 7      extrathoracic deposition. The amount of particles that enters the lung depends upon the
 8      breathing mode. Normally, more particles are collected via the nasal route than the oral route
 9      because of the nasal hairs and the more complex air passages of the nose.  Since the residence
1 0      time of diesel particles in the head region during inhalation is very small (about O.ls for human
1 1      adults at normal breathing), diffusional deposition is insignificant and the major deposition
12      mechanism  is impaction. The following empirical formulas derived by Yu et al. (1981) for
1 3      human adults are adopted for deposition prediction of DEPs:
14      For mouth breathing:
                          in
                            = 0,          for d2a< 3000                                  (C-26)
           (DF}H in  = -1.117 + 0.324 Iog(rfa20, for d*Q > 3000                          (C-27)

                              (W* « = °>                                             (C-28)

1 5      and for nose breathing:
           (DF)H iR =  -0.014  + 0.023  log(Ja20, for d*Q < 337                           (C-29)

           (DF)H tn -  -0.959  + 0.397  log(d*Q), for d*Q > 337                           (C-30)

            (DF)H ia = 0.003 + 0.033 log(d*Q), for d^Q z 215                            (C-31)

           (DF)H ei =  -0.851  +0.399  log(d2aQ), for d*Q > 215                           (C-32)
16
1 7      where (DF)jj is the deposition efficiency in the head, the subscripts in and ex denote inspiration
1 8      and expiration, respectively, da is the particle aerodynamic diameter in ^m, and Q is the air
1 9      flowrate in cmVsec.
20             Formulas to calculate deposition of diesel particles in the head region of children are
21      derived from those for adults using the theory of similarity, which assumes that the air passage in
22      the head region is geometrically similar for all ages and that the deposition process is
23      characterized by the Stokes number of the particle.  Thus, the set of empirical equations from C-
24      26 through C-32 are transformed into the following form:
        2/1/98                                    C-9        DRAFT-DO NOT CITE OR QUOTE

-------
       For mouth breathing:
                 (DF)ff .n = 0,         for d]Q  < 3000                              ,   (C-33)
                   (DF)H  .n  = -1.117 + 0.972 logK  +
                                •}   '      7                                             (C-
                     0.324 log(<0, for d\Q > 3000                                    V
                                    « = u-                           .                  (C-35)
       and for nose breathing:
          (£>F)W   = - 0.014  +  0.690 log K + 0.023  Iog(rfa2g),
                                  2                                                     (C-36)
                            for daQ < 337

                   (DF)H .n  =  -0.959  +  1.191  log K +
                     0.397  log (d^Q), for d*Q > 337                                     (  -  )

                   (DF)H el = 0.003  + 0.099 log K +
                      0.033 \og(d Q), for d Q
 6                 (DF)W    = 0.851 + 1.197 log K +
                                                                                       CC
                      0.399 Iog(^a20, for d^Q >215                                      V

 7     where K is the ratio of the linear dimension of the air passages in the head region of adults to that
 8     of children, which is assumed to be the same as .the ratio of adult/child tracheal diameters.
 9            For rats, the following empirical equations are used for deposition prediction of DEPs in
10     the nose:
                     (DF)H in  =  (DF)H „  = 0.046  +
                     0.009  \og(d*Q), for d^Q <  13.33                                    (     )

11     C.4.2. Determination of (DF)T and (DF)A
12            The deposition model adopted for DEPs is the one previously developed for
13     monodisperse and (Yu, 1978) and polydisperse spherical aerosols (Diu arid Yu, 1983). In the
14     model, the branching airways are viewed as  a chamber model shaped like a trumpet (Figure C-2).
15     The cross-sectional area of the chamber varies with airway depth, x, measured from the
16     beginning of the trachea. At the last portion of the trumpet, additional cross-sectional area is
17     present to account for the alveolar volume per unit length of the airways.
18
       2/1/98                                   C-10       DRAFT-DO NOT CITE OR QUOTE

-------
                        Summed Alveolar Cron Sectional Ant A/i)
                  Tndiet
                    Aim? Laftk z
                          •Summed Ainny
                    CTOB Section*] Area A,(x)
      Figure C-2. Trumpet model of lung airways.
1     Inhaled diesel particles that escape capture in the head during inspiration will enter the trachea
2     and subsequently the bronchial airways (compartment T) and alveolar spaces (compartment A).
3           Assuming that the airways expand and contract uniformly during breathing, the equation
4     for the conservation of particles takes the form;
                                 - a
                             dx
                                      (C-42)
      2/1/98
C-ll
DRAFT-DO NOT CITE OR QUOTE

-------
 1      where c is the mean particle concentration at a given x and time t; A[ and A, are, respectively,
 2      the summed cross-sectional area (or volume per unit length) of the airways and alveoli at rest; T)
 3      is the particle uptake efficiency per unit length of the airway; P is an expansion factor, given by:
16

                                                                                         (C-43)
        and Q is the air flow rate, varying with x and t according to the relation
                                          v
                              ^. = 1  --i
                              Q           V,
                              *£• o           I

 5      where Q0 is the air flow rate at x = 0. In Equations C-43 and C-44, V, is the volume of new air in
 6      the lungs and Vx and Vt are, respectively, the accumulated airway volume from x = 0 to x, and
 7      total airway volume at rest.
 8            Equation C-42 is solved using the method of characteristics with appropriate initial and
 9      boundary conditions.  The amount of particles deposited between location xt and x2 from time tt
10      to tj can then be found from the expression
                                  '2 X2
                            DF = f  fQcT]dxdt                                           (C-45)
                                    xi
1 1             For diesel particles, r| is the sum of those due to the individual deposition mechanisms
12      described above, i.e.,
                         T! = TI, + TIS  +  TI,  + %                                         (C_46)
1 3      where % T)S, rip, and D are, respectively, the deposition efficiencies per unit length of the airway
14      due to impaction, sedimentation, interception, and diffusion. On the basis of the particle model
1 5      described above, the expressions for r)t, r|s, rip, and T)D are obtained in the following form:
                                                '                                         (C.47)
      2
is ~  '
               = -±-[26 ^1 - e(2/3)  - e17Vl -  e2/3  +. sin'1 e1/3]                           (C-48)
                  "KL                                 ».

                                   4       F3
                            -                  )                                          (C-49)
                                  3TTI      32
        2/1/98   .                                 C-12       DRAFT-DO NOT CITE OR QUOTE

-------
                     r\D = —[l-0.819exp(-14.63A) -
                         0.0976  exp(-89.22A) -                                         (C-50)
               0.0325  exp(-228A)  - 0.0509  exp(-125A2/3)]
 1      for Reynolds numbers of the flow smaller than 2000, and
                        T]  =  —A1/2 (1  - 0.444Al/2)
                             L
 2      for Reynolds numbers greater than or equal to 2000, where ST=d2au/(18juR) is the particle Stokes
 3      number, 6 = L/(8R),  6 = 3puJL/(32uR), r= dJR, and A = DL/(4R2u). In the above definitions u
 4      is the air velocity in the airway; (i is the air viscosity; L and R are, respectively, the length and
 5      radius of the airway; us = Cj£j(18fii) is the particle settling velocity; and D = CekT(3 npd) is the
 6      diffusion coefficient with k denoting the Boltzmann constant and T the absolute temperature. In
 7      the deposition model, it is also assumed that r)I and % = 0 for expiration, while % and % have
 8      the same expressions for both inspiration and expiration.
 9            During the pause, only diffusion and sedimentation are present. The combined deposition
10      efficiency in the airway, E, is equal to:    -
                       E = 1 -(1  - Es) (1 -  ED)   .                                      (C-52)
11      where ED and Es are, respectively, the deposition efficiencies due to the individual mechanisms
12      of diffusion and sedimentation over the pause period.  The expression for Ep and Es are given by

        i  -   E -
             / =  la<          "      /  = !<            •.„     ^                        (C-53)

13      where TD = Dt/R2 in which t is the pause time and a,, a2, and a3 are the first three roots of the
14      equation:    ,
                              •/„(«) = 0  .                                               (C-54)
15      in which J0 is the Bessel function of the zeroth order, and:
                E, =  1.1094T, - 0.1604l4 for 0 < !„  < 1.                                (C-55)
                 S          £>          •-> ^         &                                     \     s
3 .
F VexD
^ + 2
i = I 0,
<
3
l/2/i V^
iz (i — y.
Es =  1  ~ 0.00691^  -0.0859V2  - 0.0582T^3,
                                                                         (  "
16      and
               Es  =  1  ~ 0.006
                               for Ts >  1,

17      where ts - ust/2R.  •


        2/1/98                                    C-13       DRAFT-DO NOT CITE OR QUOTE

-------
 1            The values of (DF)T and (DF)A over a breathing cycle are calculated by superimposing DF
 2      for inspiration, deposition efficiency E during pause, and DF for expiration in the
 3      tracheobronchial airways and alveolar space. It is assumed that the breathing cycle consists of a
 4      constant flow inspiration, a pause, and a constant flow expiration, each with a respective duration
 5      fraction of 0.435, 0.05, and 0.515 of a breathing period.
 6
 7      C.5.3. Lung Models
 8            Lung architecture affects particle deposition in several ways:  the linear dimension of the
 9      airway is related to the distance the particle travels before it contacts the airway surface; the air
10      flow velocity by which the particles are transported is determined by  the cross-section of the
11      airway for a given volumetric flowrate; and flow characteristics in the airways are influenced by
12      the airway diameter and branching patterns.  Thus, theoretical prediction of particle deposition
13      depends, to a large extent, on the lung model chosen.
14
15      C.5.3.1.  Lung Model for Rats
16            Morphometric data on the lung airways of rats were reported by Schum and Yeh (1979).
17      Table C-1 shows the lung model data for Long Evans rats with a total lung capacity of 13.784
18      cm3. -Application of this model to Fischer rats is accomplished by assuming that the rat has the
19      same lung structure regardless of its strain and that the total lung capacity is proportional to the
20      body weight. In addition, it is also assumed that the lung volume at rest is about 40% of the total
21      lung capacity and that any linear dimension of the lung is proportional to the cubic root of the
22      lung volume.
23
24      C.5.3.2.  Lung Model for Human Adults
25            The lung model of mature human adults used in the deposition calculation of DEPs is the
26      symmetric lung model developed by Weibel (1963). In Weibel's model, the airways are assumed
27      to be a dichotomous branching system with 24 generations.  Beginning with the 18th generation,
28      increasing numbers of alveoli are present on the wall of the airways and the last three generations
29      are completely aleveolated. Thus, the alveolar region in this model consists of all the airways in
30      the last seven generations. Table C-2 presents the morphometric data of the airways of Weibel's
31      model adjusted to  a total lung volume of 3,000 cm3.
        2/1/98                                    C-14       DRAFT-DO NOT CITE OR QUOTE

-------
      TABLE C-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
16a
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
Volume1* (cm)
0,243
0338
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
"Terminal bronchioles.
"Including the attached alveoli volume (number of alveoli = 3 x 107, alveolar diameter = 0.0086 cm);
2/1/98
C-15
DRAFT-DO NOT CITE OR QUOTE

-------
          TABLE C-2.  LUNG MODEL BY WEIBEL (1963) ADJUSTED TO
                           3,000 CM3 LUNG VOLUME
Generation
Number
0
2
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16a
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
Volumeb (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
"Terminal bronchioles. .
Including the attached alveoli volume (number of alveoli = 3 x 103, alveolar diameter = 0.0288 cm).
2/1/98
C-16
DRAFT-DO NOT CITE OR QUOTE

-------
  1      C.5.3.3. Lung Model for Children
  2             The lung model for children in the diesel study was developed by Yu and Xu (1987) on
  3      the basis of available morphometric measurements.  The model assumes a lung structure with
  4      dichotomous branching of airways, and it matches Weibel's model for a subject when evaluated
  5      at the age of 25 years, the age at which the lung is considered to be mature.  The number and size
  6      of airways as functions of age t (years) are determined by the following equations.
  7
  8      C.5.3.3.1,  Number of airways and alveoli
  9             The number of airways Nj(t) at generation i for age t is given by
                    Nt(t)  = 2',           for 0 *  i <20                                    (C-57)

                    (        r        fnr A7 (f\ < ?2'                                    (C* *I8^
                      \r  ft\ = fj  CA=n   1UI 'Yrw - *                                     (^-jo)
                    _ ^21
            (  N22(t) = Nf(t)  -221,          for 221  < Nf(t) < 222                    '       (C-59)
              N23(t) = 0,
            N,,(t) = 221,
              21V
                      22                  x«_  AT /t\  **. l21
             N22(t) = 2",                 for Nf(t) > 221  +  2",                          (C-60)
                           -221 -222
10      where Nr(t) is the total number of airways in the last three airway generations. The empirical
11      equation for Nr which best fits the available data is
                      { 2.036, 10;(1-0.926.^1),  ',8     '                           (c
                pW     1.468 x 107,                  t > 8   .                           ^
12      Thus, Nr(t) increases from approximately 1.5 million at birth to 15 million at 8 years of age and
13      remains nearly  constant thereafter. Equations C-58 to C-60 also imply that in the last three
14      generations, the airways in the subsequent generation begin to appear only when those in the.
15      preceding generation have completed development.
16            The number of alveoli as a function of age can be represented by the following equation
17      according to the observed data:
                   NA(t) = 2.985 x 108(1  -0.919e~045f)                *                   (C-62)

18            The number of alveoli distributed in the unciliated airways at the airway generation level
19      is determined by assuming that alveolization of airways takes place sequentially in a proximal
        2/1/98                                    C-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1      direction. For each generation, alveolization is considered to be complete when the number of
 2      alveoli in that generation reaches the number determined by Weibel's model.
 3
 4      C. 5.3.3.2. Airway size
 5            Four sets of data are used to determine airway size during postnatal growth: (a) total lung
 6      volume as a function of age; (b) airway size as given by Weibel's model; (c) the growth pattern
 7      of the bronchial airways; and (d) variation in alveolar size with age. From these data, it is found
 8      that the lung volume, LV(t) at age t, normalized to Weibel's model at 4,800 cm3 for an adult (25
 9      years old), follows the equation
              LV(f) = 0.959  x 105(1  - 0.998e"00020   (cm3).                              (C-63)
10            The growth patterns of the bronchial airways are determined by the following equations
                      £>.(0 - D.w  = a.[H(t)  - tf(25)],                                     (C-64)

                      I.(0 - L.w = p.[tf(0  -  tf(25)],                                     (C-65)
1 1      where Dj(t) and L;(t) are, respectively, the airway diameter and length at generation i and age t,
1 2      Diw and Liw the corresponding values for Weibel's model, a; and P, are coefficients given by
                                                                                        (C-66)
                   a. = 3.26 x  l(T2exp[-1.183  (/ + 1)05]

                    p. =  l.OSxKT6 exp [10.1] O' + l)'02]
                                                                                         (C-67)
1 3

1 4
1 5
1 6
17
18
        and H(t) is the body height, which varies with age t in the form
                 H(t) = 1.82 x 102(1 - 0.725e~0140 (cm).
              For the growth patterns of the airways in the alveolar region, it is assumed that

               _L = _L  = _f
                                                                                         (C-68)

                        iw

                                           for 17 < i < 23
                                                                                         (C-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(f) ~ E ~Di
i = 04
23
^-v 71 2
.f17T (W IW ''




36 aw A
                                                                                         (C-70)
        2/1/98
                                                 C-18
                                                            DRAFT-DO NOT CITE OR QUOTE

-------
 1      C.6.  TRANSPORT RATES
 2            The values of transport rates 4r for rats have been derived from the experimental data of
 3      clearance for diesel soot (Chan et al, 1981; Strom et al., 1987, 1988) and for the particle
 4      associated organics (Sun et al., 1984; Bond et al., 1986; Yu et al., 1991).  These values are used
 5      in the present model of lung burden calculation and are listed below:
                        A.jjc - 1.73  (/ = 1,2,3)                                        (C-71)

                   *S = *S =  *2  =  *£•= 0.00018                                   (C-72)

                                    =tfi = 0-0129    •      .-              '         (C-73)
                    1(3) _  1(3) _ j(3) _ ,(3)  _
                    Afffl ~  Affl - *-LB ~~ A-AB  ~  12'55                                    (C-74)

                      A.%  = 0.693     (/ = 1,2,3)                                     (C-75)

                 XfL  = 0.00068  [1 - exp(-0.046mj62)]                                 (C-76)

                      ^i  -j A«       (/ = 2,3) -                                     (C-77)

                    Xjl =  0.012 exp(-0.11sij'76) +
                                      ,62                                             (C-78)
                 0.00068 exp(-0.046m;  )   (i  = 1,2,3)
                       i(i>  _  iW ...  iO) .  id)
                     !  A/(   ~~  A;»/   A>!r   A/IH  ~
                                    xr         ,            .
                    0.012 exp(-0.11j»x   )  + 0.00086
                 X«  + X® + AS =  0.012  exp(-0.11Wj-76)  +                       •   '
                                      1 62                                             (C-oU)
                 0.00068 exp(-0.046ma  )  + 0.0161
                                        162                                           (C-81)
                   0.00068 exp(-0.046mj)  + 15.7
1
2     where ^ is the unit of day"1, and mA = m f^is the particle burden (in mg) in the alveolar
3     compartment.
4            Experimental data on the deposition and clearance of DEPs in humans are not available.
5     To estimate the lung burden of DEPs for human exposure, it is necessary to extrapolate the
6     transport rates /i§ from rats to humans.  For organics, we assume that the transport rates are the
7     same for rats and humans. This assumption is, based upon the observation of Schanker et al.
8     (1986) that the lung clearance of inhaled lipophilic compounds appears to depend only on their

      2/1/98                                  C-19      DRAFT-DO NOT CITE OR QUOTE

-------
 1     lipid/water partition coefficients and is independent of species. In contrast, the transport rates of
 2     diesel soot in humans should be different from that of rats, since the alveolar clearance rate, A,A,
 3     of insoluble particles at low lung burdens for human adults is approximately seven times that of
 4     rats (Bailey et al., 1982), as previously discussed in Chapter 4.
 5            No data are available on the change of the alveolar clearance rate of insoluble particles in
 6     humans due to excessive lung burdens. It is seen from equation C-79 that X J^for rats can be
 7     written in the form
                         A.J? = a exp(-6iO  + d      •                 .                  (C-82)

 8     where a, b, c, and d are constants.  The right-hand side of equation C-82 consists of two terms,
 9     representing, respectively, macrophage-mediated mechanical clearance and clearance by
10     dissolution. The first term depends upon the lung burden, whereas the second term does not. To
11     extrapolate this relationship to humans, we assume that the dissolution clearance term was
12     independent of species and that the mechanical clearance term for humans varies in the same
13     proportion as in rats under the same unit surface particulate dose. This assumption results in the
14     following expression for A ^in humans
                       tf} = j e*v(-b(mAlsr) +  d                                      (C-83)

15     where P is a constant derived from the human/rat ratio of the alveolar clearance rate at low lung
16     burdens, and S is the ratio of the pulmonary surface area between humans and rats, equation C-
17     83  implies that rats and humans have equivalent amounts of biological response in the lung to the
18     same specific surface dose of inhaled DEPs.
19            From the data of Bailey et al. (1982), we obtain a value ofJ.% = 0.00169 day'1 for
20     humans at low lung burdens.  This leads to P = 14.4. Also, we find S=148 from the data of the
21     anatomical lung model of Schum and Yeh (1979) for rats and Weibel's model for human adults.
22     For humans less than 25 years old, we assume the  same value for P, but S is computed from the
23     data of the lung model for young humans (Yu and Xu 1987).  The value of S for different ages is
24     shown in Table C-3.
25            The equations for other transport rates that have a lung-burden-dependent component are
26     extrapolated from rats to humans in a similar manner. The following lists the values of X^ (in
27     day"1) for humans used in the present model calculation:
28                        A<£'=  1.73 (/ = 1,2,3)                                         (C-84)
        2/1/98                                    C-20       DRAFT-DO NOT CITE OR QUOTE

-------
  TABLE C-3. RATIO OF PULMONARY SURFACE AREAS BETWEEN HUMANS
             AND 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
2/1/98  -                       C-21     DRAFT-DO NOT CITE OR QUOTE

-------
                     I *• •* —  i'-"-'  — i'-"-'  — i1---' — r\ r\r\/\ 1 o                                    //~^ O C\

                     A™ = A?* = AS  = AS  = 0-0129                                     (C-86)

                     A2s =  AS = AS = AS = 12-55                                     (C-87)

                       A(^  - 0.693      (i -  1,2,3)                                       (C-88)

              2  = 0.00068  {1  -  0.0694 exp[-0.046(mx/5)162]}                             (C-89)

                       Ai'i  = ~ A5l      (f = 2,  3)                                       (C_90)

                AjJ. = 0.0694 {0.012 exp[-0.11(OT75)176] +
                                                                                        CC-91)
                0.00068 exp[-0.046(7n/S)176]} (i  = 1,  2, 3)


              0.0694 {0.012 exp[-0.11(OTy5)176]}  + 0.00086
                        ^(2)  _ ^(2) + ^(2) + ^(2)

                   0.0694{0.012 exp[-0.110«/4)176] +                                    (C-93)
                  0.00068  exp[-0.046(mx/1S')176]} + 0.016


                   0.0694  {0.012 exp[-0.11(7n AS)176]  +                                   (C-94)
                   0.00068 exp[-0.046(7«yS)176}  + 15.7
                                            x
                                           1
 ^      C.I. RESULTS
 2      C.7.1. Simulation of Rat Experiments
 3            To test the accuracy of the model, simulation results are obtained on the retention of
 4      diesel soot in the rat lung and compare with the data of lung burden and lymph node burden
 5      obtained by Strom et al. (1988).  A particle size of 0.19 um MMAD and a standard geometric
 6      deviation, og of 2.3 (as used in Strom's experiment) are used in the calculation.
 7            The respiratory parameters for rats are based on their weight and calculated using the
 8      following correlations of minute volume, respiratory frequency, and growth curve data.
 9                                 Minute Volume = 0.9W (cnrVmin)                      (C-95)
10                              Respiratory Frequency = 475W03 (1/min)                   (C-96)
11      where W is the body weight (in grams) as determined from the equation
12              .          •        W = 5+537T/(100+T), for T>56 days                     (C-97)
13      in which T is the age of the rat measured in days.
        2/1/98   - .                              C-22        DRAFT-DO NOT CITE OR QUOTE

-------
  1             Equation C-95 was obtained from the data of Mauderly (1986) for rats ranging in age
  2      from 3 mo to 2 years old: equation C-96 was obtained from the data of Strom et al. (1988); and
  3      equation C-59 was determined from the best fit of the experimental deposition data. Figures C-3
  4      and C-4 show the calculated lung burden of diesel soot (m %+ m ($ and lymph node burden,
  5      respectively, for the experiment by Strom et al. (1988) using animals exposed to DEPs at 6
  6      mg/m3 for 1,3,6 and 12 weeks; exposed in all cases was 7 days/week and 20 h daily. The solid
  7      lines represent the calculated accumulation of particles during the continuous exposure phase and
  8      the dashed lines indicate calculated post exposure retention. The agreement between the
  9      calculated and the experimental data for both lung and lymph node burdens during and after the
10      exposure periods was very good.
11             Comparison of the model calculation and the retention data of particle-associated BaP in
12      rats obtained by Sun et al. (1984) is shown in Figure C-5. The calculated retention is shown by
13      the solid line. The experiment of Sun et al, consisted of a 30 min exposure to diesel particles
14      coated with [3Hl benzo[a]pyrene ([3H] - BaP) at a concentration of 4 to 6 |jg/m3 of air and
15      followed by a post exposure period of over 25 days. The fast and slow phase of ([3H] - BaP)
16      clearance half-times were found to be 0.03 day and 18 days, respectively. These correspond to A
17     A20 = 0.0385 day"1 and A% = 23.1 day"1 in our model, where Ac/0 is the value of A & at mA - 0.
18      Figure C-5 shows that the calculated retention is in excellent agreement with the experimental
19      data obtained by Sun etal. (1984).
20
21      C.7.2.  Predicted Burdens  in Humans
22             Selected results of lurig burden predictions in humans are shown in Figures C-6 to C-9.
23      The particle conditions used in the calculation are 0.2 urn MMAD with og = 2.3, and the mass
24      fractions of the rapidly and slowly cleared organics are each 10% (f, = f2 = 0.1).  Figures C-6
25      and C-7 show, respectively, the lung burdens per unit concentration of diesel soot and the
26      associated organics in human adults for different exposure patterns at two soot concentrations,
27      0.1 and 1 mg/m3. The exposure patterns used in the calculation are (a) 24 h/day and 7 days week;
28      (b) 12 h/day and 7 days/week; and (c) 8 h/day and 5 days/week, simulating environmental and
29      occupational exposure conditions.  The results show that the lung burdens of both diesel soot and
30      the associated organics reached a steady state value during exposure. Due to differences in the
31      amount of particle intake, the steady state lung burdens per unit concentration were the highest
32      for exposure pattern (a) and the lowest for exposure pattern (b). Also, increasing soot
33      concentratiion from 0.1 to 1 mg/m3 increased the lung burden per unit concentration. However,
34      the increase was not noticeable for exposure pattern (c). The dependence of lung burden oh the
35      soot concentration is- caused by the reduction of the alveolar clearance rate at high lung burdens
36      discussed above.
        2/1/98                                   C-23       DRAFT-DO NOT CITE OR QUOTE

-------
                                                         - J2wk
                                             - - .Q - « _ _J wk

                                        26           39
                                         Time, week
                            52
Figure C-3.  The experimental and predicted lung burdens of rats to DEPs 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).
2/1/98
C-24
DRAFT-DO NOT CITE OR QUOTE

-------
                       <3
26           39
 Time, week
                         5E
                           €5
Figure C-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);
2/1/98
C-25
DRAFT-DO NOT CITE OR QUOTE

-------
           0,8
        G
        C
           0.6  -
           04
          0.2 u
                                   10
         15
     Time, day
20
30
Figure C-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.
2/1/98
C-26      DRAFT-DO NOT CITE OR QUOTE

-------
Figure C-6.  Calculated lung burdens of diesel soot per unit exposure concentration in
            human adults exposed continuously to DEPs 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.
2/1/98
C-27
DRAFT-DO NOT CITE OR QUOTE

-------
                                     4             6
                                       Time, year
Figure C-7.  Calculated lung burdens of the particle-associated organics per unit
            exposure concentration in human adults exposed continuously to DEPs at
            two different concentrations of 0.1 and 1.0 mg/rf. 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.       •
2/1/98
C-28      DRAFT-DO NOT CITE OR QUOTE

-------
                                                 1 mg/cu. m
                                        10           15
                                          Age, year
                            20
                               25
Figure C-8.  Calculated lung burdens of diesel soot per gram of lung per unit exposure
            concentration in humans of different ages exposed continuously for one year
            to DEPs of two different concentrations of 0.1 and 1.0 mg/nf for 7
            days/week and 24 h daily.
2/1/98
C-29
DRAFT-DO NOT CITE OR QUOTE

-------
 c
 .2
 73
 w

 o
 jc  a
 o o

•i-i
 o it.
 C  OJ

 I!
 0
 •o
 *»
 o
 »
 3
                                    1 mg/cu. m
O.OOS  -
0.006  -
0,004
0,002  -
                                    10           15
                                       Age, year
                                                      20
25
 Figure C-9.   Calculated burdens of the particle-associated organic* per gram of lung per
              unit exposure concentration in humans of different ages exposed
              continuously for one year to DEPs of two different concentrations of 0.1 and
              1.0 mg/m3 for 1 days/week and 24 h daily.
  2/1/98
                               C-30
                                                DRAFT-DO NOT CITE OR QUOTE

-------
       Figures C-8 and C-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 one 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 the 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.

C.8. PARAMETRIC STUDY OF THE MODEL
       The deposition and clearance model of DEPs in humans, presented above, consists of a
large number of parameters which 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.

C.8.1.  Effect of Ventilation Conditions
       The change in lung burden due to variations in tidal volume and respiratory frequency are
depicted in Figures C-10 and C-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 DEPs 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.

C.8.2.  Effect of Transport Rates
       Transport rates have an obvious effect on the retention of DEPs 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 A j^and A, J^are studied.

2/1/98   .                                C-31       DRAFT-DO NOT CITE OR QUOTE

-------
                         0.4             0.5             0.6
                                 Tidal Volume, liter
                                                         0.7
Figure C-10.
Calculated lung burdens in human adults versus tidal volume in liter for
exposure to DEPs at 0.1 mg/m3 for 10 years at 7 days/week and 24 h daily.
Parameters used in the calculation are: (a) MM AD=0.2 jon, og=2.3,
/2=0.1, /3=0.1, (b) respiratory frequency = 14 min*1, and (c) lung volume
= 3000 cm3.
2/1/98  •
                        C-32
DRAFT-DO NOT CITE OR QUOTE

-------
      60
       50
       40
    O)

    —• 30
    O
    O
    w
      20
       10  -
      ' 0
         10
                                  Soot
          Orgarucs
12             14             16
  Respiratory Frequency, 1/min
                                                                       1.4
                                                                       1.2
                       0.8 w
                        •   "c.

                       °'6 §


                       0.4


                       0.2
                      18
Figure C-ll.  Calculated lung burdens in human adults versus respiratory frequency in
             bpm for exposure to DEPs at 0.1 mg/m3 for 10 years at 7 days/week and 24
             h daily. Parameters used in the in the calculation are: (a) MMAD=0.2
             jim, og=2.3, /2=0.1, /3=0.1, (b) tidal volume =  500 cm3, and (c) lung
             volume = 3200 cm3.
2/1/98
              C-33
DRAFT-DO NOT CITE OR QUOTE

-------
1      Experimental data of A r|'from various diesel studies in rats have shown that A ^/;can vary by a
2      factor of two or higher.  We use a multiple of 0.5 to 2 for the uncertainty in A ^and A(%\o
3      examine the effect on lung burden.  Figures C-12 and C-13 show respectively, the lung burden
4      results for diesel soot and the associated organics versus the multiples of A ^and A ^used in the
5      calculation. As expected, increasing the multiple of A ^reduced the lung burden of diesel soot
6      with practically no change in the organics burden (Figure C-12), while just the opposite occurred
7      when the multiple of A ^\vas increased (Figure C-13).
       2/1/98                                   C-34       DRAFT-DO NOT CITE OR QUOTE

-------
                       0.8
   1.2
Multiple of
1.4      1.6
 i 0)
1.8
Figure C-12.  Calculated lung burdens in human adults versus multiple of /! f^for
             exposure to DEPs at 0.1 mg/n% for 10 years at 7 days/week and 24 h daily.
             Parameters used in the calculation are: (a) MMAD=0.2 tan, og=2.3,
             /2=0.1, /3=0.1, (b) tidal volume = 500 cm3, respiratory frequency = 14
             min-V and (c) lung volume = 3,200 cm3.
2/1/98
 C-35
DRAFT-DO NOT CITE OR QUOTE

-------
             0.6      0.8
1        1.2      1.4      1.6
      Multiple of XA(2)
              1.8
Figure C-13.  Calculated lung burdens in human adults versus multiple
             exposure to DEPs at 0.1 mg/m3 for 10 years at 7 days/week and 24 h daily.
             Parameters used in the calculation are: (a) MM AD=0.2 /on og=2.3,
             /2=0.1,/3=0.1,  (b) tidal volume = 500 cm3, respiratory frequency = 14
             min-1, and (c) lung volume = 3,200 cm3.
2/1/98
        C-36
DRAFT-DO NOT CITE OR QUOTE

-------
  1      C.9, REFERENCES                •   •    •

  2
  3      Amann, C.A.; Siegla, D.C. (1982) Diesel particles -- What are they and why. Aerosol Sci. Tech. 1:73-101
  4  -
  5      Bailey, MR.; Fry, F.A.; James, A.C. (1982) The long-term clearance kinetics of insoluble particles from the human
  6      lung, Ann. Occup. Hyg. 26:273-289.
  7
  8      Bond, J.A.; Sun, J.D,; Medinsky, M.A.; Jones, R.K.; Yeh, H.C. (1986a) Deposition, metabolism and excretion of 1-
  9      [14C]nitropyrene and l-[HC]nitropyrene coated on diesel exhaust particles as influenced by exposure concentration.
 10      Toxicol. Appl.  Pharmacol. 85:102-117.
 11
 12      Chan, T.L.; Lee, P.S.; Hering, W.E. (1981) Deposition and clearance of inhaled diesel exhaust particles in the
 13      respiratory tract of Fisher rats. J. Appl. Tox. 1:77-82.
 14
 15      Diu, C.K.; Yu,  C.P. (1983) Respiratory tract deposition of polydisperse aerosols in humans. Am. Ind. Hyg. Assoc. J.
 16      44:62-65.
 17
 18      ICRP Publication 30, part 1. 1979. Limits for intakes of radionuclides by workers. Ann ICRP 2.
 19
 20      Schanker, L.S.; Mitchell, E.W.; Brown, R.A. (1986) Species comparison  of drug absorption from the lung after
 21      aerosol inhalation or intratracheal injection. Drug Metab. Dispos. 14(l):79-88.
 22
 23      Scheutzle, D. (1983) Sampling of vehicle emissions for chemical analysis and biological testing. Environ. Health
 24      Perspect. 47:65-80.
 25
 26      Schum, M.; Yeh, H.C. (1979) Theoretical evaluation of aerosol deposition in anatomical models of mammalian lung
 27      airways. Bull. Math. Biol. 42:1-15.'
 28
 29      Snyder WS.  1975. Report of task group on reference man. pp. 151-173. Pergamon Press, Oxford, London.
 30
31      Solderholm SC. 1981. Compartmental analysis of diesel particle kinetics  in the respiratory system of exposed
 32      animals. Oral presentation at EPA Diesel Emissions Symposium, Raleigh, NC, October 5-7. In; Vostal JJ, Schreck
 33      RM, Lee PS, Chan TL, Soderholm SC. 1982. Deposition and clearance of diesel particles from the lung. In:
 34      Toxicological Effects of Emissions from Diesel Engines (Lewtas J, ed.) pp, 143-159. Elsevier, New York, NY.
 35
 36      Strom, K.A.; Chan, T.L.; Johnson, J.T. (1987) Pulmonary retention of inhaled submicron particles in rats: diesel
 37      exhaust exposures and lung retention model. Research Publication GMR-5718, Warren, MI: General Motors
 38      Research Laboratories.
 39
 40      Strom, K.A.; Chan, T.L.; Johnson, J.T. (1988) Inhaled particles VI. Dodgson, J.; McCallum, R.I.; Bailey, M.R.;
 41      Fischer, D.R., eds. London: Pergamon Press, pp. 645-658.
 42
 43      Sun, J.D.; Woff, R.K.; Kanapilly, G.M.;  McClellan, R.O. (1984) Lung retention and metabolic fate of inhaled  '
 44      benzo(a)pyrene associated with diesel exhaust particles. Toxicol. Appl. Pharmacol. 73:48-59.
 45
 46      Weibel,  E.R. (1963) Morphometry of the human lung. Berlin: Springer-Verlag.
 47
 48      Yu, C.P. (1978) Exact analysis of aerosol deposition during steady breathing. Powder Tech. 21:55-62.
 49
 50      Yu, C.P.; Diu, C.K.; Soong, T.T. (1981) Statistical analysis of aerosol deposition in nose and mouth. Am. Ind. Hyg.
 51   '   Assoc. J. 42:726-733.
 52
 53      Yu, C.P.; Xu, G.B. (1986) Predictive models for deposition of diesel exhaust participates in human and rat lungs.
 54      Aerosol Sci. Tech. 5:337-347.
 55

         2/1/98                                          C-37        DRAFT-DO NOT CITE OR QUOTE

-------
1      Yu. C.P.; Xu, G.B, (1987) Predicted deposition of diesel particles in young humans. J, Aerosol Sci. 18:419-429.
2
3      Yu, C.P., Yoon, K.J., Chen, Y.K. (1991) Retention modeling of diesel exhaust particles in rats and humans. J.
4      Aerosol Medicine. In press.
        2/1/98                                      C-38       DRAFT-DO NOT CITE OR QUOTE

-------
                                       TECHNICAL REPORT DATA
                               (Please read Instructions on the reverse before compl
 1. REPORT NO.
   EPA/600/8-90/057C
                                2.
4. TITLE AND SUBTITLE

  Health Assessment Document for Diesel  Emissions
   (SAB  Review  Draft)
                                                                  5. REPORT DATE
                                                                    February 1998
                                                                  6. PERFORMING ORGANIZATION CODE
                                                                    EPA/600/021
 7. AUTHOR(S)

   See Author's Page
                                                                 8. PERFORMING ORGANIZATION REPORT NO
                                                                   NCEA-W-0114
9. PERFORMING ORGANIZATION NAME AND ADDRESS
                                                                  10. PROGRAM ELEMENT NO.
  National  Center  for Environmental  Assessment-Washington
  Office of Research  and Development
  U.S.  Environmental  Protection Agency
  Washington.  DC 20460
                                                                 11. CONTRACT/GRANT NO.
 12. SPONSORING AGENCY NAME AND ADDRESS
                                                                  13. TYPE OF REPORT AND PERIOD COVERED
                                                                    Health Assessment,
                                                                  14. SPONSORING AGENCY CODE
 15. SUPPLEMENTARY NOTES
16. ABSTRACT
 The U.S. EPA has prepared an assessment of the possible health hazards from human exposure to diesel
 engine exhaust emissions.  This 1998 assessment is an update of a December 1994 assessment which was
 reviewed in 1995 by the Agency's Clean Air Scientific Advisory Committee (CASAC). The 1998 assessment
 focuses on health hazards (hazard identification and dose-response analysis for the purpose of recommending
 measures of risk/hazard), and also provides background information about diesel engine emissions that is
 useful for putting the health information into context. EPA risk assessment methods and practice have been
 followed in identifying possible human chronic health hazards for adverse noncancer effects as well as
 carcinogenicity hazards.  This assessment will under go a second review by the CASAC'in the Spring 1998,
 and following a consideration of comments, a final document will be issued later in 1998. Notice of the
 CASAC review, opportunity for public comment, and availability of the draft document will be made in a
 separate notice. The assessment document will be available via internet (grww_.ej^tj•%oy/nc_ea ) and from NTIS
 in mid to late March, 1998.
 7.
                                   KEY WORDS AND DOCUMENT ANALYSIS
                    DESCRIPTORS
                                                   b.IDENTIFIERS/OPEN ENDED TERMS
                                                                               c.  COSATI Field/Group
 8. DISTRIBUTION STATEMENT

 Release to the Public
                                                  19. SECURITY CLASS (ThisReport}
                                                    Unclassified
21. NO. OF PAGES

   573	
                                                   20. SECURITY CLASS (Thispage)
                                                     Unclassified
                                                                                22. PRICE
EPA Form 2220-1 (R«v. 4-77)   PREVIOUS EDITION is OBSOLETE

-------