0 CQA                              EPA #9345.4-06
                                              October 2003
                      FINAL DRAFT:

        TECHNICAL SUPPORT DOCUMENT FOR A
       PROTOCOL TO ASSESS ASBESTOS-RELATED
                           RISK
                         Prepared for:

               Office of Solid Waste and Emergency Response
 ~    r rf          U.S. Environmental Protection Agency
V £ y&9C              Washington, DC 20460

-------
                                    NOTICE

This document provides guidance to EPA>taff.  It also provides guidance to the public and to the
regulated community on how EPA intends to exercise its discretion in implementing the National
Contingency Plan. The guidance is designed to implement national policy on these issues. The
document does not, however, substitute for EPA's statutes or regulations, nor is it a regulation
itself. Thus, it cannot impose legally-binding requirements on EPA, States, or the regulated
community, and may not apply to a particular situation based upon the circumstances.  EPA may
change this guidance in the future, as appropriate.

-------
                  U.S.  Environmental Protection Agency
                                            Authors
 D. Wayne Herman
 Aeolus, Inc.
 751  TaftSt.
 Albany, CA 94706
    Kenny S. Crump
    Environ Corporation
    602 East Georgia Ave.
    Ruston, LA 71270
                                    Contributors/Reviewers
 Affiliation: McGill University
 Name: Bruce Case, Associate Professor
 Location:  462 Argyle Avenue Westmount; Quebec
 H3Y 3B4 Canada
 Phone: 514-398-7192 #00521
 Fax: 514-398-7446
 Email: bruce.case@mceill.ca

 Affiliation: Pathology & Physiology Research Branch
 National Institute for Occupational Safety & Health
 Name: Vincent Castranova, Chief
 Location:  1095 Willowdale Road (L 2015)
 Morgantown, WV 26505
 Phone: 304-285-6056
 Fax: 304-285-5938
 Email: vicl@cdc.gov

 Affiliation: Department of Medicine National Jewish
 Medical Research Center
 Name: James Crapo, Chairman
 Location:  1400 Jackson Street Denver, CO 80206
 Phone: 303-398-1436
 Fax: 303-270-2243
 Email: crapoi@nic.org

 Affiliation: Medical University of South Carolina
 Name: David Hoel, Professor
 Location:  36 South Battery Charleston, SC 29401
 Phone: 843-723-1155
 Fax: 843-723-7405
 Email: whitepoint@aol.com

 Affiliation: New York University School of Medicine
 Name: Morton Lippmann, Professor
 Location:  57 Old Forge Road Tuxedo, NY 10987
 Phone: 845-731-3558
 Fax: "845-351-5472
 Email: lippmann@env.med.nvu.edu

Affiliation: Toxicology & Human Health Risk
Analysis
Name: Roger McClellan, Advisor
Location:  13701 Quaking Aspen Place, NE
Albuquerque, NM 87111
Phone: 505-296-7083
Fax: 505-296-9573
Email: roger. o. mccl el lan@att. net
Affiliation: Price Associates, Inc.
Name: Bertram Price
Location:  1 North Broadway - #406 White Plains, NY
10601
Phone: 914-686-7975
Fax: 914-686-7977
Email: bprice@priceassociatesinc.com

Affiliation: California Environmental Protection Agency
Name: Claire Sherman, Biostatistician
Location:  1515 Clay Street, 16th Floor Oakland, CA
94612
Phone: 510-622-3214
Fax: 510-622-3211
Email: csherman@oehha.ca.gov

Affiliation: Risk Evaluation Branch National Institute for
Occupational Safety & Health
Name: Leslie Thomas Stayner, Chief
Location:  Robert Taft Laboratories, C15 4676 Columbia
Parkway Cincinnati, OH 45226
Phone: 513-533-8365
Fax: 513-533-8224
Email: Its2@cdc.gov

Affiliation: Rollins School of Public Health, Emory
University
Name: Kyle Steenland, Professor
Location:  1518 Clifton Road Atlanta, GA 30322
Phone: 404-712-8277
Email:  nsteenl@snh.emorv.edu

Affiliation: Exponent, Inc.
Name:  Mary Jane Teta, Principal Epidemiologist
Location:  234 Old Woodbury Road Southbury, CT
06488
Phone: 203-262-6441
Fax: 203-262-6443
Email:  iteta@exponent.com

-------
                        ACKNOWLEDGMENTS

We wish to acknowledge the guidance and support for this project provided by Aparna
Koppikar, Richard Troast, Chris Weis, and Paul Peronard (all of U.S. EPA).

We wish to gratefully acknowledge the assistance of Nick de Klerk, FDK Liddell, J. Corbett
McDonald, Terri Schnoor, and John Dement for graciously providing the raw epidemiology data
from several key studies, without which we could not have completed our analysis.

We also wish to thank John Addison, John Dement, Agnes Kane, Bruce Case, Michel Camus,
and Vanessa Vu for their input and assistance.

Finally, we wish to acknowledge Eric Hack and Tammie Covington for their assistance with
calculations and data management and Mark Follansbee and Joanne White  for their assistance
with technical editing and managing production of the document.
                                         tv

-------
                                   ACRONYMS
AM          alveolar macrophages
AP           alkaline phophatase
ATSDR       American Toxic Substances
              Disease Registry

BAL          bronchio-alveolar lavage
BrdU         bromodeoxyuridine

CalEPA       California Environmental
              Protection Agency
CFE          colony-forming efficiency
CHO         Chinese hamster ovary

EDXA        energy dispersive X-ray analysis
EGFR        Epithelial Growth Factor
              Receptor
ERK          EGFR-regulated kinase
ESR          electron spin residence

FBP's        fibrin breakdown products

HAF          human amniotic fluid
HBE          human bronchiolar epithelial
HNE          human neutrophil elastase
HTE          hamster tracheal epithelial

IFF           idiopathic pulmonary fibrosis
IRIS          Integrated Risk Information
              System

KGF          kertinocyte growth factor

LDH          lipid dehydrogenase
LPS          lipopolysaccharide

MAPK        mitogen activated protein kinase
MI           midget impinger
MLE          maximum likelihood estimate
MMVF's      man-made vitreous fibers
MnSOD       manganese-containing
           .   superoxide
mpcf         millions of particles per cubic
              foot
mppcf        millions of dust particles per
              cubic foot
MSHA        Mine Safety and Health
              Administration
              bronchioepithelial
NHIS         National Health Interview
              Survey
NIOSH        National Institute for
              Occupational Safety and Health

OSHA        Occupational Safety and Health
              Agency
PARS         poly-ADP-ribosyl transferase
PCM         phase contrast microscopy
PE            pulmonary epithelial
PMN         polymorphonucleocyte

RBC's        red blood cells
RCF          refractory ceramic fiber
RNS          reactive nitrogen species
ROS          reactive oxygen species
RPM         rat pleural mesothelial
RR           relative risk

SAED        selected area electron diffraction
SEM         scanning electron microscopy
SHE          Syrian Hamster Embryo
SMG         small mucous granule
SMRs         standardized mortality ratios

THE          tracheal epithelial cells
TEM         transmission electron
              microscopy
TGF-p        transforming growth factor beta
TNF-a        tumor necrosis factor alpha

uPA          urokinase-type plasminogen
              activator
uPAR         urokinase-type plasminogen
              activator
U.S. EPA      U.S. Environmental Protection
              Agency

VC AM-1      vascular cell adhesion molecule
NAC          N-acetylcysteine
NHBE        normal human

-------
                      TABLE OF CONTENTS

1.0 EXECUTIVE SUMMARY  	  1.1

2.0 INTRODUCTION 	2.1

3.0 OVERVIEW	3.1

4.0 BACKGROUND	4.1
      4.1   THE MINERALOGY OF ASBESTOS	4.2
      4.2   MORPHOLOGY OF ASBESTOS DUSTS	4.3
      4.3   CAPABILITIES OF ANALYTICAL TECHNIQUES USED TO MONITOR
           ASBESTOS 	4.4
      4.4   THE STRUCTURE AND FUNCTION OF THE HUMAN LUNG	4.12
           4.4.1  Lung Structure	4.12
           4.4.2  The Structure of the Mesothelium .'.	4.16
           4.4.3  Cytology	,	4.17
           4.4.4  Implications	4.18

5.0 THE ASBESTOS LITERATURE  	5.1
      5.1   HUMAN EPIDEMIOLOGY STUDIES 	5.2
      5.2   HUMAN PATHOLOGY STUDIES  	5.6
      5.3   ANIMAL STUDIES  	5.8
      5.4   IN VITRO STUDIES 	5.9

6.0 SUPPORTING EXPERIMENTAL STUDIES	6.1
      6.1   FACTORS AFFECTING RESP1RABILITY AND DEPOSITION 	6.2
           6.1.1  Respirability of Spherical Particles	6.4
           6.1.2  Respirability of Fibrous Structures 	6.5
           6.1.3  The Effects of Electrostatic Charge on Particle Respirability	6.10
           6.1.4  General Conclusions Concerning Particle Respirability	6.11
      6.2   FACTORS AFFECTING DEGRADATION, TRANSLOCATION, AND
           CLEARANCE	6.12
           6.2.1  Animal Retention Studies	6.17
                 6.2.1.1    Studies involving short-term exposures	6.18
                 6.2.1.2    Studies involving chronic or sub-chronic exposures	6.28
           6.2.2  Animal Histopathological Studies	6.36
           6.2.3  Human Pathology Studies  	6.41
           6.2.4  Studies of Dissolution/BioDurability	6.46
           6.2.5  Dynamic Models 	6.50
           6.2.6  General Conclusions Regarding Deposition, Translocation, and
                 Clearance	6.53
      6.3   FACTORS GOVERNING CELLULAR AND TISSUE RESPONSE	6.59
           6.3.1  The Current Cancer Model 	6.64
           6.3.2  Evidence for Transformation	6.66
           6.3.3   Evidence that Asbestos Acts As a Cancer Initiator	6.79
                  6.3.3.1    Interference with mitosis	6.79
                                     VI

-------
                   6.3.3.2    Generation of reactive oxygen species (ROS)	6.82
                   6.3.3.3    Generation of reactive nitrogen species (RNS)	6.89
                   6.3.3.4    Conclusions concerning asbestos as a cancer initiator ... 6.91
             6.3.4  Evidence that Asbestos Acts As a Cancer Promoter	6.92
                   6.3.4.1    Asbestos-induced proliferation .'..;	6.93
                 '  6.3.4.2    Asbestos induced cell signaling	6.97
                   6.3.4.3    Asbestos-induced apoptosis	6.102
                   6.3.4.4    Asbestos-induced cytotoxicity	6.104
                   6.3.4.5    Association between fibrosis and carcinogenicity	6.106
                   6.3.4.6    Interaction between asbestos and smoking	6.107
                   6.3.4.7    Conclusions concerning asbestos as a promoter	6.108
             6.3.5  Evidence that Asbestos Induces an Inflammatory Response	6.109
             6.3.6  Evidence that Asbestos Induces Fibrosis	6.109
             6.3.7  Evidence that Asbestos Mediates Changes in Epithelial Permeability 6.109
             6.3.8  Conclusions Regarding the Biochemical Mechanisms of Asbestos-Related
                   Diseases	*.... 6.109
      6.4    ANIMAL DOSE RESPONSE STUDIES 	.... 6.111
             6.4.1  Injection-Implantation Studies	6.111
             6.4.2  Animal Inhalation Studies	 6.115
             6.4.3  Supplemental Inhalation Study  	6.117
             6.4.4  Conclusions Concerning Animal Dose-Response Studies  	6.126
      6.5    CONCLUSIONS FROM AN EVALUATION OF SUPPORTING
             STUDIES	6.127

7.0. EPIDEMIOLOGY STUDIES  	'	7.1
      7.1    APPROACH FOR EVALUATING THE EPIDEMIOLOGY LITERATURE  . 7.1
      7.2    LUNGCANCER	 7.4
             7.2.1  The Adequacy of the Current U.S. EPA Model for Lung Cancer	7.4
                   7.2.1.1    Exposure Dependence	7.5
                   7.2.1.2    Time Dependence 	7.8
                   7.2.1.3    Smoking-Asbestos Interaction With Respect to Lung
                             Cancer		7.12
                   7.2.1.4    Conclusions Concerning the Adequacy of the U.S. EPA  Lung
                             Cancer Model	 .*.'		7.14
             7.2.2  Estimating KL values from the Published Epidemiology Studies .... 7.15
      7.3    MESOTHELIOMA	7.23
             7.3.1  The Adequacy of the Current U.S. EPA Model for Mesothelioma ... 7.23
                   7.3.1.1    Time Dependence 	7.28
                   7.3.1.2    Exposure Dependence	7.28
                   7.3.1.3    Discussion of Adequacy of Mesothelioma Model  	7.31
             7.3.2  Estimating KM Values from Published Epidemiology Studies  	7.33
      7.4    EVALUATION OF ASBESTOS EXPOSURE INDICES	7.35
             7.4.1  Fiber Type and Size Distribution Data Available for Deriving Exposure
                   Indices	7.36
             7.4.2  Modification of Existing KL and KM to Conform to a New Exposure
                   Index  	;... 7.38
             7.4.3  Derivation of an Improved Exposure Index for Asbestos	 7.42
                   7.4.3.1    Optimizing the Exposure Index for Lung Cancer	7.43
                                        vii

-------
                7.4.3.2    Optimizing the Exposure Index for Mesothelioma	7.48
     7.5   THE OPTIMAL EXPOSURE INDEX	7.50
          7.5.1  Definition of the Optimal Index and the Corresponding Exposure-
                Response Factors	7.50
          7.5.2  Evaluation of the Optimal Exposure Index	7.50
     7.6   GENERAL CONCLUSIONS FROM QUANTITATIVE ANALYSIS OF
          HUMAN EPIDEMIOLOGY STUDIES 	7.59

8.0 DISCUSSION, CONCLUSIONS, AND RECOMMENDATIONS	8.1
     8.1   DISCUSSION AND CONCLUSIONS .	8.1
          8.1.1  Addressing Issues  	8.1
          8.1.2  Comparison with Other, Recent Risk Reviews	8.5
     8.2   RECOMMENDATIONS FOR ASSESSING ASBESTOS-RELATED
          RISKS	8.7
     8.3   RECOMMENDATIONS FOR FURTHER STUDY	8.13

9.0 REFERENCES	9.1

APPENDIX A:  UPDATE OF POTENCY FACTORS FOR LUNG CANCER (KL) AND
            MESOTHELIOMA (KM) 	A.I
     A.1   LUNG CANCER MODEL 	A.I
     A.2   MESOTHELIOMA MODEL	'.	'.	A.2
     A.3   STATISTICAL FITTING METHODS 	A.3
     A.4   SELECTION OF A "BEST ESTIMATE" OF KL AND KM	A.4
     A.5   UNCERTAINTY IN KL AND KM ...'.	A.4
          A.5.1  The Factor Fl ...	_	A.6
          A.5.2  The Factor F2	'	A.7
          A.5.3  The Factor F3	A.8
          A.5.4  The Factor F4L for Lung Cancer and F4M for mesothelioma 	A.8
          A.5.5  Combining Individual Uncertainty Factors into an Overall "Uncertainty
                Range"	;... A.8
     A.6   ANALYSIS OF INDIVIDUAL EPIDEMIOLOGY STUDIES 	A.8

APPENDIX B:  REPORT ON THE PEER CONSULTATION WORKSHOP TO DISCUSS A
            PROPOSED PROTOCOL TO ASSESS ASBESTOS-RELATED RISK ... B.I

APPENDIX C:  COMPENDIUM OF MODEL FITS TO ANIMAL INHALATION DATA IN
            SUPPORT OF THE BERMAN ET AL. (1995) STUDY AND POST-STUDY
            WORK  	C.I

APPENDIX D:  THE VARIATION IN KL VALUES DERIVED FOR CHRYSOTILE MINERS
            AND CHRYSOTILE TEXTILE WORKERS  	  D-l

APPENDIX E:  CALCULATION OF LIFETIME RISKS OF DYING OF LUNG CANCER OR
            MESOTHELIOMA FROM ASBESTOS EXPOSURE :	E-l
                                  VIH

-------
                             LIST OF TABLES

Table 4-1.    Capabilities and Limitations of Analytical Techniques Used for Asbestos
             Measurements	4.5

Table 4-2.    Comparison of Applicable Methods For Measuring Asbestos in Air	4.6

Table 6-1.    Estimation of Lung Volume and Lung Surface Area Loading Rates for Rats and
             Humans  	6.10

Table 6-2.    Relative Rates, Half-lives for Particles Cleared by the Varous Operating
             Mechanisms of a Healthy Lung	6.13

Table 6-3.    Fraction of Fibers Retained Following Chronic Exposure  	'... 6.31

Table 6-4.    Measured in vitro Dissolution Rates for Various Fibers	6.48

Table 6-5.    Putative Mechanisms by Which Asbestos May Interact with Lung Tissue to
             Induce Disease Following Inhalation  	6.60

Table 6-6a.   Sources of Various Cytokines and Other Chemical Transmitters	6.67

Table 6-6b.   Effects of Various Cytokines and Other Chemical Transmitters  	6.74

Table 6-7.    Summary Data for Animal Inhalation Experiments Conducted by Davis and
             Coworkers	6.119

Table 7-1.    Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among
             Wittenoom, Australia Miners (Deklerk 2001) Categorized by Cumulative
             Exposure Lagged  10 Years 	'	7.6

Table 7-2.    Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among
             South Carolina Textile Workers (Schnoor 2001) Categorized by Cumulative
             Exposure Lagged  10 Years 	7.7

Table 7-3.    Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among
             Wittenoom, Australia Miners (Deklerk 2001) Categorized by Years Since Last
             Exposure  	.7.9

Table 7-4.    Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among
             South Carolina Textile Workers (Schnoor 2001) Categorized by Years
             Since Last Exposure	 7.10

Table 7-5.    Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among New
             Jersey Factory Workers (Siedman et al. 1986) Categorized by Years Since First
             Exposure	7.11
                                        ix-

-------
Table 7-6.    Lung Cancer Exposure-Response Coefficients (KL) Derived from Various
             Epidemiological Studies  ........... '. ............. ................ 7.1 7

Table 7-7.  .  Fit of EPA Mesothelioma Model to Observed Mesothelioma Mortality Among
             Wittenoom, Australia Miners (Deklerk 2001) Categorized by Years Since First
             Exposure  [[[ 7.24

Table 7-8.    Fit of EPA Mesothelioma Model to Observed Mesothelioma Mortality Among
             Wittenoom, Australia Miners (Deklerk 2001) Categorized by Average Value of
             Integral (Equation 6-12) .......................................... 7.25

Table 7-9.    Mesothelimoa Exposure-Response Coefficients (KM) Derived from Various
             Epidemiological Studies  ......................................... 7.27

Table 7-10.   Fit of EPA Mesothelioma Model to Observed Mesothelioma Mortality Among
             Wittenoom, Australia Miners (Deklerk 2001) Categorized by Years Since Last
             Exposure  ......................................... ............ 7.29

Table 7-1 1 .   Fit of EPA Mesothelioma Model to Observed Mesothelioma Mortality Among
             Quebec Miners (Liddell 2001) in Each of Three Mining Areas, Categorized by
             Years Since Last Exposure ......................... .............. 7.30

Table 7.12.   Comparison of Wittenoom, Australia (DeKlerk 2001) Mesothelioma Deaths to
             Predicted Deaths Assuming Risk Varies Linearly with Exposure Intensity After
             Controlling for Years Since First Exposure and Duration of Exposure  ..... 7.31

Table 7-13.   Correlation Between Published Quantitative Epidemiology Studies and Available
             Tem Fiber Size Distributions ....................................... 7.37

Table 7-14.   Representative KL and KM Values Paired with Averaged TEM Fiber Size
             Distributions From Published Papers  ............................... 7.39

Table 7-15.   Estimated Uncertainty Assigned to Adjustment for Fiber Size ............ 7.40

Table 7-16.   Estimated Fraction of Amphiboles in Asbestos Dusts ................ ... 7.44

Table 7-17.   Results from Fitting Exposure Indices Defined by Equation 7.12 and Pcme to
             Lung Cancer and Mesothelioma Exposure-response Coefficients Estimated from
             Different Environments . . ......... . .............................. 7.47

Table 7-18.   Optimized Dose-Response Coefficients for Pure Fiber Types . . .......... 7.50

Table 7-19.   Study Specific K,, K^ K,,, K^.,  Kla, IC^, Kk> and K™ Values ............ 7.57
Table 7-20.
             Comparison of Spread in Range of Original and Adjusted K, and K,,, Values for
             Specific Fiber Types ............................................ 7.59

-------
Table 8-2.    Estimated Additional Deaths from Lung Cancer or Mesothelioma per 100,000
             Persons from Constant Lifetime Exposure to 0.0001 TEM f/cc Longer than 10 um
             and Thinner than 0.4 um - Based on Optimum Risk Coefficients
             (Table 7-18)	8.9

Table 8-3.    Estimated Additional Deaths from Lung Cancer or Mesothelioma per 100,000
             persons from Constant Lifetime Exposure to 0.0001 TEM f/cc Longer than 10 um
             and Thinner than 0.4 um - Based on Conservative Risk Coefficients
             (Table 8-1)	8.9
                                         xi

-------
                            LIST OF FIGURES

Figure 4-1.    The Structure of Lung Parenchyma Showing Alveoli and Alveolar Ducts ... 4.13

Figure 4-2.    The Structure of the Inter-Alveolar Septa	4.14

Figure 4-3.    The Detailed Structure of an Alveolar Wall that is Part of an Inter-Alveolar\
             Septum	'	4.15

Figure 6-1.    Fractions of Respirable Particles Deposited in the Various Compartments of the
             Human Respiratory Tract as a Function of Aerodynamic Equivalent
             Diameter	6.5

Figure 6-2.    Fractions of Respirable Particles Deposited in the Various Compartments of the
             Human Respiratory Tract as a Function of the True Diameter of Asbestos
             Fibers	6.7

Figure 6-3.    [Chrysotile/Lung Burden Concentration] vs. Time of Exposure	6.32

Figure 6-4.    Key for Putative Mechanisms for Clearance and Translocation of Fibers in the
             Lung	6.54

Figure 6-5.    Fit of Model. Tumor Incidence vs. Structure Concentration by TEM (Length
             Categories 5-40 ^m, >40um, Width Categories: <0.3 \im and >5 urn)  	6.122

Figure 6-6.    Fit of Model. Tumor Incidence vs. Structure Concentration by TEM (Length
             Categories 5-40 urn, >40um, Width Categories: <0.4 |am)  	6.125

Figure 7-1.    Plot of Estimated KL Values and Associated Uncertainty Intervals by Study
             Environment	7.20

Figure 7-2.    Plot of Estimated KM Values and Associated Uncertainty Intervals by Study
             Environment	7.34

Figure 7-3.    Plot of Estimated (Adjusted) KL Values and Associated Uncertainty Intervals by
             Study Environment	7.52

Figure 7-4.    Plot of Estimated (Adjusted) KM Values and Associated Uncertainty Intervals by
             Study Environment 	7.53

Figure 7-5.    Plot of Estimated KLA KLC Values and Associated Uncertainty Intervals by Study
             Environment	7.54

Figure 7-6.    Plot of Estimated KMA KMC Values and Associated Uncertainty Intervals by Study
             Environment	•.	7.55
                                          XII

-------
                      1.0 EXECUTIVE SUMMARY

The purpose of this report is to provide a foundation for completing a state-of-the-art-protocol to
assess potential human-health risks associated with exposure to asbestos.  Such a protocol is
intended specifically for use in performing risk assessments at Superfund sites, although it may
be appl icable to a broad range of situations.

The current report is a revision to a version originally submitted on September 4, 2001 (Berman
and Crump 2001), which was the subject of a peer-review consultation held in San Francisco on
February 25-26, 2003. In general, the expert panel endorsed the overall approach to risk
assessment proposed in this report, although they highlighted areas where controversies persist.

The current report incorporates the changes recommended by the peer review consultation panel
to correct minor problems with internal consistency and the overall transparency of the
discussion that are needed to improve readability.  Although some of the research and analyses
recommended by the peer consultation panel are not complete, it is anticipated that the current
document can be distributed for broader review and comment. Thus, the recommended approach
to risk assessment can be considered for use  in the interim,  while the additional research and
analyses recommended by the expert panel are completed.  At that point, a final revision of this
document will be developed and it is expected to serve as a component of a broader effort by the
U.S. Environmental Protection Agency (U.S. EPA) to revise the Agency's current approach for
assessing asbestos-related risks.

The approach currently employed at the U.S. EPA to evaluate asbestos-related risks (IRIS 1988)
is based primarily on a document completed  in 1986 (U.S. EPA 1986) and has not been changed
substantially in the past 15 years, despite substantial improvements in asbestos measurement
techniques and in the understanding of the manner in which asbestos exposure contributes to
disease. Therefore, this document provides an overview and evaluation of the more recent
studies and presents proposed modifications to the protocol for assessing asbestos-related risks
that can be justified based on the more recent work.

As reported in several  recent technical meetings and reinforced by information gleaned from the
literature, the following were identified as issues that need to be addressed to develop a protocol
for evaluating asbestos-related risk:

      •     whether the exposure-response models currently in use by the U.S. EPA for
             describing the incidence of asbestos-related diseases adequately reflect the time-
             and exposure-dependence for  the development of these diseases;

      •     whether different potencies need to be assigned to the different  asbestos mineral
             types to adequately predict risk for the disease endpoints of interest;

      •     to the extent that different asbestos mineral types are assigned distinct potencies,
             whether the relative in vivo durability of different asbestos mineral types
             determines their relative potency;
                                          1.1

-------
       •      whether the set of minerals included in the current definition of asbestos
              adequately covers the range of minerals that potentially contribute to asbestos-
              related diseases;

       •      whether the analytical techniques and methods currently used for determining
              asbestos concentrations adequately capture the biologically relevant
              characteristics of asbestos (particularly with regard to the sizes of the structures
              counted using the various analytical methods) so that they can be used to support
              risk assessment; and

       •      whether reasonable confidence can be placed in the cross-study extrapolation of
              exposure-response relationships that are required to assess asbestos-related risks
              in new environments of interest.

These outstanding issues (and other related considerations) are addressed in this document to
provide a foundation for proposing a new approach for assessing asbestos-related risks.
Although the objective of this evaluation was to identify the single best procedure, when current
knowledge is inadequate for distinguishing among alternatives, options are presented along with
a discussion of their relative advantages and limitations.  In a few cases, limited and focused
additional research studies are recommended, which may enhance the current state of knowledge
sufficiently to resolve one or more of the important, remaining issues.

Background

Inhalation of asbestos dusts has been linked to several adverse health effects including primarily
asbestosis, lung cancer, and mesothelioma (U.S. EPA 1986). Asbestosis, a chronic, degenerative
lung disease, has been documented among asbestos workers from a  wide variety of industries.
Although asbestosis cases have been observed at some locations of current interest to the U.S.
EPA, the disease is generally expected to be associated only with the higher levels of exposure
commonly found in workplace settings and is not expected to contribute substantially to
potential risks associated with environmental asbestos exposure. Therefore, asbestosis is only
considered in this document to the extent required to address its putative association with lung
cancer. Overall, the majority of evidence indicates that lung cancer and mesothelioma are the
most important risks associated with exposure to low levels of asbestos.

The Asbestos Literature

A variety of human, animal, and tissue studies have provided insight into the nature of the
relationship between asbestos exposure and disease. Ideally, human epidemiology studies  are
employed to determine the quantitative exposure-response relationships and the attendant risk
coefficients for asbestos exposure.  Exposure-response coefficients have been estimated for
asbestos from approximately 20 epidemiology studies for which adequate exposure-response
data exist.  Such coefficients vary widely, however, and the observed variation has not been
reconciled.  Among the objectives of this study is to evaluate and account for the sources of
uncertainty that contribute to the variation among the exposure-response coefficients derived
from the literature so that these estimates can be reasonably interpreted and recommendations for
their use in risk assessment developed.

                                            1.2

-------
 Animal and tissue studies indicate that asbestos potency is a complex function of several
 characteristics of asbestos dusts including fiber size and fiber type (i.e., fiber mineralogy).
 Moreover, the influence of fiber size is a complex function of both diameter and length.
 Therefore, whenever the goal is to compare across samples with differing characteristics, it is not
 sufficient to report asbestos concentrations simply as a function of mass (or any other single
 measure), which is in stark contrast to the treatment of chemical toxins. It has generally been
 difficult to distinguish among the effects of fiber size and type in many studies because such
 effects are confounded and the materials studied have not been adequately characterized.

 The Epidemiology Studies

 The existing epidemiology studies provide the most appropriate data from which to determine
 the relationship between asbestos exposure and response in humans. As previously indicated,
 however, due to a variety of methodological limitations, the ability to compare and contrast
 results across studies needs to be evaluated to determine the confidence with which results from
 existing epidemiology studies may be extrapolated to new environments where risk needs to be
 assessed.  This requires both that the uncertainties contributed by such methodological
 limitations and that several ancillary issues be addressed.

 Briefly, the major kinds of limitations that potentially contribute to uncertainty in the available
 epidemiology studies include:
       •     limitations in air measurements and other data available for characterizing
              historical exposures;

       •     limitations in the manner that the character of exposure (i.e., the mineralogical
              types of fibers and the range and distribution of fiber dimensions) was delineated;

       •     limitations in the accuracy of mortality determinations or incompleteness in the
              extent of tracing of cohort members;

       •     limitations in the adequacy of the match between cohort subjects and the selected
              control population; and

       •     inadequate characterization of confounding factors, such as smoking histories for
              individual workers.

The existing asbestos epidemiology database consists of approximately 150 studies of which
approximately 35 contain exposure data sufficient to derive quantitative exposure/response
relationships. A detailed evaluation of 20 of the most recent of these studies, which includes the
most recent follow-up for all of the cohorts evaluated in the 35 studies, was completed.  The
following conclusions result from this evaluation:

       (1)     To study the characteristics of asbestos that relate to risk, it is necessary to
              combine results (i.e., in a meta  analysis) from studies of environments having
              asbestos dusts of differing characteristics. More robust conclusions regarding risk
                                           1.3

-------
       can be drawn from an analysis of the set of epidemiology studies taken as a whole
       than results derived from individual studies.

(2)    By adjusting for fiber size and fiber type, the existing database of studies can be
       reconciled adequately to reasonably support risk assessment.

(3)    The U.S. EPA models for lung cancer and mesothelioma both appear to track the
       time-dependence of disease at long times following cessation of exposure.
       However, the relationship between exposure concentration and response may not
       be adequately described by the current models for either disease. There is some
       evidence that these relationships are supra-linear.

(4)    Whereas the U.S. EPA model for lung cancer assumes a multiplicative
       relationship between smoking and asbestos, the current evidence suggests that the
       relationship is less than multiplicative, but possibly more complex than additive.
       However, even if the smoking-asbestos interaction is not multiplicative as
       predicted by the U.S.  EPA model, exposure-response coefficients estimated from
       the model are still likely to relate to risk approximately proportionally and,
       consequently, may be used to determine an exposure index that reconciles
       asbestos potencies in different environments.  However, adjustments to the
       coefficients may be required  in order to use them to estimate absolute lung cancer
       risk for differing amounts of smoking. This issue needs to be investigated further
       in the next draft of this document.

(5)    The optimal exposure index that best reconciles the published literature assigns
       equal potency to fibers longer than 10 ^m and thinner than 0.4 u.m and assigns  no
       potency to fibers of other dimensions.

(6)    The optimal exposure index also assigns different exposure-response coefficients
       for chrysotile and amphibole both for lung cancer and mesothelioma.  For  lung
       cancer the best estimate of the coefficient (potency) for chrysotile is 0.27 times
       that for amphibole, although  the possibility that chrysotile and amphibole are
       equally potent cannot be ruled out.  For mesothelioma the best estimate of the
       coefficient (potency) for chrysotile is only 0.0013 times that for amphibole and
       the possibility that pure chrysotile is non-potent for causing  mesothelioma cannot
       be ruled out by the epidemiology data.

(7)    Using the approach recommended in the U.S. EPA (1986) update, the lung cancer
       exposure-response coefficients (KL values) estimated from 15 studies vary by a
       factor of 72 and these values are mutually inconsistent (based on non-overlap of
       uncertainty intervals). Using the approach based on the optimal exposure  index
       that is recommended herein, the overall variation in KL values across these studies
       is reduced to a factor of 50.

(8)    Using the approach recommended in the U.S. EPA (1986) update, the
       mesothelioma exposure-response coefficients (KM values) estimated from  10
       studies vary by a factor of 1,089 and these values are likewise mutually

                                     1.4

-------
              inconsistent. Using the approach based on the optimal exposure index that is
              recommended herein, the overall variation in KM values across these studies is
              reduced to a factor of 30.

       (9)    The exposure index and exposure-response coefficients embodied in the risk
              assessment approach developed in this report are more consistent with the
              literature than the current U.S. EPA approach.  In particular, the current approach
              appears highly likely to seriously underestimate risk from amphiboles, while
              possibly overstating risk from chrysotile.  Furthermore, most of the remaining
              uncertainties regarding the new proposed approach also apply to the current
              approach. Consequently, it is recommended that the proposed approach begin to
              be applied in assessment of asbestos risk on an interim basis, while further work,
              as recommended below, is conducted to further refine the approach.

       (10)   The residual inconsistency in both the lung cancer and mesothelioma potency
              values is primarily driven by those calculated from Quebec chrysotile miners and
              from South Carolina chrysotile textile workers.  The difference in the lung cancer
              potency estimated between these studies has long been the subject of much
              attention.  A detailed evaluation of the studies addressing this issue, the results of
              our analysis of the overall  epidemiology literature, and implications from the
              broader literature, indicate that the most likely cause of the difference between
              these studies is the relative distribution of fiber sizes in the two environments.  It
              is therefore likely that the variation between these studies can be further reduced
              by developing improved characterizations of the dusts that were present in each of
              these environments (relying  on either archived samples, or newly generated
              samples using technologies similar to those used originally).

Recommendations for Risk Assessment

Although gaps in knowledge remain, a review of the literature addressing the health-related
effects of asbestos (and related materials) provides  a generally consistent picture of the
relationship between asbestos exposure and the induction of disease (lung cancer and
mesothelioma).  Therefore, the general characteristics of asbestos exposure that drive the
induction of cancer can be inferred from the existing studies and were applied to define
appropriate procedures for evaluating asbestos-related risk.

Optimum values for exposure-response coefficients for lung cancer and mesothelioma were
derived in this analysis and can be combined with appropriately defined exposure estimates as
inputs for the U.S. EPA lung cancer and mesothelioma models (respectively) to assess risk.
Although these values are optimized within the constraints of the current analysis and reduce the
apparent variation across published studies substantially, the need to manage and minimize risk
when developing a general  approach for assessing risk, is also recognized. Thus, to reduce the
chance of under-estimating risks, a conservative set of potency estimates were also developed
and are also presented. To assess risk, depending on the  specific application, either the best-
estimate risk coefficients or the conservative estimates can be incorporated into procedures
described herein for assessing asbestos-related risks.
                                           1.5

-------
Tables are also provided that present estimates of the additional risk of death from lung cancer,
from mesothelioma, and from the two diseases combined that are attributable to lifetime,
continuous exposure at an asbestos concentration of 0.0001 f/cm3 (for fibrous structures longer
than 10 \im and thinner than 0,4 jim) as determined using TEM recommended methods. The risk
estimates in these tables can be combined with appropriately determined estimates of exposure
to develop estimates of risk in environments of interest.

Recommendations for Limited, Further Study

The two major objectives identified for further study are:

       (1)    to evaluate a broader range of exposure-response models in fitting the observed
             relationship between asbestos exposure and lung cancer or mesothelioma,
             respectively. For lung cancer models, this would also include an attempt to better
             account for the interaction between asbestos exposure and smoking; and

       (2)    to develop the supporting data needed to define adjustments for exposure-
             response coefficients that will allow them to be used with an exposure index that
             more closely captures the criteria that determine biological activity. Among other
             things, this work should focus on obtaining data that would permit more complete
             reconciliation of the exposure-response coefficients derived for Quebec miners
             and South Carolina textile workers.
                                           1.6

-------
                            2.0 INTRODUCTION

 The purpose of this report is to provide a foundation for completing a state-of-the-art-protocol to
 assess potential human-health risks associated with exposure to asbestos. Such a protocol is
 intended specifically for use in performing risk assessments at Superfund sites, although it may
 be applicable to a broad range of situations.

 The current report is a revision to a version originally submitted on September 4,2001 (Herman
 and Crump 2001), which (among other things) includes both an extensive review of the general
 literature and a detailed analysis of the existing epidemiology studies. These are reproduced in
 the current report in Chapter 6 and Chapter 7/Appendix A (respectively).

 The September 4,2001 version was also the subject of a peer-review consultation held in San
 Francisco on February 25-26,2003. The comments of the expert panel convened to conduct the
 peer review are included in this report as Appendix B.

 In general, the expert panel endorsed the overall approach to risk assessment proposed in this
 report, although they highlighted areas where controversies persist. They also suggested
 additional research and analyses to attempt to resolve some of the outstanding controversies and
 to refine several of the details of the approach.  In addition, they offered recommendations for
 modifications to improve the overall transparency and readability of the earlier version of this
 report.

 The current report incorporates the changes recommended by the peer review consultation panel
 to correct minor problems with internal consistency and the overall transparency of the
 discussion that are needed to improve readability. Although some of the research and analyses
 recommended by the peer consultation panel are not complete, it is anticipated that the current
 document can be distributed for broader review and comment. Thus, the recommended approach
 to risk assessment can be considered for use in the interim, while the additional research and
 analyses recommended by the expert panel are completed. At that point, a final revision of this
 document will be developed and it is expected to serve as a component of a broader effort by the
 U.S. Environmental Protection Agency (U.S. EPA) to revise the Agency's current approach for
 assessing asbestos-related risks.

 The approach currently employed by U.S. EPA to evaluate asbestos-related risks (IRIS, 1988) is
 based primarily on a document completed in 1986 (U.S. EPA  1986) and has not changed in the
 past 15 years, despite substantial improvements in asbestos measurement techniques and in the
 understanding of the manner in which asbestos exposure contributes to disease.  Therefore,
 among other things, this document provides an overview and evaluation of more recent studies
 and presents proposed modifications to the current approach for assessing asbestos-related risks
that can be justified based on the more recent work.

In May 2001, the U.S. EPA along with the California Environmental Protection Agency
(CalEPA), the National Institute  for Occupational Safety and Health (NIOSH), the American
Toxic Substances Disease Registry (ATSDR), and the Mine Safety and Health Administration
(MSHA) hosted an international  conference on asbestos in Oakland, California that was attended
                                          2.1

-------
by leading international experts on asbestos. The state of knowledge concerning such issues as
the nature of asbestos, the measurement of asbestos, and the relationship between asbestos
exposure and the induction of disease was reviewed during this conference.  Particular emphasis
was placed on identifying important knowledge gaps in these areas.
                                                        /
By coupling the outstanding issues identified at the Oakland meeting with additional information
gleaned from the literature, the following set of issues was identified as risk-specific issues of
current interest:

       •      whether the exposure-response models currently in use by the U.S. EPA for
              describing the incidence of asbestos-related diseases adequately reflect the time-
              and exposure-dependence for the development of these diseases;

       •      whether different potencies need to be assigned to the different asbestos mineral
              types to adequately predict risk for the disease end points of interest;

       •      to the extent that different asbestos mineral types are assigned distinct potencies,
              whether the relative in vivo durability of different asbestos mineral types
              determines their relative potency;

       •      whether the set of minerals included in the current definition of asbestos
              adequately covers the range of minerals that potentially contribute to asbestos-
              related diseases;

       •      whether the analytical techniques and methods currently used for determining
              asbestos concentrations adequately capture the biologically relevant
              characteristics of asbestos (particularly with regard to the sizes  of the structures
              counted using the various analytical methods) so that they can be used to support
              risk assessment; and

       •      whether reasonable confidence can be placed in the cross-study extrapolation of
              exposure-response relationships that are required to assess asbestos-related risks
              in new environments of interest.

These outstanding issues (and other related considerations) are addressed in this document to
provide a foundation for  proposing a new approach for assessing  asbestos-related risks.
Compared to the current  U.S. EPA approach, it is shown that the  new approach better predicts
risks among the environments in which asbestos-related risks have been previously evaluated
(i.e., the published epidemiology studies) so that the new approach can be used to predict risks in
unstudied environments of interest with greater confidence than predictions based on the current
approach. Moreover, completing the additional research and analysis recommended by the
expert panel (Appendix B) should facilitate further refinement while providing additional
opportunities to better evaluate and validate the proposed approach.
                                            2.2

-------
The remainder of this document is divided into 6 chapters:

       •     Chapter 3 presents an overview of the general considerations that need to be
              addressed to assess asbestos-related risks (including considerations associated
              with the manner in which asbestos exposures are characterized, the manner in
              which risk is modeled from existing data, and the manner that risk models are
              then applied to estimate risk in new environments).  The nature of the diseases
              commonly attributed to asbestos exposure are also briefly described;

       •     Chapter 4 presents a background discussion that addresses the definition of
              asbestos, the mineralogy of asbestos, the morphology of asbestos-containing dusts
              to which people are typically exposed, the capabilities and limitations of
              analytical techniques and methods used to determine airborne asbestos
              concentrations, and the structure and function of the human lung;

       •     Chapter 5 provides a description of the kinds of literature studies that are
              commonly used to support development of a protocol to assess risk, with
              particular emphasis on identifying their relative strengths and weaknesses;

       •     Chapter 6 presents a review of the literature with particular emphasis on studies
              published since the Health Effects Assessment Update (U.S. EPA 1986).
              Combined with a description of supporting analyses, the review is focused on
              reconciling apparently conflicting studies and hypotheses (when possible) and
              identifying the best candidate procedures for assessing asbestos-related risks. To
              reconcile studies, the strengths and weaknesses common to various types of
              studies are explicitly considered;

       •     Chapter 7 presents a reevaluation of the published epidemiology studies that,
              among other things, is designed to address and (when possible) resolve the
              outstanding issues of current interest; and

       •     Chapter 8 presents a proposed, new approach for assessing asbestos-related risks.

Regarding Chapter 8, although the objective of this document was to identify the single best
procedure, when current knowledge is inadequate for distinguishing among alternatives, options
are presented along with a discussion of their relative advantages and limitations. A few limited
and focused additional research studies are recommended, which have the potential to enhance
the current state of knowledge sufficiently to resolve one or more of the important, remaining
issues. These recommended studies parallel those identified by the peer review panel (Appendix
B).

This report is part of a series of documents developed as part of a multi-task project to develop a
set of mutually consistent methods for determining asbestos concentrations in a manner useful
for assessing risk and a companion protocol for conducting such risk assessments.  A method for
the determination of asbestos in air (Chatfield and Herman 1990) and a companion technical
background document (Berman and  Chatfield 1990) were published by the U.S. EPA in 1990.
The air method has since been superceded (improved) by the ISO Method (ISO 1995). A

                                           2.3

-------
method for the determination of asbestos in soils and bulk materials (Berman and Kolk 1997)
was also published by the U.S. EPA and the draft of an improved version was also recently
completed (Berman and Kolk 2000). The recommendations in this document should serve as the
basis for development of the companion risk-assessment protocol.
                                         2.4

-------
                                3.0  OVERVIEW

 Inhalation of asbestos dusts has been linked to several adverse health effects including primarily
 asbestosis, lung cancer, and mesothelioma (U.S. EPA 1986). The kinds of lung cancers linked to
 asbestos exposure are similar to those induced by smoking and a greater-than-additive effect has
 been observed from combined exposure (see, for example, Liddell and Armstrong 2002).
 Mesothelioma is a rare cancer of the membranes that line the pleural cavity (containing the heart
 and lungs) and the peritoneal cavity (i.e., the gut). Although there is some evidence of a low
 background incidence of spontaneous mesotheliomas, this cancer has been associated almost
 exclusively with exposure to asbestos and certain other fibrous substances (HEI-AR, 1991).

 Asbestosis, a chronic, degenerative lung disease, has been documented among asbestos workers
 from a wide variety of industries.  Although asbestosis cases have been observed at some
 locations of current interest to the  U.S. EPA, the disease is generally expected to be associated
 only with the higher levels of exposure commonly found in workplace settings and is not
 expected to contribute substantially to risks potentially associated with environmental asbestos
 exposure. Therefore, asbestosis is only considered in this document to the extent required to
 address its putative association with lung cancer. Overall, the majority of evidence indicates that
 lung cancer and mesothelioma are the most important risks associated with exposure to low
 levels of asbestos.

       The primary route of exposure of concern in association with asbestos is
       inhalation. There is little evidence that ingestion of asbestos induces disease (see,
       for example, IRIS 1988; U.S. EPA 1986). Therefore, this study is focused on
       inhalation hazards, and other routes of exposure are not addressed.

 Gastrointestinal cancers and cancers of other organs (e.g., larynx, kidney, and ovaries) have also
 been linked with asbestos exposures (by inhalation)  in some studies. However, such associations
 are not as compelling as those for  lung cancer and mesothelioma and the potential risks from
 asbestos exposures associated with these other cancers are much lower (U.S. EPA 1986).
 Consequently, by addressing the more substantial asbestos-related risks associated with lung
 cancer and mesothelioma, the much more moderate risks potentially associated with cancers at
 other sites are also addressed by default. Therefore, this document is focused on the risks
 associated with lung cancer and mesothelioma.

 A variety of human, animal, and tissue studies have provided insight into the nature of the
 relationship between asbestos exposure and disease. Ideally, human epidemiology studies are
 employed to determine the quantitative exposure/response relationships and the  attendant risk
 coefficients for  asbestos exposure. Risk coefficients have been estimated for asbestos from
 approximately 20 epidemiology studies for which adequate exposure-response data exist.
 However, such coefficients vary widely (for lung cancer, coefficients vary by more than a factor
 of 70 and, for mesothelioma, they  vary by more than a factor of 1,000) and this variation has not
 been reconciled. Among the objectives of this study, one is to evaluate and account for the
 sources of uncertainty that contribute to the variation among the risk coefficients derived from
the literature so that these estimates can be reasonably interpreted and recommendations for their
 use in risk assessment developed.
                                           3.1

-------
Animal and tissue studies indicate that asbestos potency is a complex function of several
characteristics of asbestos dusts including fiber size and fiber type (i.e., fiber mineralogy).
Moreover, the influence of fiber size is a complex function of both diameter and length.
Therefore, whenever the goal is to compare across samples with differing characteristics, it is not
sufficient to report asbestos concentrations simply as a function of mass (or any other single
parameter), which  is in stark contrast to the treatment of chemical toxins.  It has generally been
difficult to distinguish among the effects of fiber size and type  in many studies because such
effects are confounded and the materials studied have not been adequately characterized.

The influence of different characteristics of asbestos dusts upon risk cannot be adequately
evaluated in the existing epidemiological studies because the analytical techniques used to
monitor asbestos exposure in these studies are not capable of resolving all of the characteristics
of asbestos dusts that other types of studies indicate are important. Moreover, the exposure
indices (the range of structure sizes and shapes used to characterize an asbestos dust) that are
employed in the existing epidemiology studies may not correspond with the characteristics of
asbestos that best relate to biological activity. This hinders the ability to reconcile risk
(exposure-response) coefficients derived from different studies. It also limits the confidence
with which risk coefficients derived from the existing epidemiology studies can be applied to
assess risks from asbestos exposure in other environments. Such limitations are explored in this
report, along with potential remedies.

Based on the current approach for evaluating asbestos-related cancer risk (U.S. EPA 1986), risk
is estimated as the product of a risk coefficient and a mathematical function that depends on the
level of exposure, the duration of exposure, and time. The risk coefficient for lung cancer is
generally denoted as, "KL" and the one for nasothelioma as "K  M". A detailed description of
both the current lung cancer and mesothelioma models is provided in Chapter 7. The models
differ depending on whether lung cancer or mesothelioma is being considered.

For lung cancer, the model estimates relative risk, which means that the increase in lung cancer
incidence that is attributable to asbestos exposure is proportional to the background lung cancer
incidence in the exposed population. The background cancer incidence is the rate of lung cancer
that would be expected to occur in the population in the absence of asbestos exposure. In other
words, background lung cancer incidence is the lung cancer rate for the exposed population that
is attributable to all causes other than asbestos.

Among the implications of the lung cancer model is that the combined effects of asbestos
exposure and smoking is multiplicative and, until recently, the majority of studies have
suggested such a multiplicative relationship (see, for example, Hammond et al. 1979). However,
newer studies (for example, Liddell and Armstrong 2002) suggest a complex relationship that is
closer to additive than multiplicative.  Such considerations are addressed further in Chapter 7.

The current EPA model for mesothelioma is an absolute risk model. This means that the
increase in mesothelioma risk attributable to asbestos is independent of the background rate of
mesothelioma, which is negligible in the general population.
                                            3.2

-------
 Ideally, the risk coefficients derived from the existing epidemiology studies can be combined
 with measurements from other exposure settings to estimate lung cancer and mesothelioma risks
 in these other exposure settings. However, such risk estimates are only valid if both of the
 following conditions are met:

       (1)    asbestos is measured in the exposure setting of interest in the identical manner in
              which it was measured in the study from which the corresponding risk
              coefficients are derived; and

       (2)    such measurements reflect the characteristics of asbestos exposures that
              determine risk.

 A growing body of evidence indicates that the way in which asbestos concentrations were
 measured in the existing epidemiology studies do not reflect the characteristics of asbestos
 exposure that determine risk. Therefore, measuring asbestos concentrations in the same way in
 exposure settings of interest may not be sufficient to assure the validity of risk estimates derived
 using the published risk coefficients (and the corresponding models).  This  is because the second
 of the above-listed conditions would not be satisfied.

 Considerations necessary to compare risk coefficients derived in different exposure settings (or
 to apply a coefficient to predict risk in a setting different from the one in which the coefficient
 was derived) have been elucidated clearly in a mathematical model (Chesson et a!. 1990). The
 consequences of the model indicate that adjusting the existing risk coefficients so that they
 reflect asbestos characteristics that determine biological activity requires knowledge of the fiber
 size distributions of the dusts studied in the original epidemiology studies.  To the extent they
 exist, such data may be used to normalize each of the published risk coefficients so that they
 relate to a common exposure index reflecting asbestos characteristics that determine biological
 activity.

 Among the goals of this evaluation is to explore the possibility of defining an improved exposure
 index (that better reflects biological activity) and to use this index to reconcile the epidemiology
 data (see Chapter 7). We also evaluated improved ways of simultaneously accounting for the
 effects of both fiber size and type.

 Unfortunately, some of the issues that need to be resolved to support development of a protocol
 for assessing asbestos-related risks cannot be entirely resolved with existing data. Therefore, in
 later chapters (i.e., Chapters 6,7, and 8), we have attempted to identify such issues, to assess
their relative importance, and, when deemed appropriate, to propose limited and focused
 research projects designed to provide the data required to reduce the impacts of such knowledge
gaps.
                                           3.3

-------
                             4.0  BACKGROUND

Asbestos is a term used to describe the fibrous habit of a family of hydrated metal silicate
minerals.  The most widely accepted definition of asbestos includes the fibrous habits of six of
these minerals (IARC 1977).  The most common type of asbestos is chrysotile, which is the
fibrous habit of the mineral serpentine, a magnesium silicate. The other five asbestos minerals
are all amphiboles (i.e., all partially hydrolyzed, mixed-metal silicates).  These are: fibrous
reibeckite (crocidolite), fibrous grunerite (amosite), anthophyllite asbestos, tremolite asbestos,
and actinolite asbestos.

All six of the minerals whose fibrous habits are termed asbestos occur most commonly in non-
fibrous, massive habits.  While unique names have been assigned to the asbestiform varieties of
three of the six minerals (noted parenthetically above) to distinguish them from their massive
forms, such nomenclature has not been developed for anthophyllite, tremolite, or actinolite.
Therefore, when discussing these latter three minerals, it is important to specify whether a
massive habit of the mineral or the fibrous (asbestiform) habit is intended.

Although other minerals may also occur in a fibrous habit, they are not generally included in the
definition of asbestos either because they do not exhibit properties typically ascribed to asbestos
(e.g., high tensile strength, the ability to be woven, heat stability, and resistence to attack by acid
or alkali) or because they do not occur in sufficient quantities to be exploited commercially.

The first four of the six asbestos minerals listed above have been exploited commercially (IARC
1977). Of these, chrysotile alone accounts for more than 90% of the asbestos found in
commercial products.

Importantly, it is neither clear whether the term asbestos maps reasonably onto the range of
fibrous minerals that can contribute to asbestos-like health effects nor whether individual
structures of the requisite mineralogy must formally be asbestiform to contribute to such health
effects.

Regarding whether the term asbestos is a useful discriminator for health effects, it is well
established that erionite (a fibrous zeolite not related to asbestos) is a potent inducer of
mesothelioma (Baris et al. 1987), which is one of the two primary asbestos-induced cancers (see
Chapter 3).  It is therefore possible that the fibrous habits of at least some other minerals not
formally included in the current definition of asbestos may contribute to the induction of
asbestos-related diseases. Therefore, an efficient procedure is needed for separating potentially
hazardous materials from those that are most likely benign.
                                                                                 j
There are two issues related to the question of whether fibers must formally be asbestiform to
contribute to health effects. The first involves the relationship between fiber structure and
disease induction and the second involves measurement.  Although the evidence is
overwhelming that the size and shape of a fiber affects the degree to which it contributes to the
induction of disease (this is addressed in detail in Chapter 6), it does not appear that sizes
inducing biological activity are well distinguished by criteria that define the asbestiform habit.
Therefore, depending on the definition employed for the fibers (or fibrous structures) that are
                                            4.1

-------
counted during an analysis, it may or may not be necessary to distinguish formally between
asbestiform and non-asbestiform structures for the concentrations derived from such a count to
adequately reflect biological activity.

The dimensions of an asbestiform fiber are determined by the manner in which the fiber grows
(Addison 2001).  In contrast, the massive forms of various minerals, when cleaved, also form
elongated particles (termed "cleavage fragments") and, depending on the definition employed for
fibrous structures during an analysis, such cleavage fragments may or may not be included along
with asbestiform fibers in a count (see Section 4.3).  Although it is clear from the manner in
which they are each formed that the surface properties of asbestiform fibers and cleavage
fragments are likely to be very different (for example, the latter will have many "unsatisified"
chemical bonds),  the degree to which such differences affect the toxic potency for comparable
sized structures is not currently known.

Although it is beyond the scope of this document to present a detailed treatise on asbestos
mineralogy, the morphology of asbestos dusts, or the nature and limitations of analytical
techniques and methods used to determine asbestos concentrations, a brief overview of these
topics is presented in the following sections both to identify issues that need to be addressed as
part of the development of an appropriate protocol for assessing asbestos risks and to provide the
background required to facilitate evaluation of the relevant issues. In that regard, a section on
lung physiology and function is also provided.

4.1    THE MINERALOGY OF ASBESTOS

As previously indicated, the six asbestos minerals can be divided into two general classes.
Chrysotile is the fibrous habit of the mineral serpentine (Hodgson 1965). The smallest fibrils of
chrysotile occur as rolled sheets or hollow tubules of this magnesium silicate mineral. The larger
fibers of chrysotile form as tightly packed bundles of the unit fibrils.

Chrysotile fibrils typically range from 20 nm to approximately 300 or 400 nm (0.02 to 0.3 or
0.4 um) in diameter.  Although slightly thicker fibrils may occasionally occur, at some point the
curvature induced by the mismatch between the magnesium and silicon layers of the fibril
becomes thermodynamically unstable, so that production of thicker fibrils is prohibited (Addison
2001).

Chrysotile bundles are held together primarily by Van der Waals forces and will readily
disaggregate in aqueous solutions containing wetting agents (e.g., soap).  They will also readily
disaggregate: in lung surfactant (Addision, 2001).

The general chemical composition of serpentine is reported as Mg3(Si2O5)(OH)4 (Hodgson
1965). However,  the exact composition in any particular sample may vary somewhat from the
general composition. For example, aluminum may occasionally replace silicon and iron, nickel,
manganese, zinc or cobalt may occasionally replace magnesium in the crystal lattice of
chrysotile (serpentine).
                                          4.2

-------
The five other common varieties of asbestos are all fibrous forms of amphibole minerals
(Hodgson 1965). These are ferro-magnesium silicates of the general composition:
                                                                          t
                           A2.3B5(Si,Al)8022(OH)2

                           where:
                                  A = Mg, Fe, Ca, Na, or K; and
                                  B = Mg, Fe, or Al.

Some of these elements may also be partially substituted by Mn, Cr, Li, Pb, Ti, or Zn.

The fibrous habits of the amphibole minerals tend to occur as extended chains of silica tetrahedra
that are interconnected by bands of cations (Hodgson 1965). Each unit cell typically contains
eight silica tetrahedra and the resulting fibers tend to be rhomboid in cross-section. Amphibole
fibers are generally thicker than chrysotile fibrils and may typically range from approximately
100 nm to 700 or 800 nm in diameter (Addison 2001).  Substantially thicker fibers have also
been observed.

4.2    MORPHOLOGY OF ASBESTOS DUSTS

Structures comprising the fibrous habits of the asbestos minerals come in a variety of shapes and
sizes. Not only do single, isolated fibers vary in length and thickness, but such fibers may be
found combined with other fibers to form bundles (aggregates of closely packed fibers arranged
in parallel) or clusters (aggregates of randomly oriented fibers) or combined with equant
particles to form matrices (asbestos fibers embedded  in non-asbestos materials). Consequently,
dusts (even of one mineral variety) are complex mixtures of structures. For precise definitions of
the types of fibrous structures typically found in asbestos dusts, see ISO (1995).

Detailed descriptions of the characteristics of dusts typically encountered at environmental and
occupational asbestos sites have been reported in the literature and the following summary is
based on a previously published review (Berman and Chatfield 1990).  Typically, the major
components of the dust observed in most environments are non-fibrous, isometric particles.
Fibrous structures consistently represent only a minor fraction of total dust.  Asbestos structures
represent a subset of the fibrous structures.

The magnitude of the fraction of total dust represented by fibers and the fraction of fibers
composed of asbestos minerals vary from site to site.  However, the fraction of asbestos in total
dusts has been quantified only  in a very limited number of occupational and environmental
settings (see, for example, Cherrie et al. 1987 or Lynch et al. 1970).

The gross features of structure size distributions appear to be similar among asbestos dusts
characterized to date (Berman  and Chatfield 1990). The major asbestos fraction of all such dusts
are small fibrous structures less than 5 um in length.  Length distributions generally exhibit a
mode (maximum) between 0.8 and 1.5 urn with larger fibers occurring with decreasing
frequency. Fibrous structures longer than 5 |im constitute no more than approximately 25% of
total asbestos structures in any particular dust and generally constitute less than 10%.
In some environments, the diameters of asbestos fibers exhibit a narrow distribution that is

                                           4.3

-------
 largely independent of length. In other environments, diameters appear to exhibit a narrow
 distribution about a mean for each specific length. In the latter case, both the mean and the
 spread of the diameter distribution increases as the length of the structures increase.  The
 increase in diameter with length appears to be more pronounced for chrysotile than for the
 amphiboles, presumably due to an increase in the fraction of chrysotile bundles contributing to
 the overall distribution as length increases.

 Only a few studies have been published that indicate the number of complex structures in
 asbestos size distributions.  The limited data available indicate that complex structures may
 constitute a substantial fraction (up to one third) of total structures, at least for chrysotile dusts
 (see, for example, Sebastien et al. 1984).  Similar results were also obtained during a re-analysis
 of dusts generated from the asbestos samples evaluated in the animal inhalation studies
 conducted by Davis et al. (Herman et al., in preparation). This is the same re-analysis used to
 support a study to identify asbestos characteristics that promote biological activity (Herman et al.
 1995), which is discussed further in Section 6.4.3.

 Historically, fibrous structures have been arbitrarily defined as structures exhibiting aspect ratios
 (the ratio of length to width) greater than 3:1 to distinguish them from isometric particles
 (Walton 1982). However, alternate definitions for fibers have also been proposed, which are
 believed to better relate to biological activity (see, for example, Berman et al. 1995 or Wylie
 et al. 1993). The degree to which fibers are combined within complex structures  in a particular
 dust may also affect the biological activity of the dust (Berman et al. 1995). Therefore, proper
 characterization of asbestos exposure requires that the relative contributions from each of many
 components of exposure be simultaneously considered.  Factors that need to be addressed
 include the distribution of structure sizes, shapes, and mineralogy in addition to the absolute
 concentration of structures.  Such considerations are addressed further in Chapter 6.  Thus,
 unlike the majority of other chemicals frequently monitored at hazardous wastes sites, asbestos
 exposures cannot be adequately characterized by a single concentration variable.

 4.3    CAPABILITIES OF ANALYTICAL TECHNIQUES USED TO MONITOR
       ASBESTOS

 Due to a complex history, a range of analytical techniques and methods have been employed to
measure asbestos in the various studies conducted over time (Walton 1982).  Use of these
various methods has affected the comparability of results across the relevant asbestos studies
(Berman and Chatfield 1990).  Therefore, the relative capabilities and limitations of the most
 important methods used to measure asbestos are summarized here. Later sections of this report
incorporate attempts to reconcile effects that are attributable to the limitations of the different
methods employed in the various studies evaluated.

Analytical techniques used to measure airborne asbestos concentrations vary greatly in their
ability to fully characterize asbestos exposure. The capabilities and limitations of four analytical
techniques (midget impinger [MI], phase contrast microscopy [PCM], scanning electron
microscopy [SEM], and transmission  electron microscopy [TEM]) are described here.  A general
comparison of the relative capabilities and limitations of the analytical techniques introduced
above is presented in Table 4-1.
                                           4.4

-------
    Table 4-1. Capabilities and Limitations of Analytical Techniques Used for Asbestos
                                       Measurements8
Parameter
Range of Magnification
Particles Counted
Minimum Diameter (size)
Midget
Impinger
100
All
1 urn
Phase Contrast
Microscopy
400
Fibrous
Structures'*
0.3 jim
Scanning
Electron
Microscopy
2,000-10,000
Fibrous
Structures1"
0:1 urn
Transmission
Electron
Microscopy
5,000-20,000
Fibrous
Structures^
< 0.01 \im
 Visible
Resolve Internal Structure
Distinguish Mineralogy*1
No
No
No
No
Maybe
Yes
Yes
Yes
 The capabilities and limitations in this table are based primarily on the physical constraints of the indicated
 instrumentation. Differences attributable to the associated procedures and practices of methods in common use
 over the last 25 years are highlighted in Table 4-2.
 ""Fibrous structures are defined here as particles exhibiting aspect ratios (the ratio of length to width) greater than
 3 (see Walton 1982).
 TEM counts frequently resolve individual fibrous structures within larger, complex structures.  Based on internal
 structure, several different counting rules have been developed for handling complex structures. See the
 discussion of methods presented below.
 dMost SEM and TEM instruments are equipped with the capability to record selected area electron diffraction
 (SAED) spectra and perform energy dispersive X-ray analysis (EDXA), which are used to distinguish the
 mineralogy of structures observed.

MI and PCM are the two analytical techniques used to derive exposure estimates in  the majority
of epidemiology studies from which the existing risk factors are derived.  SEM is an analytical
technique that has been employed in several key animal studies. TEM provokes interest because
it is the only analytical technique that is potentially capable of distinguishing all of the
characteristics of asbestos that potentially affect biological activity.

Although PCM was (and still is) widely used to characterize occupational exposures, its inability
to distinguish between asbestos and non-asbestos and its lack of sensitivity  limits its usefulness
in environmental settings (Berman and Chatfield  1990).  In fact, PCM analyses and  TEM
analyses showed no correlation among measurements collected during the cleanup of the 1991
Oakland Hills fire (Berman, unpublished data). Such lack of correlation is expected to be
observed generally whenever measurements are collected at sites where asbestos concentrations
are low enough that a substantial fraction of the structures counted by PCM are not asbestos.
Consequently, TEM is the technique that has been recommended for use at  Superfund sites
(Berman and Kolk 1997; Chatfield and Berman 1990).

Importantly, the physical limitations of the various analytical techniques is  only part of the
problem.  To provide reproducible results that can be compared meaningfully to other analyses
viv other studies, one must also consider the choice of procedures (methods) that address
everything from sample collection and preparation to rules for counting and quantifying asbestos
structures.

                                             4.5

-------

                      CO
S o  'S u

-------
I


c
S
v>

2

g
A

^
ac
a
•c

w
a
v

s


I
w
•O
o

"3

S
jj

2

2

"a.

t
t«
o
a
A
O.


O

U


r4

•4

pi
S


^2
g»
  '2
         j
    •^   to .-$

  Ia-i-41U
  11<^^5 3 § £


  I? Si I 111 It
           sw^iSSsKas;
            !
            «
    <


    ii

    Ji
    S c»
            «   s  1
            CO   •=  K
      •si*
      W B
      ^C *•£


      11
      <» u
I

M
a
e

s
o
U

-------
O m E/J 3 "• —i
  fl> C O ^ 'O
P £ *> ° .S
3 -° > "C c o S
en c o o •»; c 8
.S 8  > "g -I
                                  w" .S
                                   1
                                    8 iS 1 '5 .
                                 1111!  11
                                                             00

-------
11 1 £  ^r
.S  «£420

-------
Multiple methods have been published for use in conjunction with several of the analytical
techniques mentioned above (particularly TEM).  Such methods differ in the procedures
incorporated for sample preparation and for the manner in which asbestos structures are counted.
The sample preparation requirements, conditions of analysis, and structure counting rules for
several of the most commonly employed methods are presented in Table 4-2 to illustrate how the
choice of method can result in substantially different measurements (even on duplicate or split
samples).

The second column of Table 4-2 describes the specifications of the PCM method currently
mandated by the Occupational Safety and Health Agency (OSHA) for characterizing asbestos
exposure in occupational settings.  Although this method is in common use today, several
alternate methods for counting fibrous structures by PCM have also been used historically.
Therefore, PCM measurements reported in earlier studies (including the available epidemiology
studies) may not be comparable to PCM results collected today.

The last four columns of Table 4-2 describe TEM methods that are in current use. Comparison
across these methods indicates:

       •      the shortest lengths included in counts using these methods vary between 0.06
              and 1 \im.  Given that structures shorter than 1 urn represent a substantial fraction
              of total asbestos structures in almost any environment (Section 4.2), this
              difference alone contributes substantially to variation in measurement results
              across methods;

       •      the definitions and procedures for counting complex structures (i.e., bundles,
              clusters, and matrices) vary substantially across methods, which further contribute
              to variation in measurement results. For example, the ISO Method requires that
              component fibers of clusters and matrices be counted separately, if they can be
              readily distinguished. In contrast, clusters are counted as single structures under
              the AHERA Method; and

       •      although all of the methods listed incorporate sample preparation by a direct
              transfer process (in  which the fibers are counted in their original position on the
              filter), several of the methods have also been paired with an optional indirect
              transfer process (which involves ashing the original air filter, mixing the residue
              in water, sonicating, and re-suspending the fibers on a new filter). Measurements
              derived from split samples that are prepared, respectively, by direct and indirect
              transfer, can vary by factors as large as several 100 (Berman and Chatfield 1990).
              Typically, counts of asbestos structures on samples prepared by an indirect
              transfer procedure are greater than those derived from directly prepared samples
              by factors of between 5 and 50.
                                          4.10

-------
Given the combined effects from the physical limitations of the various techniques employed to
analyze for asbestos and the varying attributes of the methods developed to guide use of these
techniques, the relative capabilities and limitations of asbestos measurements derived,
respectively, from paired methods and techniques in common use can be summarized as follows:

       •      all four techniques are particle counting techniques;

       •      neither MI nor PCM are capable of distinguishing asbestos from non-asbestos
              (i.e., they are incapable of determining structure mineralogy);

       •      counting rules used in  conjunction with MI do not distinguish isometric particles
              from fibers;

       •      counting rules used in  conjunction with PCM limits counting to fibrous structures
              longer than 5 Jim with aspect ratios greater than 3:1;

       •      the range of visibility associated with PCM limits counting to fibers thicker than
              approximately 0.3 jim;

       •      under conditions typically employed for asbestos analysis, the range of visibility
              associated with SEM limits counting to fibers thicker than approximately 0.1 urn,
              which is only marginally better than PCM;

       •      SEM  is capable of distinguishing asbestos structures from non-asbestos
              structures;

       •      TEM is capable of resolving asbestos structures over their entire size range (down
              to thicknesses of 0.01  um);

       •      TEM is capable of distinguishing the internal components of complex asbestos
              structures; and

       •      TEM is capable of distinguishing asbestos structures from non-asbestos
              structures.

More detailed treatments of the similarities and differences between asbestos techniques and
methods can also be found in the literature (see, for example, Herman and Chatfield 1990).

Due to the differences indicated, measurements from a particular environment (even from
duplicate samples) that are derived using different analytical techniques and methods can vary
substantially and are not comparable. In fact, results can differ by two or three orders of
magnitude (Berman  and Chatfield  1990). More importantly, because the relative distributions of
structure sizes and shapes vary from environment to environment, measurements derived using
different analytical techniques and methods do not even remain proportional from one
environment to the next. Therefore, the results from multiple asbestos studies can only be
meaningfully  compared if the effects that are attributable to use of differing analytical techniques
and methods can be  quantified and reconciled. Few of the existing studies, however, document
analytical procedures in sufficient detail to reconstruct exactly what was done.

                                           4.11

-------
 4.4    THE STRUCTURE AND FUNCTION OF THE HUMAN LUNG

 4.4.1  Lung Structure

 The lungs are the organs of the body in which gas exchange occurs to replenish the supply of
 oxygen and eliminate carbon dioxide.  To reach the gas exchange regions of the lung, inhaled air
 (and any associated toxins) must first traverse the proximal conducting (non-respiratory) airways
 of the body and the lung including the nose (or mouth), pharynx, larynx, trachea, and the various
 branching bronchi of the lungs down to the smallest (non-respiratory) bronchioles. Air then
 enters the distal (respiratory) portion of the lung, where gas exchange occurs.

 In humans, the respiratory portion of the lungs are composed of the respiratory or aveolarized
 bronchioles, the alveolar ducts, and the alveoli (or alveolar sacs). There are approximately 3xl08
 alveoli in human lungs (about  20 per alveolar duct) with a cumulative volume of 3.9x103 cm3
 (Yeh and Harkema 1993). This represents approximately 65% of the total volume capacity of
 human lungs at full inspiration.

 Yeh and  Harkema (1993) also report that the ratio of lung volume to body weight is
 approximately constant across a broad range of mammals (from a shrew, 0.007 kg to a horse,
 500 kg).  Human lung volumes average a little more than 5 L.

 Each human alveolus has a diameter of approximately 0.03 cm (300 p,m) and a length of
 0.025 cm (Yeh and Harkema 1993). The typical path length from the trachea to an alveolus is
 approximately 25 cm. The bronchi leading to each alveolus have branched an average of 16
 times from the trachea (with a  range of 9 to 22 branches). Importantly, the mean path length, the
 number of branches between trachea and alveolus,  and the detailed architecture (branching
 pattern) of the respiratory region of the lung vary across mammalian species. For example, rats
 and mice lack respiratory bronchioles while macaque monkeys exhibit similar numbers of
 respiratory bronchiole generations as humans (Nikula et al. 1997).  Furthermore, branching in
 humans tends to be symmetric (each daughter branch being approximately the same size and the
 angle of branching  for each is similar but not quite  equal) while rodents tend to exhibit
 monopodal branching in which smaller branches tend to come off at angles from a main trunk
 (Lippmann and Schlesinger 1984).

 The gas exchange regions of the lung are contained within the lung parenchyma, which
 constitute approximately 82%  of the total volume of the lungs (Gehr et al. 1993).  Importantly,
 the lung parenchyma is not a "portion" of the lung; it fills virtually the entire volume of the
 organ traditionally visualized as the "whole" lung.  Embedded within the parenchyma are the
 larger conducting airways of the lungs and the conducting blood vessels that transport blood to
 and from the capillaries that are associated with each alveolus. In the human lung,
 approximately 213  ml of blood are distributed over 143 m2 of gas-exchange (alveolar) surface
 area (about the size of a tennis court).  The gas-exchange surface area of lungs scale linearly with
 body weight over most mammalian species. The slope of the "reduced" line (where the Y-Axis is
the ratio of the surface area to the surface area in a reference species and the X-Axis is the ratio
of the body mass to the body mass of the same reference species) is 0.95.  Figure 4-1 is a
 photomicrograph showing two views of a portion of lung parenchyma in the vicinity of a
terminal bronchiole and an avleolar duct, which are labeled.  The circular spaces in the left

                                         4.12

-------
portion of the figure and the cavities in the right portion of the figure are alveoli. Note the
thinness of the walls (septa) separating alveoli.


   Figure 4-1. The Structure of Lung Parenchyma Showing Alveoli and Alveolar Ducts
                             (Source: St. George et al. 1993)
       «."•*•« ufifuE-SBB^ iV-='*' • *fi

    FIG. 10.  LM and SEM comparison of the centnacmar region with two different organizations
    and B: The sHort or poorly developed  respiratory bronchiole of the rat; C and D, the w
    aiveolarized respiratory bronchiole of tne eat Terminat bronchiole (TB). respiratory bronchi
    (RB). and alveolar duct (AD). For details  see ref. 1O6,
                  Confidential: Need Permission to Reproduce this Figure
                                            4.13

-------
Alveoli are separated from each other by alveolar septa that average only a few micrometers in
thickness (Gehr et al. 1993). Gas-exchange capillaries run within these septa and the air-blood
barrier, which averages only 0.62 \im in thickness, is composed of three layers: the alveolar
epithelium, the interstitium, and endothelium. The epithelium is described in detail below. The
interstitium is primarily composed of a collagenous, extracellular matrix, which constitute about
two-thirds of the interstitial volume. There is also a collection of fibroblasts, macrophages, and
other cells interspersed within the matrix (Miller et al. 1993). The cells of the interstitium
constitute about one third of its volume. The endothelial layer is composed of the smooth
muscle cells and connective tissue that constitute the walls of vascular capillaries. The amount
of connective tissue in  the septa between alveoli also varies between animal species; small
rodents have less and primates more (Nikula et al. 1997).

Figures 4-2 and 4-3 show, respectively, a typical alveolar septum and a closeup of one portion of
such a septum. In Figure 4-2, one can see that the septa themselves are thin and are filled almost
entirely with capillaries.  Figure 4-3 shows that the epithelial lining of an alveolus is no more
than 1  u.m thick, that the underlying interstitium is no thicker, and that the endothelium of the
adjacent capillary is similarly thin. These three layers constitute the major tissues of the air-
blood barrier (the rest of the barrier includes the limited quantity of blood plasma between the
endothelial wall of a capillary and a red blood cell and the outer membrane of the red blood cell
itself).

     Figure 4-2. The  Structure of the Inter-Alveolar Septa  (Source: Gehr et al.  1993)
       10. Transmission electron micrograph of an alveolus from a dog lung fixed by Intravascular
  perfuslon. The lung was inflated with air at a pressure of S cm of water on the deflation limb of the
  last of three hysteresis cycles (approximately 6O% TUG). It shows irrteratveoiar septa that are
  folded at this degree of Inflation. The surface lining layer (SLL) smoothes out every depression of
  the alveolar surface (arrows). A, alveoli: C, capillary, x 2.1OO. fmset: High power view of an
  epithelial depression filled with a fluid surface lining layer (SLL). x 1-+.6OO.
                 Confidential: Need Permission to Reproduce this Figure
                                           4.14

-------
Figure 4-3. The Detailed Structure of an Alveolar Wall that is Part of an Inter-Alveolar
                         Septum (Source:  Gehr et al. 1993)
FIG. 19. FradiOT ^squareilaitice test grid superimpos^onaJveolar(^|lla^tomeasilHBinte^
cept lengths in tissue (I,) and plasma (lp) for the calculation of Ihe harmonic mean barrier thick-
ness. Note the erythrocyte (EC) in capillary (C) and that the barrier separating blood from alveo-
lar air (A) is made of the three layers epithelium (EP), endothelium (EN), and interstitium (IN).
x 29,000. (From ref. 80.)
               Confidential: Need Permission to Reproduce this Figure


                                       4.15

-------
Gehr et.al. (1993) also report that interalveolar septa constitute approximately 14% of the
volume of the gas-exchange region of the lung (i.e., the lung parenchyma). Of this tissue mass,
approximately 20% is endothelium, 55% is interstitium, and the rest is composed of cells
associated with the alveolar epithelium. The remainder of the gas-exchange region is air space.
Gehr et al. (1993) also report that the interalveolar septa fold to accommodate changes in lung
volume during respiration, although the major contribution to lung volume changes (over the
range of normal inspiration) appears to be the collapsing (folding) of alveolar ducts ('Mercer and
Crapo 1993).

According to Gehr et al. (1993), Type I epithelial cells, which constitute approximately 95% of
the surface area of alveolar epithelium, are flat and platy (squamous), and average less than 1  ym
in thickness (except where cell nuclei exist, which are approximately 7.5 urn in diameter and
protrude into the alveolar space). Type II epithelial cells, which are cuboidal, constitute no more
than 5% of the epithelial surface. Despite the small fraction of surface that they occupy, Type II
cells serve to maintain the integrity of the overall epithelial  lining so that, for example, they limit
the tissue's permeability and control/prevent transport of macromolecules from the interstitium
to the alveolar space, or the reverse (Leikauf and Driscoll 1993). Type II cells also secrete lung
surfactant. The basement membrane of alveolar epithelium is a collagenous structure.

In contrast, the epithelial cells lining the trachea and bronchi (including the respiratory
bronchioles) are ciliated and columnar and averages between 10 and 15 urn in thickness (based
on photomicrographs presented in St. George et al. 1993). Tracheobronchial epithelium
reportedly contains at least 8 distinct cell types (St. George  et al. 1993): ciliated epithelium,
basal cells (small flat cells situated along the basal lamina and not reaching the luminal surface),
mucous goblet cells, serous cells, nonciliated bronchiolar (clara) cells, small mucous granule
(SMG) cells, brush cells, and neuroendocrine cells.  The relative abundance of the various cells
varies across mammalian species as well as across the various airway generations (branches) and
even the opposing sides of specific airways. The number of cells per length of basal  lamina also
varies across mammalian species.

4.4.2   The Structure of the Mesothelium

The mesothelium is a double layered membrane and each layer is a single-cell thick.  The two
layers of the mesothelium are separated by a space (e.g., the pleural space), which contains
extracellular fluid and free macrophages (Kane and McDonald 1993).  The pleural  space is
drained at fixed locations by lymphatic ducts.  Each layer of the mesothelium overlies a
collagenous basement membrane containing dispersed spindle cells. Depending on the location
of the mesothelium, the basement membrane may overlie the skeletal muscle of the diaphragm or
the rib cage (in the case of the parietal pleura, which is the outer layer). For the inner layer or
visceral pleura, the basement membrane overlies visceral organs (including the lungs) within the
rib cage. Healthy mesothelium is quiescent, meaning that cells are not actively dividing.

The relative size and thickness of mesothelial tissue varies across mammalian species (Nikula
et al. 1997). For example, rats have relatively  thin pleura with limited lymphatic ducts.  In
contrast, nonhuman primates have thicker pleura with greater lymphatic drainage than rats and
humans have even thicker pleura and relatively abundant lymphatic ducts.
                                          4.16

-------
4.4.3  Cytology

Alveolar Epithelium. In the respiratory region of the lung, Type II epithelial cells are
progenitor cells for Type I epithelium (Leikauf and Driscoll 1993). Type 1 epithelial cells are
not proliferation competent (Nehls et al. 1997). After injury to Type I cells, Type II cells
proliferate and reestablish the continuous epithelial surface. Type II cells also secrete surfactant.
It appears that the identity and location of the progenitor cells for Type II epithelial cells are not
currently known.

Injury or alteration of Type II cell function are associated with several diseases included
idiopathic pulmonary fibrosis (Leikauf and Driscoll 1993).  Also, crystalline silica and other
toxic agents have been shown to directly modify Type II cellular activity. For example,
crystalline silica (at sub-cytotoxic levels, <100 ng/ml) stimulates Type II proliferation in tissue
culture. In contrast, neither titanium dioxide nor aluminum coated silica induce proliferation at
corresponding concentrations.

In culture, Type II epithelial cells transform slowly into Type I cells and thus have limited
population doubling capacity (Leikauf and Driscoll 1993). Rarely can the number of Type II
cells expand past 3-10 passages (20^-30 doublings). During this time, cells terminally
differentiate, develop cross-linked  envelopes, and appear squamous, enlarged, and
multinucleated.  The process is noted to be accelerated by the presence of transforming growth
factor beta (TGF-P).

Macrophages.  Alveolar macrophages (the largest population of macrophages in the lung) are
mobile, avidly phagocytic, present antigens, and release cytokines that trigger various other
immune responses (Leikauf and Driscoll 1993). They also initiate inflammatory responses and
other repair mechanisms that are designed to restore tissue homeostasis.

The next largest population of macrophages in the lung are the interstitial macrophages (Leikauf
and Driscoll 1993). These are localized in the peribronchial and perivascular spaces, the
interstitial spaces of the lung parenchyma, the lymphatic channels, and the visceral pleura.  The
various populations of macrophages in the lung express different surface proteins,  show different
proliferative capacity, and show differences in metabolism.

The sizes of alveolar macrophages varies substantially across mammalian species (Krombach
et al. 1997). Krombach and co-workers provide a table summary of the relative sizes across
several species of interest:

        Animal	Mean diameter (nm)       Mean Volume (n,m3)
Rats
Syrian Golden Hamsters
Cyanomolgus Monkeys
Healthy Humans
13.1±0.2
13.6±0.4
15.3±0.5
21.2±0.3
1166±42
1328±123

4990±174
        Note: as indicated later (Section 6.2), the relative size of various macrophages has direct
        implications regarding the dependence of clearance mechanisms on fiber size.
                                            4.17

-------
Mesothelium.  Importantly, mesothelial cells are proliferation competent and may be their own
progenitor cells (Kane and McDonald 1993). It is also possible, however, that as yet-to-be-
identified progenitor cells are located along opposite walls of the pleura or at other locations
within the pleura! space.

Tracheo-bronchiolar epithelium. As previously indicated, tracheo-bronchiolar (i.e., non-
respiratory) epithelium is composed primarily of ciliated, columnar cells that are 10 to 15 um
thick.  Although some report that Clara cells serve as progenitor cells for tracheo-bronchiolar
epithelium (Finkelstein et al. 1997), others report that both Clara cells and bronchiolar
epithelium are proliferation competent (Nehls et al.  1 997). It appears that the identity and
location of the progenitor cells for Clara cells are not currently known.

4.4.4  Implications

The potential implications of the above observations concerning lung structure and cytology are:

       •      that the thicknesses of Type I epithelial cells, endothelial cells, and the
              interstitium in the alveolar septa are all very small relative to the lengths of the
              putative asbestos fibers that contribute to disease;
that an entire alveolus is only 300
radius by <1 um thick;
                                                  across and a typical Type I cell is 46 |im in
              that distances across alveolar septa are only on the order of a few jim and such
              septa contain both the endothelium and the interstitum. Thus, the distances that
              have to be traversed to get to these structures are also small relative even to the
              length of a fiber;

              that the alveolar septa and the walls of the alveolar ducts fold during respiration,
              which may provide mechanical forces that facilitate movement of fibers into and
              through the alveolar epithelium;

              that Type I epithelium do not proliferate so they cannot be the cells that lead to
              cancer. It is the Type II epithelial cells (and potentially macrophages, basal cells,
              or endothelial cells) that contribute to cancer in the pulmonary portion of the lung.
              Type II cells eventually terminally differentiate to Type I cells;

              that other cells in the lung that have variously been reported to be proliferation
              competent (so that they potentially serve as target cells for the induction of
              cancer) include Clara cells and bronchiolar epithelial cells;

              that the distance from the most distal airways and alveoli to the pleura is small
              relative to the lengths of a fiber; and

              that mesothelial cells are proliferation competent and thus serve as potential
              targets for the induction of cancer.
                                           4.18

-------
                5.0  THE ASBESTOS LITERATURE

This is a description of the common types of studies in the asbestos literature and an overview of
the sources of potential uncertainty typically associated with each.  Such limitations must be
considered when drawing conclusions from these studies and, more importantly, when deriving
inferences based on cross-study comparisons.  Throughout this document, we have endeavored
to identify the major sources of uncertainty in the studies we examined and have endeavored to
account for such uncertainties during our evaluation and interpretation of study results.

The types of studies available for examining relationships between risk and asbestos exposure
include human epidemiology studies, human pathology studies, a broad variety of animal
studies, and a broad variety of in vitro studies in both tissue cultures and cell-free systems. To
properly compare and contrast the results from such studies:

       •     the method(s) employed for asbestos characterization in each study need to be
             reconciled;

       •     the procedures employed for evaluating study endpoints need to be compared and
             contrasted;

       •     the relationship between the route of exposure employed in each type of study and
             the exposure route of interest (i.e., human inhalation) needs to be examined; and

       •     other major, study-specific sources of uncertainty need to identified and
             addressed.

Among study conditions and procedures that must be considered before evaluating study
conclusions, it is particularly important to address the analytical methodologies employed to
characterize the nature of exposures (or doses) in each study and such considerations are
common to virtually all of the various types of studies of interest.

As indicated in Section 4.3, the only instrument capable of completely delineating asbestos
structure-size distributions is TEM (or TEM combined with other techniques). Thus, for
example, conclusions regarding variations in biological effects due to differences in such things
as fiber size must be viewed with caution when fiber sizes are characterized using only PCM,
SEM, or other, cruder analytical techniques. Similarly, the ability to adequately determine fiber
mineralogy (fiber type), particularly of what may be minor constituents of various dusts or bulk
materials, also depends strongly on the instrumentation employed for analysis as well as the
strategy for sampling and for conducting the actual structure count. All of these factors must be
considered.

Not only does the specific instrumentation (analytical technique) that is employed in an asbestos
measurement affect the outcome of that measurement, but the particular method employed to
guide the measurement affects the outcome. As previously indicated  (Section 4.3), details
concerning the definition of structures to be included in a count, the strategy for counting, and
the minimum number of specific types of structures to be included in a count (all features that


                                           5.1

-------
vary across analytical methods) affect the precision with which fiber concentrations (particularly
of longer and thinner fibers) are delineated. It is not uncommon, for example, that asbestos
concentrations measured in the same sample may vary by several orders of magnitude, due
simply to a difference in the analytical method employed for the  analysis (even when the same
analytical instrument is employed, see Section 4.3).

Other important sources of uncertainty tend to be study-type specific and are thus addressed
separately below.

5.1    HUMAN EPIDEMIOLOGY STUDIES

A good overview of the kinds of limitations that contribute to uncertainty in the available
epidemiology studies was presented in the Health Effects Assessment Update (U.S. EPA, 1986).
As described in Appendix A of this document, while evaluating exposure-response factors
derived from the human epidemiology studies, an attempt was made to address most of the major
sources of uncertainty commonly associated with such studies, which are described briefly
below.

Epidemiology studies, which track the incidence of disease (or mortality) within a defined group
(cohort) sharing comparable exposures, have been performed on  cohorts of workers exposed to
asbestos and other mineral fibers in a variety of occupational and environmental settings.
Among these, studies that include quantification of exposures are particularly useful  for
evaluating exposure-response relationships and deriving risk factors.

Generally, the most severe limitations in an epidemiology study involve the exposure data. Both
estimates of the level of exposure and determination of the character of exposure are affected by
such limitations. Regarding the character of exposure, because exposure measurements from
most of the available quantitative epidemiology studies are based on MI measurements or PCM
measurements, detailed characterization of the size distribution or the mineral type of fibrous
structures (particularly of minor constituents) that contributed to  exposure  in such  studies is
generally lacking (Appendix A). This is particularly important because of the evidence that
neither MI nor PCM are capable of providing measurements that remain proportional (across
study environments) to  the biologically-relevant characteristics of an asbestos dust (Herman et
al. 1995). this limits the ability both to compare results across the existing studies and to
extrapolate such results to new environments for which risks need to be estimated. At the  same
time, effects on the ability to observe exposure-response trends within a single study are not
typically impaired.

Samples collected prior to the mid-1960s were often analyzed  by measuring total dust in units of
millions of particles per cubic foot (mpcf) using impingers or thermal precipitators. A
description of the relative strengths and weaknesses of these techniques is provided in
Section 4.3. The fibrous portion of the dust was not monitored.  Impinger measurements are
sometimes related to fiber counts (based on PCM) using side-by-side measurements of total dust
and fiber counts collected during a relatively brief period of time (e.g.,  Dement et al. 1983a;
McDonald et al. 1980b). However, the correlation between fiber counts and total dust is
sometimes poor within a plant (i.e., a single study environment) and generally poor between
plants (see, for example, U.S. EPA 1986).  Thus, conversions based on limited sets of paired

                                          5.2

-------
measurements are of questionable validity. In some studies (e.g., McDonald et al. 1983b) the
only available measurements are MI measurements (in mpcf) and these have been related to f/ml
by PCM using conversion factors derived in other plants, which raises further questions
concerning validity.

Even if all measurements could be adequately converted to PCM, this may still not be adequate
for assessing risk in a manner that allows extrapolation across exposure environments or studies.
Comparing exposure-response  factors derived in different exposure environments (or
extrapolating to new environments to predict risk) requires that asbestos measurements reflect
the characteristics of asbestos structures (size, shape, mineralogy) that determine biological
activity. If surrogate measures are employed (e.g., measures of asbestos structures displaying
characteristics other than those that determine biological activity), there is no guarantee that
concentrations of such surrogate measures and the true biologically active structures will remain
proportional from one environment to the next. As a consequence, the relationship between
exposure (measured by surrogate) and risk may not remain constant from one environment to the
next. Importantly, several studies suggest that PCM may, at best, be no more than a surrogate
measure (see, for example, Herman et al. 1995). Moreover, the technique was adapted to
asbestos in an ad hoc fashion with only limited thought given to biological relevance (Walton
1982).

Use of surrogate measures of asbestos exposure may be less of a problem within a single
exposure environment (where airborne asbestos structures likely have been generated in a similar
manner from similar source material).  Thus, surrogate measures of asbestos exposure may
remain approximately proportional  to the true biologically active structures, which suggests why
monotonically increasing exposure-response relationships have likely been observed with PCM-
measured concentrations in single exposure environments. In different exposure environments,
however, the distribution of fiber sizes and types of airborne asbestos structures are likely
different, since they are generated in different processes from different source material. There  is
thus little reason to expect surrogate measures of exposure to remain proportional across such
environments.

None of the published epidemiology studies incorporate TEM measurements of asbestos and
such measurements are not widely available in occupational settings (Appendix A). However,
TEM is the method currently used (and recommended) to assess exposure in environmental
settings, due both to questions  concerning biological relevance (Herman et al. 1995  and
addressed in detail in Chapter 6) and to problems with measuring environmental asbestos
concentrations by PCM (Section 4.3).

In some cases, the limited exposure characterization presented in specific epidemiology studies
can be augmented by pairing such studies with published TEM characterizations of dusts from
the same or similar exposure settings, to the extent the appropriate supplemental studies are
available.  In fact, this is the procedure adopted in this document to adjust the existing risk
factors to exposure indices that are  thought to better relate to biological activity (described in
detail in Section 7.4).  Such an approach is limited, however, to the extent that the published
asbestos characterizations actually represent exposure conditions in the corresponding
epidemiology studies. To the extent reasonable, the limitations of this approach have been
                                           5.3

-------
addressed in this study by assigning and incorporating additional factors into the calculation of
uncertainty intervals (defined in Appendix A) that are associated with the adjusted potency
factors.

Regarding levels of exposure in the epidemiology studies, in most cases, air measurements were
collected only infrequently and measurements may be entirely lacking from the earliest time
periods, when exposures may have been greatest.  In such cases, exposures are typically
estimated either by extrapolation from periods when measurements are available or by expert
judgement based on personal accounts and records of changes in plant operations, industrial
hygiene procedures, air standards, etc. Moreover, the majority of exposure measurements used .
in these occupational studies are based on area (ambient) rather than personal samples.
Typically, only a few areas of a plant have been sampled so that levels in other areas must be
approximated using expert judgement by persons familiar with operations at the plant.

It is difficult to judge the degree that available asbestos concentration measurements are
representative of actual exposures in the existing studies.  In some cases, it seems likely that
operations were shut down or otherwise modified in preparation for sampling.  Likewise, in
some operations there are brief episodes of very intense exposure and it is questionable whether
such episodes are adequately represented in the available data.

Most of the asbestos measurements used in the published epidemiology studies were collected
for insurance or compliance purposes.  They were not intended to provide representative
estimates of the direct level of exposure to workers. Some of the published epidemiology studies
lack any direct exposure data. For example, exposures were estimated for the cohort studied by
Seidman (1984) based on conditions simulated many years later in a similar plant to the one
from which Seidman studied the original cohort. In fact, the equipment in the plant from which
Seidman obtained exposure estimates came originally from the plant where  Seidman studied the
workers; it was purchased and moved. Recently, an epidemiology  study was also completed for
a cohort working at that new plant (Levin et al. 1998).

In addition to problems with the actual analysis of asbestos concentrations, individual exposures
are generally  estimated in the existing epidemiology studies by relating ambient asbestos
measurements to job descriptions and integrating the duration of exposure over the recorded time
that each worker spent in each job category.  However, sometimes there are no records of
specific areas in which an employee worked, so that work areas must be assumed based on job
title.  Some types of workers (e.g., maintenance workers) may have spent time in many different
areas of a  plant so their exposure varies from what might otherwise be assumed.

Although the  greatest problems with the data in existing epidemiology studies likely lies within
the estimates  of exposure, problems with disease-response data also exist. Mesothelioma is rare
and this disease may have been under-reported as a cause of death in  older studies.  This is
probably less of a problem in more recent studies, since the association of mesothelioma with
asbestos exposure is now well known.  In fact, the opposite tendency (over-reporting) may now
be occurring because of increased  sensitivity by examiners (an asbestos worker with
mesothelioma is now more likely to be eligible for compensation).  Some studies have '
re-diagnosed  causes of death from all of the available data (e.g., Selikoff et al. 1979); however,
                                           5.4

-------
this creates the problem of lack of comparability to control populations (for which such re-
diagnosis is not generally performed).

The choice of an appropriate control population is also an important consideration. Local cancer
rates may differ substantially from regional or national rates and the choice of an appropriate
control is not always clear.  A related problem is the lack of smoking data in many of the studies.
Because of the interrelation between smoking and asbestos in lung cancer, errors could occur in
lung cancer risk estimates if the smoking patterns of the cohort are substantially different from
those of the control population.                                                (

In some of the studies, ^substantial portion of the population is lost to follow-up (e.g.,
Armstrong 1988), and this adds additional uncertainty to the analysis.  Also, the effect of
exposure may be inaccurately evaluated if the follow-up of the population is too brief. This may
be a limitation,  for example, of the Levin et al. (1998) study.

Another problem frequently associated with these studies is that available data are not reported
in a form that is well-suited to risk assessment.  The EPA lung cancer model, for example,
requires that exposure be estimated as cumulative exposure in f/ml-years excluding the most
recent 10 years  (U.S. EPA 1986, also described Section 7.2); generally the data are not published
in this form.  The data are also frequently not available in a form that permits study of the shape
of the lung cancer exposure-response curve, so it is not possible to determine how well the EPA
model describes the data. The reporting of the mortality data for mesothelioma is generally even
less appropriate for risk assessment. Ideally, what is needed is the incidence of mesothelioma
subdivided according to exposure level, age at beginning of exposure,  and duration of exposure
(U.S. EPA 1986, also described in Section 7.3).  Such data are almost never available in
published studies and crude approximations must be made to account for this lack of
information.

It is important to understand the type and magnitude of effect that each of these sources of
uncertainties are likely to have on the distribution of potency estimates derived from the set of
available studies for lung cancer and mesothelioma, respectively. Some of the above-described
limitations likely introduce random errors that simply decrease the overall precision of a potency
estimate.  However, other types of limitations may cause systematic errors in particular studies,
which potentially bias the potency estimate either high or low. Some of the limitations may only
affect between-study comparisons and some may introduce a systematic bias between either
industry types or fiber types. Examples of some of these types of variation are provided in
Section 7.1.

We also note that, although individual estimates of potency factors from individual studies may
be highly uncertain, by combining results across multiple studies while properly addressing such
uncertainties, it may be possible to draw conclusions with greater precision than reasonable for
any individual study. This is the essential advantage of the type of meta analysis discussed in
this document (see Chapter 7).
                                           5.5

-------
5.2    HUMAN PATHOLOGY STUDIES

Human pathology studies provide a characterization of disease morphology and correlations
between causes of death and the types of asbestos fibers retained in the lungs and other bodily
tissues. These studies generally involve microscopic examination of tissue samples for
indications of morphologic changes characteristic of disease and/or microscopic examination of
digested tissue specimens to characterize the mineral fibers extracted from such tissue.

The results of human pathology studies need to be evaluated carefully by addressing effects that
are attributable to:

       •      the way tissue samples are fixed for preservation;

       •      the way tissue samples are prepared for analysis (e.g., ashing, bleach digestion,
              digestion in alkali, or some combination);

       •      the choice of methods employed for characterization of asbestos; and

       •      the choice of locations within tissues from which samples are collected for
              analysis.

Because tissue samples obtained from deceased individuals are typically stored for long periods
of time before they may be analyzed as part of a human pathology study, such  samples are
commonly fixed by treatment with chemical preservatives prior to storage.  However, Law et al.
(1991) studied the effects of two common fixatives (Karnovsky's fixative and  formalin fixative)
on asbestos fibers and concluded that such fixatives degrade and dissolve asbestos fibers
(including both chrysotile and crocidolite) at measurable rates. Therefore, particularly for
samples that are stored "wet" (as opposed, for example, to storage in paraffin blocks), the
concentrations and character of the tissue burden of asbestos may be altered during storage.
Even for studies in which relative (as opposed to absolute) concentrations are being compared,
alterations associated with preservation may limit the ability to make such comparisons,
particularly among samples stored for widely disparate periods of time or stored using widely
disparate procedures.

Fiber concentrations in tissue samples have also been shown to vary as a function of the method
employed for preparing such samples for analysis. Historically, samples that are digested in
bleach or alkali have tended to exhibit lower recovery of asbestos fibers than samples that are
ashed.  However, more recent studies suggest that improving technique has narrowed these
differences so that this is no longer a major consideration . Thus, when comparing results across
studies, due consideration needs to be given for the time frame during which such studies were
conducted and the comparability/differences in the techniques employed for tissue sample
preparation.

Once prepared, both the character and the concentration of the tissue burden measured in a tissue
sample will also depend heavily on the particular analytical method employed  to characterize
asbestos and differences attributable to such techniques must be reconciled before measurements
across  samples or conclusions across studies can be reasonably compared. A more detailed

                                           5.6

-------
description of the effects attributable to asbestos measurement was presented in the previous
section on human epidemiology studies (Section 5.1) and the same issues obtain for human
pathology studies. Unless measurements are made using comparable instrumentation with
comparable methodology, comparisons across such measurements can be very misleading.

Perhaps the biggest limitation hindering the kinds of evaluation that can be conducted based on
human pathology studies is that due to the strong dependence of asbestos concentrations on the
specific location within a tissue from which a sample is obtained.  Numerous authors have
reported that asbestos is non-uniformly distributed in lung parenchyma and other tissues
following exposure (see, for example: Bignon et al. 1979; Davis et al. 1986a; Pooley 1982). The
incidence of lesions and other pathological effects attributed to asbestos exposure
correspondingly exhibit a non-uniform distribution.

For lung tissue samples (which tend to be among the primary interests in human pathology
studies) the relationship between sample location and asbestos concentration is particularly
important. To sample deep lung tissue reproducibly, it has been shown necessary to select a
specific section of lung parenchyma from a defined portion of the bronchio-alveolar tree.
Pinkerton et al. (1986) showed that the deposition of asbestos in the lungs is an inverse function
both of the path length and the number of bifurcations between the trachea and the site.  Thus,
analyses of samples from different animals of the same species can only  be compared
meaningfully if the samples are collected from identical locations in the bronchio-alveolar tree.
Similar, nonuniform depositional patterns have also been observed in humans (Raabe 1984).
Furthermore, due to the complex branching and folding pattern of the lung, adjacent sections of
lung parenchyma frequently represent disparate portions of the bronchio-alveolar tree (Brody
et al. 1981; Pinkerton et al. 1986).  Consequently, lung burdens derived even from adjacent
samples of lung parenchyma can show broadly varying concentrations (differing by orders of
magnitude).

Unfortunately, however, tissue samples that are available for analysis in  support of a human
pathology study are typically "opportunistic" samples, which means that they were selected and
stored for an entirely different purpose than the study at hand and, although there may sometimes
have been attempts to sample comparable locations across lungs in a general way, this is not
adequate for assuring that comparable portions of the respiratory tree are being sampled. It is
therefore seldom possible to address the effects of sample location directly.  Consequently,
comparisons of tissue burden concentrations across samples from different individuals in a
human pathology study are at best qualitative and may only be useful when averaged over large
numbers of individuals and only when large differences in concentrations (several orders of
magnitude) are being distinguished. Moreover, because the parts of a tissue that undergoes
morphologic changes induced by asbestos typically corresponds to the parts of a tissue where
asbestos burdens are the highest, even comparison of morphologic effects across tissue samples
requires proper consideration of the effects of the locations from which such tissue samples were
derived.
                                           5.7

-------
5.3    ANIMAL STUDIES

In animal studies, members of one of various species (generally rodents) are exposed to
measured doses of size-selected mineral fibers and the resultant biological responses are
monitored. Animals may be dosed either by inhalation, ingestion, intratracheal installation,
implantation, or injection (U.S. EPA 1986). Such studies are conducted for several purposes.
As with human pathology studies, animal pathology studies are those in which the transport of
asbestos structures is tracked through the various organs and tissues of the animal and the
attendant cellular and molecular changes are characterized. In parallel with quantitative
epidemiology studies, animal dose/response studies track the incidence of disease among a
population that has been exposed in a controlled manner. One of the advantages of animal
dose/response studies over epidemiology studies is that exposures are controlled and can be well
characterized. The major disadvantage is that there are many uncertainties introduced when
extrapolating the results of animal data to predict effects in humans. Therefore, attempts to
adapt such things as animal-derived dose-response factors to humans are not generally
recommended.

As with human epidemiology and pathology studies, the validity of conclusions drawn from
animal studies depends strongly on the techniques and methods used to characterize and quantify
asbestos structures either in the delivered dose or in the tissues of the dosed animals (see
Section 5.1).  The ability to reconcile conclusions derived from many animal studies with the
rest of the asbestos literature is limited because SEM was commonly employed to measure
asbestos in animal studies, but not other kinds of studies. Even many of those studies in which
TEM was employed for asbestos analysis suffer from use of non-standard methods that cannot
be easily reconciled with the more traditional TEM methods, particularly because such
specialized methods are seldom adequately documented to allow comparison.
                                                                               (
As with human pathology studies, the location of a tissue sample  excised for analysis is a critical
factor that also governs the quality of an animal pathology study (Section 5.2). However, one
potential advantage frequently available in animal pathology studies over human studies is the
ability to carefully identify and select the precise tissue samples to be analyzed. The extent that
a particular animal pathology study exploits this capability can affect the overall utility of the
study. Thus,  such issues need to be addressed carefully when evaluating and comparing the
results across animal pathology studies..

The route of exposure employed in a particular animal study is also important to consider.  Each
of the routes of exposure commonly employed in these studies (inhalation,  ingestion,
intratracheal installation, and injection or implantation) delivers different size fractions of
asbestos to a target tissue with varying efficiencies. For example, injection or implantation
studies deliver 100% of all size categories of structures to the target tissue.  However, the
efficiency that each size category is delivered by inhalation is a function of the aerodynamic
properties of the asbestos structures and the air  flow characteristics of the lungs (see, for
example, Yu et al. 1991, 1994).  Thus, the relationship between dose and exposure depends upon
the route of exposure employed. Importantly, because the ultimate goal is to understand the
effects that inhaled fibers have on humans, differences between the character of the delivered
dose in an animal study and the character that such a dose would have, had it originally been
inhaled, typically need to be addressed.

                                           5.8

-------
Regarding the measurement of health effects, many of the results in animal studies suffer from a
lack of statistical significance because of small numbers of observed tumors. Consequently,
trends cannot be established conclusively.

For animal inhalation studies, meaningful comparison of the relative deposition of asbestos dusts
across species is not direct. To extrapolate results across species, the detailed differences in the
physiology of the respiratory tracts between the species need to be addressed (Section 4.4).
However, if measurements are available for both species, differences in physiology are
addressed, and the manner in which tissue burdens are analyzed is  considered, it may be possible
to compare relative tissue doses (mass of asbestos per mass of tissue) across species.

5.4    IN VITRO STUDIES

A broad range of in vitro studies provide useful insight on the effects of asbestos.  These include,
for example, studies in cell-free systems (which have been used to evaluate such things as
asbestos dissolution rates or the kinetics of free radical formation on the surface of asbestos
fibers) and studies of the effects of asbestos on cultures of a broad  variety of cell types and
tissues.

As with other studies, the potential limitations and sources of uncertainty associated with in vitro
studies need to be considered when evaluating the validity of study results or comparing such
results to those of other studies, particular studies  of varying type.  Also, as with other studies,'
among the primary sources of uncertainty that need to be addressed for in vitro studies is the
manner in which asbestos doses are characterized and quantified (Section 5.1). For in vitro
studies (as with animal studies dosed by routes other than inhalation), this also extends to the
need to consider the relationship between the character of the asbestos dose applied in vitro and
the character that a similar exposure might possess following inhalation exposure in vivo
(Section 5.3). This can be particularly problematic for studies in tissue cultures because it is not
clear how the application of a suspension of fibers (with known concentration) to a dish
containing cultured cells can be related to doses that reach corresponding tissues following
administration to whole animals.

In vitro studies, of necessity, represent isolated components of living systems observed under
conditions that may vary radically from those under which such components operate in vivo.
Consequently, the behavior of such components may also vary radically from the behavior of the
same components in vivo. Therefore, additional study-specific considerations (concerning the
design of a study and the conditions under which a study is conducted) also need to be addressed
before evaluating the validity or relevance of the results from an in vitro study to what might
otherwise be observed in a whole animals. Examples of such considerations include:

       For cell-free systems

       •      whether conditions under which the study is conducted are sufficiently similar to
              conditions in vivo to expect that the observed effect is likely to occur in vivo; and
                                            5.9

-------
       •      if the observed variables describing the nature or magnitude of the effect are also
              likely to reflect what may occur in vivo;

       For tissue cultures

       •      whether conditions under which the study is conducted are sufficiently similar to
              conditions in vivo to expect that the observed effect is likely to occur in vivo;

       •      whether responses by specific tissues or cells in culture are likely to behave
              similarly in vivo where their behavior may be suppressed, enhanced, or modified
              in some other manner due to additional stimuli provided by responses of other
              tissues and cells that are components  of a complete organism but that may be
              lacking in culture; and

       •      whether the conditions required to establish and maintain a tissue culture for
              experimentation sufficiently alters the characteristics and behavior of the cells
              being studied to minimize the relevance of results from such a study to conditions
              in vivo.

Among the most important examples of the last consideration relates to the general need to
create immortalized cells to maintain tissue cultures. Thus, questions must always be raised
concerning whether the alterations required to create immortalized cells for culture (from what
are normally mortal cells in vivo) also alter the nature of the responses being studied.

Note, many studies in the current literature also incorporate combined aspects of several of the
four general study types described in this chapter. For these studies, a corresponding
combination of the considerations described must therefore be addressed when evaluating such
studies and comparing their results with inferences derived from the rest of the literature.
                                           5.10

-------

-------
        6.0 SUPPORTING EXPERIMENTAL STUDIES

To evaluate the plausibility of cancer risk models for asbestos it is useful to examine the current
state of knowledge regarding (1) the mechanisms that facilitate the transport of asbestos to the
various target tissues of interest (i.e., the lungs and mesothelium) and (2) the mechanisms that
contribute to the development of cancer in these target tissues. Accordingly, a review of the
relevant literature is provided in this chapter to assure that the quantitative analysis in Chapter 7,
and the proposed approach for assessing asbestos-related risks in Chapter 8, are qualitatively
consistent with general implications from the broader literature. This review, although
extensive, is not exhaustive.  A much larger number of studies was reviewed than are actually
cited. Studies that are not cited are largely confirmatory or redundant. In selecting studies for
review, special effort was expended to assure that opposing views (particularly for controversial
issues) were adequately represented.

Although much progress has been made over the last decade toward elucidating the fiber/particle
mechanisms that contribute to transport and subsequent cancer induction, at least two critical
data gaps remain:

       •      no one has yet been able to track a specific lesion induced by asbestos in a
              specific cell through to the development of a specific tumor.  There have been
              experiments that show altered DNA and other types of cellular and tissue damage
              that are produced in association with exposure to asbestos. Other studies have
              demonstrated that various tumors of the kinds that result from asbestos exposure
              exhibit patterns of DNA alteration (or other kinds of cellular damage) that are
              sometimes (but not always) consistent with the earlier cellular changes associated
              with asbestos exposure. There are also studies that show that exposure to
              asbestos can lead ultimately to development of tumors. However, these types of
              studies have yet to be linked; and

       •      the specific target cells that serve as precursors to tumors in various target tissues
              are not known with certainty.

Because of the first of the above limitations, researchers have tended to report on a broad range
of tissue and cellular effects induced by asbestos  that may lead generally to various kinds of
cellular damage or injury. Cytotoxicity, for example, is one of.the endpoints typically tracked as
a marker for asbestos-induced injury. However, not all of these effects necessarily contribute
(either directly or indirectly) to the development of cancer. Therefore, one of the goals of the
following discussion is to distinguish among effects that likely contribute to the development of
cancer from those that are less likely or unlikely to contribute. .Of course, delineating such
distinctions are subject to the limitations of the current state of knowledge.

In addition, because the relative effects of fiber size, shape, and mineralogy need to be elucidated
to better indicate how asbestos concentrations should be characterized to support risk
assessment, studies that address.these topics are highlighted. Of particular interest are studies
that (1) contrast the effects of different sized fibers, (2) contrast the effects of fibers and non-
                                          6.1

-------
fibrous particles of similar mineralogy, and (3) contrast the effects of fibers of comparable
morphology (size and shape), but differing mineralogy.                      ,

The types of studies that have contributed to the state of knowledge of the effects of asbestos (in
addition to the human epidemiology studies that are evaluated in Chapter 7) include:

              whole animal inhalation studies;              •       '
              whole animal instillation studies;
              whole animal injection, implantation studies;
              human pathological studies;
              in vitro studies in cell cultures; and
              in vitro studies in cell-free systems.

Depending on the outcome(s) monitored, the animal studies may alternately be categorized as
retention studies, histopathology studies, or dose-response studies.

Each type of study possesses certain advantages and exhibits certain limitations, which have
previously been described (Chapter 5) along with descriptions of the nature of each of these
study types. In addition to the advantages and limitations that are attributable to the type of
study, the quality of the characterization of asbestos (or other particulate matter) determines the
utility of the study for addressing issues associated with fiber morphology and mineralogy.
Unfortunately, for many published studies, both the characterization of the asbestos (or other
particulate matter) and descriptions of the manner in which such materials were handled are
insufficient to establish the detailed morphology or mineralogy.  Such limitations need to be
considered when comparing across study results or evaluating the validity of study conclusions.

The rest of this chapter is divided into separate sections that address the set of factors that have
been previously identified (Chapter 3) as those that determine the biological activity of inhaled
asbestos (which is the exposure route of primary concern for humans). These are:

       •     the extent that asbestos structures are respirable and the pattern of deposition of
              inhaled structures;

       •     the extent that deposited structures are subsequently cleared or degraded;

       •     the extent that deposited structures are transported or migrate to the various target
              tissues; and

       •     the extent that retained structures induce a biological response in each target
              tissue.

6.1    FACTORS AFFECTING RESPIRABILITY AND DEPOSITION

Discounting systemic effects resulting from other forms of exposure, factors affecting
respirability are common to  all of the toxic endpoints associated with asbestos exposure
considered in this study (asbestosis, lung cancer, and mesothelioma).  Moreover, respirability is
common to the factors affecting the toxicity of inhaled, insoluble particles in general.  To be

                                            6.2

-------
respirable, an inhaled particle must pass the blocking hairs and tortuous passageways'of the nose
and throat and be deposited in the lungs.  Particles deposited in the naso-pharyngeal portion of
the respiratory tract are not considered respirable.                                      -

Not all of the inhaled particles that reach the lungs will be deposited.  Small particles may not
impact lung surfaces during inhalation and are subsequently exhaled.  Once a particle impacts oh
a surface, however, it is likely to remain because the surfaces of the lungs are wetted with a
surfactant (Raabe 1984).

Adverse health effects potentially result when particles that are deposited in the lungs remain in
contact with the tissues in the lung for a sufficient period of time to provoke a biological
response. To affect the mesothelium, an offending particle may-also need to migrate or be
transported from the lung to this surrounding tissue.  However, due to the proximity of the
mesothelium to peripheral portions of the lung parenchyma (which include locations where
particles are typically deposited), it is also possible that diffusable molecules produced in lung
tissue in response to deposited particles can have an adverse effect on the mesothelium (see, for
example, Adamson 1997).  Such effects are considered among the mechanisms of disease
induction addressed in the discussion of biological responses (Section 6.3):

The interplay between deposition and removal (clearance) of inhaled particles is an important
determinant of biological activity and separating the influence of these two processes in the
pathology of asbestos-induced disease is difficult. The term "retention" is used here to represent
the fraction of particles remaining in the lungs beyond the time frame  over which only the most
rapid removal processes (i.e., muco-ciliary clearance) are active.  The factors affecting retention
are addressed further in Section 6.2.

Published inhalation studies divide the respiratory tract into three units (see, for example, Raabe
1984). The naso-pharyngeal portion of the respiratory tract extends from the  nares in the nose
through the entrance to the trachea.' The tracheo-bronchial portion of the respiratory tract
includes the trachea and all of the branching bronchi down to the terminal bronchioles.  The
respiratory bronchioles and the alveoli, which are collectively referred to as the "deep lung", are
the bronchio-alveolar (or pulmonary)'portion of the respiratory tract. For a more detailed   '  .';•
description of the features of the respiratory tract, see Section 4.4.                           ;

The dimensional requirements for respirability have been studied  and reviewed in several studies
(see, for example, Raabe 1984 or U.S. EPA 1986). A more recent review is also presented by
Stober et al. (1993).  Much of the data reviewed in these studies is based on earlier studies in '
which researchers exposed animals or human volunteers to a series of monodisperse spherical-
particles (although Stober et al. also reviews fiber experiments).  In this manner, the impact of' :
the diameter of spherical particles on respirability was elucidated. The respirability of fibrous
materials (such as asbestos) tends to be described in terms closely associated with those
employed for spherical particles, but with adjustments for density and shape.  Importantly,
because respirability is a mechanical process: the size, shape, and density of a particle (or fiber)
determine its respirability along with the morphometry of the airways through which the particle
passes (Stober et al. 1993). Other than affecting the particle's density  or the distribution of fiber
shapes, the chemical composition (mineralogy) of a particle (or fiber) does not influence
respirability.

                                           6.3

-------
diameter of a hypothetical spherical particle of unit density that would exhibit the same settling
velocities and aerodynamic behavior as the real, non-spherical particle of interest.  Factors that
affect the aerodynamic equivalent diameter are density, true diameter, true length (for elongated
particles such as fibers), and the regularity of the particle shape.

Harris and Timbrell (1977) Findings. Because fibrous particles tend to align primarily along
the axis of travel under the flow conditions found in the lungs, respirability is predominantly a
function of the diameter of a fiber and the effect of length is secondary (Harris and Timbrell
1977). Fibrous structures with aspect ratios (ratio of length to width) >3:1 behave like spherical
particles (of similar density) with diameters up to 3 times larger and exhibit only a very weak
dependence on length. As previously indicated, however, the aerodynamic equivalent diameter
of a fibrous structure must also be adjusted for the  effects of density. This is demonstrated in
Figure 6-2  where the true diameter of a fiber is graphed on the top horizontal axis against
spherical (aerodynamic equivalent) diameters  on the bottom horizontal axis.  Figure 6-2 is an
overlay of Figure 6-1. Note that, to adjust for the density of asbestos, the true diameters listed in
the figure have been shifted to the right of where they would appear if the relationship was
exactly 1/3 of the aerodynamic equivalent diameter.

Two vertical dashed lines in Figure 6-2 represent effective limits to the range of respirable
asbestos. The line on the left  side in the figure represents the limiting diameter of the smallest
chrysotile fibril (about 0.02 \im true diameter) and thus represents a lower limit to the diameter
that is of concern when considering asbestos.  The vertical line to the right represents the cutoff
where deposition in the deep lung becomes unimportant due to removal of such  particles by the
naso-pharyngeal passageways. This latter cutoff corresponds to a true fiber diameter of 2.0 um,
which theoretically represents the upper limit  to the size of asbestos that is respirable.  As
indicated in the figure, however, deposition in the deep lung drops precipitously for fibers
thicker than about  0.7 um so that no more than a.few percent of asbestos fibers thicker than
approximately 1 \im actually reach the deep lung.

Harris and Timbrell (1977) also evaluated the relationship between the overall shape of a particle
and the extent of deposition.  Over the range of diameters that potentially represent the range of
asbestos fibers likely to be encountered, pulmonary deposition decreases with increasing
complexity of shape beyond simple cylinders  (such as clusters and matrices (see Section 4.2) at
the expense of increasing naso-pharyngeal or  tracheo-bronchial deposition.  This change also
becomes increasingly important as the length  of the structure increases.  For structures <25 urn
in length,,the difference in deposition between simple fibers and complex clusters or matrices
may vary by up to a factor of 2 with the complex structures being more likely to be removed in
the naso-pharyngeal portion of the  respiratory, tract and the fibers more likely to be deposited in
the deep lung. At  100 jim lengths,  the fraction of complex structures that survive passage
through the nose and throat in comparison with  simple fibers may vary by a factor of 5. This
means that large structures become relatively  less respirable as their complexity increases.
However, during mouth breathing large clusters and matrices may enter the deep lung.
                                            6.6

-------
 Figure 6-2.  Fractions of Respirable Particles Deposited in the Various Compartments of
  the Human Respiratory Tract as a Function of the True Diameter of Asbestos FibersS|M
         ,002
                             The Diameter of Asbestos Fiber
                                 .02
      t,o
     0,01
                                 • t  ,                  1.0              ,-   1] 10,0
            Diameter of Spherical Particle with Density Equal to One gfcm (urn)
'Source of original: Raabe 1984                                               |
bAssumes a typical tidal value of 1,450 cm3 and a rate of 15 breaths per minute      [•
The relationship between true diameters and aerodynamic equivalent diameters derived from
Harris and Timbrell (1977). Diameters adjusted for shape and density of asbestos fibers.
dAerodynamic equivalent diameter

                 Confidential: Need Permission to Reproduce this Figure
When all of the factors that Harris and Timbrell (1977) addressed are considered, the efficiency
of the deposition of asbestos structures in the deep lung is maximal for short, thin, single fibers
(<10 nm in length with a true diameter <0.7 ^m). The efficiency decreases slowly with
increasing length (up to an effective limit of 200 urn), moderately with increasing complexity of
shape, and rapidly with increasing diameter (up to an effective limit of 2.0 jjtm, true diameter).
Thinner fibers, down to the lower limit of the range for asbestos fibers (0.02 \im, true diameter),
are deposited with roughly the same efficiency. Approximately 20-25% of the fibers between
0.7 and 0.02 |im in diameter (and <10 \im in length) are deposited  in the deep lung.
                                         6.7

-------

Sussman et al. (1991a,b) Findings.  Based on a series of experiments on human tracheal
bronchial casts, Sussman et al. (1991a,b) also developed models of fiber deposition in the human
lung.  Such experiments are in fact illustrative of several research groups who have developed
deposition models based on results from experiments on airway casts (for a review, see Stober et
al. 1993).

The results reported by Sussman et al. (1991a,b) appear to be generally consistent with the
results reported by Harris and Timbrell (1977) and Yu and coworkers (described below),
although the manner in which their results are reported make them somewhat less directly
comparable. Briefly, Sussman et al. (1991a,b) report that deposition increases along most
generations of the bronchial tree with increasing fiber length and increasing airflow rate for any
fixed aerodynamic diameter.  This increased deposition efficiency is demonstrated for airway
generations at least through the ninth bifurcation and is implied to continue through to airway
generations that would be representative of the respiratory (pulmonary) portion of the lung (i.e.,
airway bifurcations greater than approximately 16 to 22). For definitions and a description of
airway generations,  see Section 4.4.

Findings of Yu and Coworkers. In a series of studies, Yu and coworkers combined an
improved model of human lung physiology (Asgharian and Yu 1988) with a series of more
rigorous equations to describe fiber mobility (Chen 1992) and used these to evaluate the
deposition of various types of fibrous materials in the lung. The trends indicated in their studies
show general agreement with those reported by Harris and Timbrell (1977), but with several
notable refinements.

In a study of refractory ceramic fibers (Yu et al. 1995a), a maximum deposition efficiency of
15% is reported for  fibers that are approximately 6 \im long and approximately 1 jim in diameter.
This is close to the fiber size  at which maximal deposition is reported by Harris and Timbrell
(1977). As with Harris and Timbrell (1977), Yu et al. (1995a) also report that deposition
efficiency decreases precipitously as diameter increases beyond 1 p.m and decreases more slowly
as diameter decreases below  1 um. For thinner structures, deposition efficiency increases with
both decreasing width and length.  As fibers get longer, optimum deposition occurs with
decreasing thickness.  Thus, for example, a maximum deposition rate of 10% occurs for fibers
that are 20 u,m long at a thickness of 0.8 urn.

In a study of silicon-carbide whiskers (Strom and Yu 1994), the deposition model is extended to
fiber widths as narrow as 0.01 fim. Results from this study indicate that fibers between 0.01 and
0.1 \im in thickness are deposited with a minimum efficiency of 5% up to lengths of
approximately 40 \im before  efficiency drops below 5%.  For thin fibers (thinner than 0.5 urn),
shorter fibers tend to be deposited in the deep lung much more efficiently than longer fibers.
More than 25% of thin fibers shorter than  1 p,m are deposited in the deep lung following
inhalation. Strom and Yu (1994) report that the efficiency of deposition in the deep lung of long
structures increases substantially during mouth breathing.

Comparing the results reported for refractory ceramic fibers (density=2.7 g/cm3) and silicon-
carbide whiskers (density=3.2 g/cm3), it also appears that the efficiency of deep-lung deposition
increases for thinner and for  longer structures as the density of the structures increases. Given
the observed density effect, chrysotile fibers that are longer than approximately 6 urn and thinner

                                           6.8

-------
                                                                             I
than 1 (Jim would be deposited in the deep lung less efficiently than (denser) amphibole fibers of
the same size. However, shorter and thicker chrysotile structures would be deposited somewhat
more efficiently than similarly sized amphiboles. This suggests that a greater fraction of the
mass of chrysotile that gets deposited in the deep lung will be composed of very shortfibers and
somewhat longer bundles than the mass fraction of short fibers or longer bundles in the air
breathed.  Also, to the extent that chrysotile fibers are curved, these would be deposited
somewhat less efficiently than straighter (amphibole) fibers of comparable size.

Based on the deposition efficiencies predicted by Yu and coworkers, fibrous structures that reach
the deep lung in humans are effectively limited to those thinner than approximately 1 urn. Given
that fibrous structures have traditionally been defined as particles exhibiting aspect (length to
width) ratios >3:1 (Walton 1982), it is clear that only particles shorter than 3 jam could
potentially be respirable and still be excluded from the definition of a fibrous structure based on
aspect ratio.  Therefore, the thickness constraint for all longer structures is best described as a
maximum width (rather than an aspect ratio) when defining the range of structures that
potentially contribute to biological activity.

Rats versus Humans. Yu and coworkers also modified their models to evaluate the rates that
fibrous materials are deposited in rat lungs and compared these with results for humans. Such
comparisons have implications for the manner in which results from animal inhalation studies
are extrapolated to humans.

Results from Yu et al. (1994) suggest that pulmonary deposition of all fibrous structures with
lengths between about 1 and  100 um and thinner than approximately 1 urn occurs at much higher
rates in rats than in humans.  Fibers as long as 90 um are deposited in rat lungs at efficiencies
exceeding 20% while fewer than 5% of structures this long are deposited in the pulmonary
region of human lungs. In fact, it is only structures between 1 and about 20 um within a very
narrow range of thicknesses (centered around I um) that are deposited more efficiently in the
deep lungs of humans than in rats.

Yu et al. (1995a) also indicate that, even when deposition efficiencies are comparable.in rats and
humans, due to differences in the total lung mass and breathing dynamics across species, the
resulting lung burdens (i.e., the mass or number of structures per mass of lung tissue) are 5-10
times higher in the rat than in humans for any given exposure. Lung burden per lung surface
area are also higher in the rat than in humans.                                    i
                                                                             i
To illustrate, assume rats and humans are similarly exposed to a concentration of 0.1 f7cm3
(100 f/L) of some fibrous material with a length at which both species retain approximately 10%
of the fibers inhaled. Table 6-1 then indicates the calculations required to determine the relative
rates at which the lung (volume and surface area) burdens in each species would develop.
                                           6.9

-------
  Table 6-1. Estimation of Lung Volume and Lung Surface Area Loading Rates for Rats
                                      and Humans
Species
Human
Rat
Species
Human
Rat
Body
Weight (kg)
70
0.15
Breathing
Rate
(L/min)
21.7
0.13
Lung Volume
(L)
5
0.01
No. Fibers
Inhaled per
Minute
(f/min)
2170
13.3
Lung Surface
Area (m2)
140
0.4
No. Fibers
Deposited per
Minute
(f/min)
217
1.3
Rest Breaths per
Minute (bpm)
15
70
Lung Volume
Loading Rate
(f/L-min)
43.4
130
Tidal Lung
Volume (L)
1.5
0.0019
Lung Surface
Area Loading
Rate
(f/m2-min)
1.6
3.3
From Table 6-1, it is clear that rats exposed to comparable airborne concentrations as humans
will increase their loading of fibers per volume (or mass) of lung at a rate that is approximately
3 times that of humans (for fibers in sizes that are deposited with 10% efficiency in both
species). Similarly, the fiber load per surface area of lung will increase in rats at a rate that is
approximately twice that of humans. Moreover, even higher relative mass or surface area
loading rates are expected for  the rat than shown in the table, due to the greater efficiency with
which most fiber sizes are deposited in rat lungs.  Data used to compute the loading rates in the
table (which are also presented) are derived from Gehr et al. (1993) and supplemented with
information from Stober et al. (1993). A more detailed description of this information is
provided in Section 4.4.

6.1.3   The Effects of Electrostatic Charge on Particle Respirability

Electrostatic charge has been shown to affect the retention of particles within the lungs (see, for
example, Vincent 1985).  Since processes that generate airborne particles generally involve some
form of abrasion, airborne dust particles frequently exhibit varying degrees of electrostatic
charge. Although this potentially leads to variation in the efficiency of particle retention in the
lungs as a function of the source of the dust, a detailed relationship between surface charge and
retention was not described in this paper. A more detailed and quantitative treatment was
developed by Chen and Yu (1993) and the implications of the Chen and Yu model are described
below (following discussion of the results of Davis et al. 1988a). Davis et al. (1988a) report that
animals exposed to dusts containing  fibrous chrysotile, whose surface charge is reduced with a
beta minus source, retain  significantly less chrysotile than animals dosed with dusts containing
particles whose surface charge has not been reduced. However, the magnitude of the difference
in the mass of fibers retained is less than a factor of 2, implying that the absolute variation due to
this effect may be small.  Further research in this area is needed.

Chen and Yu (1993) report that, based on modeling of lung deposition, overall deposition
increases with increasing charge density on the particles inhaled. However, due to the pre-
                                           6.10

-------
iltering by the naso-pharyngeal and tracheo-bronchial portions of the respiratory tract; the effects
of electrostatic charge on deep lung deposition appear to be only slight to modest.    |

Given the results of the above studies, the overall effects of electrostatic charge on particle
deposition in the deep lung appear to be relatively minor. Therefore, such effects do hot need to
be considered explicitly when evaluating the health consequences of asbestos.

6.1.4   General Conclusions Concerning Particle Respirability

Based on the information provided in the last several sections, it is apparent that in humans:

       •     deposition of asbestos fibers in the pulmonary portion of the lung occurs
             primarily at alveolar duct bifurcations;

       •     electrostatic effects on pulmonary deposition are likely minor;

       •     fibers that are deposited  in the pulmonary portion of the lung are largely thinner
             than approximately 0.7 jim and virtually all are thinner than 1 urn (except during
             mouth breathing, when thicker and more complex structures may be respired);

       •     the length of a fiber has limited impact on respirability up to a length of
             approximately 20 jim, but the efficiency of deposition of longer fibers decrease
             slowly with increasing length for longer fibers;

       •     as the length of the fibers that are inhaled increases, the thinner fibers are
             deposited with greater efficiency.  Thus, the longer the fibers inhaled, the thinner
             the fibers retained;

       •     due to differences in density, shorter and thicker chrysotile structures will be
             deposited more efficiently in the pulmonary portion of the lung than
             corresponding amphibole structures and longer and thinner amphibole .structures
             will be deposited more efficiently than corresponding chrysotile structures;

       •     curly chrysotile structures are less likely to reach the pulmonary portion of the
             lung than straight amphibole (or chrysotile) structures;

       •     except for a very narrow range of fiber sizes (centered around 6 urn hrlength and
             1 jim in diameter), virtually all size fibers are deposited with greater efficiency in
             rat lungs than human lungs;

       •     due to body morphology and the dynamics of breathing, rats exposed to similar
             air concentrations will accumulate fiber burdens both per mass (volume) of lung
             tissue and per lung surface area at a rate that is several times the rate such burdens
             accumulate in humans; and

       •     the dynamics of fiber lung deposition can now be accurately predicted in great
             detail using currently available models.
                                           6.11

-------
6.2    FACTORS AFFECTING DEGRADATION, TRANSLOCATION, AND
       CLEARANCE

Degradation and clearance mechanisms compete with deposition to determine the fraction of
asbestos that is retained in the lungs. Other (translocation) mechanisms mediate the movement
of asbestos from sites of initial deposition to various target tissues within the lung and
mesothelium. These factors affect all of the toxic endpoints of interest. Studies indicating the
dependence of the various contributing mechanisms on fiber size and mineralogy are
highlighted,  as well as studies indicating differences between mechanisms in humans and
laboratory animals.

The three units of the respiratory tract defined in the last section (naso-pharyngeal, tracheo-
bronchial, and bronchio-alveolar units) differ primarily by the types of clearance (and
translocation) mechanisms operating in each unit (Raabe 1984). These are summarized in
Table 6-2 along with rough estimates of the time frames over which each mechanism may
operate (to the extent that such estimates are available in the literature).

Briefly, the structures of the nose and throat are bathed in a continual flow of mucous, which is
ultimately swallowed or expectorated. The mucous traps deposited particles and carries them
out of the respiratory tract. The air channels of the tracheo-bronchial section of the respiratory
tract are lined with cilia and mucous secreting cells. As in the nose and throat, the mucous traps
particles deposited in these air pathways and the ciliary escalator transports the mucous up to the
throat where it may be swallowed or expectorated.  Neither the alveolar ducts nor the alveoli of
the pulmonary compartment of the lung are ciliated (inferred from St. George et al. 1993).
Therefore, particles deposited in this section of the respiratory tract can only be cleared by the
following mechanisms:

       •     if the deposited particles are soluble, they may dissolve and be transported away
             from the lungs in blood or lymph; or

       *     if they are sufficiently compact, they may be taken up by alveolar macrophages
             and transported outward to the muco-ciliary escalator of the tracheo-bronchial
             portion of the respiratory tract.

Due to a combination of chemical and physical stresses in the environment of the  lung, deposited
asbestos structures may degrade by  splitting.  Longitudinal splitting, primarily of bundles,
produces thinner structures and transverse splitting produces shorter structures. In both cases,
the number of structures produced may be larger than the number  of structures initially
deposited.
              s
By changing the size and number of structures that were initially deposited in the  lungs, splitting
may affect the rates and efficiency with which the various other degradation and clearance
mechanisms operate.
                                          6.12

-------



be
ES
S

^
£
13
V
£
1
1
.ss
'B
1
u>
_a
6S
S
D.
0
9
O

V
ja
^
S
i
0
0>
1
£
L.
£
I
3
1
J
1
1
1




V
1

o>


01
W

"S
11
is
i

1 *

B IS
go

9> "
2
(Component




































asal-pharyngeal
Z




CO
CT\
1>
1

1
O
Z

i
w
o
Z



• S


13
SCO
.„ (S
c o
S d

en
JD
"o
1
CU


w rt
S g
3 3
a a:
M
1
13 t
£ 1
•73 C
Expectoration ani
Muco-ciliary Tra




































racheo-bronchial
H




oo

~
u


•*_>
1
tt
o
12

-*-*
o
1
o



z
u
N

cs
9
o
. o

JS
o
1
CL


g
1
X


I
g
Muco-ciliary Tra

































^
>
1
!e
ulmonary (Bronc
a-
£
o r^ o
C^N ON OS
2 ~ o CM 2; — —
T3 Irt ^ Ov —j -3 -jri
CSU _^ T3 — etttSU
*J U . g _ <^ 4-> t_i
^X>^-N'« ^i2 JJt.^J^^
JO ;« S 'S gc S^^ON
o^osp S.s '52^2

If, SI ill ' 1
^^QrS^^^S1^ -
"•Cotl '>+*'•' "Si. W
co c4 J. 4« ^ • a> o
QL 04 iS £s K >\ &
-D €
3 ^.| I
I ^ § |: 1 1
S" "3 -5 S'1 * -o
1 ^ '§ •§!'•« .sa"
5. ' -1 1 f-i 1
l§ i ! I'll
|| | | Hi
5 O t« '3 o3
•< S . 3 E ' H

to
CD

-------

Ml
a
5
ja
^x
CQ
£ •
«
•s

(^
mm
X
I
•.c
J3
«
z
es


o
C
£
a!




v •§

85 w

8
IN
? s2S
a w
i

.a >.

EE?


IB
.S
I



i
'S
ftl
B
4
BA
.2







































Mechanisms































I ft
•tS "2 >
'a' & 13 o

•jjj € &. u
"5 ao >, "^
'(X 3 -0 g
UP c O
w 5 o '5
f i 11
II 11
(Diffusive transport thi
(Forced mechanical tra
Seiquestration
(Phagocytosis and inte
(AM phagocytosis, inn








































Imonary (Interstitial)
£
1 ~
o **

^ "^ 	 JH ^y\
•w o 

D c




• 'v
I
S
O O O e
o o o g
Js" ^" M S2-

CO
JW '*"**
1 1 ' ^ 1
* < 1^1 2-
•2 "S" .§ 13
 ^ o
h .55 O
U ^ S /— N -3
ti '3 *3 aj w
,g To _— 3 S
.£ S3 •§> S3 ^
+- M WD e 3 'fl «
c5 O 3 .S ®
Dd Q ^ 1> ^ 3 c
i o s iiij !i
1 flfill l|
s 1. 1 11 11 ••"!•
1 I 1 1 1 11 B 1 1
*^s_ cti ^^ ^-i Q CO '^ ^3 O ^ 03
^* Jr i U^ C *5 <2 '^ )rt • ^" Ctc E
?s *^^ fti t^ ^^ 'R* y ftf ^
_C ^S ^ ^S •™- d< ^^j ^J ^r Jf» ^J ^^

<£>

-------
00
a
a
1
B

CM
O
<*>
B
.22
ja
z
§
W)
.5
55
S?
a
O
at
i
£
Js -s
leared by I
(continuet
U
S
V*
1
j>
1

1

$
•I
S

V
8



w
O) o
••• s2
1/5 w

y
2^ ^^
"M
.S W

*

.a u
1"S
20

m u
*3 *S sfl
•s^l*
fifl®
^ flW
v &
. s£


<«
a.



issue/Lung Regime
H































Mechanisms
































he Pleura
H





















.


1
2
0 3
S '-s ^
1 3 =3
>, 2 w

2 '3 o 15
PM Phatogyctosis, Transport
Dissolution in extracellular fl
Sequestration
(Encapsulation by granul
'(Phagocytosis by mesoth

«i
«3
- '
1-d
O O
f\ .f
^ T"
.2 a
4-» 4>
TO ^
*C w
i '§
1. i-
w .2
i_ O
0 **
 o.
fli "3
4^ W
1) '^S

g |
U-, (,_
 u
S §*
*3 W
*<-i C
o ,g
11
A ,
If
0 (>
ofl rt
ty c^
.-^ O

-------
Particles and fibers that are deposited in the pulmonary portion of the lung may also be
transported by a variety of mechanisms into and through the epithelium lining, the alveolar
ducts, and alveoli to the underlying interstitium and endothelium that are located within the
interalveolar septa (see Section 4.4). In those portions of the lung parenchyma that lie proximal
to the pleura, such mechanisms may also facilitate transport to the mesothelium.  Putative
mechanisms by which such transport may occur include:

       •      if particles are sufficiently compact to be phagocytized by alveolar macrophages,
              they may be transported within macrophage "hosts" through the epithelium to the
              interstitium;

       •      if particles are sufficiently compact to be phagocytized by the epithelial cells
              lining the air passageways of the deep lung, they may be transported into cell
              interiors or transported through to the basement membrane, the interstitium, the
              endothelium, and (eventually) the pleura;

       •      particularly when associated biological effects that cause changes in the
              morphology of epithelial cells, particles may diffuse between the cells of the
              epithelium to underlying tissues; and/or

       •      particles may be transported through respiratory epithelium mechanically due to
              physical stresses associated with respiration within the lung.

Although the transport of fibers and particles from airway luniena to the interstitium is apparent
in many studies (see below), the precise mechanisms by which such transport actually occurs has
yet to be delineated with certainty.

Particles deposited in the  interstitium can also be cleared and the processes by which these
particles are ultimately cleared are similar to,  but may be substantially slower than, the
mechanisms by which particles deposited in airway spaces can be cleared.  Such mechanisms
include:

       •      if the deposited particles are soluble, they may dissolve and be transported  away
              from the lungs in blood or lymph; or

       •      if the particles of the interstitium are sufficiently compact to be phagocytized by
              interstitial  macrophages, they may be taken up and transported to the lymphatic
              system for removal.

The mechanisms by which particles that reach the pleura and mesothelium may be cleared are
also similar to those operating in the interstitium:

       •      if the deposited particles are soluble, they may dissolve and be transported  away
              from the lungs in blood or lymph; or
                                          6.16

-------
       »     if the particles that reach the pleura are sufficiently compact to be phagocytized
              by pleural macrophages, they may be taken up and transported to the lymphatic
              system for removal.                                       •       '

Particles cleared from the pleura by macrophages appear subsequently to be deposited at sites of
lymphatic drainage along the pleura (i.e., at lymphatic ducts) from where they are .ultimately
cleared in lymph (Kane and MacDonald 1993).                                  ''
                                                •   -    .                     il
The various degradation, clearance, and transport mechanisms that affect the retention of
asbestos in the lung and other target tissues (identified above) exhibit disparate kinetics that may
be further altered by the size, shape, mineralogy, and concentration of the particles affected.
Therefore, the kinetics of these mechanisms are considered below. The mechanisms evaluated
include:                                                                       \

       •     dissolution;
       •     muco-ciliary transport;                                           '
       •     macrophage phagocytosis and transport; and                       ;
       •     diffusiona! transport.       .                                     i

Evidence for the existence of these mechanisms and inferences concerning their kinetics derive
primarily from retention studies, which may include both studies of retained structures in
animals following either short-term or chronic exposure, or human pathology studies in which
the lung burdens of deceased individuals are correlated with their exposure history.  Other
information also comes from in vitro studies. Various, increasingly sophisticated models have
also been developed to predict the individual and combined effects of these mechanisms.
                                                                 1           !'
6.2.1   Animal Retention Studies

Retention studies track the time-dependence of the lung burden of asbestos or other particulate
matter (i.e., the concentration of particles in the lung) during or following exposure. Thus, such
studies are designed to indicate the degree to  which inhaled structures are retained. Depending
on the time frame evaluated, however, effects due to deposition and those due to clearance may
not easily be distinguished in such studies. Moreover, due to the near impossibility of isolating
the various compartments of the lung when preparing for quantitative analysis of tissue burden
(e.g., the pure respiratory components vs. the larger airways or the tissues directly associated
with airway lumena vs. the underlying interstitium or endothelium), it is nearly  impossible to
separate the effects of the various clearance mechanisms, which typically operate over vastly
different time scales (Table 6-2). This is why modeling has proven so important to  '
distinguishing effects attributable to individual mechanisms.

Results from retention studies must be evaluated carefully. In addition to the limitations
highlighted above, the lung burden estimates  from such studies may be affected by the manner in
which asbestos is isolated from lung tissue for measurement and the manner in whichJthe
concentration of asbestos is quantified (Chapter 5). For example, lung burden estimates may
vary substantially depending on what portions of lung parenchyma are sampled  or whether
whole lungs are homogenized.  Results may also vary depending on whether lung tissue is ashed
or dissolved in bleach  during sample preparation. More importantly, because several Clearance

                                          6.17

-------
mechanisms are affected by the size and even the mineralogy of the structures being cleared,
studies (particularly older studies) that track lung burden by mass or by total fiber number may
not adequately capture such distinctions.

6.2.LI       Studies involving short-term exposures ,

The latest retention studies tend to focus oh the fate of long fibers (typically those longer than
20 um) in support of the generally emerging recognition that these are the fibers that cannot be
readily cleared from the pulmonary compartment of the lung and that, not coincidentally,
contribute most to disease (further addressed in Section 6.4).

Hesterberg et al. (1998a), for example, tracked the time-dependent retention in rats of two fiber
categories: (1) WHO fibers1 and (2) WHO fibers longer than 20 |im for a range of man-made
vitreous fibers (MMVF's), a refractory ceramic fiber (RCFla), and amosite following a 5-day
(6 hr/day), nose-only exposure. Rats were sacrificed at intervals up to a year following
exposure. The amosite was size-selected to contain a high proportion of fibers longer than 20
u.m.  Aerosol concentrations were also adjusted to maintain target concentrations of 150 fXcm3
for long fibers for each sample tested. Airborne mass concentrations varied between  17 mg/m3
for amosite to as much as 60 mg/m3 for the other fiber types.  Lungs (without trachea or main
bronchi) were weighed and stored frozen. For analysis, each lung was dried to constant weight,
minced, and a portion was ashed. The ashed portion was further washed with filtered, household
bleach, then filtered and applied to an SEM stub. Fiber numbers and dimensions (in both
aerosols and tissue) were determined by SEM with a minimum of 200 fibers counted. In
addition, analysis continued until a minimum of 30 fibers longer than 20 (im were counted.

In their study, Hesterberg et al. (1998a) tracked the ratios of retained fiber concentrations with
time to the concentration retained 1-day following cessation of exposure. The observed time-
dependent decay in these ratios were then fit to one-pool (single first order decay) or two-pool
(weighted sum of two first order decays) models.  With zero time assumed to be the time
immediately following cessation of exposure. The authors recognize that at least some clearance
likely takes place during the 5 days of exposure so they expected the assumption that retained
concentrations at the end of exposure on day  5 to be equal to deposited concentrations would
cause their analysis to slightly underestimate  clearance rates. They also recognized that waiting
24 hours after cessation of exposure to measure retention allows some short-term clearance of
upper airways, so that they expected their analysis would better focus on slower clearance from
deeper in the lung.

Results reported by Hesterberg et al. (1998a)  indicate that the dimensions and concentrations of
fibers in aerosols from the five synthetic fibrous materials were all similar, but that the amosite
aerosol contained a substantially  greater number of fibers (including the longest fibers) and that
the fibers, on average, were somewhat shorter and substantially thinner than the other aerosols.
Of the fibers initially deposited in the lung (based on measurements made 1 day following
cessation of exposure), comparable fiber numbers of long (>20 um) fibers were retained across
        1 WHO fibers are those longer than 5 (im, thinner than 3 ^m with an aspect (length to width) ratio greater
 than 3 (WHO 1985).

                                           6.18

-------
all six fiber types.  Deposited concentrations of fibers 5-20 u.m in length were more variable, but
values within one standard deviation still overlapped. About 6 times as many short amosite
fibers (<5 ^m) were initially deposited than for any of the other fiber types. The authors also
indicate that the dimensions of retained fibers were generally shorter and thinner than the
original aerosol and were much more similar across retained fiber types than the original
aerosols.

Clearance of long fibers (>20 urn) for all six fiber types could best be described using a two-pool
model. The first pool cleared relatively rapidly (within the first 90 days) and represented a
minimum of 65% of the lung burden observed 1 day following exposure. The secondjpool
cleared much more slowly.  For amosite fibers in the second pool, during the approximately 275
days of clearance, retention was only reduced to 80% of the 90-day value. In contrast, all five of
the synthetic fibers were reduced to less than 30% of their 90-day value during this period. For
amosite, the first pool decayed with a half-life of 20 days (90%CL:  13-27) and all of the other
fibers with half-lives of 5-7 days (with varying confidence bounds). For the slower pool,
amosite fibers exhibited a half-life of 1,160 days (90% CL: 420-»)  with the other fibers showing
half-lives varying between 24 and 179 days. The combined, weighted half-life for amosite was
418 days (90%CL: 0-1060). The authors also note that data reanalyzed from an earlier study
(Hesterberg et al. 1996) indicate a corresponding weighted half-life for crocidolite of 817 days
(246-°°)*and indicate that this was best fit using a single exponential (a one-pool model).
                                                                             i
Hesterberg et al. also indicate that in this and previous studies approximately 20-60% of long
fibers typically clear from the lung within 2 weeks post-exposure.  They further suggest that this
rapid clearance  may be attributable to muco-ciliary clearance from the upper respiratory tract.
They further report from the  present study that short amosite fibers cleared much more rapidly
than long fibers. Fibers <5 u.m in length were reduced by 90% in the first 90 days (in-
comparison to 65% for long fibers).  However, from 90 to 365 days, little or no clearance was
observed for amosite fibers of any length.

For four of the synthetic fibers, long fibers cleared at the same rate as short fibers (alljmore
rapidly than amosite) and the authors report that the data suggest transverse breakage for these
fibers. Moreover, they attribute the more rapid clearance of long fibers among the MMVF's to
dissolution, since these fibers exhibit in vitro dissolution rates that are rapid relative to the time
scale of macrophage clearance. One synthetic fiber MMVF34, which is a stonewool,|
disappeared much more rapidly than any other fiber and the long fibers disappeared more rapidly
than the short fibers. MMVF34 shows the greatest in vitro dissolution rate at neutral pH for any
of the fibers tested in ths study and dissolves particularly rapidly at pH 4.5 (the pH  found in the
phagosomes of macrophages). The authors postulate that clearance of all of the synthetic fibers
are enhanced over amosite by dissolution and breakage.

In summary, Hesterberg et al. (1998a) observed that:

       •     multiple clearance mechanisms (operating over multiple time scales) contribute to
             clearance;

       •     for sufficiently soluble fibers, long fibers clear more rapidly than short fibers;
                                          6.19

-------
       •      for insoluble fibers, a subset of long fibers clears rapidly within the first few
              months following exposure and the remaining long fibers clear only extremely
              slowly, if at all;

       •      short fibers of all types are cleared at approximately the same rate (much more
              rapidly than long, insoluble fibers);

       •      a small, residual concentration of short fibers may not always clear and may
              remain in the lungs (sequestered in alveolar macrophages) for extended periods;
              and

       •      in this study, there js some suggestion that short amosite fibers clear somewhat
              more slowly than short fibers of the other, non-asbestos mineral types studied.

Regarding the last observation, whether this is attributable to differences in fiber thicknesses
among the various mineral types, due to partial contributions (even among short structures) to
dissolution, or due to a unique, toxic effect of amosite is unclear. However, the likeliest of these
candidate hypotheses is that the effect is due to partial dissolution.

This general pattern of observations are consistent with the findings of most, recent retention
studies following short-term exposure.

In an earlier study of similar design, Bernstein et al. (1996) evaluated the deposition and
clearance of a series of 9 glass and rock wools. These authors similarly found that clearance
could be modeled using a double exponential for all length fibers (in similar length categories of
<5, 5-20, and >20 urn) and that for soluble  fibers, long  fibers clear more rapidly than short fibers
(with the intermediate length fibers in between).

For the Bernstein et al. (1996) study, if one  assumes that the pool of longer-lived fibers is
representative of macrophage clearance, this suggests that the efficiency of clearance by
macrophages decreases with increasing fiber length and that the longest structures are not
phagocytized at all, so that they remain exposed to the extracellular medium where dissolution
occurs. Lending further support to this interpretation, the authors also report that the clearance
rate for long fibers correlate with measured in vitro dissolution rates at neutral pH while the
clearance rates for short fibers neither correlate with in  vitro dissolution rates at neutral pH or at
pH 4.5. Although the latter pH corresponds to the pH found in the phagosomes of macrophages,
there is likely too little fluid available in such organelles to support efficient dissolution. The
authors also indicate that a sufficient number of fibers were counted during the study to suggest
that breakage is not playing a role in clearance (except at very early times) and that the clearance
rate for short fibers appears to be the same or slower than that observed for nuisance dusts.

In another, earlier study of similar design Bastes and Hadley (1995) evaluated four types of
MMMF's and crocidolite. All of the samples (including crocidolite) had been size-selected to
assure a large fraction of fibers longer than  5 um. Unfortunately, due to differences in reporting,
it is not possible to compare the initial loading of crocidolite fibers to those reported for amosite
in the Hesterberg et al. (1998a) study. However, results from this study further support the
physical interpretation of clearance suggested in the studies discussed above. In fact, the authors

                                           6.20

-------
report that the timeTdependent size distribution of retained fibers observed in this study agree
well with a computer simulation of fiber clearance.  The simulation assumes that longjfibers
dissolve at the rate measured for such fibers in vitro and that short fibers of every type are
removed at the same rate as short fiber crocidolite (which is practically insoluble).  This is strong
evidence that short fibers are cleared by macrophage phagocytosis and that long fibers cannot be
cleared by macrophages, but may dissolve in extracellular fluid provided that they are
sufficiently soluble.                                                            •'

Regarding crocidolite; the data from the Bastes and Hadley (1995) study suggest that short
crocidolite fibers appear to clear at a rate that is somewhat slower than observed for any of the
short MMVF fibers.  Importantly, however, the interpretation of short fiber clearance In this
paper is somewhat confounded because, unlike the studies discussed above, short fibers in this
paper are defined as all fibers <20 pm, so there may be some confounding with MMVF fibers
that are dissolving. As previously indicated, the Hesterberg et al. (1998a) work also suggests
that short asbestos (amosite) fibers may clear more slowly than short fibers of differing
mineralogy and Hesterberg et al. only includes fibers <5 um in their definition. Nevertheless, it
is still possible that some effects due to dissolution may still be affecting the clearance of these
shorter fibers.-

Surprisingly, a visual inspection of the data presented in Bastes and Hadley (1995) table suggests
a lack of any long-term clearance for long fiber crocidolite (>20 |j.m). Yet, the authors model
long fiber crocidolite clearance using a single exponential  (suggesting no rapidly clearing
compartment). The long-term half-life reported for crocidolite in this study  is approximately
220 days (with estimated CIs of 165-566 days).  This overlaps with the long-term clearance half-
life reported by Hesterberg et al. (1996) for crocidolite of approximately 820 days (246-«>).

Equally surprising, Hesterberg et al.  (1998a) also modeled crocidolite clearance as a single
exponential, which might suggest better penetration to the deep lung by crocidolite, less
clearance by muco-ciliary transport or alveolar macrophage transport, or better penetration to the
interstitium than other fibers.  More likely, however, it may simply indicate  that the two-pool
model does not represent a statistically significant improvement in model  fit over the one-pool
model.  However, relative size distributions would need to be evaluated carefully before drawing
any such conclusions. Bastes and Hadley (1995) also report clearance of short fiber crocidolite
is modeled as a double exponential with short and long half-lives of 25 and  112 days,:
respectively.  Since this fiber category contains fibers up to 20 urn in  length (in this study only),
this does suggest at least some contribution from muco-ciliary and alveolar macrophage
mediated clearance for crocidolite.

In two studies, Coin et al. (1992, 1994) evaluated the fate of chrysotile fibers in rats exposed for
3 hours to 10 mg/m3 (reportedly containing >5,000 fibers longer than  5 u,m/cm3). For,ilung
analysis, the left lung was separated into peripheral and central regions under a dissecting
microscope. Slices of peripheral and central portions were separately weighed and minced.
Tissue was digested in sodium hypochlorite and then filtered. A quality control test indicated
that the digestion process caused a slight (-10%) decrease in fiber number and slight decreases
in fiber diameter and fiber length.  Fiber-size distributions were evaluated by SEM. A stratified
counting procedure was employed to assure equal precision for each length category of interest.
Measurements for each category were then converted to mass equivalents.

                                           6.21

-------
Results from the Coin et al. (1992,1994) studies indicates no difference between deposition in
central or peripheral regions of the lung. They also confirm that chrysotile splits longitudinally
in the lungs with a half-life that is competitive with the clearance rates measured in this study.
Clearance was found to be very length-dependent, so that rates decrease from a half-life of about
10 days for fibers about 4 um in length, through 30 days for fibers 8 um, to 112 days (which is
no different from zero) for fibers longer than 16 jim (all after adjusting for longitudinal splitting).
Importantly, the brief follow-up period (30 days) is too short to provide an adequate evaluation
of the longer term clearance pools observed in other studies and certainly too short to evaluate
any effects potentially associated with chrysotile dissolution. Also, that the decay curves for
clearance were limited to four points, makes evaluation of the slopes for these curves highly
uncertain.

Coin et al. (1992,1994) also report that the mass of chrysotile deposited during these short
exposures (i.e., no more than 20 u.g) is very small compared to levels at which overload has been
reported to occur (approximately 1 mg, see, for example, Yu and Yoon 1991) and that the
volume of the  16 um fibers, which have an average diameter of 0.2 ^im and therefore a mean
volume of 0.5 |im3, is small relative to the volume at which macrophage clearance of non-fibrous
particles is reported to be hindered (Morrow 1988). Thus, the authors  conclude that fiber length
presents an additional constraint on macrophage clearance, independent of any other overload.
They also indicate that inhibition of clearance due to fiber length is independent of fibrosis.

Coin et al. (1992,1994) also discuss the effect of fibrosis on clearance. They indicate that,
although increased concentrations of short fibers are observed in  focal areas of fibrosis, it is
more likely that such fibers accumulate because clearance is hindered by fibrosis in these areas
than the hypothesis that the short fibers are causing fibrosis. This is because, as they point out,
there are too many studies demonstrating the lack of ability of short fibers to induce fibrosis.

Evidence in the Coin et al. (1992,. 1994) studies suggests that no translocation from central to
peripheral regions of the lung were detected. An upper bound rate that is about 20% of clearance
is reported.  However, the short follow-up time in this study would have precluded slower
processes from being detected. Despite the lack of evidence of translocation, the authors report
that duct bifurcations in peripheral regions of the lung where fibers are deposited are no more
than 1-2 mm from the visceral pleura. In fact, in the 1994 study, the authors show that 50% of
the primary duct bifurcations in the peripheral portion of the rat lung occur within 1 mm of the
visceral pleura and some occur as close as 220 um. Deposited fibers may also affect the pleura
by inducing generation of diffusable, inflammatory agents.

A short-term inhalation study by Warheit et al.  (1997) evaluated retention of chrysotile and
aramid fibers. In  this study,  rats (and hamsters) were exposed, nose only, for 6 hours/day,.
5 days/week for 2 weeks by inhalation to UICC chrysotile and p-aramid fibers (each at two doses
of 460 or 780 fibers/ml, although the size range of these fibers is not stated nor is the manner in
which they were analyzed).  Fixed lungs were digested in chlorox during preparation for
asbestos analysis. Animals were followed for up to a year post-exposure.

As in studies described above, results from the  Warheit et al. (1997) study indicate rapid
clearance of short chrysotile fibers, but slow to non-existent clearance of fibers longer than
20 ^m. In contrast, aramid fibers apparently degrade and are subsequently cleared fairly rapidly

                                           6.22

-------
in vivo. Based on the data provided in figures, although the reported concentrations of chrysotile
and aramid fibers to which animals were exposed were equivalent, at both the lower and higher
concentrations, it appears that rats initially retain 3 to 5 times as many aramid fibers as chrysotile
fibers (at least for the size range counted, which was not reported). For both fiber types,
clearance appears rapid for an initial period of approximately 90 days post-exposure.; During
this time, the mean length of chrysotile fibers also appears to increase steadily, which!!suggests
rapid, preferential clearance of short structures. After the initial period, it appears that (as the
authors suggest), a residual concentration of longer fibers are cleared only very slowly, if at all.
                                                                             !'
Oberdorster et al. (1988) instilled a 3 ml suspension of irradiated amosite into the bronchio-
alveolar space of the right diaphragmatic lobe of the lungs of dogs to evaluate clearance and
transport. The amosite used was modified by sedimentation from UICC amosite to contain only
fibers shorter than 20 \im. One dog also had unmodified UICC amosite instilled directly into a
lymph node in the thigh.  The dogs had been cannulated to allow collection of lymph from the
right lymph duct-RLD and the thoracic duct-TD (both in the neck).

Results from Oberdorster et al. (1988) indicate that within 4 hours following instillation in the
lung, low activity was noted in postnodal lung lymph, but not in either the RLD or TD. Within
24 hours, however, activity and fibers (determined by SEM) were observed in both the RLD and
the TD. The median length of fibers observed in the lymph were significantly longer than the
instilled material, although there appeared to be a cutoff length of 16 urn in fibers observed at
nodes and 9 \im in fibers observed directly in lymph.  Fibers recovered from  lymph were also
significantly thinner and appeared to exhibit an absolute cutoff at a maximum width of 0.5 u,m.
Fibers recovered from the TD and RLD in the dog that had unmodified'UICC amosite instilled
directly into leg lymph were all short (with a maximum length of 6 p,m). Since collection times
were all short, the authors indicate that it is unknown whether longer fibers would have been
observed at later times. The authors also note the almost total absence of fibers shorter than
1 um in lymph, which they assume are cleared rapidly and efficiently by alveolar macrophages.

Oberdorster et al. (1988) also report that a rough calculation,  based on the fraction of the
material originally instilled that was recovered in the  first 24 hours, it would  take approximately
6 years to clear all of the instilled asbestos (assuming no other clearance mechanisms were
active).

Everitt et al. (1997) performed a short-term inhalation study that is interesting, particularly,
because it focused on pleural (as opposed to lung) fiber burden. The authors exposed! rats and
hamsters to one type  of refractory ceramic fiber (RCF-1) by nose-only inhalation for periods of
0,4, and 12 weeks and animals were held for observation for up to an additional 12 weeks post-
exposure. Exposures were conducted for 4 hours/day, 5 days/week, at 45.6±10 mg/m?. Groups
of 6 animals were held for 0,4, 12, and 24 weeks to determine pleural fiber burden. An agarose
casting method was reportedly used to recover fibers  from the pleura.  Analysis was by electron'
microscopy. Fibers were observed in the pleura at each time point examined (including samples
from rats sacrificed immediately following the last day of a 5-day exposure). Fibers were all
reported to be short and thin (geometric mean length: 1.6 u.m with GSD: 1.8, geometric mean
diameter: 0.1 um with GSD: 1.5).  Concentrations averaged approximately 40,000 fibers (per
whole pleura, units not reported). The authors indicate that such fibers would not typically be
                                          6.23

-------
visible by optical microscopy.  They also indicate that use of casts may be a more efficient
method of recovering fibers from the pleura.

Everitt et al. (1997) indicate that observation of rapid translocation of short, thin fibers to the
pleura has also been observed in studies of chrysotile so these results are not unique. Although it
is stated that the mechanisms facilitating transtocation are currently unknown, the authors
indicate that their finding of site-specific mesothelial proliferation supports observations by
Boutin et al. (1996) that asbestos fibers accumulate in the parietal pleura of humans at sites
associated with lymphatic drainage. Kane and MacDonald (1993) have suggested that fibers are
transported to these locations by pleural macrophages.  However, the mechanisms by which
fibers are transported from the lung to the pleura are still unconfirmed.

Older Retention Studies. Although the older retention studies generally support the results of
newer studies (such as those cited above), older studies are sometimes limited by such things as
the tracking of lung burden in terms of fiber mass or use of analytical techniques such as infrared
spectroscopy for detection of asbestos, which are neither capable of distinguishing individual
fibers nor provide any information on their sizes. Tracking of lung burdens in terms of mass
may not reflect the fate of long, thin fibers, which (by increasing concurrence) appear to be the
legitimate focus of studies evaluating biological hazards attributable to asbestos.
                                           -i
In two studies (Roggli and Brody 1984 and Roggli et al.  1987), Roggli and coworkers tracked
the behavior of chrysotile (not UICC) and UICC crocidolite in rats following 1 hour exposure by
inhalation to 3.5-4.5 mg/rn3 dusts. The authors indicate that this results in deposition of
approximately 21 ng of dust.  Portions of the lower lung lobes of selected rats were collected and
digested for asbestos analysis using a scheme that was shown to be representative. To evaluate
size distributions, more than 400 fibers from each sample were characterized by SEM. Fiber
dimensions were then used to estimate total fiber mass.

Based on their study, Roggli and coworkers indicate that similar fractions of inhaled chrysotile
and  crocidolite dust are deposited in the lung during inhalation (23 and 19%, respectively).  The
authors therefore concluded that respirability and deposition do not depend on fiber type.
Importantly, however, the manner in which this study was conducted does not facilitate
distinguishing deposition in the deep lung from deposition in the upper respiratory tract.

Roggli and coworkers further indicate that clearance rates for the two fiber types appear
comparable.  Of the chrysotile initially deposited, they report that 81% of this material is cleared
after 4 weeks.  Similarly, 75% of the crocidolite is cleared.  Importantly, because this is based on
total mass (estimated by summing volume contributions from observed fibers), it may not reflect
the specific behavior of long, thin structures. Therefore, it is difficult to compare such results
with those of more recent studies.  However, the authors do report that short structures are
cleared more readily than long structures and that chrysotile is observed to split longitudinally
in vivo (based on observation that the total number of chrysotile structures initially increases and
the mean length increases). The authors further conclude that clearance rates appear to be
independent of fiber type.

In another short-term study, Kauffer et al. (1987) report that the average length of retained
chrysotile structures increases in rat lungs following 5-hours inhalation of chrysotile dust. Based

                                           6.24

-------
 on their results, the authors report that fibers shorter than approximately 8 \im are preferentially
 cleared. Kauffer and coworkers also confirm that chrysotile fibers split longitudinally in the
 lung. In fact, several other studies (Kimizuka et al. 1987;Le Bouffant 1980) also provide
 supporting observations that chrysotile fibers (or bundles) split longitudinally in the lung.

 In two studies (Morgan et al. 1978, 1980), Morgan and coworkers report on the fate of fibers
 following short-term inhalation of radio-labeled fibers by rats. In the first study (Morgan et al.
 1978), rats inhaled UICC anthophyllite at 35 mg/m3 for a total of 8.4 hours spread over 3 days.
 The authors report that the rats retained approximately  190 ug of dust at the end of exposure,
 mostly in the alveolar region; the authors assumed that conducting airway clearance is
 sufficiently rapid to clear this portion of the lung within a few days.  Beginning about 7 days '
 following exposure, the rats were then sacrificed serially for a period up to 205 days following
 exposure. Because anthophyllite fibers are relatively thick, fibers were analyzed by optical
 microscopy.  Fibers were determined both in free cells  (mostly macrophages) recovered in
 bronchopulmonary lavage and  in lung tissue. Tissue samples and cells were digested™ KOH
 and peroxide in preparation for fiber analysis.                                     !

 Based on this first study, Morgan et al. (1978) report that anthophyllite lung content declined
 steadily by a process that could be described as a simple first order decay with a half-life of
 approximately 76 days. Free macrophages recovered by lavage,  initially contained about 8 pg
 and this too declined steadily with a half-life of about 49 days. The authors further indicate that,
 if the number of macrophages remains constant with time (i.e., they are replaced at the same rate
 they are cleared), then the decay of the load in the macrophages should match what is observed
 in the rest of the lung. They suggest that the discrepancy may be due either to an influx of an
 increasing number of macrophages in response  to injury with time and/or to transfer of some
 fibers through the alveolar wall. They also cite unpublished work indicating that uptake of fibers
 by alveolar, macrophages  is essentially complete within hours after cessation of exposure.
                                                                               i'
 The authors also report that, initially, the lengths of fibers recovered in lung lavage was  greater
 than in the original aerosol, but that the prevalence of the longest fibers decreased after the first
 7 days.  In lung tissue, however, the fraction of longer fibers (among total fibers) steadily
 increased with time.  This suggests rapid clearance of naked fibers by muco-ciliary transport
 (which is a length independent process) with later times dominated by slower clearance  in the
 deep lung by alveolar macrophages, which is a length-dependent process.           '

 Morgan and coworkers also radiometrically determined the fraction of fibers in rat feces (prior to
 sacrifice). They assumed that after 14 days, this would represent the fraction of asbestos cleared
 primarily from the alveolar region of the lung. Initially 1.4% of lung anthophyllite content was
 excreted daily, but this fell to 0.5% after 120 days. The authors indicate that this suggests that
 the elimination from asbestos in lung tissue cannot be described by a single exponential  because
 multiple processes are involved and that, over longer periods of time, the slower processes
. become increasingly important.

 In the second study, Morgan et al. (1980) track the fate of several size-selected radiolabeled
 glasses in rats, again following short-term inhalation. From an analysis of the size dependence
 of deposited fibers in this study, the authors suggest that alveolar deposition in the rat is limited
 to structures with aerodynamic equivalent diameters less than about 6 ^m and that deposition in

                                           6.25

-------
this region of the lung falls precipitously for fibers with thicknesses between about 2 and 3
(aerodynamic equivalent diameter). For fibers that are the density of asbestos, this represents an
upper bound limit to alveolar deposition for the absolute thickness of a fiber of approximately
1.5 urn with fibers deposition of fibers thicker than approximately 0.7 urn being drastically
reduced. This is in concordance with conclusions concerning deposition provided in
Section 6.1.4.  Alveolar deposition efficiency is also shown to decrease with increasing fiber
length, at least for fibers longer than approximately 8 urn, also in concordance with findings
presented in Section 6.1.4.

Intratracheal Instillation. The fate of fibers following tntratracheal instillation into the lungs
has also proven informative in some studies.  For example, Wright and Kushner( 1975)
intratracheally instilled paired samples each of several types of glass fibers, fluoramphibole, and
crocidolite into guinea pigs. For each mineral tested, a sample with predominantly short
structures (25 mg total dose for crocidolite, reportedly 99% <5 \im) and another with
predominantly long structures (4 mg total dose for crocidolite, reportedly 80% >10 urn) were
evaluated. Unfortunately, the authors do not report how fibrous structures were characterized.
Results in the cited paper report observations only after 2 years following the last injection.

Wright and Kuschner (1975) report that long structures uniformly caused fibrosis (primarily
involving the respiratory bronchioles  and alveoli and abutting the terminal bronchioles) while
the short structures were uniformly phagocytized and generally removed to thoracic lymph
nodes. Among other things, clearance to lymph suggests that fibers reached the interstitium
(Section 6.2.5). It is interesting that, even after 2 years of recovery, the authors observe elevated
levels of macrophages in the alveoli of animals dosed with short structures. Based on the
relative size distributions of the samples analyzed, the authors report that structures up to 10 um '
in length appear to be efficiently scavenged by macrophages.  Based on the observation of larger
numbers of short structures than expected in comparison with their fractions in the original
samples, the authors further conclude that glass structures underwent biodegradation so that
longer structures broke down into shorter structures that could be phagocytized.

Wright  and Kuschner (1975) also  report that long fibers are occasionally visible within the
fibrotic interstitium of dosed animals. The long-fiber dosed animals also show macrophages in
hilar lymph nodes containing fibers that are too small to resolve and all of them are short.  In
short-fiber dosed animals, some fibers are seen to remain in the lung within aggregates of
macrophages, both in alveoli and the  interstitium. Short-fiber dosed animals also show many
more macrophages within the hilar lymph nodes than  long-fiber dosed animals.

In two reports of the same study (Bellman et al. 1986, 1987),  Bellman and coworkers followed
the fate of UICC chrysotile, UICC crocidolite, several fibrous glasses, and other manmade
mineral fibers following a single intratracheal instillation of 0.3 ml of fibrous material in rats.
Groups of rats were then sacrificed at 1, 6, 12, 18, and 24 months following instillation. Lungs
were low temperature ashed and the resulting, filtered suspension analyzed by transmission
electron microscopy.  Some of the fiber types were also acid treated with 0.1 M oxalic acid for
24 hours prior to instillation.

Bellman and coworkers (1987) report that short fibers (<5 urn) from all of the fiber types were
shown to be cleared from the lungs with half-lives of approximately 100 days, with the asbestos

                                           6.26

-------
varieties tending to exhibit slightly longer half-lives than the other fibers.  Short crocidolite
fibers exhibited a half-life of 160 days. The half-life for clearance of short chrysotile was
reported to be 196 days (the longest of all).  However, this was attributed to positive l;
contributions from breakage of longer fibers.

Bellman and coworkers (1987) report that the behavior of the different long fibers (>5 jam) for
the different fiber types was radically different.  The authors report no observed net decline in
long crocidolite fibers over the 2 years of follow-up. The also report no observable changes in
width of these fibers with time. In contrast, long chrysotile fibers increased in number with time
throughout the 18-month follow-up  period and this was attributed to longitudinal splitting. The
width of these fibers reportedly decreased with time.

Bellman et al. (1987) also report that a more detailed examination of the time dependence of the
width of chrysotile fibers indicates a rapid increase in the  number of thin fibrils (<0.05 urn in
width) and thin bundles (<0.1 urn in width)  within 100 days (at the expense of thicker.bundles).
The authors suggest that this would  result in rapid decrease in the number of chrysotile structures
visible by optical microscopy and, possibly, increased clearance of the thinnest fibrils: by
dissolution, but this study shows no increased rate of clearance for thinner chrysotile structures
compared to thicker structures (when viewed by electron microscopy). In contrast,  long
chrysotile fibers that were acid-leached prior to instillation reportedly disappeared with a half-
life of 2 days.

Generally, the rate of clearance of the long fractions of the other fibers reported in the.Bellman et
al. (1986,1987) papers varies as a function of solubility and overall thickness. Importantly, all
half-lives are reported to have  high standard errors in this  study, due to the small number of
animals included for examination.

In summation, virtually all short-term retention studies indicate that:

       •      fibers retained in the  lung tend to be shorter and thinner than the aerosols from
              which they derive and the size  distributions of retained structures tend to be more
              similar overall than the size distributions observed in the original aerosols;
                                                                               i

       •      chrysotile asbestos undergoes rapid, longitudinal splitting in the lung while
              amphiboles do not;

       •      by mass, chrysotile and amphibole asbestos are deposited in the lung with
              comparable efficiencies, although it is not clear whether chrysotile dusts tend to
              contain sufficient numbers of curly fibers to limit deposition in the deep lung;

       •      multiple clearance processes operate over different time frames and some of these
              processes are strongly length-dependent. Fibers shorter than approximately
              10 u.m appear to be cleared rapidly relative to longer fibers and those longer than
              approximately 20 fim are not cleared efficiently at all (if the fibers are insoluble).
              The Bellman et al. (1986, 1987) studies appear to contrast with other studies in
              this regard in that they suggest fibers longer than 5 |im do not readily clear;
                                           6.27

-------
       •      the quickest clearance process (presumably muco-ciliary clearance) is not
              dependent on length; and

       •      the effects of fiber diameter on clearance have not been well delineated overall,
              although fibers that reach the deep lung appear to be largely limited to those
              thinner than approximately 0.7 |im.

These findings are in addition to those mentioned previously from the newer studies:

       •      multiple clearance mechanisms (operating over multiple time scales)  contribute to
              clearance;

       •      for sufficiently soluble fibers, long fibers clear more rapidly than short fibers;

       •      for insoluble fibers, a subset of long fibers clears rapidly while the remaining long
              fibers clear only extremely slowly, if at all;

       •      short fibers of all types are cleared at approximately the same rate (much more
              rapidly than long, insoluble fibers);

       •      a small fraction of short fibers may be retained for long periods under certain
              circumstances (sequestered in alveolar macrophages) despite overall rapid
              clearance of these structures; and
                                                t                 '
       •      there is some suggestion that short asbestos fibers clear somewhat more slowly
              than short fibers of the other, non-asbestos mineral types studied.

Regarding specifically the clearance of long fibers, it appears that a component of all. such fibers
clears rapidly within the first 2 weeks and this likely represents muco-ciliary clearance. A
second component (representing as much as 60% of the fibers) clears within 90 days and this
likely represents clearance by alveolar macrophages.  The remaining long fibers are  cleared only
very slowly, if at all, and this likely represents fibers that are sequestered in granutomas or that
escape into the interstitium.
6.2.L2
Studies involving chronic or sub-chronic exposures
Although the results of older retention studies following longer term (sub-chronic or chronic)
exposure were difficult to reconcile with the results following shorter-term exposures, newer
studies suggest greater consistency and a clearer picture of the fate of fibers in the lung.
Moreover, although there are further suggestions of mineralogy (fiber type) dependent effects
with some clearance mechanisms, it is important that size effects be considered simultaneously,
if the dynamics of these processes are to be understood.

In some of the latest studies, for example, Hesterberg et ali (1993, 1995, and 1998b) exposed rats
(nose only) by inhalation to a series of man-made vitreous fibers (including a variety of fibrous
glasses, rock wools, and refractory ceramic fibers) and two kinds of asbestos: chrysotile
(intermediate length NIEHS fiber) and crocidolite (size selected). Animals were dosed for 6

                                           6.28

-------
hours/day, 5 days/week for up to 2 years at target concentrations of 10-60 mg/m3.  The target
concentration for chrysotile and crocidolite was 10 mg/m3.  Animals were periodically sacrificed
during the exposure regimen to determine the character of the retained fibers.  Vitreous fiber
aerosols were characterized by PCM, SEM or, for chrysotile, by TEM. The right accessory lung
lobe of sacrificed animals was tied off, frozen, and stored for lung burden analysis.   '

For analysis, lung lobes were dried to constant weight, ashed, the residue suspended in distilled
water, and then filtered on Millipore filters (for examination by optical microscopy) or
Nuclepore filters (for analysis by SEM or TEM for chrysotile). Approximately 100 fibers were
reportedly characterized to establish fiber size distributions. However, this is problematic for
this study because chrysotile asbestos concentrations in the aerosols to which the animals were
exposed contained approximately 100 times as'many fibers as the other aerosols. Thus, although
no fibers longer than 20 u.m were observed during characterization of the chrysotile, the •
concentration of such long fibers could still have been larger in this aerosol than the other
aerosols and it would not necessarily have been observed.  This is also true of lung burden
analyses especially because indirect preparation tends to magnify the number of short'chrysotile
structures observed in a sample.                                                 '•

A comparison of the retention patterns of chrysotile and RCF-1 from the Hesterberg et al.
(1998b) study is particularly instructive.  First, it should be noted that,  in contrast to the values
reported by the authors of this study, chrysotile and RCF-1  in fact appear to exhibit comparable
in vitro dissolution rates (12.7 vs. 8 ng/cm2-hr, respectively) when rates are measuredjusing
comparable techniques (see discussion in Section 6.2.4). The dissolution rates quoted in the
Hesterberg et al. study are not derived in comparable studies.

Although  a full set of time-dependent analyses are apparently not available for chrysotile, it is
reported that approximately 14% of those chrysotile WHO fibers observed to be retained after
104 days of exposure continue to be retained after 23 days of recovery. Under the same
conditions, it is reported that 43% of RCF-1 fibers are retained, which suggests more rapid
clearance  for chrysotile. Even adjusted for the fraction of RCF-1 structures that are longer than
20 urn (and that are presumably cleared even more slowly), approximately 37% of the RCF-1
WHO fibers (<20 urn) are apparently retained over this period, which is still more than twice the
rate reported for chrysotile. Still, more detailed characterization of the size distributions of these
two fiber types would need to be evaluated before it could be concluded with confidence that
chrysotile is cleared more rapidly than RCF-1 or that dissolution plays a role.  In fact,';dissolution
would tend to cause more rapid clearance only of the longest fibers (i.e., the ones that:cannot be
cleared by macrophages [see Section 6.2.1.1]), which would further reduce the apparent
retention rate of the shorter RCF fibers, making it even more comparable to the chrysotile
number.

Based on these studies, Hesterberg et al.  (1998b) report that fibers deposited and retained in the
lung tend to be shorter and thinner on average than the sizes found in the original aerosol.  It is
also apparent from their data that long RCF-1 fibers clear more rapidly than short RCF-1 fibers
(although  a small fraction of long structures are retained at all time points following a recovery
period after cessation of exposure), which is consistent with observations in other studies for
fibers that dissolve at moderate rates. In  the short-term study performed by the same laboratory,
Hesterberg et al. (1998a), long fiber (>20 jim) RCF-la appears to clear at approximately the

                                          6.29                                '..

-------
same rate as the shorter structures (<5 um), although the scatter in the data (and an unexplained
initial rise in long-fiber RCF) prevent a more careful comparison. Similar results are also
apparent in the data presented for MMVF21. It should be noted that the dissolution rates for
RCF-1 and MMVF21 bracket the estimated dissolution rate for chrysotile asbestos (when the
three are derived from comparable studies [see Section 6.2.4]).

The data presented in Table 3 of Hesterberg et al. (1998b), which are reproduced in Table 6-3,
can also be used to evaluate the time-trend of retention during chronic exposure.  The values
presented in Columns 2, 4, and 6 of Table 6-3 present, respectively, measurements of the lung
burden for chrysotile WHO fibers, RCF-1 WHO fibers, and RCF-1 long WHO fibers (>20 urn)
in animals sacrificed immediately following cessation of exposure for the time period indicated
in Column 1.  Unlike results reported in some earlier chronic studies based on mass (see below),
there is no evidence from this table (based on fiber number) that chrysotile lung burdens reach a
plateau.  Rather chrysotile lung burdens (as well as RCF-1 lung burdens) continue to increase
with increasing exposure.

The data presented in Table 6-3 can also be used to gauge the relative efficiency with which the
chrysotile and RCF fibers are retained. Considering that the number of fibers inhaled (Ninh) over
the period of exposure would be equal to the product of the aerosol concentration (C^ in f/cm3),
the breathing  rate of the exposed animal (RB in cmVweek), and time (in weeks):

                                      NurC./RBt                              (Eq.6-1)

and the efficiency of retention is simply equal to the quotient of the number of fibers retained
(Hung) a°d the total number inhaled: N^g/N^, then the efficiency of retention is estimated by the
following simple relationship:

                          Efficiency of retention=Nlung/(Cair*RB*t)                  (Eq.6-2)

By rearranging Equation 6-2, one obtains:

                        Efficiency of retention*t=(Nlung/(CJ*(l/RB)                (Eq. 6-3)
                                          6.30

-------
           Table 6-3. Fraction of Fibers Retained Following Chronic Exposure"'
                   Chrysotile
RCF-1
Exposure
Period
Weeks
WHO
Fibers
f/lung x
10A6
Lung/
Aerosol
Ratio

WHO
Fibers
f/lung x
10A6
Lung/
Aerosol
Ratio

Long
WHO
fibers
f/lung x
10A6
' Lung/
i Aerosol
' Ratio
1.
0.0357
13
26
52
78
104
Aerosol
(ffml)
'Source:
250
180
1020
853
1600
Concentration
10600
Hesterberg et al.
0.024
0.017
0.096
0.080
0.151


(1998b)
0.009
39
56
119
173
143

187
^
4.81E-05
0.209
0.299
0.636
0.925
0.765



0.002
3
6
20
21
25

101

',! 1.98E-05
! 0.030
; 0.059
0.198
0.208
0.248
i
'

Because the breathing rate for the rats in the Hesterberg et al. (1998b) study can be considered a
constant for all experiments, Equation 6-3 indicates that the slope of a plot of N^^C^ versus
time should yield estimates of the relative efficiency of retention for each of the fiber (types
evaluated.  The plot for chrysotile is presented in Figure 6-3. Results from this plot and similar
plots for RCF-1 WHO fibers and long WHO fibers (data not shown), result in the following
estimates of the relative efficiencies of retention (along with the corresponding R2 value for the
fit of the linear trend line):

                        chrysotile WHO fibers:    0.0014, R2=0.856            /

                        RCF-1 WHO fibers:       0.0095, R2=0.694            ;'

                        RCF-1 long, WHO fibers:  0.0026, R2=0.880

Thus, it appears that chrysotile WHO fibers are retained somewhat less efficiently than either
RCF-1 WHO fibers or RCF-1 long WHO fibers. However, whether this  is due to less efficient
deposition or more efficient clearance cannot be determined from this analysis. It is also not
possible to determine whether such differences are due to the effects of differences in size
distributions among the various fiber types.  Interestingly, based on the data presented by
Hesterberg et al., which indicates that long RCF-1 WHO fibers clear more rapidly than regular
RCF-1 WHO fibers, the differences in the relative retention of these two  length categories of
fibers is due primarily to relative efficiency of clearance.                          ;

                                          6.31

-------
                                    Figure 6.3:
                       [Chrysotile Lung Burden/Aerosol
                     Concentration] vs. Time of Exposure
        o
        CO
        2
        G)
        C

        d,
0.16
0.14-
0.12-


0.08-
0.06
0.04-
0.02
   0
y=0.0014x
   = 0.8556
                           20      40       60       80
                                  Time of Exposure (weeks)
           100
                                                           120
In a similar study involving chronic exposure to Syrian golden hamsters (Hesterberg et al. 1 997),
fiber retention and biological effects associated with exposure to amosite and a series of MMVFs
were evaluated. The amosite was size selected and hamsters were exposed to one of three levels
(0.8±0.2, 3.7±0.6, and 7.3±1.0 mg/m3).  Amosite lung burdens were shown to increase regularly
with dose and time of exposure. The time dependence for accumulation of some of the MMVF's
was more complicated. None of the animals were apparently followed for any recovery periods
following cessation of exposure. The authors also indicate that the severity of the effects
observed (inflammation, cellular proliferation, fibrosis, and eventually several mesotheliomas),
appear to correlate well with the concentration of fibers longer than 20
Earlier Studies. Earlier studies, in which asbestos concentrations tend to be monitored as total
mass tend commonly to show that chrysotile asbestos is neither deposited as efficiently as
various amphibole asbestos types nor is it retained as long (i.e., it is cleared much more rapidly
from the lungs). In fact, several such studies tend to show that chrysotile asbestos concentrations
eventually reach a plateau despite continuing exposure, which suggests that clearance and
deposition come into balance and a steady state is reached. In contrast, amphibole asbestos
concentrations continue to rise with increasing exposure, even at the lowest exposure levels at
which experimental animals have been dosed.  Such observations do not appear to be entirely
consistent with those reported in newer studies (see above) that track fiber number
concentrations (in specific size categories). In these newer studies, chrysotile retention is not
observed to level off, but continues to increase in a manner paralleling amosite or other fibers.
As indicated below, however, the limitations associated with these older studies suggest that,
                                          6.32

-------
although it may not be easy to reconcile them quantitatively with the newer studies, results from
these studies are not necessarily inconsistent with those of the newer studies. Moreover, the
trends observed in the newer studies are likely more directly relevant to issues associated with
the induction of asbestos-related disease. The problems with the older studies are:

       •      the trends seen in the older studies (based on mass) may mask the more important
              trends associated with deposition and retention of long, thin fibers. Thus, results
              from such studies may not be directly relevant to considerations of risk; and

       •      the observed differences between chrysotile  and the amphiboles may be attributed
              to differences in size distribution (among other possibilities). Thus, lacking
              detailed information on size distributions, it  is difficult to reconcile the results
              from the older studies with results from the newer studies, which explicitly track
              specific size ranges of fibers.                                      ;

Given these limitations, the earlier studies are only mentioned briefly.               '•
                                                                              i
Middletori et al. (1979) tracked the  fate of asbestos (as mass measured by infrared spectroscopy)
in rats following inhalation of several asbestos aerosols (UICC chrysotile A, UICC amosite, and
UICC crocidolite) at multiple concentrations (reported at 1, 5, or 10 mg/m3).  To account for
possible differences in the nocturnal (vs. daytime) activity level of rats, several groups of rats
were also exposed in a "reversed daylight" regimen (in which cages were darkened during the ,
real day and bathed in light during the real night). Acclimatized rats in these groups were thus
dosed at times corresponding to their night.  Exposure continued for 7 hours/day, 5 days/week,
for 6 weeks.

Results from the Middleton et al. (1979) study were fit to a  three compartment model; (originally
proposed by Morgan et al. 1978) and the authors concluded that clearance was independent of
fiber type, but that the initial deposition of fibers was very dependent on fiber type.  This was
indicated by a "K-factor" representing the efficiency of initial deposition.  Chrysotile, showed K
factors that range between 0.17 and 0.36 and vary inversely with the initial exposure ;
concentration. In contrast, amosite exhibits a K factor of 0.69 and crocidolite a K factor of 1.0
and both are independent of exposure level. Although the design of this experiment precluded
fitting of the shortest two compartments of the model (with  half-lives of 0.33 and 8 days, from
Morgan et al. (1978), they did optimize the half-life of the longest compartment. Fibers in this
compartment were cleared with  a half-life of 170 days.                            ;

In a series of studies, Davis and coworkers (1978, 1980,1988a, and 1988b) report that retention
of asbestos (measured in terms of mass) appears to be a function of fiber type and surface charge
in addition to fiber size. With regard to fiber type, for example, Davis et al. (1978) report that
substantially more amphibole (amosite) asbestos appears to be deposited and retained in the
lungs of exposed rats than chrysotile. Chrysotile is also apparently cleared more readily than
amosite.  However, mineralogical effects should only be judged after adjusting for fiber size.
                                                                              i
Rats in the studies by Davis and coworkers were dosed at 6 hours/day, 5 days/week for up to
1 year at dust concentrations of 2, 5, or 10 mg/m3 (depending on the specific experiment). Right
lungs (used for determining lung burden) were ashed and the residue was washed in distilled

                                          6.33

-------
water and filtered.  The residue was formed into a potassium bromide disc and asbestos (mass)
content was determined by infrared spectroscopy.

Jones et al. (1988) report that the lung-tissue concentration of amosite increases continually with
exposure (at 7 hours/day, 5 days/week for up to 18 months) and the rate of increase is
proportional to the level of exposure. A leveling off of amphibole concentrations in lung tissue
was not observed in this study as long as exposure continued, even for the lowest level of
exposure (0.1 mg/m3) studied. The lowest exposure concentration evaluated in this study is only
1% of the concentration at which chrysotile lung burdens were shown to reach equilibrium in
other retention studies (see below). Importantly, however, these are only the older studies in
which fiber burden is tracked by mass. The newer studies don't show this effect.

The authors also report lack of any apparent change in size distribution with time among the
fibers recovered from the animal's lungs, which suggests lack of substantial clearance even of
short fibers. However, the longest recovery period following the cessation of exposure evaluated
in this study is only 38 days, which may be too short to.allow evidence for differential clearance
as a function of size to become apparent (at least in a chronic study; the time dependence in
chronic studies such as this are more complicated than for short-term  studies). Moreover, the
apparent inclusion of lymph nodes as part of the lung homogenate may have caused short fibers
initially cleared from the lung to be added back in. In this study, lungs were recovered intact
including the associated mediastinal and hilar lymph nodes, which were ashed in toto. Ash
residue was washed in acid and water, ultrasonicated and filtered for electron microscope
analysis. Note that such a procedure would include any fibers cleared to local lymph nodes.

In a widely cited study, Wagner et al. (1974), report that amphibole lung burdens increase
continually as long as exposure to amphiboles continues and that amphibole concentrations in
lung tissue decrease only slowly following cessation of exposure.  In  contrast, chrysotile lung
burdens reach a plateau despite continued exposure. Importantly, asbestos content was estimated
by determining total lung silica content and adjusting for similar analysis on filtered samples of
the original aerosols. Thus, in addition to suffering from the limitations associated with tracking
fiber burden by mass, there are questions concerning the validity of using total silica to represent
asbestos content. Therefore, for these reasons and the additional reason of the lack of controlling
for fiber size, the ability to interpret this study and reconcile its conclusions with those of newer
studies is severely limited.

Chronic Inhalation of Non-fibrous Particulate Matter. A recent study involving chronic
inhalation of non-fibrous materials is helpful at elucidating the relative localization of particles
in rats and primates. Nikula et al. (1997) studied lung tissue from  a 2-year bioassay, in which
Cynomolgus monkeys and F344 rats were exposed to filtered, ambient air or air containing one
of three particulate materials: diesel exhaust (2 mg/m3), coal dust (2 mg/m3, particles <7 \im in
diameter), or a 50/50 mix of diesel exhaust and coal dust (combined concentration: 2 mg/m3).

Results from Nikula et al. (1997) indicate that responses to all three particulate materials were
similar. The particles tended to localize in different compartments of the lung in a species-
specific manner:
                                           6.34

-------
       •      73% of particles remain in the alveolar lumen of rats, but only 43% in monkeys.
              The remainder can be found in the interstitium;                     :'

       •      in both the alveolar lumen and in the interstitium, virtually all of the particles are
              observed to be isolated within macrophages; and
                                                                              i.
       •      the particles in the interstitium reside in macrophages within the alveolar septa,
              the interstitium of respiratory bronchioles, the adventitia and lymphatic capillaries
              surrounding arterioles and veins of pulmonary parenchyma, or in the pleura.

It is not known whether free particles penetrate the epithelial lining of the airway lumena and
escape into the interstitium or whether such particles are first engulfed by macrophages and then
transported in their macrophage "hosts" into the interstitium.

Importantly, even after 2 years of exposure, the particles in the interstitium do not appear to have
elicited a tissue response.  Also, the aggregates of particle-laden macrophages observed in
alveolar lumena elicited significantly less of a tissue response im monkeys than in rats. Such
responses included: alveolar epithelial hyperplasia, inflammation, and focal septal fibrosis.
                                                                              r
The authors further indicate that "epithelial hyperplasia concomitant with aggregation" of
particle-laden macrophages in alveolar lumen is a characteristic response to many poorly soluble
particles in the rat lung, both at exposure concentrations that result in lung tumors and at
concentrations below those resulting in tumors. Such a response, however, was not  ];
characteristic of what was observed in monkeys.  Among other things, these differences in
responses suggest that rats may not represent a good model for human responses to inhalation of
poorly soluble particulate matter. It would also have been interesting had they tested a "benign"
dust such as TiO2.
In summation:
              results of (newer) sub-chronic and chronic retention studies are generally
              consistent with those of retention studies that track lung burden following short-
              term exposure (Section 6.2.1.1);

              there is some indication in these sub-chronic and chronic studies that chrysotile
              asbestos may not be retained as efficiently as amphibole asbestos. It is likely,
              however, that such distinctions are due more to fiber size than fiber type so that
              definitive conclusions concerning such effects cannot be reached until1 better
              studies that properly account for both size and type are conducted; and

              although earlier studies that track mass instead of fiber number suggest otherwise,
              chrysotile and amphibole asbestos concentrations (when measured by fiber
              number) continue to increase with time as long as exposure continues. • Due to a
              lack in the ability to distinguish among size-dependent effects when lung burdens
              are tracked by mass, the results of the earlier studies are not necessarily
              inconsistent with the results of the later studies.                     '.
                                           6.35

-------
6.2.2   Animal Histopathological Studies  •

Studies in which the lungs of dosed animals are examined to determine the fate and effects of
inhaled asbestos  are helpful for understanding the movement and distribution of retained
particles within the lung and surrounding tissue. Both the newest retention studies and the older
retention studies  tend to include at least some of this type of information.  They also tend to
indicate a consistent picture of the fate and effects of asbestos. While such studies tend to
confirm that translocation in fact occurs, they are  less helpful for elucidating the specific
mechanisms by which translocation occurs.

Newer Studies.  Ilgren and Chatfield (1998) studied the biopersistence of three types of
chrysotile ("short chrysotile" from Coalinga in California, "long" Jeffrey fiber from the Jeffrey
mine in Asbestos, Quebec, and UICC-B (Canadian) Chrysotile, a blend from several mines in
Quebec).  Both the Coalinga-fiber and the Jeffrey-fiber were subjected to  further milling prior to
use. In this study, rats were exposed via inhalation for 7 hours/day, 5 days/week for up to
2 years. Concentrations were: 7.78±1.46 mg/m3 for Coalinga-fiber, 11.36±2.18 mg/m3 for
Jeffrey-fiber, and 10.99±2.11 mg/m3 for the UICC-B fiber.  An additional group of rats was also
dosed for a single 24-hour period with Jeffrey-fiber at a concentration of 5,000 f/ml >5 \im,
Estimates of lung content of chrysotile were based on measurements of total silica content.  The
character of the three chrysotile types evaluated in this study was previously reported (Campbell
et al. 1980; Pinkerton et al. 1983). The animal studies were conducted previously with the
overall approach reported by McConnell et al. (1983a,b) and Pinkerton et al. (1984). Based on
the characterization presented in these papers, Coalinga-fiber is short, but not as extremely short
as suggested by the authors:

          Ratio of fibers longer than:   5 u.m:   10 Jim:   20 jam

                      Coalinga-fiber:    200:       78:     0.98

                        Jeffrey-fiber:    591:     220:       78

Such calculations also suggest that the single, "high" dose in this experiment was equivalent
only to a concentration that is approximately 10 times the other concentrations studied, so that it
is equivalent only to a 10-day exposure and small relative to the longer term (up to 2 years)
exposures considered in this study.

Results reported  by Ilgren and Chatfield (1998) indicate that the lung burden for short fiber
chrysotile initially increases with exposure, reaches a steady state, and then decreases steadily
following cessation of exposure. Approximately 95% of this  material is cleared within 2 years.
Thus, short fibers appear to exhibit the trend suggested by the older chronic retention studies that
tracked burden by mass and is consistent with the newer studies indicating rapid clearance of
short structures (Section 6.2.1.2).

In the studies reported by Ilgren and Chatfield (1998), most short fibers are found initially within
alveolar macrophages and, while concentrations in Type I epithelium, interstitial cells,  and
interstitial matrix increase with time, little appears to be taken up by Type II epithelium.  Small
                                           6.36

-------
amounts of short fiber chrysotile are also observed to be taken up by endothelial cells* There is
also little sign of inflammation or fibrosis following exposure to the Calidria-chrysotije.

Jeffrey-chrysotile also initially appears to be taken up primarily by alveolar macrophages, but
later becomes most prevalent in the interstitial matrix and, to a lesser extent, in interstitial cells.
Substantial numbers of fibers are also taken up by Type I epithelium and small amounts by
endothelial cells. Similar, but slightly delayed effects were seen for UICC-B chrysotile (which
has a smaller fraction of long fibers and therefore, the authors suggest, takes longer to
accumulate). Note that, by 12 months, the majority of long fibers from both these types were
found in interstitial matrix while the majority of Calidria-material was still found in alveolar
macrophages. This is consistent with observations concerning behavior between short and long
fibers reported by Wright and Kushner (1975),  Section 6.2.1.1. At this point, the Jeffrey-
material also caused substantial thickening of the basement membrane and most fibers in the
interstitium were trapped within the collagenous matrix. The authors note that some of the most
severe interstitial changes occurred adjacent to areas of bronchiolar metaplasia. Such effects
were not seen with Calidria-exposure.

Thickened basement membranes, calcium deposits, metaplastic changes, and structural
abnormalities were all observed with long fiber exposure, but not with Calidria-exposure.
Interstitial macrophages also showed morphological changes following phagocytosis of fibers.
While Calidria-material was about evenly distributed between interstitial matrix and cells, the
vast majority of long fiber material was found in the matrix. Movement into the matrix was also
observed to increase even after exposure ceased.  With time, the number of long fibers in
interstitial cells declined modestly, but declined precipitously for Calidria-material.

Jeffrey-fibers accumulated in Type I epithelial cells during exposure and then levels decreased
slowly after exposure ceased. UICC-B fibers accumulated more slowly, never reaching the same
levels as for Jeffrey and decreased more rapidly.  Concentrations of Calidria-fibers inlType I
epithelial cells was low at all time points.

Type II epithelial cells accumulated very few fibers of any type (although they took up slightly
more Jeffrey-fiber than the others). All three fiber types caused substantial increases in
interstitial cells (mostly macrophages) at 3 months and this increase persisted for the Jeffrey-
fiber, but decreased to background after 24 months for the other two fiber types.  Fibroblast
numbers also increased with the long fiber types, but not Calidria-chrysotile.        j

Type II epithelial cells showed decreases in volume and number that persisted until exposure to
long fiber.ceased and these cells displayed dramatic structural aberrations despite absence of a
fiber load.  One possible explanation for the observed changes in Type II cells, especially the
reduction in their number, is that they were undergoing terminal differentiation to Type I cells
(see Section 4.4). In  fact, the apparent absence of fibers observed within Type II cells might be
explained by  such cells taking up  fibers, but being induced to terminally differentiate^ once fibers
are accumulated. 'Overall, Type II cells displayed greater cellular response than Type I cells
(which might also suggests a role for cytokines).  All effects were observed to be fiber length
dependent and all were exaggerated following exposure to long fiber material.
                                           6.37

-------
Rats exposed to long fiber had numerous accumulations of dust-laden interstitial macrophages
and/or small focal accumulations of dust within the interstitium at the end of the lifetime study,
but such changes were not observed for Calidria-exposed animals.

The lung burden for rats exposed to the single, "high" Jeffrey-fiber exposure (based on total
silica) at 12 months (i.e., 12 months post-exposure) was not different from controls. Therefore,
the authors conclude that short-term, "high" exposures are rapidly cleared (even exposures
containing substantial quantities of long fibers).  Other changes induced by the single, short-term
high exposure of Jeffrey-fiber that was followed for 24 months also showed reversion to close to
background status. Importantly, these observations are not based on quantitation of fiber burden
in lung tissue. Rather they are inferred by observing the effects caused by the presence of fibers.
This may suggest, for example, that more than 10 day Is worth of exposure would be required at
this level of exposure before irreversible lesions develop.

In the study by Hesterberg et al. (1997), which was previously discussed (Section 6.2.1.2), the
authors note (among  dosed hamsters) that the magnitude of cellular effects appeared to differ
among the fibrous glasses as a function of their relative biodurability. For animals dosed with
the least biopersistent glass, only transient effects (influx of macrophages, development of
microgranulomas) were observed and these did not progress further.  For the more persistent
glasses, injury progressed through more intense inflammation, interstitial fibrosis, pleura!
collagen deposition, mesothelial hypertrophy and hyperplasia and eventually, mesothelioma.
The authors also suggest that amosite appears to be more potent than chrysotile, even when
aerosols contain comparable numbers.of fibers longer than 20  u.m.

Choe et al. (1997) exposed rats to chrysotile and crocidolite (both NIEHS samples) by inhalation
for 6 hours/day,  5 days/week, for 2 weeks. The rats were then sacrificed and their pleural
cavities lavaged. Results indicate that significantly more pleural macrophages were recovered in
plueral lavage fluid at one and 6 weeks  following exposure than sham exposed rats. The
centrifuged pellet from pleural lavage fluid from one of four rats also exhibited long (>8 u.m),
thin (<0.5 (Jim) crocidolite fibers (1 week following exposure). The concentration of fibers in
this pellet suggested  approximately 1 f per 4,000 cells in the pleura. Note that chrysotile rats
were not examined for fiber content.

Older Studies.  In the series of studies by Davis et al. (1978, 1980, 1985,1986a), the authors
generally report similar histopathological observations that emphasizes a marked distinction
between effects from long and short fibers. From the 1986 study, for example, Davis et al.
report that at the end of 12 months  of exposure, rats exposed to long fibers (amosite in this case)
exhibited deposits of granulation tissue around terminal and respiratory bronchioles.  They
further indicate that the granulation tissue consists primarily of macrophages and fibroblasts with
occasional foreign body giant cells.

As the animals aged, there was increased evidence of collagen deposits in these lesions and  the
oldest lesions consist mainly of acellular, fibrous tissue. The alveolar septa in these older
animals showed progressive thickening. Initially, this was apparently due primarily to
hyperplasia of Type  II epithelial  cells, but with time was increasingly due, first, to reticulin and,
later, to collagenous  deposits in the septal walls. Asbestos dust was frequently visible in these
deposits.  Epithelial cells lining alveoli adjacent to the oldest lesions also tended to become

                                           6.38

-------
cuboidal in shape. As the animals aged, these areas of interstitial flbrosis became more
extensive.

In contrast, animals exposed to short fibers (also amosite in this case) showed no suchi lesions
(peribronchia! fibrosis) at any point in time. At the end of exposure, the lungs of these animals
contained large numbers of pulmonary macrophages packed with fibers, but these cells remained
free in the avleotar spaces. The authors report that large numbers of laden macrophages
sometimes aggregated in  alveoli close to respiratory bronchioles, but that there would' be no
formation of granulation tissue or thickening of alveolar septa at these locations. Thus, with the
exception of the presence of dust-laden macrophages, the structure of the lung of these rats was
not altered.

In the Davis et al. (1987)  study of chrysotile, a slight variation of the above scenario is worth
noting. In this study, the  development of peribronchial fibrosis was reported for animals dosed
both with the long fiber material and with the short fiber material.  However, the authors also
report that the short fiber chrysotile in this study in fact contains a sizable fraction of longer
structures and this finding was corroborated by more formal size characterization (Berman et al.,
unpublished) later conducted in support of a study to evaluate the effects of size (Berman et al.
1995).

In this study, Davis et al.  also reported observations on the morphological changes observed in
the mesothelium during these studies. The authors indicate that the older animals in these
studies exhibit "areas of vesicular pleural metaplasia consisting of loose, fibrous tissue
containing large vesicular spaces lined with flattened cells". The authors also report that
examination in previous studies indicates that these cells are of a mesothelial type.

Davis et al. report that, "...occasionally the walls between vesicular spaces were so thin that they
consisted of two closely opposed layers of extended and flattened cells with no  basement
membrane in between them.  Where cells were supported by areas of fibrous tissue, abasement •
membrane was present. While no method for the direct quantification of this pleural metaplasia .
has been developed, its occurrence is closely related to the presence of advanced interstitial
fibrosis or adenomatosis in the lung tissue and it is particularly common where patches of this
type of parenchymal lesion have reached the surface. It is not known whether such lesions are
precursors to mesothelioma.  Davis et al. also note that neither of the two mesotheliomas
observed in this study showed histological patterns consistent with the observed vesicular
hyperplasia.

Brody  et al. (1981) tracked the distribution of chrysotile following inhalation by rats.'Asbestos
was initially deposited almost exclusively at alveolar duct bifurcations.  In agreement1 with
Pinkerton et al. (1986), the degree of deposition appeared to be an inverse function of the path
length  and bifurcation number for each alveolar duct. Uptake by macrophages and type  1
epithelial cells were observed following deposition. Asbestos was observed both in lipid
vesicles and free in the cytoplasm of type 1 cells. After 8 days, alveolar duct bifurcations
became thickened with an influx of macrophages. Asbestos was also observed in basement
membrane below  the epithelium. Apparently, structures had been transported through type 1
cells to the basement membrane. Once in the basement membrane, asbestos may enter the
interstitium. Predominantly short structures were monitored in this study. Long structures were

                                          6.39

-------
not readily observed (but this is likely a counting problem; under such circumstances, short
structures may serve as surrogates for the presence of other structures).

Intratracheal Instillation. Bignon et al. (1979) studied the rate of translocation of various
materials in rats. Chrysotile, crocidolite, and glass fibers were intrapleurally injected into rats
and their concentration was monitored as a function of time in lung parenchyma and other tissues
removed from the pleura.  Within 1 day following injection, asbestos was detectable in lung
parenchyma. After 90 days, asbestos was found in all of the tissues analyzed. Based on the rate
of translocation to the lung, crocidolite migrates about 10 times more rapidly than chrysotile (on
a mass basis). The rate of migration of glass is in between the two asbestos types.  Structures
initially found in the lung were significantly shorter than the average size of structures injected.
After 7 months, however, the average lengths of structures in all tissues monitored were longer
than the average length of structures originally injected. Thus, short structures migrate more
rapidly than longer structures (possibly by a different mechanism), but long structures eventually
translocate as well.  Within a target tissue, preferential clearance of short structures also
contributes to observed increases in the average length of the structures with time.

Studies of Non-Fibrous Particulate Matter.  Studjes of the fate and effects of respirable, non-
fibrous paniculate matter provide evidence for at least one mechanism by which particles (and
fibers) may be transported to the interstitium.

Li et al. (1997) evaluated the effect of urban PM 10, carbon black, and ultrafine carbon black on
rats following intratracheal instillation (0.2 ml volume instilled containing between 50 and 125
(Ag of particles).  After 6 hours, there was a noted influx of neutrophils (up to 15% of total cells
observed in bronchioalveolar lavage-BAL fluid) and increases in epithelial permeability was
surmised based on increased total protein (including increases in levels of lactate dehydrogenase,
which is a marker for cell membrane damage) in BAL fluid.

Conclusions. Overall, observations among both the newer and the older studies (including the
study by Wright and Kuschner 1975, see Section 6.2.1.2) tend to be highly consistent,
particularly with regard to the  distinction between the effects of short and long fibers.  Typically,
long fibers initially produce substantial inflammation characterized by an influx of macrophages
and  other inflammatory cells.  Ultimately, exposure to such fibers cause thickening of alveolar
septa (particularly near avleolar duct bifurcations) due to a combination  of epithelial hyperplasia
and  deposition of reticulin and, later, collagen resulting  in interstitial fibrosis. In contrast, short
fibers cause an initial influx of macrophages and, long-term, show persistent accumulations of
fiber-laden macrophages both  in alveolar lumena and in pulmonary lymph nodes, but otherwise
no structural changes are observed in the lung tissue of these animals.

These studies also provide ready evidence of the effects of fiber translocation, but generally offer
only limited evidence for elucidating the mechanisms by which such translocation occurs. There
is evidence that Type I (and possibly Type II) epithelium phagocytize particles and fibers and it
is possible that such fibers may be passed through to the basement membrane and the
interstitium.  For Type I cells,  the distance between the alveolar lumen and the basement
membrane averages less than 1 urn in any case (Section 4.4). Certainly  fibers are also observed
in the interstitium.  Particulate studies also indicate that oxidative stress induced by particulate
matter and fibers may cause morphological changes in Type II cells with consequent loss of

                                            6.40

-------
integrity of the epithelium, which increases it's permeability overall and may also allow
diffusional passage of particles and fibers.                                       I-

6.2.3  Human Pathology Studies

Human pathology studies provide additional information concerning the nature of asbestos
deposition, clearance, and retention. These are the studies in which lung burdens are measured
in samples of lung tissue and correlated with the exposures received by the individuals from
which the lung samples derive.
                                                                              r
Among the advantages of human pathology studies is that they provide direct insight into the
behavior of asbestos in humans. They are also limited, however, by the lack of ability to obtain
time-dependent estimates of lung burden (because samples are derived from deceased,
individuals),  by the manner in which lung tissue is stored (several of the fixatives employed to
store tissue samples have been shown to enhance dissolution of asbestos (Law et al. 1990,1991),
by the manner in which samples are prepared for asbestos analysis, by the manner in which
asbestos is analyzed, and by the limited ability to re-construct the uncontrolled exposures
experienced by study subjects (Section 5.2).
                                                                              !
Perhaps most importantly, the ability to construct anything but the coarsest quantitative
comparisons  across subjects is also typically limited by use of "opportunistic" tissue samples
(i.e., use of samples that happened to have been collected and stored during autopsy or necropsy)
because such samples are not controlled for location on the respiratory tree (i.e., the linear
distance and branch number from the trachea) that is represented by the sample. Because it has
previously been shown that deposition is a strong function of such location (see, for example, Yu
et al. 1991), comparisons across lung samples not controlled for these variables are problematic.
It has also been shown that samples collected from adjacent locations in lung parenchyma can in
fact exhibit strikingly different fiber concentrations due specifically to the differences)in the
location of the respiratory tree represented by the alveoli and respiratory bronchioles in the
spatially adjacent samples (Brody et al. 1981; Pinkerton et al. 1986 [Section 5.2]).   |

Despite the above indicated cautions, when interpreted carefully, human pathology studies can
provide useful evidence regarding fiber deposition, clearance, and retention in the human lung.

Newer Studies.  Among  the most recent studies, Finkelstein and Dufresne (1999) evaluated
trends in the relationship  between lung burdens for different fiber types and different size ranges
as a function of historical exposure, the duration of such exposure, the time since lastiexposure,
and other variables. The  analyses were performed among 72 cases from which tissue'samples
could be obtained (including 36 asbestosis cases, 25 lung cancer with asbestosis cases, and 11
mesothelioma cases).

Due to the excessive scatter in the data, most of the analyses presented depend on "Lowess
Scatterplot Smoothers". Moreover, although not stated, it is likely that the tissue samples
obtained were "opportunistic" in that they were not matched or controlled for relative position in
the respiratory tree.
                                          6.41

-------
Finkelstein and Dufresne (1999) employed the multi-compartment model developed by Vincent
et al. (1985) to evaluate trends in their data. The features of this model include:

       •     a compartment representing conducting airways that are cleared within minutes to
             hours by muco-ciliary transport;

       •     a compartment representing the subset of fibers reaching the pulmonary portion of
             the lung that are cleared by alveolar macrophages and transported to the muco-
             ciliary escalator.  This type of clearance is also considered relatively rapid with
             half-lives of no more than several days to several weeks.  Macrophage clearance
             is also considered size-dependent and long fibers are cleared less efficiently than
             short fibers;

       •     when sufficient dust is inhaled (or dust is sufficiently cytotoxic) to impair the
             motility of macrophages (either by volumetric overload or by toxicity), a
             sequestration compartment forms that consists of laden, but immobile,
             macrophages. Although this compartment may ultimately be cleared to lymphatic
             drainage, such clearance is assumed to be slow and size dependent (with half-
              lives of 2 or 3 years for short fibers and 8 years for fibers longer than 10 (j.m); and

       •     once the macrophage system is overloaded, fibers may  cross the alveolar
             epithelium and reach the interstitium and this compartment must be cleared by
             transport to  lymphatic drainage, which is assumed to be an extremely slow
             process.

Finkelstein and Dufresne (1999) indicate that chrysotile splits both longitudinally and
transversely in the lung and that chrysotile lung burdens decrease significantly with time since
last exposure (with short fibers clearing even  faster than long fibers), while tremolite burdens do
not appear to decrease with time since last exposure.  They also suggest that smoking does not
appear to affect clearance rates.

Finkelstein and Dufresne (1999) indicate that these type of studies are not useful for examining
the behavior in rapidly clearing compartments of the lung, but they may  provide insight
concerning the more slowly clearing compartments.  Based on their modeling, they suggest that
tremolite is transferred to the sequestration compartment at rates that are 6-20 times that of
chrysotile (which they indicate is comparable to what was found for crocidolite by de Klerk).
The authors suggest that retained chrysotile concentrations tend to plateau  after accumulation of
about 35 years of exposure, while tremolite concentrations continue to increase. They also
suggest, however, that chrysotile concentrations may begin to increase again after 40 years
(suggesting the overload is eventually reached for chrysotile as  well).  Reported half-lives from
the long-term compartment are:
                                          6.42

-------
        Chrysotile
               fibers shorter than 5 \im
               fibers 5-10 ^m in length
               fibers longer than 10 \im
        Tremolite
               fibers shorter than 5 fim
               fibers 5-10 (am in length
               fibers longer than 10 jim
3.8 years
5.7 years
7.9 years

14.3 years (not different from «>)
15.8 years (not different from •»)
150 years (not different from °°)
In a case-control study, Albin et al. (1994) examined the lung burdens of deceased workers from
the asbestos-cement plant previously studied for mortality (Albin et al. 1990). In this study,
details of the procedures used to prepare lung tissue for analysis were not provided.  It is also'
assumed that available tissue samples were "opportunistic" in that they were not matched or
controlled for relative position in the respiratory tree.

Results from Albin et al. (1994) are consistent with (but do not necessarily support) the
hypothesis that chrysotile is cleared more readily from a long-term sequestration compartment
than amphiboles.  The authors also report that chrysotile fibers observed in this study are much
shorter than the amphibole fibers observed, so that differences in clearance rates might be
attributable to size differences. The authors also suggest that clearance is impaired by fibrosis.
                                                                              <;
                                                                              1.
Studies of Quebec Miners.  Several authors also studied the lung content of various groups of
deceased chrysotile miners in Quebec and found, despite overwhelming exposure to chrysotile
from the ore (which contains only trace quantities of tremolite, see Case et al. 2000 and
Sebastien et al. 1986), a substantial number of fibers (in some cases the majority of fibers)
observed in the lungs of deceased miners from this area are tremolite. Thus, for example:

       •      in a study of lung burdens in 6 mesothelioma victims, Churg et al. (1984) showed
              that amphiboles structures were 5-15 times as plentiful as chrysotile despite the
              predominantly chrysotile exposure;                                 j
                                                                              i
       •      in a study of lung tissue from 20 asbestosis cases, Pooley (1976) found' substantial
             concentrations of tremolite in the lungs of deceased Quebec chrysotile workers;

       •      in a much larger study comparing lung burdens of Quebec workers with those
             from the South Carolina textile mill, which has also been extensively studied for
             asbestos-related mortality (Section 7.2.3), Sebastien et al. (1989) examined  161
             lung tissue samples (89 from the Quebec mines). Results from this study indicate
             that geometric mean tremolite fiber concentrations were more than 3 times mean
             chrysotile fiber concentrations (18.4 vs. 5.3 #ng dry lung tissue) among the
             deceased Quebec miners evaluated.  It was also found that, despite these
             differences, the overall size distributions of tremolite and chrysotile fibers
             observed in lung tissue were approximately the same, although this conclusion is
             suspect.  A more detailed discussion of the results of this study is provided in
             Section 7.2.3; and
                                          6.43

-------
       •      in a more focused study using a subset of the lung samples evaluated by Sebastien
              et al. (1989), Case et al. (2000) found that the majority of long fibers (longer than
              18 um) in the lungs samples from the deceased Quebec miners that he examined
              were in fact composed of tremolite.  These authors also found substantial
              concentrations of long tremolite fibers (relative to chrysotile fibers) in deceased
              workers from South Carolina and even higher concentrations of commercial
              amphibole fibers (amosite and crocidolite) in the lungs of these workers. Thus, in
              addition to suggesting the relative persistence of amphibole asbestos compared to
              chrysotile in vivo, this finding also suggests that the accepted notion that the
              South Carolina cohort studied by Dement et al. and McDonald et al. (see
              Appendix A) was exposed almost exclusively to chrysotile may not be correct.
              This study is discussed more fully in Section 7.2.3.

These observations provide evidence that either amphibole (tremolite) asbestos is deposited
more efficiently in the compartments of the lung where clearance is slow or chrysotile asbestos
is cleared more rapidly and efficiently from even the slowest clearing compartments of the lung
(or both). Moreover, this conclusion appears to apply similarly to both short and long fibers.

In another study, McDonald et al. (1993) suggest more  specifically that lung burden data from
Quebec indicate little evidence of decreasing chrysotile concentration with time since last
exposure. Rather they suggest simply that tremolite is initially deposited in the deep lung more
efficiently. These authors also indicate that tremolite fibers are mostly optical while chrysotile
are mostly "Stanton" or thinner.

McDonald et al. (1993) report good correlation of both tremolite and chrysotile with estimated
past exposures, which contrasts with the findings of the evaluation we conducted on the data
from Sebastien et al. (Section 7.2.3).  In McDonald et al.(1985), reported geometric mean
measurements for the various fibers in lung are: tremolite: Ixl0*-18.2xl06, chrysotile :
1.5xl06-15.7xl06, respectively, when exposure varied from <30 mpcf to >300 mpcf. However,
note that, if these are PCM measurements, this may not be telling the whole story.  The authors
also report that 66% of those who died 10 years since first exposure and half of those who died
30 years since first exposure showed high chrysotile concentrations in their lungs.
Unfortunately, without access to the raw data from this study, it is not possible to identify the
route of the apparent discrepancy between the findings  of this study and those reported above for
ostensibly similar studies.

Some of the pathology studies that have been published suggest that at least some clearance
mechanisms show a dependence on fiber size, which is consistent with what is observed in
animal studies (Section 6.2.1).  Notably, for example, Timbrell (1982) studied deceased workers
and relatives from the Paakkila anthophyllite mine in Finland. He found that structures shorter
than 4 um and less than  0.6 um in  diameter are completely cleared from healthy lungs. The
efficiency of clearance decreases slowly with increasing size.  Structures longer than 17 u.m and
thicker than 0.8 um in diameter are not significantly cleared. The study is based on a
comparison of structure  size distributions in lungs compared to the structure size of the material
in the original dust exposure.  Timbrell also noted that asbestosis suppresses the removal
process.
                                           6.44

-------
When considering the dependence of clearance on size (particularly via mechanisms involving
phagocytosis), it is necessary to address differences in human and animal physiology;' Due to
differences in the morphology, for example, human macrophages have been shown capable of
phagocytizing larger particles and longer fibers than macrophages found in mice and ;rats
(Krombach et al.  1997 [for details, see Section 4.4]). Thus, the range of fibrous structures that
are efficiently cleared from human lungs is expected to include longer fibers than theirange
efficiently cleared in mice or rats. Unfortunately, given the limited precision of the available
data, the size ranges that are reported to be cleared efficiently in rats and humans, respectively,
cannot be easily distinguished.

Several human pathology studies also support observations from animal studies indicating that
clearance may be inhibited by the development of fibrosis (Albin et al. 1994;Churg et al. 1990;
Morgan and Holmes 1980) or by heavy smoking. However, other studies do not indicate such
hindered clearance either with smoking (Finkelstein and Dufresne 1999) or with fibrosis.
                                                                              I
Older Studies. Morgan and Holmes (1980) examined tissue samples from 21 patients in
England (10 who died of mesothelioma, 3 who died of lung cancer, and 8 who died of other
causes). In this study, formalin-fixed tissue samples were digested with hypochlorite.'' The
residue  was then rinsed, diluted, and an aliquot filtered.  The filter was mounted on a microscope
slide and clarified for analysis by phase contrast optical microscopy. Importantly, the.authors
note that chrysotile fibers were ignored in this study because they would not generally have been
detected by this technique.  Portions of the filters were also carbon coated and prepared for TEM
analysis.  Based on the observation that only 19% of the fibers observed in this study were
between 2.5 and 5 um, when the authors expect airborne distributions to contain closer to 90% of
the fibers within this size range, the authors conclude that short fibers are preferentially cleared
from the lung. They also conclude, based on one subject with asbestosis whose lung tissue
exhibited 72% short fibers, that asbestosis hinders clearance. The authors also note that fewer
than 1% of ferrugenous bodies (iron-coated asbestos bodies) are <10 urn in length, which
indicates (in agreement with previously published work) that such bodies seldom form on short
fibers. They also suggest that virtually all fibers longer than 20 nm tend to be coated in the
distributions they observe.                                                      '

Le Bouffant (1980) studied the concentrations, mineralogy, and size distributions of asbestos
fibers found in the lungs and pleura of deceased asbestos workers. Based on the analysis,
Le Bouffant (1980) found that the average ratio of chrysotile fiber concentrations found in the
lung versus the pleura is 1.8 while for amosite the ratio is 34.  This indicates that chrysotile
migrates from the lung to the pleura more rapidly than amphiboles resulting in a higher fraction
of total fibers in the pleura being composed of chrysotile (3% in the lungs versus 30%iin the
pleura). With regard to size, the researchers found the size distribution of amosite is virtually
identical in the lung and pleura while chrysotile fibers found in the pleura are much shorter than
chrysotile fibers found in lung tissue. This suggests that the movement  of chrysotile is a result of
a combination of translocation and degradation to shorter fibers (or that tissue samples have been
contaminated with environmentally ubiquitous short, chrysotile structures).  The authors indicate
that chrysotile fibers apparently degrade to shorter fibers more rapidly than amosite and
translocate to the pleura more rapidly than amosite. Thus, a greater fraction of chrysotile fibers
(albeit short fibers) reach the pleura than amosite fibers over fixed time  intervals.  However, the
                                          6.45

-------
results of this study also confirm that the longer araosite fibers do eventually translocate,
although on a much more extended time scale than the translocation of chrysotile.

Importantly, the results of this study need to be evaluated carefully. Boutin et al. (1996) showed
that the majority of asbestos fibers in the pleura (particularly the long fibers) are aggregated in
localized "black spots" (which surround the sites of lymphatic drainage). Thus, if the tissue
samples analyzed by Le Bouffant (1980) do not contain representative sets of such spots, the
conclusions drawn by Le Bouffant (1980) may be subject to question.

In summation, human pathology studies tend generally to support the findings of other studies
regarding the size effects of asbestos (i.e., short fibers tend to clear more rapidly than long fibers,
which can be retained in pulmonary tissues for extended periods).  They also appear to highlight
drastically different behavior between chrysotile and the amphibole asbestos types (particularly
tremolite) regarding the heavily favored retention of the latter, which has also been indicated in
animal studies. Unfortunately, the ability to draw quantitative conclusions from human
pathology studies is hampered by the severe limitations of these studies (Section 5.2).

6.2.4  Studies of Dissolution/BioDurability

Although asbestos minerals are relatively insoluble in vivo in comparison, for example, to
various fibrous glasses or other man-made mineral fibers (see, for example, Hesterberg et al.
1998a or Bastes and Hadley 1996), they do eventually dissolve in the body.  Therefore, this
pathway may contribute importantly to the overall biological clearance of asbestos. Moreover, it
has been suggested by several researchers (see, for example, McDonald 1998a and other
references cited below) that differences in biodurability between chrysotile and the amphiboles
may at least partially explain the disparate potencies observed for these  fiber types toward the
induction of mesothelioma and, potentially, lung cancer (see Sections 7.2.4.2 and 7,3.3.2).

       Note that the term "biodurability" is used here to indicate the persistence of a
       particle or fiber attributable specifically to solubility (in the absence of other
       clearance or degradation  mechanisms). In contrast, the term "biopersistence" is
       used to indicate the overall persistence of a particle or fiber in the body
       attributable to the combined effect of all mechanisms by which it might be
       removed. Thus, for example, while biopersistence can be evaluated in vivo,
       biodurability can best be  inferred from in vitro dissolution studies so that effects
       from other clearance mechanisms can be eliminated.

Several studies further indicate that both the in vivo biopersistence and the bio-activity
(including carcinogenicity) of various fiber types may be linked to their observed, in vitro
dissolution rates (Bernstein et al. 1996; Bastes and Hadley 1995, 1996; Hesterberg et al. 1998a,
1998b).  Such studies, however, typically involve fiber types with dissolution rates that are rapid
relative to the rates of clearance by other mechanisms (Sections 6.2.1.1 and 6.2.1.2). In such
studies, moreover, the various types of asbestos are typically employed as negative (insoluble)
controls. In fact, most of these studies are based on experiments with rats and the 2-year lifetime
of a rat is comparable to the anticipated lifetime of chrysotile asbestos in the body and short
compared to the anticipated lifetimes for the amphiboles (see below). Therefore, such studies
are not particularly sensitive to differences in the relative biodurability of the different asbestos

                                           6.46

-------
types. In fact, in the majority of these studies, the dissolution rates reported for asbestos were
derived indirectly by analogy with other minerals or are quoted from other studies that derive
rates similarly and may therefore be somewhat unreliable. Nevertheless, a review of a subset of
these studies is instructive.                                                        i:

Bastes and Hadley (1995 and 1996) report a simple model that reasonably predicts the relative
fibrogenicity and tumorigenicity for a range of synthetic fibers based on the dissolution rates of
the fibers measured  in vitro. The authors found that they could explain observations by
assuming that the effects of the various fibers are a function of an adjusted dose that accounts for
biodurability. Thus,

                                          F=f(ax)

where "F" is the observed incidence of the endpoint, "f' is the dose-response function proposed
for the effect, "x" is the measured dose, and "a" is an adjustment factor that accounts for
durability.
                                                                                i
In the model, "a" is  determined simply as "td/tL" where td is the time that a fiber of diameter, "D"
remains in the lung and tL is the lifetime of the exposed animal (e.g., 2 years for rats)J This
simple model reasonably reconciles the results observed in animal inhalation and injection
studies of MMVF's, RCF's, and asbestos for endpoints including lung tumors, degree of fibrosis,
and (for intrapleural injection studies) mesothelioma.  Based on a chi-square test, the simple
model is shown to adequately fit the data to a number of databases reviewed.  In contrast, the
unadjusted doses do not. Importantly, the dissolution rates used for the various asbestos
minerals in this  study were  estimated by analogy with similar minerals and  therefore may be
unreliable.

In studies comparing in vivo biopersistence with dissolution rates measured in vitro, Bernstein et
al. (1996) and Hesterberg et al. (1998a), indicate that it is necessary to consider only Jong fibers
(typically longer than 20 urn), because shorter structures are typically cleared by other
mechanisms.  They also indicate, at least for this type of study in which whole lungs were
homogenized and dissolved prior to preparation for asbestos analysis,2 that clearance is initially
rapid. This reflects muco-ciliary clearance from the upper respiratory tract. Therefore, it is
clearance of long fibers from a longer term pool that tracts in vitro dissolution rates.

These authors also report that long, soluble fibers (longer than 20 (im) are actually cleared more
rapidly than short fibers in these studies. They indicate that this is likely due to long fibers being
too long to be effectively phagocytized by macrophages so that they are left to dissolve in the
extracellular fluid at neutral pH. Shorter fibers are effectively taken up by macrophages so that
dissolution is hindered by the more acidic environment of the phagosomes (pH 4.5)  and by the
limited volume of fluid within which to dissolve.
        When whole lungs are homogenized to determine lung burden, this includes the largest airways, which
initially contain substantial concentrations of material that is rapidly cleared by the muco-ciliary escalator.
However, because this material dominates the quantity of material observed, such studies are not useful for tracking
the longer term clearance processes that occur in the deep lung.

                                            6.47

-------
Law et al. (1991) studied the dissolution of a range of fibers in solutions used as common
fixatives for biological samples.  The authors report that chrysotile and crocidolite, as well as
many other fibers, dissolve at measureable rates in the fixatives studied (Karnovsky's fixative
and formalin fixative). They therefore recommend that fiber concentrations and size
distributions obtained from tissue samples stored in such fixatives should be evaluated carefully
to account for the possible effects of the fixatives.

Although Coin et al. (1994) reported seeing no effective reduction in long fiber (>16 p,m)
chrysotile (nor other evidence of dissolution) in their study of fiber biopersistence, the limited
time frame of this study (30 days) may have been too short to allow detectable changes to
accumulate.

The most consistent data for the comparative biodurability of chrysotile and the amphiboles
(specifically crocidolite) is found in two in vitro studies of the dissolution rates of fibers that
were conducted under comparable conditions.  In the first of these studies, Hume and Rimstidt
(1992) measured the dissolution  rate for chrysotile asbestos at neutral pH under conditions
analogous to biological systems. The  dissolution rate that they report for chrysotile converts to:
KdisS~12.7 ng/cm2-hour and this is reportedly independent of pH.  In a comparable study Zoitus
et al. (1997) report the following dissolution rate for crocidolite: Kdiss=0.3 ng/cm2-hour, which is
40 times slower than for chrysotile.  Dissolution rates for several MMVF's and RCF-1 are also
reported in the latter paper, which are  listed from fastest dissolving to slowest in Table 6-4.

            Table 6-4.  Measured in vitro Dissolution Rates for Various Fibers*
Fiber Type K,^ (ng/cm2-hr)
MMVF 10
MMVF 1 1
MMVF 22
MMVF 21
Chrysotile
RCF1
Crocidolite
259
142
119
23
12.7b
8
0.3
  'Source: Zoitus et al. (1997)
  bSource: Hume and Rimstidt (1992)

  Note that dissolution rates for other amphiboles, such as amosite are probably no more than a factor of
  two or three different than that reported above for crocidolite (see, for example, Hesterberg et al.
  1998a,b).
                                            6.48

-------
To compare the effect of biodurability on the in vivo biopersistence of asbestos and other fiber
types, both the detailed kinetics of dissolution and the distribution of fiber sizes must be
considered.                                                                    :.

As reported by Zoitus et al. (1997), at a sufficiently high rate of fluid flow, the rate of mass loss
from a fiber is proportional to its surface area, A. Thus:
                                        dM/dt=-kA.

This means that for a uniform mass fiber dissolving congruently:

                                    l-(M/M0)°-5=2kt/D0p.

       where:
   (Eq.6-4)
i   (Eq.6-5)
              M     is the mass at time t;
              M0    is the initial mass at time t=0;
              D0     is the initial diameter of the fiber; and
              p      is the density of the fiber.

Substituting the equation relating the mass and the diameter of a fiber (M=p7id2h/4) into the
above equation, cancelling terms, and rearranging indicates that (during dissolution) the diameter
of a fiber decreases linearly with time:
                                       D=D0 -2kt/p.
   (Eq. 6-6)
       where:
              D is the diameter at time t; and
              all other terms have been previously defined.
                                                                              (
Furthermore, the rate of reduction in radius is given by: k/p. Based on the dissolution rates given
above for chrysotile and crocidolite, the radius reduction rates (v,^) for these fiber types are
determined to be: 1.26xlO"8 um/sec and 2.6xlO"10 urn/sec, respectively. Thus, the dissolution of
each fiber is a zero order process (i.e., the rate is constant with time and independent of
concentration). Given these rates, a chrysotile fiber 1 um in diameter will disappear in
approximately a year (3.9xl07 sec) and a crocidolite fiber of the same diameter in approximately
60 years (1.9xl09 sec).

The number rate of disappearance of a population of fibers due to dissolution is a function of the
rate of radial reduction for the fiber type and the distribution of fiber diameters in the population.
The time at which the entire population finally dissolves can be estimated simply by dividing the
radius of the largest fiber by the radius reduction rate, vrad, that is appropriate for the fiber type.
The number of fibers remaining from the population at time t will be equal to the number of
fibers in the original  distribution with radii larger than vradt for the reduction rate that is
appropriate for the fiber type.
                                           6.49

-------
       Note that dissolution will not cause an immediate reduction in fiber concentration.
       The number of fibers will not begin to decrease until sufficient time has elapsed
       for the thinnest fibers to completely dissolve. Bastes and Hadley (1994) therefore
       recommend tracking the time dependence of the mode of the distribution of fiber
       diameters to best gauge the effects of dissolution in vivo.

Importantly, fibers in vivo will only dissolve at the rates predicted by the above equations if the
fluid in which they are dissolving flows past the  fibers sufficiently rapidly to prevent saturation
from limiting the rate (Mattson 1994).  Especially for slow dissolving materials of limited
solubility like asbestos, it is expected that the observed dissolution rate in vivo will generally be
slower than the rates predicted based on in vitro  measurements. Even for more rapidly
dissolving fibers like most fibrous glasses and manmade mineral fibers, dissolution is hindered
in compartments of the body in which the volume of available solute is limited.

In summary, dissolution is a zero-order (i.e., constant with time, independent of concentration)
clearance mechanism that is dependent on fiber mineralogy, that the effect it has on fiber
populations (concentrations) is a function of the  distribution of fiber diameters within the
population, and that the theoretical rate of dissolution may not be achieved in all tissues in all
compartments of the lung or mesothelium due to limits in the  rate  of in vivo solute flow.

6.2.5   Dynamic Models

Unlike particle deposition in the lungs, which is  an entirely mechanical process, clearance,
transport, and degradation mechanisms tend to be complex biochemical processes. Due to the
incomplete understanding of such processes, state-of-the-art modeling of degradation and
clearance is not as advanced as that for deposition. Even the most sophisticated of degradation
and clearance models remain semi-empirical.  Although, current models in this area show
general agreement with the sparse, available data, there are clearly areas of weakness that require
additional research. Nevertheless, the  models provide a good indication of the kinds of
processes that are important in the body and their overall constraints.  It is also  noted that models
for degradation and clearance in humans tend to be better developed than those for animals,
primarily due to  confounding  uncertainties associated with animal ventilation rates. An
overview of the state of the art, which  was current as of the date of publication, can be found in
Stoberetal.(1993).

According to Stober et at. (1993), the general conclusions that can be drawn from the current
models are that:

       •      clearance from ciliated  airways is rapid, independent of particle/fiber type,
              apparently independent of particle size, and can be described as the sum of two,
              weighted exponentials (i.e., an assumed combination of two first-order decay
              processes), although the process may in fact be zero order (i.e., the reduction in
              concentration with time is constant and independent of concentration) with rates
              that differ primarily by  the distance that a particle must traverse  to return to the
              trachea.
                                           6.50

-------
              Based on studies of particle clearance reported by Raabe (1984), mucofciliary
              transport in the nose and throat generally exhibits a half-life for clearance of
              4 minutes. Clearance of the tracheo-bronchial section of the respiratory tract is a
              function of the distance from the trachea and generally varies from a half-life of
              30 minutes for the largest bronchi to approximately 5 hours for the smallest and
              most remote bronchi. In healthy humans, material deposited in this region is
              generally cleared within 24 hours.  In contrast, the clearance mechanisms
              operating in the deep lung, beyond the muco-ciliary escalator, operate over time
              frames of many days to years (see Table 6-2);                       -•

       •     clearance of insoluble particles from the pulmonary portion of the lung occurs
              primarily by macrophage transport and such transport has several components.
              One component represents the population of "free" macrophages located within
              the alveoli that engulf particles and transports them to the muco-ciliary, escalator.
              Macrophages are also renewed at some rate of recruitment that may be dependent
              on particle concentrations. In fact, numerous studies have demonstrated that
              macrophage recruitment is induced by the deposition of asbestos and other
              particles in the lung (Section 6.3.5).  Particles also migrate into the interstitium
              where another population of macrophages clears these particles to lymph. This
              second component (interstitial clearance) is much slower than the first (see
              Table 6-2);       ,                                               :

       •     each macrophage can carry a maximum load and the mobility of each macrophage
              decreases with increasing load. At sufficient loading, macrophages become
              immobile and aggregates of overloaded macrophages in the alveoli may then
              sequester particles for some period of time as this clearance mechanism  is shut
              down. In the interstitium, masses of immobile macrophages may trigger
              development of granulomas that sequester particles for extended periods of time
              by effectively preventing clearance of the particles within such tissue,j'at least
              until or unless the granulomas resolve. Thus, these models incorporate overload
              mechanisms and the incorporation of such overload mechanisms are required to
              explain observed trends in experimental results; and

       •     in various published studies, overload (immobilization of laden macrophages) has
              been modeled as dependent on the total volume or mass of phagocytized material
              (for compact particles) and (additionally) on the length of phagocytized  material
              (for fibers). It is also possible that the motility of macrophages and the
              consequent overall rate of this clearance process is additionally a function of fiber
              diameter and/or particle toxicity (the latter for special cases).

Interestingly, while it is reported that large, spherical particles are not readily cleared by this
mechanism, the range of sizes over which clearance becomes hindered corresponds reasonably
well to the limits of overall respirability. In contrast, fibers that are clearly too long to be cleared
by macrophages, if they are.sufficiently thin, are quite respirable.  Thus fibrous materials present
a unique challenge to the respiratory tract based solely on the dimensions of these materials.
                                          6.51

-------
Stober et at. (1993) also notes that many models incorporate the assumption that most clearance
processes are first order (i.e., that the rate of reduction of mass or fiber number is proportional to
the remaining mass or fiber number, respectively, and independent of other factors). Thus, the
combined effects of multiple clearance processes can be expressed as a weighted sum of
exponentials and this approach has been fairly successful at mimicking actual processes. This
means, however, that the half-lives "t,/2 's" attributed to the various first order decays are
empirical and do not necessarily correspond to any specific physiological or biochemical
features of the processes being modeled. Depending on the specific process, clearance rates may
be zero order or may be a more complicated function of multiple variables  than can be described
by a first order decay. Nevertheless, models incorporating these simplifying assumptions have
shown good success at adequately describing observed effects.

Note that half-lives for first order decay processes represent the time required for half of the
initial mass to decay (or be transported or whatever) and can be estimated as: tl/2=(ln2)/k with k
being the first order rate coefficient or proportionality constant between rate and mass. This is
why so many of the retention studies cited above provide estimates of a series of decay constants
or half-lives that are assumed to correspond approximately to the major clearance processes
contributing to the observed, overall reduction in concentration.

Due to the complexity of the processes involved, only a small number of dynamic models for
fiber retention have been developed. Interpretation of the results of these models requires that
the meaning of the term "retention" first be reconciled across studies.

Dement and Harris (1979) report that, based on a mathematical model, the  fraction of structures
retained in the deep lung is unlikely to vary by more than a factor of 2 for different asbestos
mineral types. In this study, however, the term retention appears to refer primarily to a very
short time period that primarily includes consideration of deposition, but not clearance processes.

Using a definition for retention that reflects long-term residence in the lung, Yu et al. (1990)
developed a model of chrysotile retention that explicitly incorporates longitudinal splitting,
dissolution, and size-dependent clearance.  Time-dependent lung burden estimates derived using
the model were shown to compare reasonably well with published data (Abraham et at. 1988, as
cited by Yu et al. 1990) both in terms of fiber concentrations and fiber size-distributions.

In a later modification of their retention model for chrysotile, Yu et al. (1991) also considered
the effect of airway asymmetry on fiber retention.  In this version of the model, Yu and
coworkers incorporated information concerning the geometry of the bronchio-alveolar tree
(including mean distance and the mean number of airway bifurcations between the trachea and
the alveoli in each section  of the lung) and studied the effects of such considerations.  The
modified model predicts a non-uniform distribution of the asbestos that is retained in the lung
and the predictions reasonably reproduce the distributions observed  by various researchers and
measured formally by Pinkerton et al. (as cited by Yu et al. 1991).

Yu and Asgharian (1990) also modeled the long-term retention of amosite  in rat lungs. In
contrast to the models employed for chrysotile, the model presented for amosite incorporates a
term for the clearance rate that is not a constant but, rather, is a function of the lung
concentration of asbestos,  which was adapted from an earlier model for diesel soot. This

                                           6.52

-------
modification was incorporated to adequately mimic the suppression of clearance with |increasing
lung burden that has been observed by several research groups (e.g., Davis et al. 1978; Wagner
et al.  1974) for amphiboles.  Conditions under which elevated asbestos (or dust) concentrations
are observed to reduce clearance are referred to as "overload" conditions. Model predictions
were  shown to reasonably reproduce the time-dependence of amosite lung burdens (in terms of
mass) in several studies.                                                        '

Importantly, the overload conditions addressed in the Yu and Asgharian (1990) model were
primarily observed among older retention studies where lung burden was tracked as total
asbestos mass (Section 6.2.1).  Such studies tend to suggest a difference in the behavior of
chrysotile and the amphiboles. As indicated in Section 6.2.1, however, later retention* studies,
which track lung burden as a function of fiber number (in specific size categories) tend to show -
this effect is a function of fiber length more than fiber type and newer models may need to
incorporate such factors that indicate reduced macrophage motility  as a function of fiber length.

Moreover, it is important to consider the major, confounding effects, if the goal is to develop a
model that not only reproduces the time-dependence of clearance, but also captures relevant
physical phenomena.  Thus,  for example, Yu et al. (1994) were able to reproduce the time-
dependence of the retention in rats of inhaled RCF-1 (as a function  of fiber size) using a model
in which macrophage  motility was limited only by total lung burden and not dependent on fiber
size.  However, these authors also failed to consider that long RCF-1 fibers in fact dissolve at
rates competitive with the clearance of short fibers (see Section 6.2.1), which is probably why
they did not find a dependence on length;  the two effects cancelled  out.

6.2.6   General Conclusions Regarding Deposition, Translocation, and Clearance
                                                                              t,
The current literature on deposition, translocation, and clearance paint a consistent picture of the
fate of fibrous structures in the lung. The ultimate fate of biodurable fibers depends
overwhelmingly on their size.  Although there may be additional effects due to mineralogy
(addressed further in Section 6.3) and for rare, special cases this may be important, generally
such effects appear to  be minor.                                                 !'

The primary effect attributable to mineralogy that is important to consider in relation to
clearance is that associated with biodurability. Fibers that dissolve  in the  lung at rates that are
competitive with the other clearance mechanisms described below may be cleared sufficiently
rapidly to preclude adverse effects, even when such fibers are too long to be cleared efficiently
by macrophages (see Section 6.2.1).                                            ::

A schematic representation of the complex set of mechanisms that contribute to the translocation
and clearance of fibrous structures that have been deposited in the deep lung is presented in
Figure 6-4. This description was developed based on the complete  spectrum of observations
reported in each of the previous sections of this chapter including, primarily, the descriptions of
the most sophisticated of the models reviewed by Stober et al. (1993).
                                          6.53

-------
                                      FIGURE 6-4:
        PUTATIVE MECHANISMS FOR CLEARANCE AND TRANSLOCATION
•*v
           Alveolar
           Lumen
    K6 =
                                                                         to Lymph
                                                              K11 * ffslze.oonc, toxldty)
        Capillary
         Lumen ,
                                                  »—>K-/-    '
  jo
Lymphatic
Drainage
Figure 6-4.  Key for Putative Mechanisms for Clearance and Translocation of Fibers in the
                                        Lung

K! =  rate constant for phagocytosis of fibers by alveolar macrophages. This mechanism is an
      inverse function of fiber length and, likely, diameter. This mechanism is likely pseudo
      first order (assuming sufficient numbers of macrophages, the rate will be proportional to
      the number of fibers);
K2 =
K =
K4 =
       rate constant for phagocytosis of fibers by Type I epithelial cells. This mechanism is an
       inverse function of fiber length and, likely, diameter. This mechanism is likely pseudo
       first order;

       rate constant for diffusion of fibers from the alveolar air space (lumen) through the
       epithelial lining to the underlying interstitium.  This mechanism is diffusion limited and
       likely independent fiber size or type. This mechanism likely parallels the behavior of
       diffusion in through a finite, column of fixed diameter.

       rate constant for phagocytosis of fibers by Type II epithelial cells. This mechanism is an
       inverse function of fiber length and, likely, diameter. This mechanism is likely pseudo
       first order;
                                         6.54

-------
 Figure 6-4 Key for Putative Mechanisms for Clearance and Translocation of Fibers in the
                                    Lung (continued)                         (

K3 =   rate constant for transport by macrophage to the muco-ciliary escalator. This mechanism
       is likely an inverse function of fiber length and the volume (or mass) of the fibers
       phagocytized.  Macrophages that become immobilized tend to aggregate in alveolar
       lumena. For macrophages with fixed loads, this mechanism may be first orderlor zero
       order.

K6 =   rate constant for putative discharge to interstitium of phagocytized fibers by Type I
       epithelial cells. There has been no direct verification of this mechanism;     *

K7 =   rate constant for phagocytosis of fibers by endothelium. This mechanism is an'inverse
       function of fiber length and, likely, diameter. This mechanism is likely pseudo first
       order;

K8 =   rate constant for putative mechanism in which fibers internalized by macrophages are
       transported through the epithelial lining of the alveolar space to the underlying
       interstitium.  There has been no direct verification of this mechanism;
                                                                             I
Kj =   rate constant for diffusion of fibers from the interstitium through the endothelial lining to
       the enclosed, capillary lumen.  This mechanism is diffusion limited and      •;

       likely independent fiber size or type. There has been no independent verification of this
       mechanism;

K,0 =  rate constant for phagocytosis of fibers by interstitial macrophages. This mechanism is
       an inverse function of fiber length and, likely, diameter. This mechanism is likely pseudo
       first order (assuming sufficient numbers of macrophages, the rate will be proportional to
       the number of fibers);

Ku =  rate constant for transport by macrophage from the interstitium to the lymphatic system.
       This mechanism is likely an inverse function of fiber length and the volume (or mass) of
       the fibers phagocytized. Macrophages that become immobilized tend to induce
       formation of interstitial granuloma.  For macrophages with fixed loads, this mechanism
       may be first order or zero order.
K,2 =  rate constant for putative discharge to capillary lumena of phagocytized fibers by
       endothelial cells. There has been no direct verification of this mechanism;

K13 =  rate constant for phagocytosis of fibers by mesothelial cells of fibers transported to the
       mesothelium. This mechanism is an inverse function of fiber length and, likely,
       diameter.  This mechanism is likely pseudo first order;
                                          6.55

-------
 Figure 6-4 Key for Putative Mechanisms for Clearance and Translocation of Fibers in the
                                    Lung (continued)

K14 =  rate constant for putative mechanism in which fibers internalized by macrophages are
       transported from the interstitium through intervening tissue and the mesothelium to the
       pleural space. There has been no direct verification of this mechanism;
K15 =
rate constant for phagocytosis of fibers by pleural macrophages. This mechanism is an
inverse function of fiber length and, likely, diameter. This mechanism is likely pseudo
first order (assuming sufficient numbers of macrophages, the rate will be proportional to
the number of fibers);

rate constant for transport by macrophage to sites of lymphatic drainage (lymphatic
ducts) along the pleura. This mechanism is likely an inverse function of fiber length and
the volume (or mass) of the fibers phagocytized. For macrophages with fixed loads, this
mechanism may be first order or zero order.

Not shown: apparently fibers that are too large to pass through lymphatic ducts attract
accumulation of macrophages at sites of lymphatic drainage (Kane and MacDonald
(1993).
rate constant for the renewal of the alveolar macrophage population. While the there is
likely a steady state  rate for background renewal, given that the average life of an
alveolar macrophage is reported to be on the order of 7 days, this rate is also stimulated
in response to insult by foreign substances in the lung;

rate constant for the renewal of the interstitial macrophage population.  While the there  is
likely a steady state  rate for background renewal, this rate is also expected to be
stimulated in  response to insult by foreign substances in the interstitial space;

rate constant for the renewal of the pleural macrophage population. While the there is
likely a steady state rate for background renewal, this rate is also expected to be
stimulated in  response to insult by foreign substances in the pleura;
                                          6.56

-------
As shown in Figure 6-4, briefly, the first reaction to the introduction of fibers (or other
participate matter) into the alveolar lumen is scavenging by alveolar macrophages. It has been
reported that the initial uptake by macrophages is a rapid process that is essentially complete
within hours after initial deposition. Rates for several of the mechanisms depicted in Figure 6-4
have been estimated in the literature and are summarized in Table 6-2.              ;

The rate of removal (to the muco-ciliary escalator) by alveolar macrophages is then determined
by a variety of effects. Macrophage motility is a size dependent-process so that only fibers that
are sufficiently compact (<~20 \im) can be removed from the lung. The rate of removal by this
process may also be suppressed both for fibers of intermediate lengths (10-20 um, which are
short enough to be phagocytized, but long enough to suppress macrophage motility) and by the
overall mass/volume of particles deposited (and, proportionally, taken up by each macrophage).
Note that the dimensions provided are the ones that are apparently appropriate for humans. For
rats, the corresponding dimensions may be somewhat smaller.

Likely competing with scavenging by macrophages are (1) phagocytosis by the epithelial cells
lining the alveolus and (2) diffusive transport to the intersititium.  Both Type I and Type II
epithelial cells appear to phagocytize fibers. Although relatively few fibers are observed to be
taken up by Type II cells, as previously discussed (Section 6.2.2), one possible explanation for
the limited observation of fibers in Type II cells is that uptake of fibers induces terminal
differentiation to Type I cells. It is expected that phagyctosis by epithelial cells is a size-
dependent process.                                                            ;

Especially when the presence of fibers (or other particulate matter) induces morphological
changes in Type II cells that increase the overall permeability of the epithelial lining (Section
6.3.7), fibers can apparently diffuse into the interstitium.  This process, potentially supplemented
with expulsion of phagocytized fibers by epithelial cells and/or transport of fiber laden
macrophages through  the epithelial lining, represents the set of putative mechanisms by which
fibers may reach the interstitium. It is expected that these processes are somewhat slower than
uptake by alveolar macrophages (see Table 6-2).

Therefore, if this latter mechanism is operating at peak efficiency, relatively few fibers may
reach the interstitium.  Note that diffusive transport is likely  independent of fiber length, but may
be dependent on fiber width with thinner fibers more rapidly diffusing to the interstitium.

Fibers reaching the interstitium are likely cleared primarily by interstitial macrophages, which
phagocytize the fibers and transport them to the lymphatic system. Both the efficiency of
phagocytosis and motility of macrophage transport in the interstitium likely depend oh fiber size
and the total volume/mass of fibers in the same way described above for transport by alveolar
macrophages.  However, all such mechanisms are substantially slower in the interstitium than in
the alveolar lumen (see Table 6-2).

Macrophages that have been immobilized (due to fiber size or volume/mass) in the interstitium
tend to aggregate and  induce formation of granulomas, which may sequester the fibers in these
cells.  Although there  is less evidence for this, fibers free in the interstitial matrix might also
trigger such a process. Such fibers would typically be too large to have been effectively
phagocytized by any of the cells of the interstitium.

                                          6.57

-------
Fibers may also reach the endothelium and be taken up by endothelial cells lining the capillaries
of the deep lung. Because fibers have also been observed in capillary lumena, mechanisms
similar to those described for transport through the alveolar epithelium to the interstitium may be
operating to transport fibers into capillary lumena. While it is expected that such mechanisms
will also show size dependence similar to that previously described, little is known about the
details or the rates of such processes.

Also by mechanisms similar to some or all of the putative mechanisms described for
translocation of fibers from the alveolar lumen to the interstitium, fibers may reach the pleura.
Whether fibers can also reach the pleura via transport in blood or lymph has not been definitively
determined. Fibers reaching the pleura may be phagocytized by mesothelial cells or may pass
through such cells to the pleural cavity. Fibers reaching the pleural cavity are apparently
phagocytized by pleural macrophages (probably showing a similar size or volume/mass effect as
described above for similar mechanisms) and are apparently transported to (and deposited at)
sites of lymphatic drainage along the pleura.  If such fibers are then too large to pass through the
lymphatic ducts, they may trigger accumulation of additional macrophages and other
inflammatory cells.

Overall, the effects of size appear to be:

       •     few fibers thicker than approximately 0.7 um and virtually none thicker than
              approximately 1.5 ^m appear to reach the deep lung. Of these, the longer the
              fiber, the thinner the fiber. Importantly, the distribution of sizes of structures
              deposited in the deep lung tend to be much more similar across studies than the
              distributions in the aerosols originally inhaled. Thus, deposition is a very
              effective filtering process; and

       •     short fibers (or compact particles) that are shorter than somewhere between 10
              and 20 u.m tend to be taken up almost entirely by macrophages and  are either
              cleared via the muco-ciliary escalator, isolated in immobilized macrophages that
              remain within alveolar lumena, or transported to lymphatics (presumably after
              first reaching the interstitum). These processes appear to be efficient for the
              shortest of the fibers in this range (and all shorter fibers) so that no further effects
              are manifest.  In contrast, longer fibers, which are not efficiently cleared or
              isolated by macrophages either in the alveolar lumen or the interstitium appear to
              trigger a range of additional responses, some of which appear to lead to disease.

Therefore, based on deposition, translocation, and clearance, it is fibers thinner than
approximately 0.7 um and longer than a minimum of approximately 10 urn (with relative
contributions increasing with increasing length up to at least 20 um) that likely contribute to
disease.  Modifications to this range of structure sizes due to effects attributable to direct
biological responses  are addressed further in Section 6.3.

Regarding putative differences in behavior between chrysotile asbestos and amphibole asbestos,
such effects are adequately addressed by the unifying discussion provided above. To make sense
of such differences, the effects of fiber size must first be explicitly considered. Thus, chrysotile
fibers may not be as efficiently deposited in the deep lung to the extent that they are curlier or

                                           6.58

-------
occur in thicker bundles than amphibole fibers. The overall load of chrysotile deposited in the
deep lung may also be cleared more rapidly than amphiboles to the extent that (1) short, thin
fibrils ultimately represent a greater fraction of the total load of chrysotile than the amphiboles
and (2) a subset of long chrysotile fibers, not sequestered in an environment with limited fluid
flow, may be cleared more rapidly than similarly long amphibole fibers due to contributions
from dissolution.

Importantly, the concentrations of asbestos to which humans are exposed are much lower than
the concentrations to which animals were exposed in the various literature studies cited.
Moreover, as indicated in Section 6.1.2, for virtually any exposure of interest, the resulting
(volumetric or surface area) lung burden will be substantially higher in rats than in humans.
Therefore, overload conditions or other processes that might impede or alter the clearance
mechanisms described above, will never occur in humans without first having affected the
results of the animal studies reported. Thus, conclusions concerning size-dependence, and related
effects (except to the extent that they need to be adjusted for cross-species differences) should
remain valid when extrapolated between animals and humans.

6.3    FACTORS GOVERNING CELLULAR AND TISSUE RESPONSE
                                                                              i
For inhaled structures that survive degradation and clearance, a series of complex reactions
between the structures retained in the lung and surrounding tissue may induce a biological
response.  Asbestosis (fibrosis), pulmonary carcinomas, or mesotheliomas may result.
Mesotheliomas are likely associated with structures that are translocated from the lung to the
mesenchyme, although diffusable growth promoters and other chemical signals produced by
asbestos exposed cells in lung tissue immediately proximal to the mesothelium may also play a
role (see Adamson 1997, as described in Section 6.3.4.1).

That the specific biochemical triggers for asbestos-related diseases have not been definitively
delineated as of yet is not surprising.  Despite great progress in elucidating candidate *
mechanisms,  the number of candidate mechanisms is large and confounded by "cross-rtalk"
between mechanisms.. Moreover, similar toxic endpoints  may result from entirely independent
mechanisms that exhibit disparate dose-response characteristics but, nevertheless, may be
triggered by the same  or similar toxins.  In such cases, however, the relative contributions from
each mechanism to a particular endpoint may vary substantially.  Unfortunately, the ability to
compare results across studies of different mechanisms is currently limited due to the inability to
reconcile the quantitative effects of dose and response across dissimilar studies.

Table 6-5 illustrates the range and complexity of the biological responses triggered by asbestos
in lung tissue. The table was developed based on the information gleaned from the studies
described in this section. Importantly, not all of the mechanisms listed contribute equally to the
toxic endpoints that are attributable to asbestos, but their relative importance has yet to be
entirely delineated. The toxic endpoints of potential interest to which each of the listed
mechanisms potentially contribute are indicated in bold italics.
                                          6.59

-------
 Table 6-5.  Putative Mechanisms by Which Asbestos May Interact with Lung Tissue to
                           Induce Disease Following Inhalation
Asbestos (in vivo)
    Generates reactive oxygen species (ROS)
       • may affect neighboring cells and tissues
    Interacts with macrophages
    Interacts with lung epithelium
Asbestos interacts with macrophages
    Induces generation and release of reactive oxygen species/reactive nitrogen species (ROS/RNS)
       • may affect neighboring cells and tissues
       • may induce inflammatory response (which may promote cancer)
       • may induce fibrogenesis (which may promote cancer)
       • induces signaling cascades in macrophages and neighboring tissues
           - mediates apoptosis (which may regulate cancer)
           - mediates proliferation (which may promote cancer)
       • causes genotoxic effects in neighboring tissues
           - may cause cancer initiation
           - may induce arrest of cell cycle
           - induces signaling cascades
               • mediates apoptosis (which may regulate cancer)
               • mediates proliferation (which may promote cancer)
       • depletes reserves of antioxidants in macrophages and neighboring tissues
           - may induce cytotoxicity (which may promote cancer, by inducing proliferation)
           - may increase susceptibility to insult by other toxic agents (which may promote cancer)
    Induces release of various cytokines
       • affects neighboring cells and tissues
       • mediates inflammatory response (which may promote cancer)
       • mediates fibrogenesis (which may promote cancer)
       • induces signaling cascades in macrophages and neighboring tissues
                                            6.60

-------
  Table 6-5.  Putative Mechanisms by Which Asbestos May Interact with Lung Tissue to
	Induce Disease Following Inhalation (continued)	^	
            - mediates apoptosis (which may regulate cancer)
            - mediates proliferation (which may promote cancer)
        * mediates proliferation in neighboring tissues (which may promote cancer)
    'Induces signaling cascades and mediates apoptosis in macrophages (which may regulate cancer)
    At high enough concentrations, promotes cytotoxic cell death (which may promote cancer by
    inducing proliferation)                                                        \
 Asbestos interacts with Lung Epithelium
    Induces release of reactive oxygen species/reactive nitrogen species (ROS/RNS)
        • may affect neighboring cells and tissues
        • may induce inflammatory response (which may promote cancer)
        • may induce fibrogenesis (which may promote cancer)                        \\
        • induces signaling cascades in epithelium and neighboring tissues
            - mediates apoptosis (which may regulate cancer)
            - mediates proliferation (which may promote cancer)
        • causes genotoxic effects in epithelium and neighboring tissues
            - may cause cancer initiation
            - may induce arrest of cell cycle
            - induces signaling cascades
               • mediates apoptosis (which may regulate cancer)
               • mediates proliferation (which may promote cancer)
        • depletes reserves of antioxidants in epithelium and neighboring tissues
            - may induce cytotoxicity (which may promote cancer by inducing proliferation)
            - may increase susceptibility to insult by other toxic agents (which may promote cancer)
    Induces release of various cytokines
        • affects neighboring cells and tissues
        • mediates inflammatory response (which may promote cancer)
        • mediates fibrogenesis (which may promote cancer)
        • induces signaling cascades in epithelium and neighboring tissues
           - mediates apoptosis (which may regulate cancer)
                                            6.61

-------
 Table 6-5. Putative Mechanisms by Which Asbestos May Interact with Lung Tissue to
	Induce Disease Following Inhalation (continued)	
          - mediates proliferation (which may promote cancer)
       • mediates proliferation in epithelium and neighboring tissues (which may promote cancer)
   Causes genotoxic effects
       • may cause cancer initiation
       • may induce arrest of cell cycle
       • induces signaling cascades
          - mediates apoptosis (which may regulate cancer)
          - mediates proliferation (which may promote cancer)
   Induces signaling cascades
       • mediates apoptosis (which may regulate cancer)
       • mediates proliferation (which may promote cancer)
   Increases epithelial permeability
       • encourages fibrosis  (which may promote cancer)
       • facilitates translocation
   At high enough concentrations, promotes cytotoxic cell death (which may promote cancer by
   inducing proliferation)
                                           6.62

-------
As indicated previously (Section 6.0), although much has been learned about specific
components of the underlying mechanisms by which asbestos causes the above-listed diseases,
substantial knowledge gaps remain. Moreover, because of these gaps, multiple candidate effects
have been explored as potential contributors to carcinogenicity (or fibrogenicity) andl'one of the
goals of this document is to distinguish among those effects that are likely to contribute to the
induction of cancer from those that are less likely or unlikely to contribute (given the:current
state of knowledge). Accordingly, an overview of recent studies is presented below following a
brief description of the current model of the general mechanism for cancer.  Note that, due to the
availability of several recent reviews (including, for example, Economou et al. 1994;'Floyd
1990; Kamp et al. 1992; Kane and MacDonald 1993; Mossman and Churg 1998; Mossman et al.
1996; Oberdorster 1994; Robledo and Mossman 1999), only the most recent primary articles are
included in the following review.                                                .

Also, as indicated below, many of the biological responses provoked by retained asbestos are in
fact dependent on fiber size and type. Therefore, studies that distinguish among effects induced
by different size fibers (or fibers and non-fibrous particles) of the same mineralogy and studies
that distinguish among effects induced by comparably sized fibers (or non-fibrous particles), but
differing mineralogy are highlighted. However, due to the limits with which fibrous materials
have tended to be characterized in many of these studies, the database from which to 'distinguish
among the  effects of size and mineralogy are limited and conclusions from differing studies must
be compared with caution.

A large body of evidence supports the conclusion that it is primarily (if not exclusively) long
fibers (those longer than a minimum of 5-10  u.m) that contribute to disease (see  Sections 6.2.1,
6.2.2, and 6.4).  Much evidence also indicates that the potency of long fibers increases with
length at least up to a length of approximately 20 um. Therefore, because short and long fibers
are both respirable (for fibers that are sufficiently thin, less than approximately 0.7 u.m  in
diameter), differences in the ultimate response to short and long fibers must be attributable to
differences in tissue and cellular responses to the retained fibers in each size range.

At least at a histopathological level, clear differences have been observed in the  responses
evoked by  short and long structures (see Sections 6.2.1 and 6.2.2).  It is a goal of this; section to
determine whether the biochemical triggers that mediate the disparate responses to short and
long fibers  can also be identified.  Unfortunately, while a large body of knowledge has  been
amassed, definitive conclusions are not yet possible. This  is because the specific mechanisms by
which asbestos acts have still not been definitively determined, although many candidate
mechanisms have been elucidated (see above). However, important inferences can still be
gleaned from the available studies.

Evidence for the relationship between fiber diameter and disease is somewhat less clear.
Although there appears to be a fairly sharp cutoff in the diameter of fibers that are respirable (see
Section 6.1.4), several studies suggest that the most potent fibers are substantially thinner than
the sharp cutoff in respirability (see Section 6.4). If these latter observations are valid then, as
with length, differences in the ultimate response to thick and thin fibers must be  attributable to
differences in the tissue and cellular responses elicited by retained fibers of each width. As
indicated with length, however, definitive conclusions regarding biochemical triggers and the
effect of size on such triggers are not yet possible because  the specific triggers that lead to

                                           6.63

-------
specific asbestos-related diseases have not been definitively identified. Still, useful inferences
can be developed.

6.3.1   The Current Cancer Model

The following description of the current model for cancer is derived from the ideas presented in
Moolgavkar et al. 1988, Mossman 1993, and Economou et al. 1994.

In the current model of cancer, normal cells must accumulate specific, multiple mutations before
a tumorigenic cell is created that can lead to the development of cancer. Each successive
mutation produces an initiated cell (a cell that is transformed from normal cells because it
incorporates one or more of the requisite mutations, but that has not yet acquired all of the
changes needed to produce cancer).  Each initiated  cell may then proliferate to generate a
population of similarly initiated cells, which increases the probability that other events will lead
to further mutation in at least one of these cells. Individual mutations may occur spontaneously
or may be induced by exposure to mutagens.

Generally, the minimum, heritable changes required before a normal cell is transformed into a
metastatic tumor cell include, but may not be limited to (see, for example, Hei et al. 1997 or
Kravchenko et al. 1998):

       •    escape from terminal differentiation or programmed cell death (especially in
              response to DNA damage);

       • .    escape from anchorage/neighbor dependent growth inhibition;

       •     development of self-promoting growth and proliferation; and

       •     active expression of cytokines needed to promote angiogenesis and allow tissue
              invasion.

It is not clear whether the mutations associated with these changes need to occur in a particular
order, although the first of the above-listed changes would facilitate accumulation of all later
changes.  It is also unclear which of the above changes contribute to the time dependence of
tumor development and this may vary among tumor types. It is likely that only a subset of the
required mutations determine the ultimate time development of the associated cancer. For
example, once a cell begins self-promoting growth, later mutations (even relatively rare ones)
may become incorporated rather quickly.  Also, some mutations may be rare or may require
intervention by a toxic agent, while other mutations may occur spontaneously and may thus
occur frequently, once a sufficient number of initiated precursor cells are generated. This is one
of the reasons that models with as few as one or two stages have proven successful at predicting
the time course of many types of cancer (see, for example, Moolgavkar et al. 1988).
                                           6.64

-------
To produce cancer, the mutations that occur must also be "heritable" meaning that they must be
preserved and passed on to daughter cells during mitosis. Thus, it is not only necessary to cause
alteration in the DNA of a cell (genetic damage), but the mechanisms by which the cell
subsequently repairs such changes or prevents cell division (e.g., arrest of the cell cycle,
programmed cell death, terminal differentiation) in the presence of such changes must'also be
defeated. Generally, if a cell proceeds through DNA synthesis (in preparation for mitosis) before
accumulated alterations to DNA are repaired, the sites of such alterations can lead to errors in
replication during synthesis, which in turn result in permanent, heritable mutations in one or both
of the daughter cells that are created from mitosis.                                 '

Toxic agents that (directly or indirectly) cause DNA alterations may contribute to the!'
development of cancer by inducing one or more of the set of requisite mutations required for
cancer to develop. In traditional parlance, such agents are termed "initiators". In addition, any
toxic agent that enhances proliferation also facilitates the development of cancer both;by
increasing the probability of creating spontaneous mutations (due to errors that infrequently, but
unavoidably, occur whenever DNA is synthesized in support of mitosis) and by  increasing the
numbers of any initiated cells that may be present, which may then serve as additional targets for
initiators or may incorporate additional spontaneous mutations. Agents that facilitate the
development of cancer by inducing proliferation are traditionally termed "promoters."

Multiple mechanisms have been identified by which both initiators and promoters may act.
Initiators, for example, may  react directly with DNA to cause genetic damage, or may induce
generation of other reactive species (such as reactive oxygen species or reactive nitrogen
species) that, in turn, react with DNA to cause genetic damage. In addition, fibrous materials
may uniquely damage chromosomes by interfering with mitosis causing aneuploidy   '•
(incorporation of an incorrect number of chromosome copies in cells) and/or various clastogenic
changes (alterations in the organization and structure of specific chromosomes).

Promoters may also induce proliferation via a variety of mechanisms.  Cytotoxic agents,  for
example, may induce proliferation in a tissue by damaging or killing cells and thereby, induce
stem cells to proliferate to replace the damaged cells.  Promoters may also induce release of
various growth factors that, in turn, induce proliferation in targeted tissues. This may>occur, for
example, as part of the inflammatory response to tissue insult.

Promoters that are biopersistent (such as long asbestos fibers) or promoters that are continually
reintroduced (through chronic, external exposure) may also chronically up regulate (or down
regulate) certain cell signaling cascades that may contribute to cancer development in -a variety
of ways including: (1) activation of genes that mediate proliferation or production of various
growth factors (including any of various oncogenes) or (2) suppression of genes that inhibit
proliferation or growth (including any of various tumor-suppressor genes).           '

There is growing evidence that all varieties of asbestos fibers (and certain other fibrous
materials) can act both as cancer initiators and promoters. However, the biological responses to
these materials appear to vary in different tissues so that it may be important to separately
evaluate  the behavior of asbestos in specific tissues. Biological responses to varying  fiber types
also appear to vary, particularly in relation to a fiber's biopersistence (Sections 6.2.1 and 6.2.4).
Accordingly, an overview of the generic evidence that specific types of asbestos may'act as an

                                           6.65

-------
initiator and, separately, as a promoter is reviewed below, followed by consideration of the more
limited data suggesting tissue-specific variation in biological responses.

6.3.2   Evidence for Transformation

Several recent studies provide evidence that specific types of asbestos can induce transformation
of cells in specific tissues (both lung epithelium and mesothelium) to create tumor cells. This
provides further, confirmatory support for the whole animal.studies in which cancer is induced
by exposure to asbestos (see Sections 6.4).

An immortalized, but non-tumorigenic, cell line of human bronchial epithelial cells (BEP2D
cells) was transformed by a single exposure to 4 ug/cm2 UICC chrysotile B (Rhodesian).
Surviving cells (the treatment caused 18% cell death) went through several transformations
including: altered growth kinetics, resistence to serum'induced terminal differentiation, and loss
of anchorage dependent growth, before becoming tumorigenic (Hei et al.  1997). Tumorigenicity
developed in the various exposed cell lines over a period  of several to 11 weeks following
exposure. When injected into nude mice, secondary tumors developed with a latency of 8-10
weeks.                                             .

The authors indicate that there were no mutations in these cells at either codon 12 or 13 or the
ras gene (mutations that have sometimes been observed in asbestos-induced lung cancers. Also,
because this cell line already contains alterations in genes for p53 (a protein that plays a role in
cell-cycle control, among other things, see Table 6-6) and Rb (Table 6-6), the authors speculate
that such changes are not rate controlling for transformation to cancer.

It should also be noted that cultures of Type II cells have been particularly difficult to maintain
due to the tendency of these cells to undergo terminal differentiation (to Type I cells) once they
are removed from their natural environment in the epithelial lining of lung alveoli (see Leikoff
and Driscoll 1993, as described in Section 4.4).

Kravchenko et al. (1998) indicates that, unlike cultures of lung epithelial cells, in vitro cultures
of rat mesothelial cells tend to transform spontaneously to tumorigenic cells. The major changes
that occur with time include: altered response to epithelial growth factor (from growth-
proliferation inhibition by this factor to growth-proliferation stimulation), morphological
changes (from polygonal epithelial-like cells to elongated fibroblast-Hke cells and, eventually,
polynucleated cells exhibiting broad polymorphism), and multi-layered growth and an ability to
grow as colonies and masses in semi-solid agar. Eventually, these cells become immortal and
induce cancer when harvested and injected into other rats. The authors indicate that asbestos-
induced transformations in such cultures proceed through identical stages, but that they occur
much more rapidly. For example, incorporation of the stimulating response to EGF occurs
spontaneously at passages 9 or 10, but at passages 6 or 7  when induced by asbestos. Asbestos
was applied at 5 jig/cm2, which was noted to be sub-lethal (95% cell survival was noted at this
rate of application).
                                           6.66

-------





r Chemical Transmitters
«
<»ri
O
•9
a
es
2
e

o
*
U
VI
M
.2
ss
^

o
W5

g
0
C/3

R
A
H




0)
CJ
Refereni


i



a
B
2J
*S
_2j
BC








JJ
3
o
E/}




1
5


'3
1
15
u
i
V
d


































g
gnal Transmitte
t/5
i1
S .S

§ oo ^
S OS "e
EM
III!


A transcription factor








_

















0.
<
M
Activator Proteii

1

•a
t3
1
1
T3
S
Binds the DNA promoter
region of genes governing
inflammation, proliferatin, i
apoptosis
Induces apoptosis

r-
o\

•—*
•2
^i
u
.§
y
J


f^
*£.
•C O
Qfi r^
a s
•« o

?1
o w
•a g
d o
« U
-
<;
a



r-

Lechner et al.


Inhibits apoptosis


























O
•«
Fifth Componen
of Compliment

o\

"S
IM


"o
1
IA
1

























S
Q
O
4-*
c
Cyclin Depende
Kinases

r-
o,

Lechner et al.

V
Inhibits advance of cell cycl

























«
S
u
"S sfl
Cyclin Depende
Kinase Inhibitor

-5

Kravchenko e
1998


Inhibits proliferation of
mesothelial cells





v



















tj
& O
U tfl
•3
Epithelial Grow
Factor


"73
Goodglick et j
-1997.


Stimulates proliferation of
— mesothelioma tumorcells
i

1


i
,
'



j

i
!
I


















Mossman et a
1997


Mediates MAPK signaling

























fT l
O
PQ
J5
o -2
O g
uu c2

(X)

-------




S
'I
c?
ical Transmitters (coi
§
JS
(J
t_
JS
6
•e
B
05

m4


ON
ON

"3
"G
Q.
Released from asbestos

[X,
1 1 W
o o 2
^





c
"^
epithelial cell detachmer
culture







activated PMN





ON
ON
13
t>
ex



With asbestos, increases
















.52
'S
—*
pulmonary epithelial eel












1
T&
S
Nario and H






r-


2
(4-1 T3
0 S
c 3
0 Ji
'•5 <*-
S <=>
.£?«
S § £
S S'51
|^
|1.l
HH — Vt •

1
1
3
•O
3
0
'i
Released by lung

P _
0* <

*

S;
2-
13
ts
£ 1
2 u
epithelial and
endothelial cells
following exposure to
silica but not TiO2
Enhanced expression by
RPM cells exposed to
chrys or crc




>


Luster and



Is a pleiotropic cytokine




T3
OS
en
.S
on
1
u
1
t—t


— VO

OO

ON
Simeonova





oo
ON
ON

5
o
I
1
•1
'o.
(U




oo
ON
ON
mm*
Luster and
Simeonova
S
«
2
Is a neutrophil chemoatt










°?
H^

"7^ O Q, *
S r? O
III 1
g s 1 1
elastase


Intercellular
C ^
o *i3
'83
11
Interleukin-1
Interleukin-6

OO
J£
J3
oo

-------






s
•5
5
•>i
i
I
a
13
u
1
J3
u

Itf
V
6
j——
a
t/j
1
o
U
CA
es of Variou!
V
S
O
CO
1
H



0
£

3


Q.
"e!
1
Jg
u

















r-

O\
. .
j=
0
— •>
e
!'i
pressed by acti
icrophages, bul
mates
i^ TO "j"
n3 E &


O
S
1
u
Inducible Nitri
Oxid Synthase

^
2
G
1
S
•a
a
o
w
=
l!
c -S
3 T3
si
.g s
I'S





2
c
o
VI
1
'^
*2
1
£
1
T3
U
U
Q


S


Keritonizing
Growth Factor


^r1  OS
•a  —
                         c  „
                         o  tJ


                        'S  §,
                         4)  01)


                         I"

                         U  •"  C  81
                         u  -o  a  2
                        .>  U  M  3




                        PI1
                        Q5  rt  c«
Oc

I
                                                                                 03

                                                                                 (O
                                en
ex
                 3
O
                                              2

                                              a
                                              u.

-------






f
3
•1
S
«
^
t
O
-*M
"**
••M
S
_
o
Cfl
a
VO
^
H





V
w
§

aj •-•
-| !
If 8
1 o -i
w is «

S
ON


!«
I
CQ

M
%
1
S
£
"8
I
13




'I
(X

T3
U
>
1
_W
"u
33
s

§^

PQ
-

1
c
D
  §

                    .2
                      l
s^
                        o\
                        ON
                        rt
                        s
                        DQ
                          
-------








a
.5
ical Transmitters (cont
§
JO
U

U
pi
•m
O
•e
a
es
_c
urces of Various Cytoh
o
03
^
V
S
H





,








c
<2
as



w



-V
u
S

b,
*S
p^


Sources

j;
|
<




i
0 OS

1^
Induces JNK arm of MAPI
cascade

r-
Os
Os
*""*
^^
«

^2
O
o
00
Q
Some particles induce
activated macrophages
to release this (shown
both in vitro and in
vivo)


bl "*
•— 13


in
'5!
0 rt
S -S
a £•
2 <
fe >-
o 5
3 «a
H u.

r~
Os
Os
^^
13

"g
U
cn
g
^
c
u
Is involved in the recruitm

r-
os
Os
'""'
•M
a

--
O
!S
O
Regulated by oxidant-


























of inflammatory cells












sensitive transcription of
NF-KB
















t-
Os
Os
*""'
13

"o
tn
g


Stimulates macrphages,
epithelial cells,endothelial
cells, flbroblasts to secrete


C"">
Os
OS
•*
,
13
w
60
1
N
Macrophages in BAL
from asbestosis and
idiopathic interstitial


























chemokines and adhesion
molecules












flbrosis patients release
TNFalpha


















r-
o\
—
13
r^
M
'C

w>
Can induce apoptosis amoi
neutrophils





























OS
OS
OS

•
~3
a3
g
"op
'(3
.s
e
Can induce ROS productio
leukocytes






























r*s
Os
Os
^-
~f»
00
1
N


Intradermal inj stimulates
local accumulation of
flbroblasts and collagen






























rri
OS
Os

13
00
1
N

S3
In vitro, stimulates fibrobh
DNA synthesis and
proliferation





























'
^t
Os1
Os

13'
^
1:
CQ:
= '"

Is a mitogen for epithelial j


!>•'
osi
ON'
^^'

13J
'Jjj
^.
li

Expressed in BA 	
epithelium exposed to
asbestos


li- rt
H "3



i.,,
oo' o
S't3
i
<2<-s
"3 ' g ^
2 'S £•
E— 'O 
-------






s-
1
•s
I
^
nicai Transmitters
S
d
1_
U
ja
•*•»
O
•9
a
es
3
_B
irces of Various Cytofc
o
CC
vg
T
^O
V
1
TO






,
1
CO
Expressed in BA
epithelium exposed to
asbestos

03
S
PL,
O
H


Ofi 0
•s 1
^ rt
i ^
n r^

6 r^
is
6-s'

1
•s a>
Does not appear to i
chemotaxis of rat lui
r~-
Os
Os

"ea
€
1
*«*»4
1*
All three isoforms
expressed in the fibrotic























fibroblasts







lesions (by hyperplatic
type II cells) of
asbestosis and pleural
flbrosis.patients from
Quebec













+j
4J
3
*
S
>
O m
'e °^
6 os
O — '
013

1 «
, u
't- +3
0 X
Induces chemotax'is
mononuclear leukoc
r-
OS
OS

•a
13
l-i
1
}
Mesothelial tumor cells
expressed only TGF-
betal and 2















             .
            "^  VI
            0>  EA

            8 a J2

            8il
            .S SL
               s
            •a 3 ••"
            .2 = x
            8 8-f

            J'S'S
            < a, c5
    > gj
     i Os
     O ON
      -a
    o t>
                        CM
                        I*-.

                        CD

e
**
'.n
a
|&

-------





5
a
.8
Transmitters (
n
B
V
S
t_
«
O
•o
d
i
i
o

U
3
*i
•5
6B
a
£
ca
so
SO
Jj
H









u
1
<£
£



1
td




t>
w
i
cS






s
Q
^^


•
J>
pfi
*

C«
• M
W
CA
1
V

U
r-
^^
•a
t>
•8
1

O.
»
tns a heterodimer
alters behavior
C T3
G C
h 
1
U

*>.
J2
.1 •£
v
!£•€
_H
S
$ . 	 1
> £
3

— *O

d
e3 O
11








































S
E
M
U




























Q
O
°?
i


'H
lla

•Sol
c & «
'S M 5





































u
s
•a
£
e
V
O
oo t- oo t-- 'i !>
22 22 i§
. « TUJ » » T3 ' ^^
^"2 ^s "H"2 ^ s '' *
Qi O G 0^  1 1 1 |: ||
'»







'i











fO t>
y. S S fc-
<2 .2, •§ S
i ? Si




!.



















CO

CD
























-------







at
l_
«
e
"g-
ta
a
2
H
"«
o
1

U
0)
€
O
•o
§
1
2
o
*
U
09
A
 ^J
""^











s
J
c
o
'E
'5
ca
1
<



—
PLH
 O -o J3 S
5 S> «• £ & 2



X *V
^"0*0
CQ m u

^^
.1
"EL

c
(j
«,_
1
|
<§








ON
•a
s
1

i












1
1
Mediates


M
^

U



CO

*2
c
'^
1
T3
Q
_c
"o
^








r-
ON
T3
£

1











f
u
"8
<*-
0
8
1-
1



£ O
o
U 0

1
1 2
>~ Q.

frt ^
C 3
2 z
III
1 11
>% >. 42
U U U






oo

- »
1 5
o ^5
u ^
1 1

t\s n



("C
i
3
W B3
"^ C
o o
(4 24
12 *^
5 &
^ c
.2 1
1 S
jS 1

1 1 s



b
O
U





(J
o
[2
|
1
&







c-«
ON
issman et a
t^/
'"CJ











I
"«
^
0,
2
1



^
o
u

1
u
0^
Ui
O
t»
c2
|
1
Q.







^
ON
issman et a

'"CjJ












I
§•
^3
V

UH

J^j V
2 ^
w o

X
•g
w

ofi
2 2
™ cL,
a i 2
"^ Kl u
S 6 »
s* fl
erf "3 -5
&L, g B
O 2 -a








fn
t>
ex

. JJ
*




"w
^
KJ
1
"5,

o
1
«
o>
i
_c
>v
CQ


oo
£ w
2 z
o S



4)

cj
S
"«>
p.
i


fs.
(O

-------
i)







a
.* .
« •

§ .
£
nsmittei
L
H
"3
.a
ytokines and Other Chem
> *
U
09
a ,
_g
A
O
i
«B .
U

>c
2
_w
2
es .
H





v


fitf











S
i
u


*
u
i
A
<







$
I
eft
Chemical





1
.
•a
n.



^5
"«
1
S.
>•.
«-
0
With asbestos, increases pulm
lysis


t*)
(»«.
cu











^
OS
Os
"2
£•
js
3C
'!"
ra
.2
'C

z
2
,1
3
' O

1
S
>.
Facilitates migration of leukoc
sites of injury



t
1 3

^^
•I
u
J2
"o
c
_o
cd
Intercellular
Interleukin-1
00
Os
*•*
2
c
.1
CO
1
&
tjn

J






Is a pleiotropic cytokine




3









Interleukin-6
00
OS
Os
*-"*
>.
g
1
Cfl
1
5j
Tn

J





•^
Is, a neutrophil chemoattractah




2 I


M

•S
J>
^
1
LH
Interleukin-8
Inducible Nil





OS
Os
c



<

!3
"3
"o
CO
u
E
<*j
o
Induces transient proliferation
cejls




|




o
1
tt(
S
ViJ
Keritonizing



r~
ON
5
^f
i
Wl
w>
O
S






<4-l
O
g
•c
CO
1


(

*4
eu
<

S
,s

t;

£
13
is
^
c
V
00
3
SJ



r-

•3
^,
i
w
fft
O
S






I
C3
I
U
[A
1










-,



























|i
,2 ca




-^
_o
c
!S
^5 A
Macrophage
Protein- 1 alp!





r-
os
•—
*3.
u
Q

8

u
XI
>, •
^ C
g -r §
_§ c
O 'CO

PO ' ; S

„ 1 T
u C j'
60 S,
H § '
~ >\ '•
A transcription factor that regi
involved with the inflammator
adhesion; and growth 'control.
Promoted by AP-1



.,
1 1 I 1





! ; o
i 
-------
n
                                 V
                                 s
                                 t
                                 2
                                 1
                         •I

                        JS
                        U
£
w
                         O



                        I"
                        U
                         
-------
In a study in which p53 deficient mice were intrapleurally injected with UICC crocidolite (200 jig/week),
Marsella et al. (1997) showed that p53 deficient mice exhibited substantially increased susceptibility to
development of mesothelioma. In this study, 12.5% of homozygous mice (p53 deficient) developed
mesothelioma and the rest died of lymphomas or hemangiosarcomas that develop spontaneously in these
mice; 76% of heterozygous mice died of mesothelioma,  and only 32% of wild-type mice (p53 competent)
died of mesothelioma. The authors suggest that p53 deficient mice are susceptible to excess proliferation
induced by crocidolite due to loss of control at the Gl/S check point that is normally mediated by p53.

In further confirmation of this hypothesis, Marsella et a!. (1997) report in the same study that 7.5 jig/cm2
of crocidolite applied to wild murine mesothelial cells in culture induced substantial apoptosis while
p53-deficient cells were resistant to apoptosis. The authors also note that most of the p53-deficient cells
are tetraploid (suggesting a loss of a spindle check point) while the wild type cells are all dipjoid.

Note that, although these studies  suggest that asbestos alone can induce complete transformation of both
lung epithelial cells and mesothelial cells, such studies must be evaluated with caution.  In the case of the
Hei et al. (1997) study, for example, the effects of the (asbestos-independent) mutations required to
initially establish the immortal line of lung epithelial cells are not entirely known. Therefore; the
response to asbestos of epithelial cells in vivo may be substantially different.               :

In the case of the Kravchenko et al. (1998) study, it is clear that mesothelial cells in vitro do not behave in
the same manner as those in vivo; in vitro, they spontaneously transform to tumorigenic cells; This
suggests that one or more growth inhibitory signals exist in vivo (which are absent in vitro) that are
critical to maintaining the health  of the mesothelium

6.3.3   Evidence that Asbestos Acts As a Cancer  Initiator

As indicated in a review by Jaurand  (1997), historically, the status of asbestos as a mutagen was
questioned, due primarily to the failure to produce detectable gene mutations in short-term assays.
However, more recent studies provide clear evidence that asbestos (and other fibrous materials) can
produce mutations. Moreover, fibrous materials may induce mutations by. multiple mechanisms
including, for example:

        •      direct interference with mitosis;

        •      production and release of reactive oxygen species (ROS); or

        •      production and release of reactive nitrogen species (RNS).

Of these, the most consistent, positive evidence that asbestos can act as a cancer initiator (i.el, a genotoxin
or mutagen) has been from assays designed to detect the kinds of genetic damage that result from
interference with mitosis (Jaurand 1997).  Although generation of ROS and RNS plays a cle|r role in
mediating asbestos-induced disease, the direct link to a role of asbestos as an initiator  is somewhat more
tenuous (see Sections 6.3.3.2 and 6.3.3.3).
6.3,3.1
Interference with mitosis
Apparently, the physical presence of asbestos fibers can interfere with proper spindle formation and the
function of other structural units required for mitosis (Jaurand 1997). This typically leads to, aneuploidy
(an incorrect number of copies of the chromosomes contained within a cell), development of micronuclei
(fragments of chromosomes enclosed in a membrane that are isolated from the main nucleus'of the cell),
and has also been shown to lead to clastogenic effects (changes in the organization and structure of the
                                              6.79

-------
chromosomes). Assays for these kinds of genetic alterations have consistently shown asbestos capable of
inducing these effects.

Jaurand (1997) also indicates that:

        •      fibers must be phagocytized by the target cells before they can interfere with mitosis;

        •      once phagocytized, all asbestos types are observed to interfere with mitotic activity; and

        •      samples enriched in long, thin fibers enhance these effects. In contrast, short fibers do
               not appear to contribute to these effects.

Jaurand (1997) notes, however, that results related to fiber size have not been entirely consistent,
primarily because not all studies have rigorously controlled for or even adequately characterized the sizes
of the fibrous structures in test materials. Jaurand (1997) also notes that this mechanism is not dependent
on the formation of reactive oxygen species or any other reactive free radicals.

Among studies that suggest that surface chemistry (or presumably fiber type) may not be an important
factor in determining the degree to which asbestos (or other fibrous materials) interfere with mitosis,
Keane et al. (1999), exposed cultured V79 cells (hamster lung fibroblasts) to untreated and HCl-treated
chrysotile asbestos. The acid treatment substantially reduces the magnesium content on the surface of the
chrysotile. The authors also  noted a "small" effect on fiber size due to treatment (treated fibers showed a
20% excess of short fibers).  Cells were exposed to doses ranging between 0.4 and 12.7 yg/cm2 (each
dose left in place for 24 hours and then rinsed off). Cells were harvested after an additional 24 hours and
evaluated for cytotoxicity and the presence of micronuclei.

Results from the Keane et al. (1999) study indicate that untreated fibers were shown to be slightly more
cytotoxic than treated fibers, but both treated and untreated fibers were observed to increase the
abundance of micronuclei in a similar, dose-dependent fashion. The induction of micronuclei appeared to
saturate at approximately 35/1,000 cells observed at an applied asbestos concentration of 40 ng/cm2. In
contrast, substantial cytotoxicity was only observed at the highest doses employed. The authors thus
concluded that surface chemistry (at least in terms of magnesium content) does not appear to have a major
affect on induction of micronuclei and that the observed genetic damage and cytotoxicity appear to occur
through entirely independent mechanisms.

In other studies, cultured cells were assayed for a variety of genotoxic effects following exposure to a
range of fibrous materials.

For example, Dopp and Shiffmann (1998) dosed human amniotic fluid (HAF) cells or Syrian Hamster
Embryo (SHE) cells with UICC amosite, Rhodesian chrysotile, crocidolite, or ceramic fibers at
concentrations of 0.5, 1.0, 5.0, and  10.0 ug/cm2 and assayed the cultures for formation of micronuclei and
a variety of clastogenic effects.

Based on their study, Dopp and Shiffmann (1998) report that all asbestos types induced formation of
micronuclei in SHE cells in a dose-dependent fashion at rates that were significantly elevated over control
animals. The effect appeared to saturate at doses between 1  and 5 ug/cm2 and rates appeared to peak at
between 48 and 66 hours post-exposure. Ceramic fibers, which were noted to be longer but thicker than
the asbestos fibers tested, also showed significantly increased induction of micronuclei, but at rates less
pronounced than for asbestos. However, it is difficult to judge whether this is due to differences in fiber
type because the data in the paper are not adequate to distinguish effects of fiber size and number from
effects of fiber type. Similar results were obtained with HAF cells, but overall rates were about one third
those observed in SHE cells.

                                               6.80

-------
                                                                                      l
The authors also note observing various disturbance of chromatin structure during interphase.' They
report observing formation of chromatin bridges and chromosome displacements in meta and anaphases
and impaired chromatin separation in mitosis. They also report that cytokinesis was frequently blocked.

In addition to the effects that they observed that are attributable to fiber interference with mitosis, Dopp
and Shiffmann report observing a variety of clastogenic effects.  In all cases where the authors labeled
specific regions  of specific chromosomes, fiber-exposed cells showed significantly greater frequencies of
DNA breaks over controls or gypsum-exposed animals. Different regions of various chromosomes also
showed significantly different frequencies of breakage with patterns that were specific to the different
fiber types.  The authors hypothesize that the observed clastogenic effects may be due to production of
reactive oxygen  species, to formation of some type of clastogenic factor, or to the direct interactions
between fibers and chromosomes during mitosis that are the apparent cause of the disturbances discussed
in the previous paragraphs. However, It was not possible to distinguish among hypothetical causes of the
observed clastogenic effects in this study.                                               •'

Kodamaet al. (1993)  exposed cultures of human bronchiolar epithelial (HBE) cells to asbestos (chrysotile
at 0 to 4 |ig/cm2  and crocidolite at 0 to 300 ug/cm2).  They then examined cells at 24, 48,72, and 96 hours
following exposure for cytotoxic effects and cytogenetic effects. Results indicate that both fiber types
induced concentration-dependent inhibition of cell proliferation and colony-forming ability, but chrysotile
was 100 to 300 times  more toxic. The authors report this translates to a 40-fold increase in toxicity on a
fiber for fiber basis (although the range of sizes  included in this count is not indicated).     ':

Kodama et al. (1993)  also report that, at 72 hours, chrysotile (4 ug/cm2) caused a 2.7-fold increase in
binuclei and a 1.6-fold increase in micronuclei.  Over the same time interval, at 300 ug/cm2, crocidolite
caused a 1.9-fold increase in binuclei, but did not cause micronuclei. They also report that chrysotile
failed to produce significant numerical chromosome changes in HBE cells and increased structural
aberrations only at the 24-hour time point. The  frequency of neither aneuploidy nor polyploidy was
increased at any time  point following exposure to asbestos in this study. The authors indicate that this
contrasts with observations of relatively high incidences of asbestos-induced chromosome changes
observed in some rodent cell cultures and clastogenic effects observed in human mesothelial 'cell lines.
They further speculate that phagocytic cells with high mitogenetic activity are likely most susceptible to
the effects of asbestos, which primarily interferes with mitosis. However, they suggest that epithelial
cells that are exposed  to fibers may undergo terminal differentiation (from Type II to Type I cells) and
thus cease mitosis. This would  effectively prevent adverse genetic effects from asbestos exposure.  Such
pathways are not available to mesothelial cells.                                          !.

Hart et al. (1994) studied the effects of a range of fiber types (with varying size distributions) on Chinese
hamster ovary (CHO) cells. The authors indicate that such cells are  very different from cells in
pulmonary tissues in that they are immortalized, aneuploid, undifferentiated, and preneoplasic.  They also
note that the responses observed in these cells differs from responses observed in cells from pulmonary
tissues.  Nevertheless, the implications concerning fiber size are instructive. Fibers evaluated included
long, medium, short, and UICC crocidolite and chrysotile along with a range of MMVF's, RCF's, and
fibrous glasses.  Exposure concentrations ranged between 10 and 225 ug/cm2.

Results from the Hart  et al.  (1994) study indicate that all of the fibers caused qualitatively similar toxic
effects: concentration-dependent reduction in cell numbers and an increase in the incidence of abnormal
nuclei with little or no loss in viability.  Fiber-induced cell death in CHO cells appears to be minor, even
at relatively high exposure concentrations. Based on mean dimensions (which is problematic), the
diameter dependence on the observed toxic effects, particularly on the formation of aberrant nuclei, was
slight or absent.  However, the effect with length was striking. For lengths up to at least 20 |im, potency
toward both cytotoxicity and the induction of aberrant nuclei increased dramatically  with increasing
length. The authors also note that the lack of an observed fiber composition associated effect on the

                                              6.81

-------
toxicity of CHO cells does not correlate with findings from recent rodent inhalation studies using the
same test fibers. The authors therefore speculate that CHO cells may not represent an appropriate in vitro
model for fiber effects. However, it may also be that effects in vitro occur over time scales that are rapid
relative to those that occur in vivo so that biodurability is not important in vitro.

In a study by Takeuchi et al. (1999) cultured human mesothelioma cells (MSTO211H) and, separately,
cultured human promyelocytic leukemia cells (HL60), which are not phagocytic, were dosed with
crocidolite between 0.6 and 6.6 ug/cm2 (no size range indicated). Studies with latex beads confirmed that
the mesothelioma cells are actively phagocytic, but that the leukemia cells are not. The authors indicate
that dosed mesothelioma cells showed significantly increased numbers of polynucleated cells, tetraploid
cells, and  cells with variable DNA content at the GO/G1  transition in the cell cycle and that the extent of
effects was dose-dependent. Leukemia cells showed no  such effects.

The authors further indicate that, when the mesothelioma cells were sorted by fiber content, those with the
highest fiber content showed the greatest effects.  The authors also indicate that cells are stimulated
neither to  release superoxide nor NO at the concentrations of crocidolite studied.  However, they
hypothesize that intracellular ROS may have been generated because they report finding increased levels
of 8-OH-dGua (and oxidized form of one of the DNA bases) following crocidolite exposure.
Nevertheless, the authors conclude that the mechanisms  by which crocidolite induce cytotoxicity and,
potentially, carcinogenicity is related to phagocytosis. Importantly, the effects described in the Takeuchi
et al. study are entirely consistent with effects attributable to interference with mitosis, despite any
speculation by the authors.

6.3.3,2        Generation of reactive oxygen species (ROS)

ROS have been implicated as mediators in a variety of toxic effects (including cancer initiation)
associated with a broad range of toxins (see, for example, Floyd 1990). Moreover, substantial evidence
indicates that asbestos can induce generation of ROS by several mechanisms  and that asbestos-induced
ROS play a role in several of the toxic effects attributable to asbestos (see below). However, whether
ROS play an important role in asbestos-associated cancer initiation is less clear and needs to be evaluated
carefully. Therefore, evidence that asbestos induces the production of ROS and that ROS contribute to
the adverse health affects attributable to asbestos are reviewed below with particular attention to effects
that may contribute to the initiation of cancer.  Contributions by ROS to other asbestos-related toxic
effects are also evaluated in later sections of this chapter.

Note, that generation of certain reactive nitrogen species (specifically the peroxinitrite ion) is closely
associated with ROS generation so that evidence for the generation of certain reactive nitrogen species
(Section 6.3.3.3) constitutes additional evidence for the generation of ROS.

Asbestos-induced Generation of ROS. Asbestos has been shown capable of generating a variety of
reactive oxygen species (ROS) including hydrogen peroxide (H2O2), superoxide (O2~), and hydroxyl
radical (OH«) via several mechanisms (see, for example, Fubini 1997; Jaurand 1997; Kamp et al. 1992)
including:

        •      catalytic production of superoxide from oxygen in aqueous solution;

        •      catalytic production hydrogen peroxide from oxygen in aqueous solution;

        •      catalytic production of hydroxyl radical by the Fenton reaction (degradation of hydrogen
               peroxide catalyzed by iron on the surface of fibers  or mobilized  from the surface of
               fibers);


                                              6.82

-------
        •      catalytic production of several ROS by redox cycling of iron on the surface of fibers or
               mobilized from the surface of fibers (Haber- Weiss type reactions);         '

        •      catalytic production of ROS by release of heme and heme protein from various cellular
               components;
                                                                                    \t
        •      by inducing release of various ROS species from phagocytes during "frustrated"
               phagocytosis; and
                                                                                    i
        •      by binding to cell receptors or other features of surface membranes that trigger signaling
               cascades mediating the production and release of various ROS (and RNS).  •
                                                                                    i
For a general review of the chemistry involved in these processes, see Floyd (1990).        j;

Fenton, Haber-Weiss, and Related Reactions. Most of the evidence for Fenton and Haberj Weiss
reactions (and related free  radical generating reactions) that take place on the surface of asbestos fibers
comes from experiments in cell-free systems (see below). Therefore, their relevance to the conditions
found in vivo may be limited. Moreover, the size of the fibers (especially in terms of their cumulative
surface area) and the history of their surfaces (in terms of metal contaminants or coatings that might be
present) may substantially alter the effects of such experiments (Fubini 1997).  Unfortunately, however,
few of the available studies report characterization of fiber sizes or surface conditions in sufficient detail
to judge the importance of these effects.

For fibers. Zalma et al. (1987) evaluated a range of fibers (UICC crocidolite, UICC amosite,1 UICC
Canadian and Zimbobwean chrysotile, industrial chrysotile, and magnetite) for their ability to produce
free radicals by the direct reduction of oxygen in aqueous solution.  In some cases, hydrogen:peroxide
was also added to the solution. Results indicate that all of the fibers tested were able to generate hydroxyl
radicals (even in the absence of hydrogen peroxide), but that the efficiency of production was a strong
function of the activation (by grinding) or pacification (by coating with benzene or other agents) of the
fiber surface. Chrysotile was reported to be the most efficient at generating radicals and the authors
assumed that this is due to iron contamination on the surface (since  iron is not a component of the "ideal"
chrysotile fiber).  However, such conclusions are difficult to evaluate in the absence of simultaneous
consideration of fiber size.

Governa et al. (1998) evaluated the ability of wollastonite fibers to generate  ROS in both a cell free
system and a suspension of polymorphonucleocytes (PMN's). The  fibers were observed to produce ROS
in both systems and that ground wollastonite produced substantially more ROS  than unground. The
efficiency of ROS generation in PMN suspension is also reported to be greater for wollastonite than for
either chrysotile or crocidolite (tested in previous studies). However, no size information is given. Based
on additional work with various inhibitors added to the system, the authors also indicate thationly a
fraction of the ROS generated was composed of hydroxyl radicals.                        '
                                               >
Brown et al. (1998) subjected several different fiber types (amosite, silicon carbide whiskers,'RCF-1, and
various fibrous glasses) to  two standard chemical assays for free-radical production (in cell-free systems).
The authors indicate that, of the fiber types tested, only amosite showed free radical activity significantly
above controls in both assays and only RCF-1 additionally showed significantly elevated free radical
activity in one assay. However, there is not enough information provided in this study to determine
whether the observed differences are due to differences in fiber sizes, sample preparation (i.e., surface
condition), or fiber type. In apparent contrast, for example, Gold et al. (1997) report that amosite and
crocidolite produce few free radicals in cell free systems, unless they are ground.
                                              6.83

-------
Weitzman and Graceffa (1984) indicate that chrysotile, crocidolite, and amosite are all capable of
catalyzing the generation of hydroxyl radicals and superoxide from hydrogen peroxide in vitro and that,
based on experiments with various iron chelators, these reactions are iron dependent. The authors further
indicate that hydrogen peroxide is produced in large quantities as a normal bi-product of tissue
metabolism, but that it is effectively scavenged by various enzymes. The authors speculate that, by
physically damaging cell membranes, asbestos may allow release of the precursor hydrogen peroxide
before it can be scavenged.

For particles. Silica, residual fly ash, and ambient air dusts also can create ROS in vitro and the
efficiency of production correlates with ionizable concentrations of various transition metals complexed
on the dust (Martin et al. 1997).  In vivo, such metal containing particles also cause release of ROS from
macrophages (Martin et al. 1997).  Additionally, binding of silica to plasma membranes  of airway lung
cells and phagosomes provokes generation of ROS.

Castronova et al. (1997) further indicate that it is the concentration of contaminating  iron on the surface
of freshly fractured quartz that enhances free radical production in aqueous solution (in cell free systems).
These authors also showed that high iron-containing (430 ppm) quartz dust inhaled by rats (at 20 mg/m3
for 5 hours/day for 10 days) induced 5 times more leukocyte recruitment, 2 times more lavageable red
blood cells, 30-90% increase in macrophage production of ROS, 71% increase in nitric oxide production
by macrophages, and 38% increase in lipid peroxidation of lung tissue than observed in rats exposed to
low iron-containing (56 ppm) quartz.

Although iron is required for the reactions considered here, studies indicate that the iron content of the
fiber itself is not a good indicator of reactivity (Gold et al. 1997).  Studies also indicate that the iron that
participates in these reactions need not originate with the fiber (Fubini 1997; Jaurand 1997) and
biological systems contain abundant sources of iron. Therefore, both iron-containing fibers and iron-free
fibers have been shown to participate in these reactions in vivo.

Release of Items and heme protein. At least one research group (Rahman et al.  1997) indicates that heme
and heme protein cause extensive DNA damage in the presence of asbestos in vitro and, based on
previous studies, that this may involve heme catalyzed production of ROS following  asbestos-induced •
release of heme from cytochrome P-450, from prbstaglandin H synthetase (or perhaps from other heme
containing proteins).  Importantly, the authors indicate that such observations relate to a nuclear pool of
heme, which suggests that ROS generation via this mechanism may occur in the immediate vicinity of
DNA.  The work by this group suggests at least one additional pathway by which asbestos may induce the
production of ROS and by which ROS-mediated damage to DNA might occur.

Frustrated phagocytosis. Numerous studies indicate that long asbestos fibers (longer than somewhere
between 10 and 20 urn) cannot be efficiently phagoycitized by macrophages (see, for example, Sections
4.4 and 6.2) and that macrophages that are damaged by such "frustrated" phagocytosis release ROS (see,
for example, Kamp et al. 1992; Mossman and Marsh 1991). Due to the differences in the size of
macrophages across species (see discussion of Krombach et al. [1997] in Section 4.4). The minimum
length beyond which phagocytosis may become frustrated may differ in different animals. However, it is
clearly longer fibers that contribute to this mechanism for generating ROS. Shorter fibers (<10 (im) are
unlikely to cause frustrated phagocytosis in any of the mammalian species of potential interest.

Lira et al. (1997) showed that alveolar macorphages in culture (after stimulation with
lippopolysaccharide) generated free radicals (ROS) when subsequently exposed to chrysotile, crocidolite,
or amosite (all UICC). They found chrysotile to be the most potent inducer of free radical activity (which
is not surprising given that UICC chrysotile contains the highest fraction of long fibers of the UICC
samples tested (Berman, unpublished).  Based on tests with various inhibitors, the authors indicated that


                                              6.84

-------
the free radicals generated by the alveolar macrophages occurred through a pathway mediated by protein
tyrosine kinase, phosphoISpase C, and protein kinase C and that the effects are dose-related. '

Kostyuk and Potapovich (1998) cultured peritoneal macrophages and showed that treatment with
chrysotile asbestos (1 jj-g, no size data given) resulted in production of frustrated phagocytosis and cell
injury (the latter as evidenced by release of lipid dehydrogenase, LDH, a marker for membrane damage).
By working with various chelators and flavonoids (natural plant products, some of which quench
superoxide and some of which chelate iron), the authors indicate that cell injury was likely induced by
superoxide and that the superoxide was likely produced by an iron-dependent mechanism.  Note that this
contrasts with the above studies that suggests production of radicals by frustrated phagocytosis in culture
is an iron-independent mechanism.

At least for one kind of phagocyte: polymorphonucleocytes (PMN's), a study by Ishizaki et al. (1997)
suggests that crocidolite and erionite may induce production of ROS from PMN's by each of two
mechanisms. The first requires phagocytosis and may represent the traditional, "frustrated" phagocytosis
pathway indicated above.  The second pathway is triggered by an interaction between the fiber and the
cell surface and is mediated by NADPH. The authors also cite evidence that chrysotile may similarly act
through both of these pathways.

Afaq et al. (1998) cultured alveolar macrophages and peripheral red blood cells (RBC's) that were
harvested from rats 30-days following a single 5 mg intratracheal instillation of UICC crocidolite, U1CC
chrysotile, or ultrafine titanium dioxide. The authors indicate that  populations of alveolar macrophages'
were significantly increased (over sham-exposed rats) for all three  particle types and that acid
phosphatase and lipid dehydrogenase (LDH), which are markers of cell membrane damage, were
observed in cell-free lung lavage from animals exposed to all three particle types.  Both alveolar
macrophages taken from asbestos-exposed animals (both types) showed significantly elevated lipid
peroxidation and hydrogen peroxide production over titanium dioxide exposed animals.  However, the
latter also showed elevated peroxidation and peroxide production that were significantly elevated over
sham-exposed animals. Similar results were observed among RBC's from asbestos-exposed|animals, but
not from titanium dioxide-exposed animals. Note, it is possible that ROS production induced by Ti02
occurs by a different mechanism (or set of mechanisms) than that for asbestos (see, for example, Palekar
etal.  1979, Section 6.3.4.4).                                                           ;
                                                                                    !'
In vivo Evidence for Asbestos-generated ROS. Several studies involving whole animals  also indicate
that asbestos exposure induces the generation of ROS. Importantly, in  such studies, evidence for
generation of ROS is generally determined based on observation of the putative effects of ROS, rather
than ROS directly.                                                                  !

Ohio et al. (1998) intratracheally instilled 500 jig of crocidolite (NIEHS) into rats.  This was/observed to
induce a neutrophilic inflammatory response within 24 hours (in contrast to saline-exposed rats). The
authors collected chloroform extracts from exposed lungs and subjected them to electron spin residence
(ESR) spectrometry.  Results indicate the presence of a carbon-centered radical adduct that Has a structure
consistent with products of lipid-peroxidation. The radical signal was only observed in asbestos-exposed
animals and persisted even after one-month following exposure. The authors also report that depletion of
neutrophils did not affect the signal and that dextrin-induced inflammation did not produce the signal.
                         •*
Yamaguchi et al. (1999) studied effects in rat lung tissue at 1,3, 5,7, and 9 days following a^ingle
intratracheal instillation of 2mg of either glass fiber or UICC crocidolite. The authors indicate
significantly increased levels of 8-OH-Guanine (an oxidized form  of the DMA base Guanine) one day
after crocidolite instillation and increasing repair activity for this oxidized form of guanine with time that
became significant at days 7 and 9 following instillation.  Glass fibers (noted to be non-fibrogenic and
                                              6.85

-------
non-tumorigenic) did not produce either increases in 8-OH-Gua or its repair activity. The effects
associated with crocidolite were all noted to be dose related.

Several of these studies also suggest distinctions in ROS generation (either the absolute generation of
ROS or generation of specific ROS species by specific tissues) due to differences in fiber (or particle) size
or type.

Nehls et al. (1997) intratracheally instilled rats either with quartz (2.5 mg) or corundum (2.5 mg).  The
latter mineral is reportedly non-tumorigenic.  Results indicate that lung epithelial cells in quartz exposed
rats exhibited increased 8-oxo-Guanine levels (a DNA adduct generated by reaction with ROS, see above)
that persisted for up to 90 days post-exposure. Elevated levels of the DNA adduct appeared in all  cell
types in all areas of the lung. The authors suggest that the observed persistence of the elevated levels of
8-oxo-Guanine suggests that it was produced at a rate in excess of the lung's capacity for repair. The
authors also report enhanced and persistent inflammation, cell proliferation, and an increase in neutrophil
population in bronchio-alveolar lavage (BAL) fluid and an increase in tumor necrosis factor alpha
(TNF-a) in BAL fluid in the quartz exposed animals.  TNF-a is a cytokine linked to a variety of effects
including the recruitment of inflammatory cells (see Table 6-6).  In contrast, exposure to corundum
produced none of the effects observed with quartz.

Timblin et al. (1998b) report that ROS induced responses by rat lung epithelial cells vary depending on
whether exposure is to crocidolite, hydrogen peroxide, or cadmium chloride (CdCl2). In response  to ROS
generation induced by the first two agents increased levels of cJun protein (a protooncogene, Table 6-6) is
observed.  Further, crocidolite, but not hydrogen peroxide, causes elevation in the levels of manganese-
containing superoxide (MnSOD) dismutase (an enzyme that catalyzes dismutation of superoxide, Table
6-6). Neither of these agents affect levels of either of two common stress proteins (Table 6-6): GRP78 or
HSP72/73, nor do they affect cellular glutathione levels.  In contrast, cadmium chloride does not alter
MnSOD levels, but increases levels of GRP78 and HSP72/73 in addition to cJun protein. Therefore,
ROS-related mechanisms may be complex and may be toxicant-specific. Thus, it may not be correct to
assume that all fibers and particles act through common ROS-related pathways.

In contrast to the results of the above study (which showed no affect on glutathione levels), Golladay et
al. (1997) showed that human lung epithelial cells (cultured A549 cells) exposed to crocidolite (NIEHS
sample) showed substantial reduction in intracellular levels  of glutathione (without increases in  the
oxidized forms of glutathione). Rather, an associated increase in extracellular, reduced glutathione was
observed, suggesting that crocidolite induces release of glutathione from the interior of these cells to the
environment.  The authors also indicate that, given that the half-life for reduced glutathione outside of
cells is on the order of an hour, while extracellular reduced glutathione levels remained elevated for more
than 24 hours following exposure, cells must have been releasing reduced glutathione continuously.
Because no concomitant release of LDH  or labeled adenine  was observed (despite loading of cells with
labeled adenine prior to the experiment), the authors conclude that release of glutathione is not associated
with membrane disruption or apoptosis (which is induced to some degree by exposure to crocidolite).
Also, all of the effects described above were similarly associated with exposure to de-ironized crocidolite.
Thus, the iron content of the fibers does not play a role in this process.

Note that the apparent difference in the reported effect of crocidolite exposure on intracellular glutathione
levels in the above two studies might be due (individually or in combination) to differences in the  cell-
types studied, differences in the size distribution of the crocidolite employed, differences in study  design,
or other factors. Insufficient information is available to distinguish among these possibilities.

Kaiglova et al. (1999) intrapleurally injected rats with 10 mg of long amosite and collected bronchio-
alveolar lavage (BAL) fluid 24 hours following exposure and at later time intervals.  They indicate that
total protein and alkaline phophatase (AP) were both elevated in BAL 24 hours after exposure and that

                                               6.86

-------
AP remained elevated for at least 3 months following exposure.  They also noted increased levels of lipid
peroxides in BAL at 24 hours, but not 3 months following exposure. The authors indicate that
antioxidants were significantly decreased following exposure: glutathione was significantly decreased in
lung tissue at both 24 hours and 3 months following exposure, but was normal in BAL fluid at all time
points; a-tokopherol and retinol were significantly decreased at 3 months in lung tissue; and ascorbic acid
was significantly decreased in both lung tissue and BAL at 24 hours and remained low at 3 months. The
authors indicate that decreases in antioxidants implies a role for ROS (or RNS) in lung tissue injury.  It is
also possible that the varied responses of specific antioxidants may suggest a role for toxin- and injury-
specific ROS/RNS.                                                                  ;

Conclusions Concerning Generation of ROS. Except for frustrated phagocytosis (which is unique to
long fibers), ROS generation by the mechanisms discussed above are considered a common response to
respiration of particles in general (see, for example, Martin et al. 1997 who suggest that ROS "...may be a
global signaling mechanism mediating response to paniculate insult mostly by activation of kinases and
transcription factors common to many response genes." They further indicate that if the load of ROS
generated is too great, or the  airway in which it is generated has been previously impaired, ".Lthese same
mechanisms can result in deleterious respiratory lesions and outright pathology").  However,';not all'non-
fibrous particles are similarly capable of inducing production of ROS. As indicated above, fbr example,
while crystalline silica is a potent inducer of ROS production, carundum is not (Nehls et al. 1997).
Moreover, the spectrum of ROS (set of species) that are induced by particular toxicants are generally
specific to the offending toxicant (Timblin et al. 1998b). Therefore, the generic grouping of ROS
mediated pathways by particles and, especially, by  particles and fibers, does not appear justified. These
mechanisms are more complex and individualized than such generic grouping suggests.    '

ROS can be generated by multiple pathways that are variously dependent on particle size, whether a
particle is a fiber, fiber size, and particle or fiber type (i.e., chemical composition).  The primary
mechanism(s) through which ROS are generated in response to one type of particle or fiber may be very
different than that through which ROS generation is induced by another and the resulting suite of ROS
(set of species) may also differ.  Importantly, the relationship between dose and  response foneach
mechanism may also differ (see, for example, Palekar et al. 1979, Section 6.3.4.4).         !.

Given the above, comparing among the ability of fibrous materials and non-fibrous analogs to induce
generation of ROS requires that such analogs be properly matched before valid conclusions can be drawn.
Thus, for example, the appropriate non-fibrous analog for crocidolite is the massive habit of reibeckite
and the appropriate analog for chrysotile is the massive habit of antigorite or lizardite.  Due to differences
in both chemistry and crystal structure, crystalline silica is not an appropriate non-fibrous analog for any
of the asbestos types. Moreover, because ROS can be generated by different mechanisms, critical
comparison across analogs requires more than the simple confirmation that ROS are generated or even
whether the relative efficiency with which ROS are generated is comparable. It is also necessary to
contrast the specific complexion of ROS (set of species) generated and the specific tissue/cellular
environments (i.e., locations within a cell) in which they are generated in response to each analog.

It is also clear that ROS are generated by both iron-dependent and iron-independent pathways. Even for
the  iron-dependent pathways, however, the source of iron need not derive from the fibers or particles
themselves. Therefore, since iron is abundant in vivo (and the environment), both iron-containing and
iron-free fibers (or particles) can potentially participate in both the iron-independent and the iron-
dependent pathways.
                                                                                   .1
Effects Mediated by ROS. ROS have been implicated as mediators in a variety of toxic effects
(including cancer initiation) associated with a broad range of toxins (see, for example, Floyd 1990;  Martin
et al. 1997). Cellular and tissue effects that have been associated with the effects of ROS include:
                                              6.87

-------
               enhancement of overall uptake of particles by epithelium;
               stimulation of inflammatory responses;
               stimulation of various signaling cascades and production of cytokines;
               inducement of apoptosis;
               cytotoxicity;
               mediation of cell proliferation;
               formation of oxidized macromolecule (including DNA) adducts; and
               induction of DNA strand breaks.

However, only the last two of the above list of ROS effects are potentially relevant to the initiation of
cancer (the topic of this section).  The other effects in the above list likely contribute to the induction of
other asbestos-related diseases and may even promote (but not initiate) asbestos-related cancer. Thus,
they are addressed further in later sections of this chapter (see below).

Although, ROS generation is associated with exposure to various particles and fibers (including all forms
of asbestos), generation of ROS does not necessarily imply carcinogenesis. For example, Zhu et al.
(1998) indicate that ROS generation is induced in response to exposure to asbestos, crystalline silica, and
coal mine dust. However, based on an extensive record of human exposure, the latter (coal mine dust) is
not carcinogenic in humans." Therefore, evidence related to the last two of ROS-associated effects listed
above, are examined  in more detail below.

Several studies indicate that, once generated by exposure to asbestos, ROS can interact with DNA in vitro
and in vivo to produce oxygenated adducts, primarily 8-oxo-Guanine.  Of the studies reviewed above, for
example, Yamaguchi et al. (1999) observed that crocidolite, but not (non-tumorigenic) glass produced
dose-dependent increases in 8-oxo-Guanine in rat lung tissue following intratracheal instillation. Also,
Leanderson et al. (1988), Park and Aust (1998), Keane et al. (1999) indicate that asbestos induces
formation of oxidized DNA adducts in in vitro assays. Brown et al. (1998) indicate that asbestos induces
ROS-mediated DNA strand breaks.

However, not all DNA strand breaks attributable to asbestos occur through pathways involving ROS.
OHikainen et al. (1999) exposed cultures of human mesothelial cells (MeT-5A, transfected with SV-40,
but nontumorigenic)  to hydrogen peroxide (100 uM) or crocidolite (2-4 ug/cm2), either alone or'in
combination with TNF-a and performed assays for DNA strand breaks. The authors note that the
concentrations of asbestos evaluated are well below those that have been associated with cytotoxic
effects.  Crocidolite alone was shown to produce DNA strand breaks at the concentrations tested.  The
authors also note that, at lower concentrations, only reversible effects were observed (presumably
indicating DNA repair). Co-exposure to crocidolite and TNF-a increased the observed incidence of DNA
damage, but the effect was less than additive. The  authors indicate that additional studies with
antioxidants indicate that the DNA damage induced by crocidolite in this study occurs through a
mechanism that does not involve ROS.  In fact, a potentially much more substantial mechanism by which
asbestos may induce  DNA breaks and various clastogenic alterations involves its ability to interfere with
mitosis (Section 6.3.3.1).

There is also evidence that different tissues may respond to asbestos-induced ROS generation differently.
For example, Zhu et  al. (1998) indicate that MnSOD is found in the mitochondria of Type 11 epithelial
cells of rats exposed  to crocidolite. The authors further indicate that, because fibroblasts, alveolar
macrophages, or endothelial cells do not display this protein when stimulated by exposure, this suggests a
difference in the susceptibility of epithelial cells to certain types of asbestos-induced injury.

Importantly, although these studies provide evidence that indicates asbestos is capable of producing
oxidized DNA adducts (or strand breaks) through ROS mediated processes, they do not  address the
question of whether such adducts can lead to heritable mutations in DNA in vivo nor do they indicate

                                               6.88

-------
whether such adducts lead to tumor production. Therefore, such studies should only be construed to
suggest a potential that asbestos can act as a cancer initiator through ROS-mediated pathways;

As previously indicated, not all mechanisms involving the generation or effects of ROS are similarly
fiber-size or fiber-type dependent and those that are do not necessarily depend on these variables in the
same way. At this time, it is not possible to distinguish among the relative importance of the different
mechanisms, so that it is difficult to judge the relative importance of the different effects.  However, the
overall general implication (with the few exceptions noted) is that ROS generation more likely contributes
to other asbestos-related diseases and to cancer promotion than to cancer initiation.         ;!

One final note concerning the effects of ROS that specifically involves the behavior of the hydroxyl
radical is also warranted. The hydroxyl radical is an extremely reactive species. Whether in vitro or in
vivo, this species will react with virtually 100% efficiency with every organic molecule it encounters.
Therefore the effects attributabfe to the hydroxyl radical are limited to those involving reactions in the
immediate vicinity of the location at which it is generated.  Thus, unless asbestos-induced generation of
this radical occurs within the nucleus and in the immediate vicinity of susceptible strands of DNA, it is
unlikely that these radicals are the direct cause  of DNA damage.                          |

Rather, hydroxyl radicals tend to react with cell membranes and other cellular components to'produce
further intermediate radicals (primarily lipid peroxides), which are much more stable than the'hydroxyl
radical and may migrate substantial distances before having an effect.  It is likely that these intermediate
radicals are ultimately responsible for any ROS-mediated DNA damage that may be attributed to
asbestos.                                                                            ••

6.3.3.3     >  Generation of reactive nitrogen species (RNS)                       "•

Nitric oxide (NO) is produced ubiquitously in biological  systems and serves many functions (Zhu et al.
1998). It is highly reactive and, therefore, generally short-lived in vivo.  However, nitric oxide has been
shown to react with superoxide (02~) to form the peroxynitrite ion (ONOO") at near diffusion-limited rates
(see, for example, Zhu et al.  1998). This the peroxynitrite ion (an RNS) may represent the primary
species responsible for the effects attributed to  ROS (at least when the primary ROS formed is
superoxide).                                                                          ;
                                                                                     i
Asbestos-induced Generation of RNS.  Alveolar macrophages, lung endothelial and epithelial cells, and
alveolar epithelium (in both  rats and humans), when stimulated by inflammatory agents, generate both
superoxide and over-produce nitric oxide. These then  combine to produce the perxoynitrite ion (Martin et
al, 1997; Zhu et al.  1998). These cells up-regulate production of NO when stimulated by various
cytokines, lipopolysaccharide, and interferon y. Because the peroxynitrite ion is a strong oxidant and
nitrating agent and is extremely reactive, evidence for its production is generally indicated in most studies
by the presence of nitrotyrosine, the stable product of tyrosine nitration, (Zhu et al.  1998). Evidence for
production of NO is frequently indicated by observation of nitrite.  Numerous studies also provide
evidence of nitric oxide and  peroxynitrite ion production specifically in response to exposurejto asbestos
in various tissues.                                                                     "•          •  .

Both chrysotile and crocidolite up-regulate the production of nitric oxide by alveolar macrophages in the
presence of interferon-y and the interaction between asbestos and interferon is synergistic. Non-
fibrogenic carbonyl iron did  not induce nitric oxide formation (Zhu et al. 1998). These authors also cite a
study in which intratracheal  instillation of silica and coal mine dust caused more inflammation and nitric
oxide formation than TiO2 or carbonyl iron (on an equal particle basis).  This suggests both tHe geometry
and chemical composition of particles determine their ability to up-regulate nitric oxide.    '.
                                               6.89

-------
Inhalation of chrysotile or crocdolite induces secretion of both TNF-a and nitric oxide by pleural
macrophages (Tanaka et al. 1998). Tanaka et al. (1998) studied the effects of RNS in rats exposed by
inhalation to 6 to 8 mg/m3 crocidolite or chrysotile (both NIEHS samples) for 6 hours/day, 5 days/week
for 2 weeks. Rats were sacrificed at 1 and 6 weeks following exposure. The authors indicate that
asbestos induces formation of stable products of nitric oxide in cells obtained by lung lavage 1 week after
asbestos exposure. Nitrotyrosine (a marker for ONOO" formation) was also observed. Also, a greater
number of alveolar macrophages and pleural macrophages were shown to express iNOS protein (the
inducible form of nitric oxide synthase, Table 6-6) than sham exposed animals. Exposed rats showed
significantly elevated immuno-staining for nitrotyrosine in the region of thickened duct bifurcations as
well as within bronchiolar epithelium, alveolar macrophages, and mesothelial cells of both the visceral
and parietal pleura. Nitrotyrosine staining was persistent, being observed at both 1 and 6 weeks following
exposure.

Quinlin et al. (1998) studied the production of nitric oxide in rats exposed to crocidolite or chrysotile
asbestos (both NIESH reference samples). Rats were exposed by inhalation at 6 hours/day, 5 days/week
and lavaged at 3,9, and 20 days. Lavaged macrophages showed significantly increased nitrite/nitrate
(indicating production of nitric oxide) and this was suppressed  with inhibitors to iNOS. Thus, nitric oxide
is produced via an iNOS pathway. The authors also note that nitric oxide production correlated
temporally with neutrophil influx in the lavage fluid. They also indicate that asbestos exposed animals
showed a 3- to 4-fold increase in iNOS positive macrophages in their lungs.

Quinlin et al. (1998) also exposed cultured murine alveolar macrophages (RAW 264.7 cells) to
crocidolite, riebeckite, and crstobalite silica in vitro. These cells showed increased iNOS mRNA
following exposure to asbestos and even greater increases if the cells were also stimulated with
lipopolysaccharide (LPS). Both crocidolite and riebeckite (but not cristobalite silica) stimulated increased
iNOS promoter activity when applied in combination with LPS. Thus, in this case, their appears to be a
mechanism that is sensitive to particle composition, but not size.

Park and Aust (1998) treated cultures of human lung epithelial  (A549) cells with crocidolite and observed
induction of iNOS and  reduction of intracellular glutathione (GSH) levels. Based on studies with
inhibitors, the authors further indicate that iron mobilized from crocidolite was required both for
formation of nitric oxide and to generate 2'-deoxy-7-hydro-8-oxoguanosine, but not for the observed
decrease in intracellular glutathione. The  authors note that approximately 5 times as much chrysotile
(containing approximately 3%  iron) as crocidolite was required to produce the same level of nitric oxide
formation. Importantly, these experiments were conducted in vitro in serum-free medium, so that no
extra-biological source  of iron  was present.

Choe et al. (1998) dosed cultured rat pleural mesothelial cells with either chrysotile or crocidolite (both
NIEHS samples) at concentrations between 1.05 and 8.4 ug/cm2 with or without co-stimulation with 50
ng/ml of interleukin-1 p (IL-lp). The authors report that mRNA for iNOS in asbestos and IL-lp dosed
cells increased progressively from 2 to 12 hours following exposure. Both types of asbestos also
stimulated production of nitric  oxide (measured as nitrite) in IL-lp stimulated cells in a dose- and time-
dependent fashion. Both types of asbestos also induced expression of iNOS protein and formation of
nitrotyrosine (based on nitrate detection) in IL stimulated cells. In contrast, carbonyl iron particles did
not induce any of the effects observed  for asbestos in IL stimulated cells.  Thus, formation of RNS
appears to be either fiber size dependent (not induced by particles) or mineralogy-dependent (or both).

As with the production of ROS, RNS production is apparently a function of multiple, complex
mechanisms. Also, as with production of ROS, the dose-response characteristics of the various
mechanisms differ.  Several of the mechanisms show a strong dependence on fiber size and some
dependence on fiber type. However, there are other mechanisms that are dependent primarily on fiber (or
particle) type, but may  not be dependent on size (or at least not dependent on fiber size).  At this point in

                                              6.90

-------
 time, it is not possible to gauge the relative importance of the various mechanisms by which RNS may be
 generated, except that it is likely that the importance of the various mechanisms likely differ in different
 cells and tissues and likely differ as a function of the specific toxin whose presence is inducing RNS
 production.

 There are also indications that production of RNS may be (animal) species specific.  For example, Jesch
 et al. (1997) report that alveolar macrophages harvested from rats expressed iNOS when stimulated with
 either LPS (Hppopolysaccharide) or interferon-y. In contrast, iNOS expression could not be induced in
 hamster, monkey or human macrophages.

 Effects Mediated by RNS. Based on Zhu et al. (1998), over-production of nitric oxide can:  .

        •      inactivate critical enzymes;                                            '
        •      cause DNA strand breaks that result in activation of poly-ADP-ribosyl transferase
               (PARS);                                                             '
        •      inhibit both DNA and protein synthesis; and                             ;
        •      form peroxynitrite by reaction with superoxide.

 In turn, peroxynitrite ion may:

        •      initiate iron-dependent lipid peroxidation;
        •      oxidize thiols;
        •      damage the mitochondrial electron transport chain; and
        •      nitrate phenolics (including tyrosine).                                   •
                                                                                   ii
 Also, some of the damage to alveolar epithelium and pulmonary surfactant system previously attributed to
 reactive oxygen species may actually be caused through ROS generation of peroxynitrite.  For example,
 Chao et al. (1996) report that crocidolite treatment of human lung epithelial cells (A549 cells) results in
 formation of 8-hydroxy-2'-deoxyguanosine (8-OHdG) in DNA and synthesis of mRNA for iNOS.  An
 iNOS inhibitor reduces intracellular nitrite and eliminates production of 8-OHdG. Addition of
 independent NO donor, recovers production of 8-OhdG.  Thus, production of the oxygenated-DNA
 adduct in this case appears to be generated by reaction with RNS.

 As with ROS, it is primarily the potential for RNS to contribute to DNA damage (e.g., strand breaks or
 generation of oxygenated adducts) that represent the primary pathways by which RNS might participate
 in cancer initiation (as opposed to cancer promotion or other asbestos-related diseases).  It appears that
 RNS-mediated DNA damage is closely associated with ROS generation and mediation of DNA damage
(see  Section 6.3.3.2). Thus, there is little to add here.                                   '
6.3.3.4
Conclusions concerning asbestos as a cancer initiator
The strongest, most consistent evidence that asbestos can act as a cancer initiator relates to the tendency
of asbestos to interfere with mitosis.  Although there is evidence that asbestos may induce production of
DNA adducts and DNA strand breaks (through ROS and RNS mediated pathways), whether such adducts
or breaks ultimately lead to permanent, heritable changes to DNA remain to be demonstrated.1 The
relative importance of the ROS/RNS mediated pathways compared to the pathway involving interference
with mitosis also remains to be determined. As indicated in later sections, however, ROS/RNS mediated
pathways may play substantial roles in cancer promotion and induction of other asbestos-related diseases.

Regarding the primary mechanism by which asbestos may initiate cancer (interference with mitosis), the
pathway is length dependent (short fibers do not appear to contribute to the effect).  Further, although
                                             6.91

-------
there may also be a dependence on fiber type (chemical composition), it is apparent that all types of
asbestos can act through this pathway. The dependence of this pathway on fiber diameter is less clear.
Thus, other than to suggest that fibers must be respirable (and therefore thinner than approximately 0.7
urn, Section 6.1) to have an opportunity to act, whether further diameter constraints are associated
specifically with the mechanism of cancer initiation is not known.

For pathways involving generation of ROS or RNS, some mechanisms (such as those associated with
frustrated phagocytosis) are length dependent, others are not. Although there appears to be some
dependence on fiber type for several mechanisms and a general dependence of several of these
mechanisms on surface chemistry, the relative importance of fiber type to the overall contributions from
these pathways remains to be determined.

Importantly, although crystalline silica may also act to  produce some of the same effects as asbestos
(including potentially induction or promotion of cancer), there is substantial evidence that this material
does not act through the same pathways and that the characteristics of its respective dose-response
relationships may differ.  Thus, for example,  while asbestos likely initiates cancer through a mechanism
that favors long (and potentially thin) fibers,  silica more likely acts through a mechanism that is
dependent on total surface area, with freshly  and finely ground material likely being the most potent.  In
contrast, grinding asbestos fibers tends to lesson carcinogenicity overall.  Due to differences in chemistry
and crystallinity (reinforced by studies indicating a lack of correspondence in behavior) crystalline silica
does not appear to be an appropriate analog for any of the asbestos types. Rather, for example, the
appropriate non-fibrous analog for crocidolite is riebeckite and the appropriate non-fibrous analog for
chrysotile is antigorite or lizardite.

There is also evidence that the relative importance of asbestos as a cancer initiator may differ in differing
tissues. For example, asbestos can only interfere with mitosis in cells that actively phagocytize fibers and
not all cell types actively phagocytize particles (although both mesomelial cells and pulmonary epithelial
cells appear to actively phagocytize fibers, Section 6.3.3.1). However, there is also evidence that
pulmonary epithelial cells (Type II) may undergo terminal differentiation to Type 1 cells and thus escape
potential cancer initiation by asbestos (Section 6.3.3.1). Such a pathway  is  not available to mesotheiial
cells. To the extent that pathways involving  generation of ROS or RNS contribute to cancer initiation, as
indicated throughout this section, the rates of generation and the spectrum of the species generated varies
as a function of cell type.

6.3.4   Evidence that Asbestos Acts As a Cancer Promoter

Primarily, asbestos may promote cancer by facilitating tissue proliferation.  However, additional
mechanisms associated with the observed  interaction between asbestos exposure and smoking (Section
6.3.4.6) also need to  be considered.

Substantial evidence exists indicating that asbestos induces proliferation in target tissues associated with
lung cancer and mesothelioma and this is summarized  below followed by an overview of studies that
suggest the various mechanisms by which asbestos may facilitate such proliferation. Evidence suggesting
the various mechanisms related to the interaction between smoking and asbestos exposure are also briefly
reviewed.

There are numerous mechanisms by which asbestos may facilitate proliferation including:

        •     direct cell signaling to  induce proliferation. This may occur by:
                       — direct interactions between fibers and receptors on the cell surface;
                       — interactions between phagocytized fibers and intracellular components of


                                               6.92

-------
                       signaling cascades; and                                         '
                       - or interactions between cells and intermediate species (e.g., ROS or RNS)
                       whose generation and release has been induced by asbestos; or
                                                                                     i
        •      response to induced cell death in target tissues, which then stimulates stem cells to
               proliferate to replace killed cells. Cell death may be induced either through:
                       - inducing apoptosis (programmed cell death); or
                       - direct cytotoxic effects, which leads to necrosis.                 •'
                                                                                     I-
These pathways are summarized in Table 6-5, which provides a perspective on the complexity of the
interactions between asbestos and the cells and tissues of the body.                        \

Regarding asbestos-induced cell death, asbestos-induced apoptosis (and all of the other effects described
above) typically occurs at exposure concentrations that are much lower than required to induce frank
cytotoxic effects (Sections 6.3.4.3 and 6.3.4.4). Therefore, it is primarily the former that is of potential
interest in terms of implications for asbestos-induced diseases in humans. Due to the high exposure
concentrations typically required, the importance of contributions from frank cytotoxicity to human
disease is unclear (Section 6.3.4.4).                                                     \
6.3.4.1
Asbestos-induced proliferation
Numerous in vivo studies indicate that all types of asbestos induce proliferation in target tissues relevant
to lung cancer and mesothelioma. Such proliferation is also suggested by the animal histopathological
observations previously described (Section 6.2.2). Moreover, many of these studies suggest that the
underlying mechanisms may be fiber size- and fiber-type specific. Responses may also be species-
specific.                                                                             '•:
                                                                                     ('

Importantly, there are some studies (primarily in vitro studies) that suggest asbestos acts to inhibit (rather
than induce) proliferation. Although it is likely that the contrasting observations found in such studies are
due to differences in conditions, timing, dose, or the type of asbestos employed, the underlying reason for
the contrasting observations is not always apparent.

The evidence for proliferation is summarized by tissue below.

In lung epithelium.  Brody et al. (1997) showed that rats and mice exposed for a brief (5 hour) period to
chrysotile asbestos at 1,000 fibers/cm3 (no size indicated) exhibited focal scarring at bronchio-alveolar
duct junctions that are identical  to those seen in asbestos-exposed humans. After 3-consecutive
exposures, the lesions persisted  for 6 months.  In regions where fibers are deposited, macrophages are
observed to  accumulate, epithelium is injured, and proliferation is observed to occur:  In this.study, the
authors also showed by immunohybridization staining that the four genes required to express three
peptide growth factors (TGF-cc,  TGF-P, and the A and B chains of PDGF) and the proteins themselves are
expressed in bronchio-alveolar tissue within 24 hours of exposure. PDGF is expressed almost
immediately and expression remains elevated for 2-weeks post-exposure, but only in regions where fibers
are deposited.

The authors report that PDGF is a potent growth factor for mesenchymal cells, TGF-ec is a potent mitogen
for epithelial cells, and TGF-p inhibits fibroblast proliferation, but stimulates synthesis of extra-cellular
matrix (Table 6-6).  The authors also report additional experiments with knockout mice indicate that
TNF-a is required to induce the  early stages of proliferation. The authors also indicate that Type II cells
produce TGF-pl and TGF-p2 (two of three isoforms of this protein) and that they are stimulated to do so
when co-cultured with macrophages.
                                              6.93

-------
Adamson (1997) intratracheally instilled size-separated (by sedimentation) long and short crocidolite
fibers into rats (0.1 mg in a single dose) and noted that the long fibers damaged the bronchiolar
epithelium and that fibers were incorporated into the resulting connective tissue; granulomas formed with
giant cells containing fibers).  The long fibers also appeared to escape into the interstitium.  Labeled
thymidine uptake (which indicates DNA synthesis and suggests proliferation) following long fiber
exposure was seen in lung epithelial cells, fibroblasts, and pleural mesothelial cells. Such labeling peaked
at 2% in mesothelial cells and 3% in epithelial cells within one week following exposure. Proliferation
appeared to end shortly beyond one week.  Short fibers were observed to have been efficiently
phagocytized by alveolar macrophages and a small increase in macrophage population appeared to have
been induced. Otherwise, none of the other effects attributable to long fibers (described above) were
observed with short fibers. Note that such observations are entirely consistent with those reported for the
range of studies described in Section 6.2.2.

McGavran et al. (1989). exposed both normal (C5+) mice and compliment deficient (C5-) mice to asbestos
(at 4 mg/m3) by inhalation and observed proliferation of bronchio-alveolar epithelium and interstitial cells
at alveolar duct bifurcations (based on incorporation of tritiated thymidine) between 19 and 72 hours after
a single, 5-hour exposure. Sham exposed rats showed fewer than 1% of epithelial and interstitial cells at
alveolar duct bifurcations incorporate labeled thymidine. In contrast, thymidine uptake in asbestos
exposed animals is significantly elevated for the first few days, begins to decrease at 8 days, and returns
to normal by one month following exposure. Both C5+ and C5- mice show similar increases in volume
density of epithelial and interstitial cells at 48 hours post-exposure. However, one month following
exposure C5+ mice developed fibrotic lesions while C5- mice were no different than controls.  The
authors conclude that the depressed macrophage response in C5- mice does not appear to change the early
mitogenic (and proliferative) response to asbestos, but apparently attenuates later fibrogenesis.

Chang et al. (1989) describes the morphometric changes observed in rats following 1 hour inhalation
exposure to chrysotile (at 13 mg/m3). Within 48 hours following exposure: the volume of the epithelium
increased by 78% and the interstitium by 28% at alveolar duct bifurcations relative to sham-exposed
animals. Alveolar macrophages increased 10-fold and interstital macrophages 3-fold. Numbers of Type I
and  Type II epithelial  cells increased by 82% and 29%, respectively.  At 1 month following exposure, the
numbers of Type I and Type II pneumocytes were still elevated, but not significantly. However, the
volume of the interstitium had increased by 67% accompanied by persistently high numbers of interstitial
macrophages, accumulation of myofibroblasts, smooth muscle cells, and an increased volume of
interstitial matrix.

In lung endothelium. In addition to the general evidence for proliferation of epithelial cells provided by
Adamson (1997), McGavran et al. (1989), and Chang et al. (1989), as cited above, a more detailed
description of the nature of asbestos-induced proliferation of endothelial cells is also available.

Proliferation of endothelial cells and smooth muscle cells of arterioles and venules near alveolar duct
bifurcations is induced in rats inhaling chrysotile (no size information given) for 5 hours at 4 mg/m3
(McGavran et al. 1990). This is based on the observed uptake of labeled thymidine (which indicates
DNA synthesis and suggests proliferation) that is significantly increased over controls between 19 and 72
hours following exposure. Twenty-eight percent of vessels near bifurcations exhibited labeled cells 31
hours after exposure.  One month following exposure, the thickness of the smooth muscle layers around
these blood vessels is  significantly increased (doubled). In contrast, labeling of these same endothelial
and  smooth muscle cells in sham exposed rats is zero. The authors indicate that endothelial cells and
smooth muscles associated with pulmonary blood vessels are normally quiescent with turnover rates on
the order of years.
                                               6.94

-------
In mesothelium.  In the second part of the study addressing epithelial proliferation, Adamson (1997)
reports that rats instilled with 0.5 mg of unmodified, UICC crocidolite were sacrificed at 1 week and
6 weeks following exposure and subjected to bronchiolar lavage and pleural cavity lavage. Lavaged
alveolar macrophages were observed to contain fibers, but pleural macrophages did not.  At 1 week,
collected pleural macrophages were shown to induce proliferation of fresh mesothelial cells in culture and
pleural lavage fluid showed an even greater effect. No effects were observed at 6 weeks.  Further work
with anti-bodies to various cytokines indicated that early, transient proliferation of mesothelial cells was
dependent on kertinocyte growth factor (KGF), but not on PDGF, FGF, or TNF-oc (Table 6-6). This
suggests that early, transient proliferation is induced by diffusing cytokines rather than direc^fiber
exposure. Adamson further reports that KGF is a fibroblast-derived cytokine that acts on epithelial cells
so that it's up-regulation likely results from epithelial injury with penetration of asbestos to the
interstitium (where fibroblasts are found, Section 4.4). A similar transient proliferative response in
mesothelial cells was also observed following exposure to chrysotile and in response to crystalline silica
exposure. Thus, it appears that the mesenchymal proliferative response may be mineralogy specific for
particles and is size-specific for fibers.                                                  '  •

Everitt et al. (1997) exposed rats and hamsters by inhalation to RCF-1 (45 mg/m3-650 f/cm3) for 12
weeks and then let them recover for up to an additional 12 weeks prior to sacrifice. The authors indicate
that both rats and hamsters showed qualitatively similar levels of inflammation at time examined (4 weeks
and 12 weeks). They  also indicate the mesothelial cell proliferation was observed in both animal types,
but was more pronounced in hamsters at all time points examined.  The greatest proliferation in both
species was in the parietal pleura lining the diaphragm. The authors also report that fibers (primarily
short and thin) were also observed in the pleural cavities of both species at all time points.   .

Several in vitro studies provide evidence that asbestos either induces or inhibits proliferation' in lung
tissues of interest and that at least some of the mechanisms involved are fiber size and mineralogy
dependent. As previously indicated, the specific reasons for the apparent'contrast between results
observed in vivo (where asbestos consistently promotes proliferation) and the inhibition sometimes
observed in in vitro studies is not always apparent. However, it must be due to the special conditions that
must be created to conduct in vitro studies, which may not support certain mechanisms that are important
in vivo.

Timblin et al. (1998a) completed a study in  which rat pleural mesbthelioma (RPM) cells in culture were
dosed with crocidolite or various cation-substituted erionites.  The  expression of several gene and gene
products were then tracked. Cultures in this study were exposed to 1,5, or 10 ug/cm2 of the various
fibrous materials. Analysis of the fibrous materials indicated that crocidolite contained many more fibers
per gram of material (probably because they are thinner) and that the preparation contains somewhat
longer fibers than the erionites evaluated.  In crocidolite, for example, 88% of the fibers are longer than 5
jim, 68% longer than 10 urn, and 37.5% longer than  20 um. All of the cation substituted erionites  showed
approximately the same size distribution: 50% longer than 5 um, 10-20% longer than 10 jini, and 1-5%
longer than 20 um.

Results indicate that the various cation substituted erionites behave differently and that the Na substituted
erionite shows the largest overall potency, at least for some endpoints, but not others. Fe, Na, and  Ca
substituted erionite all appear to induce c-fos expression in a dose-dependent fashion (increasing regularly
among the 1, 5, and 10 jig applications). 'K-erionite may also show the same pattern, but the'changes
were not indicated as significant over controls (apparently due to greater variability). Only Na-erionite
showed significantly increased expression of c-jun (at 1 and 5 ug applications, but not significantly at 10
jig, apparently due to greater variability. However, the mean result for 10 jig shows a consistent trend
with the lower concentration application results).
                                              6.95

-------
Na-erionite induces c-fos at the same or greater rates as crocidolite asbestos for the same mass application
(but not the same fiber number).  Crocidolite also appears to show a dose-response trend for c-jun
expression, but only the result for the highest application (10 ug) is significantly different from controls.
In contrast, Na-erionite appears to show greater induction of c-jun expression at lower dose than
crocidolite, but the increase with increasing dose is much lower for Na-erionite. Crocidolite also appears
to induce substantial apoptosis (even at the low dose of 5 ng) and that the induction is dose-dependent.  In
contrast, non-fibrous riebeckite does not appear to induce apoptosis.  Comparison between crocidolite
dosed cultures and Na-erionite dosed cultures indicate that crocidolite induces substantial apoptosis at all
time periods following application, but that Na-erionite induces little apoptosis even at higher mass dose
and longer time periods  than crocidolite. The authors also indicate that crocidolite and Na-erionite appear
to stimulate DNA-synthesis, which appears to be a compensating mechanism to fiber cell toxicity. The
authors indicate that chemistry is important in fiber toxicity as Na-erionite was a strong inducer of c-jun,
even at relatively low concentrations, but several of the other cation substituted erionites (including
Fe-erionite) were not.  They further suggest that, given the difference in the fiber lengths of the
crocidolite samples and  Na-erionite samples, that fiber length may be a less important consideration than
fiber surface chemistry.  However, despite the author's assertion, considering that non-fibrous riebeckite
does not induce any of the effects observed for crocidolite, there appears to be a clear size effect. It may
simply require that a defined, minimum length is necessary to induce the effect.

The authors also indicate that balance between proliferation and apoptosis is required to maintain
homeostasis in healthy tissue.  They further indicate that other studies suggest that c-Jun expression is
linked to proliferation and induction of cancer, while c-fos expression is linked to apoptosis.  Thus,
suppression of c-fos may be linked to carcinogenesis by allowing establishment or maintenance of a
transformed cellular phenotype.  This is in fact an early step in carcinogensis. Many environmental
agents stimulate both apoptosis and proliferation and, depending on the degree, may cause imbalances
that lead to disease. Relative stimulation of c-fos and c-jun may reflect some of these pathways.  Since
crocidolite induces both c-fos and c-jun in this study, the implication is that it mediates both apoptosis
and proliferation.

Wylie et al. (1997) dosed hamster tracheal epithelial (HTE) cells and rat pleural mesothelial (RPM) cells
with various asbestos  and talc samples  and-evaluated proliferation based on a colony-forming efficiency
(CFE) assay. The samples were NIEHS crocidolite and chrysotile and three different talcs. Samples were
characterized in the paper by mineralogical composition, surface area, and size distributions.

The authors indicate that both asbestos samples increased colony  formation of HTE cells (suggesting
induction of proliferation), but talc samples did not.  RPM cells, in contrast,  showed only dose-dependent
decreases in colony forming efficiency for all samples, which the authors indicate is a sign of
cytotoxicity.  The authors report that all samples show corresponding effects when concentrations are
expressed as fibers longer than 5 um or by total surface area. They also suggest that the "unique"
proliferative response by HTE cells could not be explained by either fiber dimension or surface area and
suggests a mineralogical effect.

Barchowsky et al. (1997) dosed cultured (low passage) endothelial cells to NIEHS chrysotile, crocidolite,
or RCF-1. After 1 to 3 hours exposure to 5 ug/cm2 (non-lethal concentrations), asbestos (but not RCF-1)
causes changes in cell morphology (cells elongate), increases in cell  motility, and increases in gene
expression.  Further work by the authors indicate that these effects are mediated by interaction between
asbestos and the receptor for urokinase-type plasminogen activator (uPAR).  The authors also suggest that
attachment of asbestos to cell membranes, internalization of asbestos fibers by the cells, and the
morphological changes  induced by asbestos are each mediated by different proteins.

Examples of in vitro studies that indicate asbestos (and other fibrous materials) may inhibit proliferation
in culture include the following.

                                               6.96

-------
In contrast to the above, Levresse et ai. (1997) found that chrysotile and crocidolite act to inhibit
proliferation in cultured rat pleural mesothelioma (RPM) cells. In this study, RPM cells (diplpid, no more
than 25 passages) were dosed with either UICC crocidolite or NIEHS Zimbabwean Chrysotile at
concentrations varying between 0.5 and 20 ug/cm2. The authors also note that the chrysotile sample
contains approximately 4 times the number of fibers as the crocidolite  samples. Cells were then examined
at 4,24, and 48 hours following treatment. In untreated cultures, the number of cells in replicative phase
decrease with time, which indicates that such cells are headed for confluence (completion of a monolayer
on the culture medium). At 48 hours, for example, 10% of untreated control cells were observed to be in
replicative stage. Chrysotile (but not crocidolite) decreased the fraction of cells in replicative phase in a
time- and  dose-dependent manner.  At 48 hours, for example, cells treated with 10 jig/cm2 chrysotile
showed only 1.5% in replicative phase. Further tests confirmed that this was due to blockage of cells at
the Gl/S boundary of the cell cycle. Both chrysotile and crocidolite appears to induce a timerdependent
increase in the number of cells at G2/M in the cell cycle, although this effect was not observed to be dose-
dependent for chrysotile. Even on a fiber-number basis, chrysotile appears to be elicit a greater response
(arrest a greater percentage of cells) than crocidolite.

The authors also indicate that chrysotile caused nuclear-localized, time-dependent increases in p53
concentrations. Crocidolite produced much lower levels that were not detectable in the nucleus.
Chrysotile was  also observed to produce blockage at the GO/G1 transition of the cell cycle, but crocidolite
did not. They also note that p53  is known to mediate arrest at this stage in the cycle so observing that
chrysotile induces arrest at this transition in the cell cycle may be consistent with the observed increased
expression of p53. The authors also note that chrysotile triggers apoptosis in this study and that
crocidolite shows a smaller, but detectable effect.  Spontaneous apoptosis in untreated cultures ran
between 0.5 and 1%  at 24-48 hours whereas chrysotile induced 4% apoptosis, peaking at 72 hours
following exposure to 10 ug/cm2. The authors indicate that the lower level of effects observed with
crocidolite could be due to the substantially smaller number of long fibers in UICC crocidolite compared
to NIEHS chrysotile.                                                                 '..

The previously reviewed (Section 6.3.3.1) study by Hart et al. (1994) also suggests that long,|medium,
short, and UICC crocidolite and chrysotile along with a range of MMVF's show a dose-dependent
inhibition on proliferation of cultured CHO cells and that potency toward the effect is a direct function of
fiber length.                                                                          r

Several of the above-described studies, in addition to providing evidence that asbestos induces
proliferation in  various lung tissues, also suggests certain mechanisms. Asbestos may induce
proliferation, for example, by inducing production of specific cytokine growth factors (Adanison 1997;
Brody et al. 1997), or by inducing certain signaling cascades (Barchowsky et al. 1997, Timblin et al.
1998b). It is also possible that the two effects may be related (i.e., that stimulation of a particular
signaling cascade may result in production of certain growth-stimulating cytokines). Other mechanisms
may also be important (Table 6-5).
6.3.4.2
Asbestos induced cell signaling
Asbestos has been shown to induce a variety of cell signaling cascades in a variety of target cell and
tissue types.  Such signaling may then trigger effects in the stimulated cells that may include:'

        •      proliferation;
        •      morphological changes;
        •      generation and release of various cytokines, enzymes, or extracellular matrix; or
        •      programmed cell death.
                                              6.97

-------
Note, due to the large number of chemical species that need to be considered in this discussion, Table 6-6
provides a summary of the sources of such species (Table 6-6A) and the effects attributable to such
species (Table 6-6B).

In specific cases, asbestos may initiate cell signaling by interacting directly with receptors on the cell
surface, by causing generation and release of intermediate species (e.g., ROS or RNS) that trigger cell
signaling, or (for phagocytized fibers) by interacting with ihtracellular components of a particular
signaling cascade. The specific responses to cell signaling induced by asbestos are frequently cell- or
tissue-type specific. Moreover, depending on the specific mechanism, cell signaling by asbestos may be
dependent on fiber size and/or type.

Barchowsky et al. (1998) showed in a set of studies that long chrysotile and long crocidolite, but not
RCF-1 fibers (at concentrations between 1 and 10 ug/cm2), which is reportedly below levels that typically
induce cytotoxic effects) induced up-regulation of urokinase-type plasminogen activator (uPA) and its
receptor (uPAR) in both lung endothelial cells (vascular cells) and lung epithelial cells. They also
showed that the increased pericellular protolytic activity (requiring cleavage of plasminogen to plasmin)
that is induced by asbestos in these cells is mediated by uPA.

In prior studies, chrysotile  has been shown to cause endothelial cells to elongate and increase expression
of adhesion molecules for phagocytes.  They also show enhanced proteolytic activity and matrix
interactions.  Chrysotile has also been shown to stimulate fibrinolytic activity in lung epithelial cells and
extravasating macrophages.  All such stimulation appears to result from up-regulation of uPA. Thus, this
mechanism may explain the observed asbestos-induced changes in lung endothelial and epithelial cells
including vascular remodeling, development of vascularized granular tissue, increased matrix turnover,
and leukocyte extravasation (which in turn may be caused by cell activation and elaboration of proteases
and adhesion molecules). The authors suggest that asbestos-induced up-regulation of uPA and uPAR
expression may represent a global mechanism for pulmonary toxicity and fibrosis induced by crystalline
fibers. Importantly, chrysotile was shown to induce uPA and uPAR expression in the absence of serum,
so the effect is apparently due to direct binding  of fibers to cell-surface receptors.

Mossman et al. (1997) observed that concentrations of 1.25-5 ug/cm2 of crodidolite (sample from  TIMA)
caused expression of c-jun and AP-1 in both cultured hamster tracheal epithelial (THE) cells and cultured
rat pleural mesothelial (RPM) cells.  Crocidolite was also observed to trigger the EGFR-regulated kinase
(ERK) and mitogen activated protein kinase (MAPK) pathways in RPM cells. The authors indicate that
these pathways are also stimulated by hydrogen peroxide and that NF-K0 induction stimulated by
crocidolite is also stimulated by crystalline silica (although silica stimulation of the MAPK pathway was
not investigated). They also note that the non-fibrous analog to crocidolite, riebeckite does not elicit
these activities.

The authors indicate that induction of the NF-KP cascade was inhibited by excess glutathione, which is
stimulated  by N-acetylcysteine (NAC),  suggesting that this pathway is induced by asbestos-caused
oxidative stress (perhaps through ROS or RNS intermediates).  Application of NAC also diminished
crocidolite induced c-fos and c-jun RNA levels and inhibited activation of the ERK-MAPK cascade.
Further work suggests that asbestos triggers the MAPK pathway by interaction with the Epithelial Growth
Factor Receptor (EGFR), either directly or by phosphorylation of this receptor by ROS. At the
concentrations examined, crocidolite induces substantial apoptosis (apparently through activation of the
ERK-MAPK cascade). In contrast, the  authors note that TNF-a induces the JNK arm of the ERK-MAPK
cascade, which leads to proliferation. Asbestos does not elevate JNK over the time-period of the study.
The authors note that it has been shown in some studies that inhibition of ERK in some cells, also  inhibits
asbestos-induced apoptosis.  Importantly, given that these processes are induced  by crocidolite, but not by
its non-fibrous analog, riebeckite, induction of the ERK-MAPK cascade appears to be a fiber-size
dependent process.

                                              6.98

-------
In a related study to that conducted by Mossman et al. (1997), Zanella et al. (1999) report that the (TIMA)
crocidolite (at concentrations of 2.5-10 ug/cm2, but not its non-fibrous analog, riebeckite, eliminated
binding of EOF to its receptor EGFR. Because EOF does not bind to crocidolite in the absence of
membrane, this is not simply a case of crocidolite tying up ligand. Crocidoiite also induces a greater than
2-fold increase in steady-state message and protein levels of EGFR.                        ;

The authors also note that the tyrphostin, AG-1478 (which specifically inhibits the tyrosine kinase
activity of EGFR), significantly mitigated asbestos-induced increases in mRNA levels of c-fos, but not
c-jun, and that the asbestos action was not blocked by a non-specific tyrphostin, AG-10. Moreover,
pretreatment of RPM cells with AG1478 significantly reduced asbestos-induced apoptosis.  Therefore, the
authors concluded that asbestos-induced binding to EGFR initiates signaling pathways responsible for
increased expression of the protooncogene c-fos and the development of apoptosis.  This apparently
occurs through the EGFR-extracellular signaling regulated kinase (ERK). It is hypothesizedjthat asbestos
may induce dimerization and activation (phosphorylation) of EGFR, which also prevents binding of EGF.
Asbestos apparently serves the same role as EGF in that it promotes aggregation of EGFR, which in turn
promotes binding to the extracellular domain of tyrosine  kinase receptors and the activation of their
intracellular kinases. The authors also indicate that other work suggests that crocidolite fiber exposure
leads to aggregation and accumulation of EGFR at sites of fiber contact and that asbestos also stimulates
biosynthesis of the EGFR and activates ERK in an EGFR-dependent manner.

The authors speculate that asbestos binding may not be ligand-site specific, but may be charge related or
may induce EGFR phosphorylation by local production of ROS, which has previously been demonstrated
to cause EGFR activation. As previously indicated, that this effect is driven by crocidolite exposure, but
not by exposure to riebeckite indicates that this mechanism is fiber-size specific.           ;'

Johnson and Jaramillo (1997) showed that UICC crocidolite, but not JM-100 glass, applied to a culture of
immortalized human Type II epithelial (A549) cells at non-cytotoxic concentrations (for 20 hours) results
in increased expression of p53, Cipl, and GADD153 in a dose- and time-dependent fashion.! Expression
was observed to be maximum at 18 hours. The crocidolite treatment was also  shown to cause an increase
in the number of cells arrested in Stage G2 of the cell cycle (with a persistent decrease in the'number of
cells in Gl). This was considered surprising because both p53 and Cipl  are known to mediate arrest in
Stage Gl. The authors suggest that these findings indicate a strong dependence on both fiber type and
fiber size (JM-100 glass contains substantially  more long fibers than crocidolite).  However, jit is not
possible to separate  the effects of fiber type versus fiber size in this study.

Luster and Simeonova (1998) indicate that at high concentrations, ROS may induce frank cytotoxicity.
At low or moderate  levels, ROS are more likely to induce cell signaling cascades that may, in turn,
contribute to asbestos-related disease. The authors dosed cultures of immortalized human Type II
epithelial (A549) cells, originally derived from a lung carcinoma, and normal human bronchjoepithelial
(NHBE) cells with long (Certain-Teed supplied) crocidolite (reported mean length: 19 um) at
concentrations ranging between 0 and 24 ng/ml. Results indicate that secretion of both Interluken-8
(IL-8) and IL-6 was stimulated by crocidolite exposure in a dose-dependent manner. In contrast,
increases in LDH levels (which indicate cell damage) was only detected at the highest exposure
concentrations tested. Further work indicates that stimulation of IL-6 and IL-8 secretion occurs through
ROS that are generated in an iron-dependent process (that may also include NF-Kp induction).  Note that
the trend of cell signal induction at low and moderate levels of asbestos exposure with evidence for
cytotoxicity observed only at the highest exposure concentrations is common to many of these kinds of
studies.                                                                            '
                                              6.99

-------
Choe et al. (1999) conducted a combined in vivo/in vitro study of the effects of low level exposure to
chrysotile or crocidolite at inducing leukocyte attachment to rat pleural mesothelial cells.  The authors
note that similar populations of rat pleural leukocytes (74% macrophages, 2% neutrophils, 10% mast
cells, and 10% eosinophils) were observed in both asbestos-exposed and unexposed rats.

In the second part of the study, cultured RPM cells were exposed to either crocidolite or chrysotile (both
NIEHS samples) at concentrations ranging between 1.25 and 10 ug/cm2, which was noted to be below
levels at which substantial cytotoxicity is observed.  Attachment of rat pleural leukocytes to RPM cells
was then observed to increase with increasing dose of asbestos to the RPM cells. In contrast, carbonyl
iron (a non-fibrous particle) also induced enhanced attachment, but at much lower levels and the effect
was not dose-dependent.  Further analysis indicated that asbestos-induced adhesion is mediated by up-
regulation of IL-lp (but not dependent on TNF-a or nitric oxide production, although it is noted that
TNF-a independently increases attachment). Asbestos also induces increased expression of vascular cell
adhesion molecule (VCAM-1).  The authors also note that rat pleural leukocytes harvested from asbestos-
exposed rats also showed increased adhesion to in vitro RPM cells over leukocytes harvested from sham
exposed rats. Thus, asbestos appears to trigger alterations in these cells as well.

In a recent paper, Driscoll et al. (1997) reviewed the role of tumor necrosis factor alpha (TNF-a) in
mediating the inflammatory response to lung insult by paniculate matter. The authors indicate that
quartz, coal dust, crocidolite, and chrysotile are all potent inducers of TNF-a production.  Titanium
dioxide (Ti02), corundum (aluminum oxide: A12O3), and latex beads are not. Pulmonary macrophages
have been shown to secrete TNF-a in vivo in response to exposure to some of the dusts listed above  .
(including asbestos). This is also observed among macrophages from asbestosis patients.

The authors indicate that rats immunized against TNF-a show reduced recruitment of neutrophils, which
demonstrates that TNF-a is involved in the recruitment of inflammatory cells.  TNF-a stimulates
macrophages, epithelial cells, endothelial cells,  and  fibroblasts to release chemokines that include
adhesion molecules (Eselectin, ICAM, VCAM). Inflammatory cells then interact with such molecules
and migrate along gradients from vascular structures in the lung to the lung interstitium and even the lung
air spaces. It has also been shown that release of TNF-a by macrophages is apparently dependent on
oxygen stress (i.e., exposure to ROS/RNS) that is induced in pathways that require iron. Production of
TNF-a and other compounds that mediate the inflammatory response are regulated by the oxidant-
sensitive transcription factor NF-Kp. This factor exists as a heterodimer in the cytoplasm in an inactive
form because it is bound to the inhibitory 1-kappaBalpha (1-KBa), which masks a nuclear translocation
signal. Appropriate stimulation of a cell induces phosphorylation of 1-KBa,  which then marks it for
proteolytic degradation. Then, NF-Kp translocates to the nucleus and induces transcription. The process
is stimulated by oxidants and inhibited by antioxidants.

In another review, Finkelstein et al. (1997) indicated that Type II epithelial cells and Clara cells (non-
ciliated bronchiolar epithelial cells) respond to and produce specific cytokines during the inflammatory
process. Early  responses to particle challenge include increases in mRNA and protein for IL-1 p, IL-6,
and TNF-a. These are also accompanied by changes in specific epithelial genes including those for
surfactant protein C and Clara cell secretory protein. The authors further indicate that these responses are
due to direct interaction with particles rather than a result of macrophage-derived mediators and they
suggest a more significant role for epithelial secretions in the overall pulmonary response than previously
suspected. Results also suggest that Type II pneumocyte-derived growth factors may play a significant
role in the pathogenesis of pulmonary fibrosis.

Also in this paper, Finkelstein et al. (1997) report that intratracheal instillation of lipopolysaccharide (a
potent inflammatory agent) caused increases in both lavage fluid and plasma levels of TNF-a and IL-6.
Intrapleural injection induced primarily increases in plasma levels.  The authors indicate that this suggests
that the observed cytokines are produced primarily at the site of injury. The authors further indicate that

                                              6.100

-------
 IL-6 is elevated in lavage fluids following exposure'to Ni2S3, a suspected human carcinogen, but not
 following exposure to Ti02 or NiO. In vitro studies indicate that release of IL-1 p and TNF-« (by Type II
 cells occurred only following exposure to crocidolite or ultrafine Ti02, but not pigment gradeTiO2. The
 authors also indicate that protein C and Clara cell secretory protein were both expressed in their
 respective source cells following exposure only to the fibrogenic of the above particles. They also report
 that crystalline silica has been shown to promote cytokine release and hypertrophy in Type II'cells.
                                                                                      i
 Jagirdar et al. (1997) used immunohistochemistry in a study to show that all three  isoforms of TGF-p
 (1,2, and 3) are expressed in the fibrotic lesions of asbestosis and pleural  fibrosis patients from the
 Quebec mines, primarily by Type II pneumocytes.  The cases examined averaged 38 years of exposure to
 the Quebec  chrysotile. The authors also indicate that the hyperplastic epithelium of silicosis patients also
 show elevated expression of all three isoforms. They further indicate  from previous studies that
 mesothelioma tumor cells frequently express TGF-P2 while the cells in the stroma of such tumors
 frequently express TGF-p 1.  It is also noted in this study that jun and fos  are both transcription factors
 that activate the TGF-pl  promoter.                                                      ;

 Zhang et al. (1993) indicate that macrophages obtained in BAL fluid from idiopathic  pulmonary fibrosis
 (IFF) patients and asbestosis patients show significantly increased secretion of TNF-a and asbestosis
 patients also showed significantly increased secretion of IL-lp. Macrophages and monocytes'obtained
 from both kinds of patients also show elevated expression of mRNA for these cytokines. In an in vitro
 part of this study, Zhang and coworkers, showed that chrysotile, crocidolite, amosite, and crystalline silica
 all stimulated IL-lp and TNF-a release and up-regulated their respective  mRNA in both macrophages and
 monocytes.  The authors also report that these two cytokines have been shown to up-regulate collagen
 Types I and III and fibronectin gene expression in human diploid lung fibroblasts after short-term, serum
 free exposure in vitro.
                                                                                      i
 Holian et al. (1997) exposed cultures of normal human alveolar macrophages (AM) (obtained by lavage)
 to varying concentrations (up to 25 |ig/ml) of short chrysotile, UICC crocidolite, ground silica,
 wollastonite, and titanium dioxide to determine whether these materials cause a phenotypic shift in
 macrophage populations by inducing selective  apoptosis.  The authors indicate that normal lungs contain:
 40-50% RFD1+7+ suppressor AM, and 5-10% RFD1+ immune activator. In this study, the pbrogenic
 subset of the particles tested (not wollastonite or titanium dioxide)  increased the ratio of
 activator/suppressor AM by a factor of 4 within a few hours and the effect was seen to increase with time.
 The authors also note that fibrogenic particles decrease the abundance of RFD7+ AM (phagocytic), but
 the consequences of this phenotypic shift are unclear.   '                                  !•

 The authors indicate that AM taken from fibrotic patients release a variety of proinflammatory mediators
 capable of stimulating fibroblast proliferation and collagen synthesis.  Even in the absence of evidence of
 fibrosis, workers who have been heavily exposed to asbestos yield similarly activated AM.  In contrast,
 they also note that, in vitro studies in which AM are stimulated with fibrogenic particles, while such AM
 are activated to release inflammatory cytokines, such releases are orders of magnitude less than that seen
 from AM derived from fibrotic patients. The authors indicate that the apoptosis-driven phenotypic shift
 in AM that is indicated by this study may explain the apparent discrepancy.

 In a previously described study, Timblin et al. (1998a), see Section 6.3.4.1, showed that crocidolite
 asbestos and several cation substituted erionites all stimulate c-jun and c-fos in rat pleural mesbthelial
 cells, but to  varying degrees  depending on  fiber chemistry. The effect also appears to be size dependent
 as crocidolite, but not its non-fibrous analog riebeckite induces the effect.                   "

 In general, the kinds of signaling cascades that  are potentially stimulated by exposure to asbestos are
 important due to their potential to contribute to the promotion of cancer.  Such pathways, for example,
 may mediate proliferation or may suppress apoptosis. Alternately, they may mediate an inflammatory

                                              6.101
       •      by causing sufficient genetic damage to trigger apoptosis; or    •           !
       •      by triggering a signal cascade that leads to apoptosis.

As previously indicated, asbestos may trigger signaling cascades by interacting directly with receptors on
the cell surface, or by inducing production of intermediate species (such as ROS or RNS) thatjmay in turn
induce cell signaling.

Some of the mechanisms by which asbestos may act to induce apoptosis may be fiber size- or,type-
dependent. Also, responses may vary in different target tissues.

Fibers. As indicated in Section 6.4.3.2, crocidolite (but not the non-fibrous analog riebeckite) induces
apoptosis in hamster tracheal epithelial cells and rat pleural mesothelial cells when applied at non-
cvtotoxic concentrations (Mossman  et al. 1997). Results from the study also indicate that apoptosis is

-------
response that in turn may lead to proliferation or to production and release of other, mutagenic agents
(e.g., ROS or RNS). Pathways that facilitate development of fibrosis may also contribute to cancer
promotion, given the apparent link between fibrosis and the development of lung cancer, which may
relate (among other possibilities) to inhibition of fiber clearance (Section 6.3.4.5).

The ERK-MAPK signaling pathway evaluated in multiple studies by Mossman and coworkers (Mossman
et al. 1997; Timblin et al. 1998a,b; Zanella et al. 1999) in rat pleural mesothelial (RPM) cells is a case in
point (see above). These studies suggest that crocidolite stimulates the ERK-MAPK cascade through
interaction with the EOF receptor. This ultimately leads to transcription of mRNA for c-fos. Crocidolite
has also been shown to induce c-jun (apparently through a separate mechanism) and the balance between
c-jun and c-fos has been implicated in guiding a cell toward either proliferation or apoptosis. Although
the direct connection between c-jun/c-fos and apoptosis has not been established, it is observed that
crocidolite induces substantial apoptosis in RPM cells at the same concentrations at which it induces
substantial expression  of c-fos and c-jun. The  link is also implied because inhibition of the ERK pathway
has been shown in some studies to inhibit asbestos-induced apoptosis. Na-erionite has also been shown to
induce c-fos at levels comparable or higher than crocidolite for comparable exposures (at least on a mass
basis) and induce c-jun at higher levels. However, it is not known whether Na-erionite and crocidolite act
via the same pathways.  Potentially due to the increased, relative expression of c-jun induced by
Na-erioinite, increased apoptosis is not observed in association with exposure to Na-erionite. However,
the link between increased c-jun expression and inhibition of apoptosis has not been demonstrated
explicitly.

Both crocidolite and Na-erionite were also shown by Mossman and coworkers to induce uptake of
bromodeoxyuridine (BrdU) by RPM cells in these same experiments. Uptake of BrdU is an indicator of
DNA synthesis.  Since it has also been shown that crocidolite is capable of damaging DNA via ROS and
other pathways (Sections 6.3.3) and both crocidolite and erionite are known to induce mesotheliomas in
any case, the balance between proliferation and apoptosis in this cell population that is struck by exposure
to these toxins may very well determine whether development of cancers are promoted or prevented.  Of
course, both proliferation and apoptosis may also  be mediated by other pathways independent of the ones
described here.

The problem is that the_range of responses that are induced by asbestos in the lung are varied and
complex (see Table 6-5) so that it has not yet been possible to definitively identify the biochemical
triggers that lead to lung cancer or mesothelioma.  It is even likely, for example, that different
mechanisms (or combinations of mechanisms) predominate under different exposure conditions or in
association with differing fiber types or particle sizes. Still, examination of the dependence of candidate
mechanisms on fiber type and particle size can be instructive, especially to the degree that such
indications are consistent with observations in whole animal studies (see, for example, Section 6.2.2). For
the signaling cascade described above by Mossman and coworkers, for example, the effects attributable to
crocidolite are clearly dependent on fiber size  because the non-fibrous analog to crocidolite, riebeckite
does not induce any of the effects.  It also appears that the chemistry of the fibers is important, given the
observed differences in responses among the various, substituted erionites.

6.3,4.3        Asbestos-induced apoptosis

Apoptosis (programmed cell death) is generally triggered when a cell accumulates certain types of genetic
damage, when cell signaling cascades are triggered by external stimuli that may occur, for example, as
part of the need to maintain tissue homeostasis or to cause a phenotypic shift in response to toxic
challenge (see, for example, Holian et al.  1997, Section 6.3.4.2), or when a cell has completed a pre-
programmed number of divisions.  Asbestos can induce apoptosis in a variety  of cells by several
 mechanisms including primarily:.


                                               6.102
effect). Although non-fibrous particles (such as crystalline silica) may also induce apoptosis, as
previously suggested, this may be through separate mechanisms from those responsible for asbestos-
related effects, even if the same endpoint results.

6.3.4.4        Asbestos-induced cytotoxicity

While there is ample evidence from various in vitro studies that asbestos is cytotoxic, such effects are
observed almost exclusively at the highest concentrations evaluated in an experiment (for example Luster
and Simeonova [1998], Section 6.3.4.2 and Choe et al. [1999], Section 6.3.4.2). Many experiments are
conducted at concentrations below those for which cytotoxicity is important because the other toxic
effects attributable to asbestos occur at substantially lower exposure levels and researchers prefer to study
such effects in the absence of potentially confounding cytotoxicity.  For in vitro studies, for example,
non-cyotoxic effects are typically studied at concentrations less than approximately 10 ue/cm2 (or

-------
The long and short samples were reportedly prepared from the UICC sample by repeated centrifugation.
Goodglick and Kane (1990) report that all three types of crocidolite stimulated release of ROS from
macrophages. At sufficiently high concentrations, all three also caused substantial cytotoxicity, although
apparently due to the longer time required to settle in culture, the foil effects from short fibers take longer
to develop. They suggest that, on a total fiber number or surface area basis, long and short crocidolite
appear to exhibit approximately equal potency toward the cytotoxicity of macrophages. Further work
with various inhibitors indicates that cytotoxicity is mediated by production of ROS and that'ROS are
produced via an iron-dependent pathway. They also indicate that, among the effects of crocidolite
exposure is that macrophage mitochondria! membranes are depolarized.
                                                                                                 %

Goodglick and Kane (1990) also evaluated the effects of long and short crocidolite in vivo. This was
done by evaluating the effects of intraperitoneal injection of the various samples (long, shortj-or mixed
crocidolite or titanium dioxide particles) in C57B1/6 mice.  Results indicate that a single injection of long
crocidolite (480  ug) induced an intense inflammatory response,  leakage of albumin, and fibers observed
scattered across the diaphragm. In contrast, a  single injection of short crocidolite  (600 jig) induced only a
relatively mild inflammatory response and only limited clusters  of fibers observed on the surface of the
diaphragm.

To test whether short fibers would show a greater response, if they were not cleared more readily than
long fibers, Goodglick and Kane (1990)  also subjected mice to 5 consecutive, daily injections of 120 ug
of short crocidolite and noted more substantial aggregations of fiber clusters along the diaphragm as well
as a more pronounced inflammatory response.  Cell injury was also assessed by Trypan blue'staining
(which indicates cell death). All mice singly or multiply injected with mixed or long crocidolite showed
marked Trypan blue  staining. Single injections of short structures showed only limited Trypan blue
staining. However, following 5 daily injections of short fibers, multiple Trypan blue stained; cells were
observed on the diaphragm in the vicinity of the locations were clusters of fibers were also observed. The
authors also indicate  (in contrast) that neither single injections of 160 or 800 ug nor 5 consecutive (160
jig) injections of titanium dioxide produced any Trypan  blue staining.
                                                                                     i
The authors conclude from this study that both short and long crocidolite fibers appear to be cytotoxic to
macrophages while titanium dioxide particles are not (suggesting that not only fiber length, tiut fiber type
is important to cytotoxicity).  They further suggest that, while short fibers tend to be cleared rapidly in
vivo, when such clearance mechanisms are overwhelmed (such as by repeated insult through repeated,
daily injections in this study), then the toxic effects of short structures becomes apparent.  As indicated  in
other studies, however, there may be multiple  mechanisms working to produce similar responses, that
such mechanisms may exhibit varying dose-response characteristics, and that cytotoxicity may not
generally be directly  related to mechanisms that contribute to carcinogenesis.  Moreover, there almost
certainly are at least some longer fibers in the  short fiber preparation and extended analysis to determine
their relative concentration with adequate precision would be helpful to see  if the  relative magnitude of
the observed effects correlate.                                                          :

Palekar et al. (1979) studied the ability of four different samples of commingtonite-grunerite, each also
subjected to varying degrees of grinding, to induce hemolysis of mammalian erythrocytes arid
cytotoxicity to Chinese hamster ovary (CHO)  cells. The samples studied include: UICC amosite
(4.13 m2/g surface area/mass), which is denoted as "asbestiform grunerite; "semi-asbestiform"
commingtonite (3.88 m2/g); acicular commingtonite (ground to three particle sizes: 3.76,2.45, and 0.82
mz/g), and acicular grunerite (2.82 m2/g).

Results from this study indicate that amosite induced the greatest hemolysis of erythrocytes by far
(approximately 50%) while acicular, unground grunerite caused no hemolysis. However,'grinding the
acicular grunerite to increasingly smaller particle  sizes and greater surface area ultimately results in some
hemolysis. Both semi-asbestiform and acicular, unground commingtonite show hemolytic activity

                                              6.105

-------
between amosite and unground,- acicular grunerite and grinding acicular commingtonite also increased its
hemolytic activity.

Similar results were also observed for cytotoxicity.  Amosite was by far the most cytotoxic and the effect
was dose-dependent. A dose of 0.05 mg/ml caused approximately 75% cell death for CHO cells. At 0.2
mg/ml, only 1% of cells survived. Acicular grunerite was nontoxic even at 0.5 mg/ml.  With grinding,
acicular grunerite cytotoxicity increased, albeit only slowly. The most heavily ground sample killed
fewer than 25% of cells at 0.2 mg/ml and killed only 65% at 0.5 mg/ml. The cytotoxicity of semi-
asbestiform commingtonite was substantially less than amosite, but greater even than ground, acicular
grunerite. For this material, 0.2 mg/ml killed approximately 65% of cells and 0.5 mg/ml killed
approximately 90%.  Interestingly, approximately the same dose-response curve for cytotoxicity was
observed for the 3.88 and 1.61 specific surface area samples of this material. The 1.21 samples was
somewhat less cytotoxic.  Acicular commingtonite was somewhat less cytotoxic (for corresponding
doses) than semiasbestiform commingtonite at the highest specific surface area (3.76) and its toxicity
decreased with decreasing surface area.

The authors also indicate that neither surface charges on crystal particles nor Magnesium ion content
appear to correlate with biological activity.  The authors conclude that the degree of "asbestiform"
character of a mineral has a dominant effect on biological activity.  Moreover, although non-fibrous
particles  may also be biologically active (and their activity increases with increasing specific surface
area), the effects of particles and fibers lie along entirely separate dose-response curves. The biological
activity of fibrous materials does not appear to depend directly on specific surface area.

Importantly, the results of the Palekar et al. (1979) study are also consistent with the possibility that
fibrous structures within a specific range of sizes and shapes contribute strongly to biological activity
while largely non-fibrous particles act through a separate mechanism that depends primarily on total
surface area, but that particle-for-particle elicits a substantially lower overall response than the
mechanism by which fibers act.  Such a scenario is supported by several studies.  Jaurand (1997), for
example, indicate that ROS are implicated in the cytotoxicity of long, but not short fibers on tracheal
epithelial cells. Although the evidence for distinct mechanisms for fibers and particles discussed here is
specific to cytotoxic and hemolytic effects, evidence in other studies suggest similar scenarios for other
toxic endpoints (potentially including endpoints that contribute to carcinogenicity).

Comparisons of the rate and extent of effects observed in epidemiology studies, whole animal  dose-
response studies, and in vitro studies suggests that cytotoxicity may not be important to human exposures.
Unfortunately, however, there is currently insufficient information to compare doses and exposures across
these studies in a more quantitative fashion. Therefore, the importance of cytotoxicity to human asbestos
exposure cannot be definitively determined at this time.

6.3.4.5        Association between fibrosis and carcinogenicity

The hypothesis that lung tumor induction is associated with the fibrosis has been examined by several
authors.  There appears to be a debate as to whether fibrosis is a necessary precursor for development of
lung cancer (associated with exposure to fibers), whether the presence of fibrosis is an additional factor
contributing to increased risk for lung cancer, or whether the two diseases are largely unrelated.  This is
an important consideration because the characteristics of the exposure-response relationship between
asbestos  and lung cancer or asbestosis (fibrosis) apparently differ (Sections 6.3.6 and 6.4).

Based on animal studies, Davis and Cowie (1990) found that rats that developed  pulmonary tumors
during inhalation experiments exhibited a significantly greater clinical degree of fibrosis than rats that did
not develop tumors.  Furthermore, Davis and Cowie (1990) reported suggestive evidence that  the


                                              6.106

-------
pulmonary tumors that did develop in the dosed rats tended to develop within portions of the rat's lungs
that were already scarred by fibrosis.
                                                                                     I
As part of a review, Mossman and Churg (1998) indicate that fibrosis of any cause (including diffuse
idiopathic fibrosis) appears to be associated with an increased risk of lung cancer and that this1 is observed
in both human and animal studies.  They also indicate that only those strains of mice and hamsters that
develop fibrotic lesions following exposure to crystalline silica show an increased risk of developing
cancer. They also report that, in parallel to what is reported for asbestos by Davis and Cowie (1990), lung
tumors that develop following exposure to crystalline silica tend to occur primarily (if not exclusively) in
those portions of the lung where fibrotic lesions predominate.
                                                                                     i
In contrast, Case and Dufresne (1997) from their study of lung burdens among Quebec miners and millers
indicate that there is high overlap in the range of concentrations that lead to both lung cancer and
asbestosis and those that lead to lung cancer alone, which the authors suggest show a lack of relationship
between the two diseases (one is not predictive of the other). The authors indicate that, based on
regression of the 111 cases they examined, the only indicator that reasonably tracks lung cancer is
severity of smoking and they indicate that this is true despite the level of fiber content in the lung.
                                                                                     i'
                                                                                     I
Although a definitive determination concerning the relationship between fibrosis (asbestosis) and lung
cancer cannot be developed at this time, it does appear that there is some association between;the two
diseases. Most likely, fibrosis is an additional risk factor for lung cancer and thus represents an additional
set of mechanisms that may contribute to the overall risk of developing lung cancer in association with
exposure to asbestos. However, based on the evidence as a whole, including the evidence that the
character of the dose-response relationship for lung cancer and asbestosis differ, it is not clear that
development of fibrosis is an absolute precursor that is required before asbestos-related lung tumors can
develop. Interestingly, based on the studies reviewed by Mossman and Churg (1998), the relationship
between fibrosis and lung cancer induced by silica may be substantially stronger than that between
fibrosis and lung cancer induced by asbestos.
6.3.4.6
Interaction between asbestos and smoking
Numerous studies have indicated a synergistic relationship between smoking and asbestos exposure
toward the induction of lung cancer (see, for example, Hammond et al. 1979; Kamp et al. 1992,1998;
Mossman et al. 1996).  Smoking is also suspected to facilitate development of asbestos-induced fibrosis
(Kamp et al. 1992). However, a more recent study (Liddell and Armstrong 2002) suggests a more
complicated relationship that is closer to additive than multiplicative (for a more detailed discussion of
this study, see Section 7.2.3).  Therefore, discussion of a more complex interaction (which may not be
specifically synergistic) is addressed below.

Putative mechanisms that may contribute to an interaction between asbestos exposure and smoking
include:

       •      facilitated transport of carcinogenic components of smoke that may be adsorbed on the
               surface of asbestos fibers, which may then serve as vehicles to transport these materials
 ,              through cell membranes to cell interiors and even to locations adjacent to or within the
               nucleus (see, for example, Fubini 1997; Mossman et al. 1996);             !

       •      asbestos-catalyzed production of more highly mutagenic metabolites of the various
               components of smoke, including benzo(a)pyrene (see, for example, Mossman et al.
               1996);
                                             6.107

-------
       •      smoke-product induced inhibition of clearance of asbestos fibers and/or asbestos induced          ^Br
               inhibition of clearance of smoke products (see, for example, Mossman et al. 1996); and

       •      smoke-product induced facilitation of uptake of asbestos by lung epithelium (see, for
               example, Mossman et al. 1996).

Although much progress has been made at elucidating the nature of these mechanisms and other candidate
mechanisms, at this point in time it is possible to indicate definitively neither what mechanisms are
important to any observed interaction between smoking and asbestos exposure nor to indicate the relative
magnitude of the contributions from such mechanisms.  Moreover, a detailed review of such mechanisms
is beyond the scope of this document.

6.3.4.7        Conclusions concerning asbestos as a promoter

There is strong evidence that asbestos acts as a promoter for cancer. While this may primarily involve
mechanisms the contribute to the induction of proliferation, mechanisms associated with the an
interaction between smoking and exposure to asbestos (for the induction of lung cancer; smoking does not
appear to affect mesothelioma) are also important.  There also appears to be an association between the
development of fibrosis and an increased risk for lung cancer.

Evidence indicates that multiple mechanisms may be involved with asbestos-induced cancer promotion
and that such mechanisms may be complex and interacting. The different mechanisms also appear to
exhibit dose-response relationships with differing characteristics. While there are indications that the
most important among these mechanisms may be strong functions of fiber size (with long fibers
contributing most to the induction of disease), mechanisms that depend primarily on surface area or total
fiber (particle) number (for any size range) may also contribute to overall cancer promotion. Importantly,
these latter mechanisms also appear to be strongly associated with the composition of fibers (particles)
and may therefore contribute more substantially to the disease induction of agents that have been shown
to be particularly toxic (such as crystalline silica), as opposed to particles, fibers, or asbestos in general.

At this point in time, the available data may not be sufficient to distinguish among the relative
contributions from the various mechanisms to the overall promotion of cancer, at least in terms of the
mechanistic data itself.  Importantly, however, the mechanistic data should not be considered to be
inconsistent with the results from whole animal studies, where there are clearer indications that fiber size
plays a major role in carcinogenicity and fiber (particle) type  is also important (see Sections 6.2 and 6.4).
Such studies indicate, for asbestos (and other biodurable fibers) that:

        •      short fibers (less than somewhere between 5 and 10 jim) do not appear to contribute to
               disease;

        •      potency likely increases regularly for fibers between 10 urn and a minimum of 20 ^im
               (and, perhaps, continues  to increase up  to lengths of at least 40 p,m); and

        •      fiber type may be important primarily in determining biodurability.

They further indicate that particularly (or uniquely) toxic particles (such as crystalline silica) may act
through a different set of mechanisms that are not dependent on fiber length, but that induce toxic
endpoints paralleling those observed for asbestos.

Importantly, the mechanisms by which asbestos may act as a promoter appear to occur in cell lines that
may contribute both  to the induction of lung cancer and mesothelioma.


                                              6.108

-------
6.3.5   Evidence that Asbestos Induces an Inflammatory Response

There is ample evidence that asbestos induces an inflammatory response in pulmonary tissues and the
pleura (see, for example, Sections 6.2.2 and 6.3). Moreover, there appears to be multiple biochemical
triggers that mediate this response and various mechanisms may be fiber size- and/or fiber type-specific
(Table 6-5). Because the role that inflammation plays in the induction of cancer has been addressed
elsewhere (Sections 6.3.3 and 6.3.4), it is beyond the scope of this document to provide a detailed review
of the mechanisms that lead specifically to inflammation.
                                                                                   !

6.3.6   Evidence that Asbestos Induces Fibrosis

There is ample evidence that asbestos induces fibrosis in pulmonary tissues.(see, for example, Sections
6.2.2 and 6.3). Moreover, there appears to be multiple biochemical triggers that mediate this.response
and various mechanisms may be fiber size- and/or fiber type-specific (Table 6-5). Because the role that
fibrosis plays in the induction of cancer has been addressed elsewhere (Sections 6.3.3 and 6.3.4), it is
beyond the scope of this document to provide a detailed review of the mechanisms that lead specifically
to fibrosis. Such mechanisms have also been the subject of recent reviews (see, for example, Mossman
and Churg 1998; Robledo and Mossman 1999).                                         :;
                                                                                   '!
6.3.7   Evidence that Asbestos Mediates Changes in Epithelial Permeability

As previously indicated (Section 4.4), maintaining the overall integrity of the epithelial surface of the
lung is among the various functions of Type II epithelial cells (Leikauf and Driscoll 1993).  It has been
shown that asbestos induces changes in the morphology of Type II epithelial cells (see, for example,
Ilgren and Chatfield 1998), which has the effect (among others) of increasing the overall permeability of
lung epithelial tissue to various macromolecules and, potentially to asbestos fibers themselves. The
former plays a role in asbestos-induced fibrosis (by allowing cytokines that stimulate fibroblast
proliferation or stimulate fibroblasts to generate extracellular matrix to pass through the epithelium and
reach the underlying fibroblasts, Section 6.3.6).  The latter may be important to facilitating transport of
asbestos from the alveolar lumen  to the interstitium (see, for example, Lippmann 1994).

Changes in epithelial permeability may be triggered by cytokines released from other cells or by the
action of asbestos fibers on epithelial cells directly. Moreover, some of the mechanisms that mediate this
response may be sensitive to fiber size and/or fiber type.  For example, Gross et al. (1994) showed that
monolayers of human bronchial epithelial cells cultured over a porous medium and exposed to
cryogenically ground chrysotile (average length: 1 urn, average aspect ratio:i4 at 15 ug/culture plate)
became permeable to fibrin breakdown products (FBP's).  The cultures were grown over human serum
with labeled fibrinogen. This was based on observed increased concentrations of FBP's (double in 24
hours) in the ablumenal chambers of exposed cells compared to cells in control cultures. Because the
epithelium showed greater permeability to all concentrations, the increased concentrations were not due to
increased breakdown.  The observed FDP flux was not vectoral, not sarurable, and required neither
proteolytic processing  nor active transport. Thus, asbestos increases the paracellular flux of intact FDP
across airway epithelium.

6.3.8   Conclusions Regarding the Biochemical Mechanisms of Asbestos-Related,Diseases

That the specific biochemical triggers for asbestos-related diseases (particularly, the asbestos-related
cancers) have not been definitively delineated as of yet is not surprising.  The detailed interactions
between fibers (and particles) and the cells and tissues of the lung are complex and there are complex,
multiple, interacting mechanisms  by which such interactions may contribute to disease. Despite  great
progress in elucidating candidate mechanisms, the number of candidate mechanisms is large and

                                             6.109

-------
distinguishing among their relative contributions has been difficult. This is because, among other things,
the ability to compare results across studies of different mechanisms is currently limited due to the
inability to reconcile the quantitative effects of dose and response across dissimilar studies.

Nevertheless, a number of important implications can be gleaned from the available literature. First, it
appears that asbestos can function both as a cancer initiator and a promoter. It also appears that both the
initiation and promotion of cancer may occur through more than one mechanism.

Regarding cancer initiation, asbestos likely acts primarily through a mechanism involving interference
with mitosis. By this mechanism, asbestos fibers are phagocytized by target cells, migrate to perinuclear
locations, and interact with the spindle apparatus and other cell assemblages required to complete mitosis.
This tends to result in aneuploidy and may cause various clastogenic effects. This mechanism is driven
by long fibers; short fibers do not appear to contribute to the effect. It also appears that all asbestos fiber
types (and potentially other durable fibers with sufficient dimensions) cause genetic damage via this
mechanisms. If there are effects due to fiber type, they appear only to play a secondary role.

Although there is also" evidence that asbestos may induce production of DNA adducts and DNA strand
breaks (through ROS and RNS mediated pathways), whether such adducts or breaks ultimately lead to
permanent, heritable changes to DNA remain to be  demonstrated. The relative importance of ROS/RNS
mediated pathways for  initiating cancer, compared to the pathway involving interference with mitosis,
also remains to be determined.

There is also some evidence that the relative  importance of asbestos as a cancer initiator may differ in
different tissues. Lung epithelial cells, for example, appear to be relatively resistant to the mechanisms by
which asbestos may initiate cancer. Mesothelial cells are not.  Among several possibilities, this may be
due to the ability of proliferation-competent lung epithelial cells (Type II cells) to undergo terminal
differentiation when challenged with certain  toxins  and this is a pathway not available to mesothelial
cells.

The mechanisms by which asbestos may promote cancer primarily involve mechanisms that contribute to
the induction of proliferation, although mechanisms associated with an interaction between smoking and
exposure to asbestos to induce lung cancer are also  important.  There also appears to be an association
between the development of fibrosis (including asbestosis) and an increased risk of lung cancer.

Evidence indicates that multiple mechanisms may be involved with asbestos-induced cancer promotion
and that such mechanisms may be complex and interacting.  The different mechanisms also appear to
exhibit dose-response relationships with differing characteristics.  While there are indications that the
most important of these may be strong functions of fiber size (with long fibers contributing the most to
carcinogenicity), mechanisms that  depend primarily on surface area or total fiber number (for any size
range) may also contribute to overall cancer promotion. These latter mechanisms also appear to be
strongly associated with the composition of fibers and may therefore contribute more substantially to the
disease induction of agents that have been shown to be particularly toxic, as opposed to particles, fibers,
or asbestos in general.

Although crystalline silica may act to produce some of the same effects as asbestos (including
carcinogenicity), there  is substantial evidence that this family of materials do not act through the same
pathways and that the characteristics of their respective dose-response relationships may differ.  Thus, for
example, while asbestos likely induces cancer through mechanisms that favor long (and potentially thin)
fibers, silica more likely acts through a mechanism that is  dependent on total surface area, with freshly
and finely ground material likely being the most potent. In contrast, grinding asbestos fibers tends to
lesson its carcinogenicity overall.  Due to differences in chemistry and crystallinity (reinforced by studies
indicating a lack of correspondence in behavior), crystalline silica does not appear to be an appropriate

                                              6.110

-------
analog for any of the asbestos fiber types. Rather, for example, the appropriate non-fibrous analog for
crocidolite is riebeckite and the appropriate non-fibrous analog for chrysotile is antigorite or Ijzardite,

6.4     ANIMAL DOSE RESPONSE STUDIES
                                                                                     *,.

Ideally, human epidemiology studies (reviewed in Chapter 7) provide the best data from which to judge
the effects of asbestos in humans and from which to derive exposure-response factors for humans.
However, animal dose-response studies have proven useful for elucidating certain features of the
relationship between asbestos dose and response that cannot be adequately explored in the human studies,
primarily due to limitations in the manner that exposures were characterized in the human studies (see
Chapter 5).

Unlike human epidemiology studies, exposures in animal studies are controlled and better quantified.
Frequently, the characteristics (in terms of fiber size, shape, and type) of such exposures have also been
better quantified and this has allowed exploration of the effects that such characteristics (fibersize, shape,
type) have on disease response. Accordingly, an overview of animal dose-response studies is'provided in
this section. Both injection-implantation studies and inhalation studies are reviewed. Particular attention
is also focused on a "supplemental" animal inhalation study that we conducted with the specific aim of
identifying the characteristics of asbestos that best relate to risk. The strengths and limitations of these
kinds of studies are described in Chapter 5.                                              •'

6.4.1   Injection-Implantation Studies                                             :

Because the fibrous materials in injection and implantation studies are placed immediately against the
target tissue, the effects of processes associated with inhalation, retention, and translocation are avoided.
The only active mechanisms that need  to be considered in these studies are those that occur directly in the
target tissue (including degradation, clearance, and biological responses of the types described in the
previous sections of this chapter). Fibrous materials placed against the tissue surface are subject to
dissolution, phagocytosis by macrophages, and phagocytosis by the cells of the target tissue.  These
mechanisms are described in greater detail in Section 6.2. A  range of biologic responses have also been
observed (described in Section 6.3).                                                    ,

Numerous researchers have performed these types of studies.

The Work of Stan ton and Coworkers. In a series of studies, Stanton and coworkers (1972,' 1977,1981)
implanted fibrous materials and induced mesotheliomas in rats.  In the studies, a pledgette composed of
coarse glass is loaded with hardened gelatin containing sample material and is surgically implanted
immediately against the left pleura of the rats.  Control  studies demonstrate that the coarse glass of the
pledgette does not induce significant tumors in the absence of other tumorigenic agents in the gelatin.

Although the mass dose of material implanted was the same for all experiments (40 mg), the observed
incidence of mesothelioma varied among samples. By characterizing the dimensions of fibrous structures
in the samples using a microscope, the  researchers were able to explore the relationship between fiber size
and the  incidence of mesothelioma.  By studying a wide range of fibrous materials, Stanton and his
coworkers concluded that the induction of mesothelioma is determined primarily by the physical
dimensions of fibers and that mineral composition is secondary. Further, potency appears to  increase
with the length and decrease with the diameter of fibrous structures.  The researchers also concluded that
the incidence of malignant tumors correlates with the degree of fibrosis induced by the presence of the
fibrous materials. This does not necessarily imply, however, that fibrosis is a necessary step in the
induction of asbestos-induced tumors (see Section 6.3.4.5).
                                             6.111

-------
Conclusions from the Stanton et al. (1972,1977,1981) studies indicating that mineralogy is not a factor
in biological response conflicts with evidence provided in Chapter 7 and implications gleaned from
mechanism studies presented in Section 6.3.  However, the studies by Stanton and coworkers have been
shown to suffer from certain methodological limitations (Herman et al. 1995) so that results from these
studies should be considered more qualitative than quantitative.

Due to limitations in the ability to produce samples composed of uniform fibers, quantitative relationships
between size and potency were explored by Stanton and coworkers using a regression analysis.
Structures longer than 8 um with diameters less than 0.25 jim or longer structures with diameters less than
1.5 jim were found to represent the range of sizes that best correlate with carcinogenicity. It was further
stated that such correlations did not eliminate the possibility that other size ranges also contribute to
potency, only that the two size ranges identified appear to correlate best. Samples that varied
significantly from the reported correlations were attributed to errors in the characterization of structure
size distributions in those samples. However, other methodological limitations might also have
contributed to the observed deviations or such "outliers" may also suggest evidence for a mineralogical
effect that is similar to what is reported in other studies (see Section 6.3 and Herman et al. 1995).

The precision of estimates for the ranges of sizes that contribute to biological activity that are derived
from the Stanton and coworkers (1972, 1977, 1981) studies is limited so that such estimates should also
be considered qualitative.  Size distributions were determined by characterizing 200 to 1,000 structures
using TEM and there is no indication that statistically balanced counting rules were employed (Section
4.3).  Under such conditions, counts of structures longer than 8 \im are likely small and subject to large
uncertainties for most of the samples characterized. Confidence intervals are not provided for any of the
exposure values presented in these studies.

Potentially larger errors in the studies by  Stanton and coworkers could have been introduced by the
method employed to relate fiber counts to sample mass.  As indicated in Chapter 5, estimating
contributions to mass by sizing total particles and assuming that this is proportional to total sample mass
is subject to error from the limit to the precision of characterizing structure dimensions (particularly
diameter) and by not accounting for nonasbestos (and possibly nonfibrous) material in the samples.  Thus,
for example, there is no discussion of the precision with which the cut point of 8 um was determined in
these studies.

Re-analysis and Extension of the Stanton Studies. Several researchers have re-evaluated data from the
implantation studies to test additional hypotheses. Using the Stanton and coworkers (1972, 1977, 1981)
data, Bertrand and Pezerat (1980) examined the relationship between mesothelioma incidence and several
characteristics not evaluated by Stanton and coworkers including: average fiber length, average fiber
diameter, average fiber aspect ratio, total fiber surface  area and total fiber volume.  Results from the
regression analysis  indicate that potency varies directly with average length and inversely with average
diameter, but that neither parameter is a good indicator alone. Combining the effects of length and
diameter, average aspect ratio is highly correlated with potency. Biological activity does not correlate
highly with structure count, surface area or volume except when fiber sizes are restricted to the long, thin
structures that Stanton and coworkers defined. Results of this study are not inconsistent with those
originally presented by Stanton and coworkers except that they emphasize a set of characteristics that
relate parametrically to biological activity rather than expressing exposure as a single restricted size range
of structures.

Bertrand and Pezerat (1980) were able to find good correlations between response and specific "average"
characteristics of the samples that are not proportional  to the quantity of the material present in the sample
("intensive" characteristics). Such intensive characteristics as average aspect ratio, average length, or
average diameter are properties that are independent of the mass of material in a sample. Since response


                                               6.112

-------
must be a function of the quantity of sample present, intensive characteristics should have to>'be multiplied
by characteristics that are proportional to the mass of a sample (e.g., fiber number, sample mass, or
sample volume) in order to relate them to response. Properties that vary with the mass of a sample are
termed "extensive" properties.                                                         '

The correlations between intensive properties and response reported by Bertrand and Pezeraf (1980)
likely succeed within the Stanton and coworkers (1972, 1977, 1981) database because a constant sample
mass (40 mg) was employed for all of the implantation experiments. However, to apply dose-response
relationships that are dependent only on intensive characteristics beyond the data presented by Stanton
and coworkers (where mass dose will not be constant), it is necessary to pair intensive characteristics with
extensive characteristics (such as mass or number of fibers per sample). Therefore, it is unclear how the
conclusions from this paper may be generalized to other data sets.
                                                                                     ;»
In a similar study, Bonneau et al. (1986) also examined parametric relationships between  structure
characteristics and mesothelioma induction. The paper examined specifically correlations between
carcinogenicity and dose in terms of two specific relationships: dose expressed as fibers longer than 8 um
that are thinner than 0.25 f±m ("Stanton" fibers) and dose expressed as mean aspect ratio.  The researchers
conclude that mean aspect ratio provides an excellent indication of carcinogenicity for individual fiber
types, but that each fiber type must be treated separately.  Poorer correlations are found for the
relationship between the concentration of "Stanton" fibers and mesothelioma, even when  fiber types are
considered independently.  Although these results appear to be consistent with findings reported from
mechanistic studies (Sections 6.2 and 6.3) in that they posit a role for fiber mineralogy, the relationships
evaluated by Bonneau et al. (1986) also suffer from the limitation of expressing dose only in terms of
intensive quantities,  as discussed above. Direct comparison with other studies is therefore difficult.

Following up on the reported problems of the studies by Stanton and coworkers in characterizing
crocidolite, Wylie et al. (1987) reanalyzed seven crocidolite samples originally studied by Stanton et al.
She and coworkers then used the new  size distributions to reevaluate the "Stanton hypothesis" (that the
concentration of "Stanton" fibers in a sample correlates with carcinogenicity). Wylie and coworkers note
that substantial deviations from the Stanton hypothesis occur for specific samples. They conclude that a
specific structure size range alone is not sufficient to characterize biological activity and that a parametric
relationship with other structure characteristics (potentially including mineral type) may be necessary to
sufficiently describe biological activity.                                                  i

Conclusions from the Wylie et al. (1987) paper must be interpreted carefully because the  researchers
evaluated only the relationship between carcinogenicity and the single specific size range indicated
("Stanton" fibers). Thus, the possibility that improved correlations exist between biological activity and
different size ranges or a combination  of size ranges cannot be ruled out.  Qualitatively, conclusions
presented in this paper  are not inconsistent with the conclusions reported by Stanton and coworkers
regarding the general relationship between response and fiber dimensions.
                                                                                      !
The  Wylie et al. (1987) study appears to suffer from several methodological problems.  These1relate to the
manner in which the sample reanalysis was performed. 'The drop method for preparing electron
microscopy grids (used in this study) is not satisfactory  for preparing grids. In fact, as reported in the
study itself, grids prepared as duplicates by this method were'shown to be non-uniform at the 95%
confidence interval using a chi-square test. In addition, only 100 to 300 fibers were counted for each
sample. Since there is no indication that statistically balanced counting was performed, the uncertainty
associated with counts of "Stanton" fibers may be substantial. Such errors would be further multiplied by
uncertainty introduced during the sizing of total particles to determine the number of fibers per unit mass.

In a later study, Wylie et al. (1993) examined the effect of width on fiber potency. In this latter study,
results from animal injection and implantation studies were pooled and subjected to regression analyses to

                                              6.113

-------
identify correlations between exposure and tumor incidence. The animal studies selected for inclusion in
this analysis were performed on a variety of tremolite samples exhibiting a range of morphological and
dimensional.characteristics.

In their regression analyses, Wylie et al. (1993) evaluated a range of exposure indices that emphasize
different morphological or size characteristics to help elucidate the characteristics of asbestos that induce
a biological response. Because all of the animal studies included in their analyses involved tremolite,
mineralogy was not an issue. Results from this study suggest that fibers longer than 5 |itn and thinner
than 1 um best correlate with tumor incidence among the animal injection and implantation studies
examined.  Further, they suggest that a width limit, rather than a limit on aspect ratio, better reflects the
bounds of the asbestos characteristics that determine biological activity. They also suggest that complex
structures (bundles and clusters) need to be evaluated as part of the determination of exposure because
such structures can breakdown and contribute to the population of thinner fibers.

Although the results of the Wylie et al. study tend to support the general conclusions in this document
related to width, if not to length (see Section 6.5 and Appendix A), as the authors themselves indicate,
such results should be considered qualitative due to the limitations imposed on their study by the
methodology employed. Their study was conducted by:

        •      combining results from multiple studies without careful consideration of variation
               introduced by methodological differences across the studies;

        •      employing asbestos concentrations determined by SEM and without careful consideration
               of differences in the counting methodologies employed by differing'research groups
               across studies; and
       .•      considering injection and implantation studies, which do not account for mechanisms
               related to inhalation and deposition that affect the exposure-response relationship in
               humans.

The limitations imposed by the above constraints are highlighted in Chapters 4 and 5 of this report.

Other Injection Studies. A series of injection studies were conducted by several research groups. In
these studies, fibrous materials were suspended in saline and injected into rats immediately adjacent either
to the pleura or peritoneum. A large number of fibrous materials have now been studied by this process,
as reported by: Bolton et al. (1982, 1984,1986); Davis et al. (1985, 1986a,d, 1987, 1988a); Muhle et al.
(1987); Pott et al. (1974,1976,1978,1982,1987); and Wagner et al.  (1976, 1982,1985).  Newer studies
are also discussed in Section 6.2. Results confirm that it is the fibrous nature of the materials that is the
primary factor leading to the induction of tumors and that potency appears to depend directly on length
and inversely on diameter.

The authors of these studies tend to indicate that, except where fibers are not persistent in vivo, due to
solubility or other degradation processes, the mineralogy of the fibers appears to play only a secondary
role in determining disease  incidence.  Researchers conducting injection experiments also tended to report
a correlation between tumor incidence and the degree of fibrosis induced by the sample.  These
observations are consistent  with the ideas originally articulated by Stanton.

Pott developed Stanton's ideas further by suggesting that carcinogenicity is a continuous function of fiber
dimensions, which decreases rapidly for lengths less than 10 um and also decreases with  increasing
diameter. The possibility was also raised that the apparent inverse dependence on diameter may be an
artifact due to the limited number of thick fibers that can be injected in a sample of fixed mass.
                                              6.114

-------
 Although the published injection studies indicate that potency decreases with decreasing length,
 researchers in these earlier studies were reluctant to identify a length below which contributions to
 carcinogenicity can be considered inconsequential. This may be due in part to the skewed distribution of
 fiber sizes typical of asbestos dusts. Thus, for example, even if structures less than 5 urn are only 1% as
 potent as structures longer than 5 urn, they may be as much as 100 times as plentiful in some asbestos
 dusts, so that the total contribution to potency would be equal for both size fractions.

 Reasonable dose-response curves have been generated using various sample masses of a single material.in
 some of these studies.  This has been demonstrated for UICC crocidolite and UICC chrysotile "A"
 (Bolton  1984). Results indicate that the relationship between tumor incidence and the log of the dose
 may be linear and there is no effective threshold. A consistent difference between the two dusts is
 apparent; the points lie along separate curves and chrysotile appears to be more potent per unit sample
 mass,                                                                                 ;:

 In general, the analytical techniques used for quantifying size distributions in these studies are not fully
 documented.  To the extent that they are, it appears that similar approaches were adopted to those
 described for the implantation studies above.  Consequently, similar limitations apply to the interpretation
 of results.  Briefly, large uncertainties are likely associated with counts of long fibers and estimates of the
 number of fibers per unit sample mass. Counts in several of the studies also suffer from limitations in the
 ability of SEM or PCM to detect thintfibers (Section 4.3); whenever SEM or PCM was employed, the
 thinner fibers were likely under-represented in reported fiber size distributions.
                                                                                      i
 Because samples are placed against mesenchyme in the published implantation and injection studies,
 results of these studies most directly represent processes associated with the induction of mesothelioma.
 Assuming, however, that clearance and degradation processes are similar in the deep lung, once a fiber
 reaches a target tissue, results from the implantation and injection studies may also provide a model of
 biological response in lung tissue and the factors that lead to the induction of pulmonary tumors.  Such a
 model must be considered qualitative at best, however, because it has been shown that the mechanisms of
 tissue response to the presence of asbestos in lung parenchyma and in the mesenchyme differ in detail
 (Section 6.3). The time periods over which the various clearance mechanisms operate in the deep lung
 and the mesenchyme also differ (Section 6.2), although, it is apparent that the general nature of the
 clearance and degradation processes in the two tissue types are generally similar.

 6.4.2   Animal Inhalation Studies

 Animal inhalation studies measure response to exposure in controlled systems that model  most of the
 relevant variables associated with asbestos disease mechanisms in humans (including respirability,
 retention, degradation, clearance, translocation, and tissue-specific response). Thus, the available
 inhalation studies are the best database from which to evaluate the integrated effects that lead to the
 development of asbestos-related disease. Such studies can be used both to identify the characteristics of
 asbestos  that determine biological activity and to qualitatively elucidate the nature of the corresponding
 relationship between exposure via inhalation and the induction of disease.                  '

 In this section, the existing animal inhalation studies are reviewed.  In the following section, a project
 undertaken to overcome the limitations of the existing animal inhalation studies is described (the
 supplemental inhalation study). Because this latter project was specifically designed to support the risk
 protocol  presented in this document, the nature and results of this project are described in  detail.

The existing animal inhalation database consists of approximately 30 studies of which approximately 20
contain dose-response information based on lifetime monitoring of exposed animals, including the work
 by: Bellman et al. (1986, 1995); Bolton et al. (1982); Davis et al. (1978, 1980, 1985, 1986a,d, 1988a,b);
                                              6.115

-------
Goldstein et al. (1983); Le Bouffant et al, (1987); Lee et al. (1981); McConnell et al. (1982); Muhle et al.
(1987); Platek et al. (1985); Smith et al. (1987); and Wagner et al. (1974,1982, 1985, 1987).  The studies
are similar in overall design, although differences in experimental details potentially affect the
comparability of results from separate studies.

In the inhalation studies, plugs formed from bulk samples of fibrous asbestos and related materials are
placed in a dust generator and aerosolized.  The generators (Beckett 1975), usually a modified version of
the apparatus originally designed by Timbrell et al. (1968), consist of a rotating brush that sweeps over an
advancing plug of bulk material and liberates fibers that are entrained in the controlled air flow passing
through the device. The airborne dust  is then passed either into a delivery system for nose-only exposure
or into an exposure chamber where animals are kept for fixed periods of time (usually 7 hours per day) on
a weekly routine (typically 5 days per week). The exposure routine is continued for as long as 2 years in
some of the studies. In some, but not all of the studies, fiber-containing air is passed through a cyclone or
elutriator prior to the exposure chamber so that exposure consists primarily of particle sizes within the
respirable range.

Asbestos concentrations in the animal inhalation experiments are monitored by a combination of
techniques.  The concentration of total dust in the chamber is generally monitored gravimetrically.
Simultaneously, membrane filter samples are collected and fibers counted by PCM.  The quotient of these
two measurements yields the number of (PCM) fibers per unit mass of dust (Section 4.3).  The
distribution of fiber sizes within the dusts introduced into the animal exposure chambers may also be
determined in these studies by any of a variety of methods. As indicated previously (Section 4.3),
however, the utility of such measurements  depends on the precise manner in which they are derived.

To derive fiber size distributions, dust  samples from these studies have generally been collected on
polycarbonate filters for analysis by SEM.  However, such distributions suffer both from the limitations
of SEM (Section 4.3) and from the manner in which they are tied to the inhalation experiments (see
Chapter 5).

Theoretically, the dose of any fiber size fraction can be estimated in a two-step process. The procedure
incorporates consideration of a size fraction termed the PCM-equivalent fraction (PCME), which is the
fraction of structures measured by SEM (or TEM) that correspond to the size range of structures known to
be visible and therefore countable by PCM. First, the concentration of the PCM-equivalent fraction of the
fiber size distribution (measured by SEM)  is normalized by dividing its value by the PCM-measured
concentration per unit dust mass observed  in the inhalation experiment. This ratio is then  multiplied by
the fractional concentration of any specified size range of interest within the distribution (measured by
SEM) to determine the exposure level  for that size fraction.  However, because bivariate (length by
diameter) size distributions have not typically been developed in the available studies and because the
number of total fibers longer than 5 jim observed by  SEM (without adjustment for width)  does not
correspond to the number of total fibers longer than 5 um observed by PCM, it is not possible to derive a
true PCME fraction from the SEM data. Therefore, the theoretical approach described above for
estimating exposure to specific size fractions cannot generally be applied in the existing studies.

Note that SEM analyses are typically conducted on limited dust samples only to provide information on
size distributions.  SEM is not used routinely to monitor daily asbestos concentrations in these
experiments. Therefore a procedure like that described above is required to link the absolute
concentrations to which rats are exposed to the measured and relative size distributions that are
determined by SEM.

As indicated above, the data within the published animal inhalation studies are further constrained by the
limitations of the analytical methods employed to generate the data  (Section 4.3). Comparison of data
between studies is also hindered by the lack of sufficient documentation to indicate the specific methods

                                               6.116

-------
                                                                                            s
and procedures employed in each study. Frequently, for example, it is unclear whether respirable dusts or
total dusts have been monitored. Also, several studies fail to report one or both of two critical pieces of
information: fiber-number-to-mass conversion factors and fiber size distributions. In addition, few
studies indicate the precise counting rules employed for generating size distributions.       j

When structure-number-to-mass conversion factors are provided, unless the conversion factor is derived
by counting fibers in a specific size range in a known mass of sample, and fiber concentrations in other
size ranges are normalized to this count, several types of error may be introduced. For example, if total
sample mass is assumed proportional to calculated mass derived from volume characterizations of the
particles counted, unless isometric particles are sized along with fibers and both asbestos and nonasbestos
particles are included in the count, a bias will be introduced in the conversion factor because total sample
mass will have been under-represented to the extent that such particles are ignored in the estimation of
fiber mass.  Even if such particles are included, significant uncertainty may result from estimating the
volumes of irregular particles and the limited precision associated with the count of the largest particles
(due to their limited number).  The uncertainty in the measurement of a fiber's diameter is squared in
contributing to the uncertainty associated with  a mass estimate.                            !

Among reported variations in study design, differences in the detailed design and operation of the
aerosolization chamber and the frequency and duration of exposure also potentially contribute to variation
in results between studies. Also, use of differing animal strains and species across the various studies
suggest the possibility that physiological differences may contribute to the observed variation in study
results. Such differences are discussed further  in Chapter 5.                               !

A small subset of the asbestos dusts evaluated in the animal inhalation studies have been analyzed by
TEM. However, even the published fiber size distributions from these TEM studies are subject to
variation from differences in procedures used for sample preparation, from differences in counting rules,
and from precision limitations due to the limited number of fibers actually characterized (Section 4.3).
This latter limitation particularly affects the precision with which longer fibers are counted. ',

Although fiber-size distributions are primarily based on SEM analyses rather than TEM analyses in the
existing animal inhalation studies, results generally echo the results of the injection and implantation
studies. Thus, longer fibrous structures are observed to contribute most to asbestos biological activity, at
least qualitatively. For example, dusts containing predominantly long amosite or long chryso'tile fibers
induce far more pulmonary tumors than samples containing predominantly short structures (Davis et a).
1986a,b). However, dusts evaluated in the existing inhalation experiments have not been characterized
sufficiently to distinguish the dependence of biological activity on fiber diameter. Neither are the existing
studies sufficient to evaluate the importance of mineralogy (or other potentially important asbestos
characteristics) in determining risk.

6.4.3   Supplemental Inhalation Study

Given the problems with the existing animal inhalation studies, a project was undertaken to overcome
some of the attendant limitations (Herman et al. 1995). To control for effects from variation in study
design and execution (including choice of animal strain, animal handling procedures, equipment design,
sample handling procedures, dosing regimen, and pathology protocols), the project focused on a set of
studies generated from  a single laboratory (i.e., the studies published by Davis and coworkers).
Ultimately, the results from six studies covering nine different asbestos samples (including fpiir types of
asbestos with samples exhibiting multiple size distributions for two asbestos  types) and a toteil of 13
separate experiments (some samples were studied at multiple exposure levels or in duplicate runs)  were
pooled for analysis. The database of experiments employed in the project is  described in Table 6-7.
                                              6.117

-------












3
HH
11
•g
e
U
•e
s
A
1
Q
>>
J3
"O
S
U
s
e
r )
.52
41
i
it
c
_g
"S
JB
s
»""l
"3
E
'£

1
S
£
i
s
5

t-^
4
H








a
B
£
>
o ? C «
1 ii '
|f ||
^ 2 * -
f-<< ^^ |M

"•« >>
e C „,
« S a a
* 'S ° B
g S S =
"5 fifi *3 £"•
z £.
u _w
lol
a a
Z <

s-s
u 1
a. w

ill?
s S S

H •—
U

8
jjjj

b
<


e
•3
B&
i_
V


U .,
II
rt
OO OO 00
r*" r*" oo
Os CT\ O\

• ^ ^
(S Q Q

— O fl




oo in o





cs oo cs






vo r~ oo




M O O


o o
m —  o ?3
Efl W) EC
> > >
Q Q Q

« o —




r*- i^1 ^5





^D OO ^D






— «•<*•




§ « ON


O O O
f*» NO f"*
« m vo
•-< rM CH
ON OV
O ON Gv







o o o
CO P P


|
1
CO
5

S8§
« P D
U U 4*
000
C/3 VI W
U U U
vg
oo
ON

13
0
m
>
5

O




oo





2






i ^ (^

•3 13 13
4) ^jj "^>
W W trt
> > >
QQQ

O m —




«N — O





O OO O






 cs" p>

0 O O







P J oo







8l?S
5^S

.« .s .-S
"S 'K *J3
<; < <
00 00
r~ p-


rt Q
"S "55
M M
> >
0 13
O O

— o





Q O Q

rs o o




00 00





VO O O






IN O O




O\ O NO
' fN C^ f*^


1


o o o







^. o u







1 II
w
I 1 1
<3 .£> 03
**S ^O OO
OO OO OO
Ch ON ON

13 73 13
4) ^[> "^
tfl OT OT
s s s
Q Q Q

O 0 O




fS fS f>»





- - -






— — —




\O VO "tj"





o o o







u u o







s g s
1 § I
u u u

AJ 
£?
5
f
u
1
u
•s
u
u
3
U
"

2
o
1
«3
g
00
.£
_o
^
^-i
•*?•
_g
**-j
•3
S
S
JS
o
u
o
1/3
a
€
*s
•o
o
y
1
Cu
Sur^
'I
_£>


S "3
>. o.
« T3
81
^•§
!S s
 s
« ?
tl
' MM
„ D,"°
v> O ^ ^
O\ -G m ^
"sli
13-3-S 5
rt 3 «) 'T-*
c§ u^^-
g §51
,'




i,


!







u '
| ^

c
.2 .

V)
^ i
~V i'
>
•a '
'5
u
| .
1 :

s» '
i '
S,
>. '
5
u '
is ,
£ '
E •-,
«s .
•s ,
.s
£ ,
o
E •
U V
S :•
s i-
E '
1'
g
i§ '
c
§
'^
^
1
BO ,
-o
Q
**-« :
"8
irt
3 '
1
1
o>

-------
In the statistical analysis performed in this study, the individual dose-response profiles from each
of the two data sets were fit to a linear dose-response model:
                                                                                 (Eq.6-7)
where:
       "Pi"    is the probability of inducing pulmonary tumors observed in the "ith" study;

       "Q0"   is a parameter that accounts for the background incidence of pulmonary tumors
              (assumed to be the same in all studies);

       "Xy"    is the concentration of the "jth" size fraction of fibers in the "ith" study;

       "aj"    is the coefficient of potency for the "jth" size fraction of fibers; and

       "b"    is a coefficient that represents the absolute potency of asbestos. In some analyses
              this coefficient is allowed to assume different values for different types of
              asbestos (e.g., chrysotile vs. amphibole), or other differences in experimental
              conditions.

The "aj"s in this analysis are constrained to be positive because it is assumed that no fiber
prevents cancer. The "aj"s are also constrained to sum to 1.0, which means that they represent
relative potency rather than absolute potency. Asbestos size fractions evaluated represent
disjoint (mutually exclusive) sets.

The model (Equation 6-7) allows separate potency coefficients to be assigned to individual size
fractions in a dose-response relationship that depends on multiple size fractions.
Simultaneously, the "b" coefficients allows separate potencies to be assigned to different fiber
types or to results from different studies performed under different experimental conditions.

Several investigators (Bertrand and Pezerat 1980; Bonneau et al.  1986; Stanton et al. 1977;
Wylie et al. 1987) have used a logit curve to investigate the dose-response relating various
measures of asbestos exposure to tumor response. The logit formula specifies that the tumor
probabilities satisfy the relation:
                                  log[P/(I-P)]=a + b-logx
(Eq. 6-8)
 where log x is some measure of asbestos exposure, such as log of concentration of fibers in some
 size range. In some instances, the logit model was expanded by replacing b-log x with a term
 representing a linear combination of exposure indices, so that multiple exposure indices could be
 explored simultaneously.  The models were fit using standard linear regression based on normal
 theory.
                                           6.120

-------
 An equivalent form for the logit model is:
                                     P=e"xb/(l+e8xb)
  (Eq.6-9)
 Written in this form, it is clear that this model does not permit a background response (i.e., P=0,
 whenever x=0).  This is not a serious limitation when there are no tumors in control iinimals,
 such as was the case in Stanton et al. (1977). However, the model will not adequately fit data in
 which tumors are found in control animals.  This was one reason for adopting the linear model
 (Equation 6-7) used in the investigation of the animal data reported in the study described here.
                                                                             i
                                                                             I
 There is no evidence from this study that the linear model is inadequate.  For cases in this study
 in which the fit between exposure and response is shown to be inadequate, the lack of fit is
 typically observed to be due to an inconsistent (non-monotonic) dose-response curves so that
 there is no indication that a non-linear model, such as the logit, would provide a better fit.

 The linear model (Equation 6-7) used in this study was fit using a maximum likelihood (Cox and
 Lindley 1974) approach that utilizes the actual underlying binomial probabilities. This is a more
 efficient estimation method than use of regression methods based on normal theory, which was
 the fitting method used in the earlier studies (described above).  In addition, the regression
 procedures applied in the earlier studies indicate only whether the exposure measures'that were
 studied are significantly correlated with tumor response. In contrast, statistical goodness of fit
 tests were applied in this  study to determine whether exposures that are described by a particular
 characteristic (or combination of characteristics) satisfactorily predict the observed turner
 incidence. To illustrate, it is apparent from Text Figure 2 of Stanton et al. (1981) that the
 exposure measure they identify as being most highly correlated with tumor incidence (fibers
 longer than 8 jim and thinner than 0.25 urn) does not provide an acceptable fit to the observed
 tumor incidence.  Similarly, although all of the univariate exposure measures listed in Table 2 of
 Berman et al. (1995) are highly correlated with tumor incidence, none of them adequately
 describe (fit) lung tumor incidence.

 To test for goodness-of-fit in this study, each relationship was subjected to a chi-square
 goodness-of-fit test in which the fit of the model was rejected if the corresponding p-value was
 less than 0.05, indicating that the true model would provide a worse fit only 5% of the time.
 Among models that were not rejected based on a goodness-of-fit test, several hypotheses
 concerning the relative merit of the various models were also examined using the method of
 maximum likelihood (Cox and Lindley 1974).

 An example of an adequate fit to the tumor response data is provided in Figure 6-5 (Figure 3 of
Berman et al. 1995).  Note that Figures 2 and 3 of the original paper were inadvertently switched
during publication; the correct Figure 3 is reproduced here. The exposure index plotted in
Figure 6-5 is the sum:
               Exposures, x^ + a2 x,2 + a3 x^O.OOHC, + 0.853C2 + 0.145C3
(Eq. 6-10)
                                          6.121

-------
s^
e?
••<•
11
ii
•*-" o
•= I
i. -S
*• a
CO (
81
S ^
**
li
 o
 g
2 8P
E^5
 . u
«? J3
^ to
2 g
01
CM
        if!   .•- '„ i   f"  8-  8

-------
where:
       "C,"   (= Xj,) is the concentration of structures between 5 and 40 \im in length that are
              thinner than 0.3 u.m;                                             :
        'C2"   (= Xj2) is the concentration of structures longer than 40 urn that are thinner than
              0.3 [im; and                                                    ;•
        C3"   (= Xj3) is the concentration of structures longer than 40 um that are thicker than 5
This index of exposure represents one of the optimum indices reported in Berman et al. 1995.

As is clear from the figure, when exposure is expressed in the manner described above, the
tumor responses observed in the 13 separate experiments that were evaluated increase
monotonically with increasing exposure. It is also apparent that the data points representing
each study fall reasonably close to the line representing the optimized model for this exposure
index.  This was verified by the fact that the corresponding goodness-of-fit test p-value was
greater than 0.05. Thus, exposure adequately predicts response.

Results obtained from completing more than 200 statistical analyses to determine whether
various measures of asbestos exposure adequately predict lung tumor response (Berman et al.
1995) indicate that:

       •      neither total dust mass nor fiber concentrations determined by PCM adequately
              predict lung tumor incidence;

       •      no univariate  measure  of exposure (i.e., exposure represented by the
              concentration of a single size category of structures as measured by TEM) was
              found to adequately predict lung tumor incidence. Of the univariate measures of
              exposure examined, the concentration of total structures longer than 20 um
              provides the best fit (although still inadequate); and

       •      lung tumor incidence can be adequately predicted with measures of exposure
              representing a weighted sum of size categories in which longer structures are
              assigned greater potency than shorter structures.

The set of analyses completed in support of this work are summarized in Appendix C.
One example of an exposure measure that adequately describes lung tumor incidence is
presented in Figure 6-5. Another exposure index shown to provide an adequate fit is: '•
                                   0.0024Ca + 0.9976Cb
(Eq.6-11)
                                          6.123

-------
where:
       "Ca"   is the concentration of structures between 5 and 40 jam in length that are thinner
              than 0.4 urn; and

       "Cb"   is the concentration of structures longer than 40 |im that are thinner than 0.4 (Am.

The fit of this index is depicted in Figure 6-6.

In addition to the above, a series of hypotheses tests were also conducted to test such questions
as: whether fiber type affects potency or whether the component fibers in complex clusters and
matrices should be counted individually. Questions concerning whether mesothelioma incidence
can be adequately described by the same measure(s) of exposure that describe lung tumor
incidence were also addressed. Taken as a whole, the results presented in Herman et al. (1995)
support the following general conclusions:

       •      structures contributing to lung tumor incidence are thin (<0.5 \im) and long (>5
                  with structures longer than 20 \im being the most potent;
              the best estimate is that short structures (<5 u.m) are non-potent.  There is no
              evidence from this study that these structures contribute anything to risk;

              among long structures, those shorter than 40 urn appear individually to contribute
              no more than a few percent of the potency of the structures longer than 40 urn;

              lung tumor incidence is best predicted by measurements in which the component
              fibers and bundles of complex structures are individually counted;

              at least for lung tumor induction in rats, the best estimate is that chrysotile and the
              amphiboles are equipotent;

              for equivalent size and shape structures, amphiboles are more potent toward the
              induction of mesothelioma than chrysotile; and

              after adjusting for the relative potencies of fiber type, the size categories that
              contribute to lung tumor incidence appear also to adequately describe
              mesothelioma incidence.
                                          6.124

-------
a S
.2 3.
«-*.
8?
« v
8.3
u »
  rf
& s
« a.
If
u  ^
in
CM
CO

-------
A number of supplemental analyses were also conducted, primarily to identify optimal
procedures for performing asbestos analysis and for estimating concentrations.  These analyses
have not yet been published, but they are included in the summaries in Appendix C. The most
important results of the supplemental analyses are that:

       •      tumor incidence can only be adequately fit by data derived from TEM analysis of
              samples prepared by a direct transfer procedure.  Measurements derived from
              indirectly prepared samples could not be fit to lung tumor incidence in any
              coherent fashion; and

       •      it was not possible to identify an exposure measure in which potency is expressed
              in terms of a single, continuous function of structure length.

Regarding the last point, although we were not able to identify a continuous function of length
that provides an adequate fit to the tumor incidence data, the general results from the above
analysis are not inconsistent with the hypothesis that potency is a continuous function of length
(i.e., the Pott hypothesis, Pott 1982). The Pott hypothesis suggests that relative potency is low
for short fibers, rises rapidly over an intermediate range of length, and approaches a constant for
the longest fibers.

6.4.4  Conclusions  Concerning Animal Dose-Response Studies

Results from our evaluation of the animal dose-response data for asbestos (including the existing
injection/implantation studies, the existing inhalation studies, and our supplemental study)
indicate that:

        •      short structures (less than somewhere between 5 and 10 urn in length) do not
              appear to contribute to cancer risk;

        •      beyond a fixed, minimum length, potency increases with increasing length, at
              least up to a length of 20 |am (and possibly up to a length of as much as 40
       •     the majority of structures that contribute to cancer risk are thin with diameters
              less than 0.5 urn and the most potent structures may be even thinner. In fact* it
              appears that the structures that are most potent are substantially thinner than the
              upper limit defined by respirability;

       •     identifiable components (fibers and bundles) of complex structures (clusters and
              matrices) that exhibit the requisite size range may contribute to overall cancer risk
              because such structures likely disaggregate in the lung. Therefore, such structures
              should be individually enumerated when analyzing to determine the concentration
              of asbestos;

       •     for asbestos analyses to adequately represent biological activity, samples need to
              be prepared by a direct-transfer procedure; and
                                           6.126

-------
       •     based on animal dose-response studies alone, fiber type (i.e., fiber mineralogy)
              appears to impart only a modest effect on cancer risk (at least among the various
              asbestos types).

Regarding the last of the above bullets, that only a modest effect of fiber mineralogy ;was
observed in the available animal dose-response studies (when large effects are observed among
human studies, Chapter 7), may be due at least in part to the limited lifetime of the rat relative to
the biodurability of the asbestos fiber types evaluated in these studies, although it is also possible
that different mechanisms drive the effects observed in the animal studies than those .that
dominate for asbestos-induced cancers in humans and that such mechanisms depend more
strongly on mineralogy. Other explanations are also possible.  Issues relating to botK fiber
mineralogy and fiber size are addressed further in Section 6.5 and Chapter 7.

6.5    CONCLUSIONS FROM AN EVALUATION OF SUPPORTING STUDIES

Although gaps in knowledge remain, a review of the literature addressing the health-i-elated
effects of asbestos (and related materials) provides a generally consistent picture of the
relationship between asbestos exposure and the induction of lung cancer and mesothelioma.
Therefore, the general characteristics of asbestos exposure that drive the induction of cancer can
be inferred from the existing studies and  can be applied to define appropriate procedures for
evaluating asbestos-related risk. Furthermore, although it would be helpful to definitively
identify the underlying biochemical triggers and associated mechanisms that drive asbestos-
induced cancer, this is not an absolute prerequisite for the development of a technically sound
protocol for assessing asbestos-related risk.

As previously indicated (Sections 6.3.8),  the biochemical mechanisms that potentially contribute
to the induction of asbestos-induced cancer are complex and varied. Moreover, different
mechanisms appear to exhibit differing dose-response characteristics (i.e., the various
mechanisms do not all show the same kind of dependence on fiber size or fiber type). Some
mechanisms, for example, suggest that fiber length is important and that only structures that are
sufficiently long induce a response.  In contrast, other mechanisms suggest that fibers (and even
non-fibrous particles) may all contribute  to response and that the magnitude of the response is a
function of the total surface area of the offending fibers (or particles). Among these ;!
mechanisms,  additionally, some suggest that fiber type (i.e., mineralogy) is not an important
determinant of potency while other mechanisms indicate that fiber type is an important
determinant of potency.                                                      '

The existing studies are not currently adequate to support definitive identification of the specific
mechanisms that drive the induction of asbestos-related cancer  (versus other mechanisms that
may contribute only modestly or not at all). However, whatever mechanisms in fact contribute
to the induction of disease, they must be consistent with the gross characteristics of exposure that
are observed to predict response in the available whole-animal dose-response studies1 and human
epidemiology studies. Therefore, the implications from these latter studies regarding  the
dependence of asbestos-induced cancers on fiber size and type are reviewed here in some detail.
Further, in Chapter 8, they are used to support development of a protocol for evaluating
asbestos-associated risks.
                                         6.127

-------
Fiber Length.  Fibers less than a minimum length between 5 and 10 \im do not appear to
contribute to risk. This is supported both by the results of our re-analysis of the animal
inhalation studies conducted by Davis and coworkers (Section 6.4.3 and Berman et al. 1995), in
which this hypothesis was tested formally, and by inferences from the broader literature. As
long as fiber size is adequately characterized, the animal inhalation studies (Section 6.4.2) and
injection/implantation (Section 6.4.1) studies consistently indicate lack of ability of short
structures to contribute to the induction of cancer. Furthermore, animal retention studies
(Section 6.2.1) and histopathology studies (Section 6.2.2) provide strong mechanistic evidence
that explains the lack of potency for short structures; they are readily cleared from the respiratory
tract. Even when sequestered in large numbers in macrpphages within the lung, there is little
indication that such structures induce the kinds of tissue damage and related mechanisms that
appear to be closely associated with the induction of cancer.

Although there are mechanism studies that may suggest a role for short fibers in the induction of
asbestos-related disease (see, for example, Goodglick and Kane 1990, Section 6.3.4.4), such
studies dp not track cancer as an endpoint.  Therefore, the relationship between the toxic
endpoint observed and the induction of cancer needs to be adequately addressed before it can be
concluded definitively that short structures can contribute to cancer.  Moreover, such a
conclusion would be surprising given the substantial evidence that exists to the contrary.

Beyond the minimum length below which structures may be non-potent, potency appears to
increase with increasing length, at least up to a length of 20 ^m and potentially up to  a length of
40 um.  The latter limit is suggested by our re-analysis of the Davis et al. studies (Section 6.4.3)
in which it was also found that structures longer than 40 p.m may be as much as 500 times as
potent as those between 5 and 40 u.m in length.  The former  limit is suggested by broader
inferences from the literature that suggest the cutoff in the length of structures that are at least
partially cleared by macrophages from the lung may lie close to 20 ^m and that the efficiency of
clearance likely decreases rapidly for structures between 10  and 20 urn in length (Section 6.2).
Such inferences are further reinforced by measurements of the overall dimensions of
macrophages in various mammals by Krombach et al. (1997), as reported in Section 4.4.

Importantly, the inferences that potency increases for structures longer than 10 urn (up to some
limiting length) from these various studies are strongly reinforcing, even though  the upper limits
to the points at which potency stops increasing do not precisely correspond. Furthermore, that
the longest structures are substantially more potent than shorter structures (and that the shortest
structures are likely non-potent) dictates that asbestos analyses performed in support  of risk
assessment need to provide adequate sensitivity and precision forcounts of the longest
structures.

Fiber Diameter. Because fibers that contribute to the induction of cancer must be respirable,
they must also be thin. The studies reviewed in Section 6.1  indicate that respirable fibers are
thinner than 1.5 \im and the vast majority of such structures are thinner than 0.7 |im.  In fact, the
results of injection, implantation,  and inhalation studies reviewed in Sections 6.4.1 and 6.4.2 and
the results of our supplemental  re-analysis of the Davis et al. studies (Section 6.4.3) indicate that
the fibers that contribute  most to the induction of asbestos-related cancers are substantially
thinner than the limit suggested by respirability alone.
                                           6.128

-------
Importantly, the results of all of the studies cited above indicate that it is a cutoff in absolute
width that defines the bounds of biological activity rather than a cutoff in aspect ratio (the ratio
of length to width) that has been used to define fibrous structures heretofore. That is why the
exposure index recommended based on our review of the studies by Davis and coworkers
incorporates a maximum width as a cutoff, rather than a minimum aspect ratio.      •'

Fiber Complexity. In our supplemental evaluation of the Davis et al. studies (Section 6.4.3), the
tumor incidence data from the animal inhalation studies were best fit (predicted) by exposure
indices in which the component fibers and bundles of complex structures (clusters and matrices)
were separately enumerated and included in the exposure index used to represent concentration
(Section 6.4.3). The appropriateness of such an approach is further supported by theiobservation
that loosely bound structures (including, for example, chrysotile bundles) readily disaggregate
in vivo (Section 6.2). Therefore, it is recommended in this report that those components of
complex structures that individually exhibit the required dimensional criteria be individually
enumerated and included as part of the count during analyses to determine the concentration of
asbestos in support of risk assessment.                                          ,'

Fiber Type (Mineralogy). The magnitude of any effect of mineralogy upon cancer risk in
rodents appears to  be modest at best. On the other hand, mineralogy appears to be an important
determinant for cancer risk in human epidemiology  studies (Chapter 7), with chrysotile
appearing less potent than amphibole for inducing mesothelioma and (with lesser certainty) lung
cancer.  This difference may be due to differences in the life spans of rats and humans compared
to the differential biodurability of the different fiber types. It must also be emphasized that, due
to confounding, the effects of fiber size and fiber mineralogy need to be addressed  '
simultaneously, if one is interested in drawing useful conclusions concerning fiber mineralogy.
                                                                             i: •
Results from some (but not all) of the animal injection, implantation, and inhalation studies
previously reviewed (Sections 6.4.1  and 6.4.2) suggest that mineralogy plays an important role
in determining biological activity. However, the nature of the effects of mineralogy are not
easily separated from size effects, due to the methodological limitations of the studies cited.
Therefore, the evidence from these studies can be considered ambiguous. Formal hypothesis
testing during the re-analysis of rat inhalation studies (Section 6.4.3) indicates that, when size
effects are addressed, chrysotile and the amphiboles exhibit comparable potency toward the
induction of lung cancer. In contrast, amphiboles were estimated to be approximately 3 times
more potent than chrysotile toward the induction of mesothelioma, once fiber size effects are
addressed.

Several of the human pathology studies cited previously (Section 6.2.3) suggest that mineralogy
is an important factor in determining cancer risk, but these studies similarly suffer from
methodological difficulties that introduce ambiguity into the inferences drawn.  However, it is
clear from the human epidemiology data (Chapter 7) that mineralogy plays a substantial role in
the determination of risk for human cancer (primarily, mesothelioma).

The underlying cause(s) for the observed difference in potency between chrysotile and the
amphiboles may relate to differences in fiber durability (Section 6.2), to size/shape related
differences in fibers that are a function of mineralogy and that cause differences in deposition,
retention, or translocation (Sections 6.1  and 6.2), and/or to the dependence on mineralogy of the

                                          6.129

-------
specific mechanisms underlying the biological responses of specific tissues (Section 6.3). The
relative magnitudes of such effects on animal and human pathology also need to be considered,
if the observed differences in potency among animal and human studies, respectively, is to be
reconciled. Such considerations are addressed further in Chapter 8.

Importantly, whether the observed differences in the role of mineralogy toward animal and
human pathology can be reconciled, the effects of mineralogy can be adequately addressed when
assessing asbestos-related cancer risk for humans by incorporating dose-response coefficients
explicitly derived from the human epidemiology data. Because this is the approach proposed in
this document, effects due to mineralogy are properly addressed.

An Appropriate Exposure Index for Risk Assessment. The optimum exposure index defined
based on the re-analysis of the animal inhalation studies conducted by Davis and coworkers
(Section 6.4.3) is a weighted sum of the concentrations of (1) structures between 5 and 40 um in
length that are thinner than 0.4 UJTI, and (2) structures longer than 40 um that are thinner than 0.4
urn (Equation 6-11).

This index was shown to adequately fit (predict) the tumor incidence data across the 13 separate
animal inhalation experiments evaluated (P=0.09).  Whether the exposure index defined in
Equation 6-11 is also optimal for capturing the relevant characteristics of fibers that contribute to
the induction of human cancer is an open question. Because it captures the major characteristics
(concerning length and diameter) identified above that are indicated to be important for human
exposures, it represents a promising candidate.  Unfortunately, however, the data required to
match this index to a set of human-derived exposure-response coefficients does not currently
exist (Section 7.4). Therefore, compromises are required to apply the general conclusions of this
Chapter to the human data. These are addressed further in Section 7.4.1.
                                          6.130

-------
                   7.0.  EPIDEMIOLOGY STUDIES

The existing epidemiology studies provide the most appropriate data from which to determine
the relationship between asbestos exposure and response in humans. As previously indicated,
however, due to a variety of methodological limitations (Section 5.1), the ability to compare and
contrast results across studies needs to be evaluated to determine the confidence with' which risk
may be predicted by extrapolating from the "reference" epidemiology studies to new
environments where risk needs to be assessed. Reliable extrapolation requires both that the
uncertainties contributed by such methodological limitations and that several ancillary issues
(identified in Chapter 2) be adequately addressed.

A detailed discussion of the methodological limitations inherent to the available epidemiology
studies was provided in the Health Effects Assessment Update (U.S. EPA 1986) and additional
perspectives are provided in this document (Section 5.1). The manner in which the uncertainties
associated with these limitations are addressed in this document are described in Appendix A.
As previously indicated (Chapter 2), the ancillary issues that need to be addressed include:

       (1)     whether the models currently employed to assess asbestos-related risk' adequately
              predict the time and exposure dependence of disease;
                                                                            i
       (2)     whether different mineral types exhibit differential  potency (and whether any
              differences in potency relate to the relative in vivo durability of different asbestos
              mineral types);

       (3)     whether the set of minerals included in the current definition for asbestos
              adequately covers the range of minerals that potentially contribute to asbestos-
              related diseases; and

       (4)     whether the analytical techniques and methods used to characterize exposures in
              the available epidemiology studies adequately capture the characteristics of
              exposure that affect biological  activity.

All but the third of the above issues are addressed in this chapter.  Currently, the third issue can
best be addressed by evaluating inferences from the broader literature (see Chapter 6). The
remaining issues are addressed separately for  lung cancer and mesothelioma following a brief
overview of the approach adopted for evaluating the epidemiology literature.

7.1    APPROACH FOR EVALUATING THE EPIDEMIOLOGY LITERATURE

To develop exposure-response relationships (and corresponding exposure-response coefficients)
for use in risk assessment from epidemiological data, two basic types of information are
necessary: information on the disease mortality experienced by each member of the study
population (cohort) and information on the asbestos exposure experienced by each member of
the cohort.  So that disease mortality attributable to  asbestos can be distinguished from other
(background) causes of death, it is also necessary to have knowledge of the rates of mortality
                                          7.1

-------
that would be expected in the study population, absent exposure. Normally, such information
must be determined based on a "reference" or "control" population.

Ideally, one would like to have complete knowledge of exposure at any period of time for each
individual in the cohort and complete access to the data to fit different types of
exposure/response models to the data so that the approach for evaluating the relationship
between exposure and response can be optimized. In most instances, unfortunately, the data
suffer from multiple limitations (see Section 5.1 and Appendix A) and the analysis is further
constrained by less than complete access to the data.

Briefly, the major kinds of limitations that potentially contribute to uncertainty in the available
epidemiology studies (and the effect such limitations likely produce in estimates of exposure-
response coefficients) include:

       •      limitations in the manner that exposure concentrations were estimated
              (contributing to variation across studies);

       •      limitations in the manner that the character of exposure (i.e., the mineralogical
              types of fibers and the range and distribution of fiber dimensions) was delineated
              (contributing to systematic variation between industry types and, potentially,
              between fiber types);

       •      limitations in the accuracy of mortality determinations or incompleteness in the
              extent of tracing of cohort members (contributing to variation across studies);

       •      limitations in the adequacy of the match  between cohort subjects and the selected
              control population (contributing to variation across studies and may have a
              substantial effect on particular studies); and

       •      inadequate characterization of confounding factors, such as smoking histories for
              individual workers (contributing to variation across studies and may have a
              substantial effect on particular studies).

More detailed discussion of the above limitations is provided in Section 5.1. The manner in
which these limitations are being addressed in this evaluation are described briefly below and in
more detail in Appendix A.

The existing asbestos epidemiology database consists of approximately 150 studies of which
approximately 35 contain exposure data sufficient to derive quantitative exposure/response
relationships. A detailed evaluation of 20 of the most recent of these studies, which includes the
most recent follow-up for all of the cohorts evaluated in the 35 studies, based on the
considerations presented in this overview, is provided in Appendix A.

This new analysis of the epidemiology database differs  from the evaluation conducted in the
1986 Health Effects Assessment Update (U.S. EPA 1986) in several ways.  It incorporates new
studies not available in the  1986 update that contain information on exposure settings not
previously evaluated as well as more recently available follow-up for exposure settings

                                            7.2

-------
previously evaluated. It also incorporates new features in the manner in which the analysis was
conducted. These new features include:

       •     estimation of "uncertainty" bounds for the exposure-response coefficients
              (potency factors) derived from each study; and                     ;

       •     for the lung cancer model,  introducing a parameter, a, which accounts for the
              possibility that the background lung cancer mortality rate in the asbestos-exposed
              cohort differs systematically from the rate in the control population.  :

The uncertainty bounds were developed to account for uncertainty contributed by the manner
that exposure was estimated, by the manner that work histories were assigned, by limitations
imposed by the manner in which results were reported in published papers, and by limitations in
the accuracy of follow-up, in addition to accounting for the statistical uncertainty associated with
the observed incidence of disease mortality.  A detailed description of how the uncertainty
bounds were constructed is provided in Appendix A.                             i

Exposure-response coefficients were estimated for each cohort both by requiring that a=l (the
approach followed in the 1986 Health Effects Assessment Update, U.S. EPA 1986) and by
allowing a to vary while fitting the lung cancer model to data. Allowing a to vary addresses
potential problems due to differences in background lung cancer rates between the cohort and the
control population due, for example, to differences in smoking habits in the two populations. As
indicated in Appendix A, this adjustment has a substantial effect on the fit of the U.S. EPA
model to the data for several specific cohorts and a corresponding effect on the estimates of the
lung cancer exposure-response coefficients for those cohorts.                     j

We were also able to obtain the original, raw data for selected cohorts from a limited number of
the more important of the published epidemiology studies.  This allowed us to more formally
evaluate the appropriateness of the existing U.S. EPA models for lung cancer and mesothelioma
(Sections 7.2 and 7.3).

Exposure-response coefficients, and corresponding risk estimates derived therefrom,'must be
based upon an "exposure index" that expresses the relative potency of asbestos fibers of different
dimensions.  For example, the exposure index utilized in the 1986 Health Effects Assessment
Update (U.S. EPA 1986) assigns equal potency to all fibers longer than 5 u.m that exhibit an
aspect ratio >3 and a thickness >0.25 ^m,  regardless of type of asbestos, and assigns zero
potency to shorter, squatter, or thinner  fibers. In this update, we evaluated a range of such
exposure indices, both with respect to agreement with evidence from the literature on the relative
potency of asbestos structures of differing types and dimensions (Section 7.4) and with respect to
overall agreement across the exposure-response coefficients derived from the available
epidemiology studies and adjusted for exposure index. This analysis led to proposed new
exposure indices that better reflect the evidence from the literature on the relative potency of
different structures and provide improved  agreement among exposure-response coefficients
estimated from different environments. Such improvement in agreement across studies
correspondingly increases the confidence with which the exposure-response factors derived from
the existing studies can be applied to new  environments.
                                           7.3

-------
7.2    LUNG CANCER

The 1986 U.S. EPA lung cancer model (U.S. EPA 1986) assumes that the relative risk, (RR), of
mortality from lung cancer at any given age is a linear function of cumulative asbestos exposure
in units of fiber-years/ml (f-y/mf) as measured by PCM, disregarding any exposure in the most
recent ten years. This exposure variable is denoted by CE10, and its use embodies the assumption
that asbestos exposures during the most recent 10 years do not affect current lung cancer
mortality risk. The mathematical expression for this model is:

                                  RR=l+KL*CEi0                          (Eq.7-1)

where the linear slope, KL, is termed the "lung cancer exposure-response coefficient." This
parameter is generally estimated by fitting the model to data from an occupational mortality
cohort  study consisting of observed and expected numbers of cancer deaths categorized by
cumulative exposure, with the expected numbers determined from age- and calendar-year-
specific lung cancer mortality rates from an appropriate control population (e.g., U.S. males).
This model predicts that the mortality rate in an asbestos-exposed population is the product of
the mortality rate in an unexposed, but otherwise comparable, population, and the RR.
Consequently the excess mortality due to asbestos is the product of the background mortality rate
and the excess RR, KL*CEi0.  Since smokers have a higher background mortality from lung
cancer, the model predicts a higher excess asbestos-related lung cancer mortality in smokers than
in non-smokers.

To account for the possibility that an occupational cohort may have a different background
mortality rate of lung cancer than the control population (e.g., due to different smoking habits or
exposures to other lung carcinogens), in the present analysis Equation 7-1 is expanded to the
form,

                               RR = a * (1+ KL * CE,0)                        (Eq. 7-2)

where a is the RR in the absence of asbestos exposure relative to the control population. This
form of the model contains two parameters, the background RR,  a, and the lung cancer
exposure-response coefficient, KL.

7.2.1   The Adequacy of the Current U.S. EPA Model for Lung Cancer

Access to the raw epidemiology data from two key studies allowed us to evaluate the adequacy
of the U.S. EPA model (Equation 7-2) for describing the time and exposure dependence for lung
cancer in asbestos-exposed cohorts. For this analysis, the raw data for the cohort of crocidolite
miners in Wittenoom, Australia was graciously provided by Dr. Nick de Klerk (de Klerk 2001)
and the raw data for the cohort of chrysotile textile workers (described by Dement et al. 1994)
was graciously provided by Terri Schnoor of NIOSH (Schnoor 2001). The Wittenoom cohort
was originally described by Armstrong et a!. (1988), but the data provided by de Klerk includes
additional follow-up through 1999.
                                          7.4

-------
 7.2.1.1
Exposure Dependence
To evaluate the adequacy of the linear exposure response relationship assumed by the U.S. EPA
lung cancer model, the lung cancer model (Equation 7-2) was fit to the raw data from both the
South Carolina and Wittenoom cohorts.  In these analyses, each person-year of follow-up was
categorized by cumulative exposure defined using a lag of 10 years.  The data were then grouped
into a set of cumulative exposure categories and the observed and expected numbers pf lung
cancers were computed for each category. For South Carolina, expected numbers were based on
sex-race-age- and calendar-year-specific U.S. rates. Separate analyses were conducted for white
males, black males, and white females, as well as for the combined group. For Wittenoom
expected rates were based on age- and calendar-year-specific rates for Australian  men. The
categorized data and the resulting fit of the model to the data from Wittenoom and South
Carolina are presented in Tables 7-1 and 7-2, respectively.1                        '

For the Wittenoom data (Table 7-1), the fit with a=l is poor (pO.OOOl), as the model
overpredicts the number of cancers in the highest exposure category and underpredicts at lower
exposures. By contrast the fit of the model with a variable is adequate (p=0.1). This model
predicts a relatively high background of lung cancer in this cohort relative to Australian men in
general («=2.1) and a correspondingly shallow slope, with the RR increasing only from 2.1 at
background to 3.6 in the highest exposure category. A test of the hypothesis that  a=l for this
cohort is rejected (p<0.01) and the model fit to the data (with a variable) predicts  Ki=f0.0047
(f/ml-yr)-1.

The fit of the U.S. EPA model to the South Carolina lung cancer data categorized by cumulative
exposure is shown in Table 7-2. The model with cc=l cannot be rejected both when the model is
applied to white males only (p=0.54) or with all data combined (p=0.92). Since the values « and
KL estimated from white males only are similar to those estimated using the complete,cohort
(black and white males and white females), the fit to the complete data is emphasized in this
analysis. This fit predicts a=1.2 and 1^=0.021 (f-y/ml)'1. A test of the hypothesis that «=1 for
this cohort cannot be rejected (p=0.21), and in this fit KL=0.028 (f-y/miy1
        The fitting of the lung cancer model to the cohort data was carried out by assuming that the observed
numbers of cancers in different exposure categories were independent, with each having a Poisson distribution with
expectation equal to the expected number based on the control population times the relative risk predicted by
Equation 7-2. In the computation of the relative risk for a cumulative exposure category, the person-year-weighted
average cumulative exposure (lagged  10 years) for the category was used to represent the exposure in that category.
With these assumptions, a and KL were estimated by the method of maximum likelihood, confidence intervals were
constructed using the profile likelihood method, and likelihood ratio tests were used to test hypotheses (Cox and
Oakes 1984;  Venson and Moolgavkar 1988).                                                   *>

                                            7.5

-------
 Table 7-1. Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among
   Wittenoom, Australia Miners (Deklerk 2001) Categorized by Cumulative Exposure
                                Lagged 10 Years
Cumulative Exposure Lagged
10 Years (f-y/ml)
Range
0
0-0.4
0.4-1.0
1.0-2.3
2.3-4.5
4.5-8.5
8.5-16
16-28
28-60
60+
Total

Test of H0: a=l
p<0.01
Average
0
0.19
0.69
1.6
3.3
6.2
11.8
21.5
41.1
142.0



Observed
Deaths
5 '
27
11
22
28
38
31
21
25
43
251
Goodness of Fit

Expected
Deaths
4.6
7.9
8.2
11.6
12.9
14.3
13.2
9.2
11.6
11.6
105.1
P-value

Predicted Deaths
by Model
(0=1)
4.6
8.0
8.3
12.1
14.0
16.7
17.4
14.5
24.5
56.5
176.6
<0.0001

(«=2.1)
9.8
17.0
17.6
24.9
27.9
31.4
29.8
21.6
29.6
41.6
251.0
0.1

Estimates of KL (f-y/ml)'1
KL=0.027
90% CI: (0.020,0.035)
      [MLE])
KL=0.0047
90% CI: (0.0017,0.0087)
                                       7.6

-------
  Table 7-2. Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among
   South Carolina Textile Workers (Schnoor 2001) Categorized by Cumulative Exposure
                                    Lagged 10 Years
Cumulative Exposure Lagged
10 Years (f-y/ml)
Range
0
0-0.5
0.5-1.0
1.0-2.5
2.5^4.5
4.5-8.5
8.5-16
16-28
28-60
60-80
80-110
110+
Total
Average
0
0.35
0.75
1.7
3.4
6.2
11.8
21.3
41.5
69.2
93.3
173

Observed
Deaths
0
4
• 7
10
13
11
11
8
14
9
10
25
122
\
Expected
Deaths
0.9
2.6
5.5
9.7
7.9
7/7
7.2
5.3
6.3
3.0
•2.9
4.3
63.4
Predicted Deaths
by Model
(a=l)
0.9
2.7
5.6
10.1
8.6
9.1
9.6
8.4
13.8
9.0
10.8
25.6
114.1
' j[a=1.2)
' ' 1.1
, 3.2'
6.8
12.1
' f ,0.2
.' 10.5 '
10.9
i
, 9.2 •
, 14.3
9.0
1 10.5
24.2
122.0
                              Goodness of Fit P-value
                     0.92
• 0.96
 Test of H0: «=1
 p=0.21

 Estimates of KL (f-y/ml)'1

 (a=l)
 KL=0.028
 90% CI: (0.021,0.037)
  (a=1.2[MLE])
  KL=0.021
.  90% CI: (0.012,0.034)
Appendix A contains fits of the model to published exposure-response data from 18 studies (all,
with the exception of Wittenoom and South Carolina, obtained from the published literature). In
each of these 18 cases the linear model (Equation 7-2) provides an adequate description of the
exposure-response. However, o was estimated as >1.0 in 15 of these cases and statistically
significantly so in six cases, compared to only one case where a was found to be significantly
<1.0. If the true background lung cancer rates in these 15 cohorts with cc>l .0 are equal to that in
the corresponding control population (i.e., so that a fit with a=1.0 is appropriate), the exposure-
response would appear to be supra-linear. At the same time, the data in  11 of the 18 studies
(more than half) can be adequately fit with o=1.0 and the studies for which a=1.0 does not
provide an adequate fit do not appear to be related by mineral type, by industry, or by the size of
the study (Appendix A).  Thus, while we completed our analysis using the current lung cancer
                                          7.7

-------
model (Equation 7-2), further evaluation of the lung cancer exposure-response relationship
appears to be warranted.

7,2 J.2              Time Dependence

The lung cancer model (Equation 7-2) was next evaluated to determine whether it adequately
describes the time-dependence of the lung cancer mortality observed in the Wittenoom and
South Carolina cohorts. Of particular interest is whether the model accurately predicts lung
cancer mortality many years after exposure has ceased. The model predicts that RR increases
linearly with cumulative exposure lagged 10 years until 10 years following the end of exposure,
after which it remains constant from that time forward. However, it has been suggested (e.g.,
Walker 1984) that RRs for lung cancer eventually decline after cessation of exposure. Any
investigation of this issue should control for exposure level, since exposures can also fall with
increasing time due to higher death rates in more heavily exposed subjects. The availability of
the raw data from the South Carolina and Wittenoom cohorts provides an-opportunity to explore
this issue in more depth.

To investigate the assumption inherent in the lung cancer model (Equation 7-2) that the RR
remains constant following the secession of exposure, bivariate tables were constructed from
both the Wittenoom and South Carolina data in which observed and expected numbers of lung
cancers were cross-classified by both cumulative exposure (using the same exposure categories
as in Tables 7-1 and 7-2) and time since last exposure categorized using 5-year intervals. The
lung cancer model (Equation 7-2) was then fit to these bivariate data. A formal statistical test
was conducted of whether the lung cancer RR changed following the end of exposure. This test
consisted of appending the multiplicative factor, exp(-K.TSLE), to the lung cancer model
(Equation 7-2), where TSLE is time since last exposure, and conducting a likelihood ratio test of
whether the estimated parameter, K, is statistically significantly different from zero. For
presentation the bivariate tables were collapsed into univariate tables (Tables 7-3 and 7-4)
categorized only by time since last exposure, by summing observed and expected cancers over
cumulative exposure categories and computing person-year-weighted averages of cumulative
exposure and times since last exposure.

Table 7-3 shows the resulting fit of the U.S. EPA lung cancer model to the Wittenoom lung
cancer data categorized by time since last exposure. The RRs rise to a maximum between 10
and 30 years from last exposure and then decline thereafter.  However, a very similar pattern is
seen with cumulative exposure, which peaks between 10 and 15 years from last exposure, and
then declines due to higher mortality among more heavily exposed workers. The U.S. EPA lung
cancer model (which  does not assume a decrease in RR with time, but does account for any
decrease in exposure with increasing time since last exposure) provides an adequate fit to these
data (p=0.13, « estimated), and there is no apparent tendency for the predicted deaths to fall
below observed at the longest times since last exposure. To the contrary, the model-predicted
number of lung cancer deaths more than 35 years since last exposure (53.3) is very close to, but
slightly larger than, the  observed number (51). Likewise, the parameter K was not significantly
different from zero (p=0.16). Thus, the lung cancer model (Equation 7-2) provides a good
description of the Wittenoom lung cancer data categorized by time since last exposure and there
is no indication of a drop in the RR up to 45 or more years after exposure has ended that cannot
be accounted for by reduced exposures in the longest time categories. It should also be noted

                                           7.8

-------
^
             that the values of a (2.1) and KL (0.0051 f7ml-y) estimated from this bivariate analysis are very
             similar to those estimated from the univariate analysis summarized in Table 7-1.

               Table 7-3. Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among
              Wittenoom, Australia Miners (Deklerk 2001) Categorized by Years Since Last Exposure
Years Since Last Exposure
Range
0-1
1-5
5-10
10-15
15-20
20-25
25-30
30-35
35^0
40-45
45+
Total
Average
0.27
3.0
7.5
12.5
17.5
22.5
27.4
32.3
37.1
43.3
51.2

Average
Exposure
Lagged 10 Years Observed
(f-y/ml) Deaths
1.2
1.2
7.7
24.8
24.1
22.8
22.0
21.7
20.0
16.7
13.0

0
1
8
19
26
42
54
50
31
20
0
251
Expected
Deaths
0.6
1.6
3.9
7.0
11.4
16.4
20.1
20.3
13.9
9.6
0.4
105.1
Predicted Deaths
Relative by Model
Risk (o=2.1)
0
0.6
2.0
2.7
2.3
2.6
2.7
2.5
2.2
2.1
0

1.5
3.9
10.3
19.2
29.1
39.8
'47.2
46.6
31.4
21.1
0.8
250.9
                                                      Goodness of Fit P-value
0.19
             Table 7-4 shows the corresponding fit of the U.S. EPA lung cancer model to the South Carolina
             lung cancer data. This table indicates a marked decrease in RR with increasing time since last
             exposure. However, there is a concomitant decrease in cumulative exposure.  The U.S. EPA
             lung cancer model provides an adequate fit to these data (p=0.31, a estimated), and the estimates
             of a (1.3) and KL (0.20 (f-y/ml)"1) are very similar to those obtained from the univariate analysis
             (Table 7-2). There is no obvious tendency for the model to underestimate risk at the longest
             times since last exposure, and.the value of K estimated for this cohort is not significantly
             different from zero (p=0.12).
                                                      7.9

-------
 Table 7-4.  Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among
          South Carolina Textile Workers (Schnoor 2001) Categorized by Years
                                    Since Last Exposure
                               Average
Years Since Last Exposure     Exposure                                     Predicted Deaths
                           Lagged 10 Years   Observed   Expected  Relative      by Model
                                (f-y/ml)        Deaths     Deaths    Risk        (o=1.3)
    Range
A
Average
0-1
1-5
5-10
10-15
15-20
20-25
• 25-30
30-35
35^0

40-45
45+
Total
0.06
3.0
7.5
12.5
17.5
22.5
27.5
32.5
37.4

43.5
50.4

21.8
13.1
16.5
16.5
11.5
9.9
9.2
8.5
7.8

7.1
31.6

15
7
14
8
3
15
13
11
17

19
0
122
4.1
2.2
3.5
3.8
3.6
5.1
7.1
9.5
10.5
t
13.8
0.0
63.4
3.7
3.2
3.9
2.1
0.8
2.9
1.8
1.2
1.6

1.4
0.0

13.4
7.3
11.5
10.4
7.0
8.8
11.7
15.0
16.2

20.8
0.0
122.0
                                              Goodness of Fit P-value
                                                                   0.31
Raw lung cancer data were not available from any additional studies, so that the corresponding
analysis could not be performed on any other cohort. However, the report by Seidman et al.
(1986) on workers at a factory in Patterson, New Jersey that utilized amosite as a raw material
contains data in a form that permit a similar analysis using time since onset of exposure rather
than time since last exposure. The result of this analysis is shown in Table 7-5, which
corresponds to Tables 7-3 and 7-4 except that Table 7-5 presents mortality by time since first
exposure rather than time since last exposure.2  Since exposures in the Seidman study averaged
       2Table 7-5 was created in the following manner.  Table XIV of Seidman et al. (1986) contains expected
numbers of deaths from all causes cross-classified by cumulative exposure and time since onset of exposure.
Assuming that in each time-since-exposure-onset category the number of deaths from all causes is proportional to
the number of person-years, an average cumulative exposure was estimated for each time-since-exposure-onset
category.  Similarly, using Table V in Seidman et al. (1986), which contains expected numbers of deaths from all
causes cross-classified by cumulative exposure and time since onset of exposure, an average duration of exposure
was estimated for each time-since-exposure-onset category. The ratio of the estimate of cumulative exposure and
duration of exposure provides an estimate of the exposure intensity (f/ml) in each time-since-exposure-onset
category.  Using this estimate of exposure and the midpoint of the duration of exposure range for each cell, an
average cumulative exposure was estimated for each cell in Table X of Siedman et al., in which lung cancer data
were cross-classified using the same scheme as was used in Table V for deaths from all causes. The EPA lung
cancer model (Equation 7-2) was then fit to the lung cancer mortality data in this table. This bivariate table was then
collapsed by summing over cumulative exposure categories to produce Table 7-5, which categorizes the Seidman
et al. (1986) lung cancer data by time-since-exposure-onset.
                                              7.10

-------
only 1.5 years, time since first exposure approximates time since last exposure. Because
Seidman et al. (1986) did not include a lag in their exposure estimates, only data for times since
first exposure >10 years are included in our analysis, since (in this range) lagged and>unlagged
exposures would be very similar.                                              :
  Table 7-5. Fit of EPA Lung Cancer Model to Observed Lung Cancer Mortality Among
   New Jersey Factory Workers (Siedman et al. 1986) Categorized by Years Since First
                                       Exposure
Years Since
First Exposure
10-14
15-19
20-24
25-29
30-34
35+
Total
Observed
Deaths
13
20
17
22
21
16
109
Expected
Deaths
8.1
8.8
9.0
8.6
6.1
4.2
44.7
Relative Risk
1.6
2.3
1.9
2.5
3.5
3.8

Predicted Deaths by Model
(cc=3.4)
16.2
i.
18.8
20.1,
20.8 '
17.8 ',
I
15.3,
109.0.
                     Goodness of Fit P-value
0.55'
The RR of lung cancer increases with time since the beginning of exposure in the Seidman
cohort.  Nevertheless, the fit of the U.S. EPA model to these data with a estimated is very good
(p=0.55) and the numbers of cancer deaths predicted by the model agree closely with the
observed numbers, even up through 35+ years from the beginning of exposure. Time since last
exposure (estimated in each cell as time since first exposure minus average duration'of exposure)
is not a significant predictor of lung cancer mortality (K is not significantly different from zero,
p=0.33). The values of a (3.4) and KL (0.01 (f-y/ml)"1) estimated from this analysis agree closely
with those estimated in Appendix A by a different approach (cf. Table A-13 in Appendix A).
The analysis in Appendix A was also repeated after eliminating workers who worked longer than
2 years (i.e., workers for whom time since first exposure was not a good approximation to time
since  last exposure). Results obtained in this analysis are very similar to those shown in Table
7-5.

These analyses of the relationship between lung cancer mortality and time after exposure ends
are based on cohorts exposed to relatively pure asbestos fiber types: crocidolite (Wittenoom,
Table 7-3), chrysotile (South Carolina factory, Table 7-4) and amosite (Patterson, New Jersey
factory, Table 7-5).  All three of these  were consistent with the assumption inherentjn the U.S.
EPA lung cancer model (Equation 7-2) that RR of lung cancer mortality remains constant after
10 years past the end of exposure. Thus, the U.S. EPA model appears to adequately'describe the
time-dependence of lung cancer mortality in asbestos exposed cohorts.
                                         7.11

-------
7.2,1.3              Smoking-Asbestos Interaction With Respect to Lung Cancer

The existence of an interaction between smoking and asbestos in causing lung cancer has been
known for many years.  For example, Hammond et al. (1979), in a study of U.S. insulation
workers, found a multiplicative relationship between smoking and asbestos exposure in causing
lung cancer. A multiplicative relationship for the interaction between smoking and asbestos is
also the relationship inherent in the U.S. EPA lung cancer model (Equations  7-1 and 7-2), by
virtue of the fact that increased risk from smoking in reflected in the background risk, which is
multiplied by the asbestos-induced RR. More recent work, however, suggests that the
interaction between smoking and asbestos exposure may be more complicated.

As described by Liddell (2001), many reviews of the interaction between smoking and asbestos
exposure have tended to conclude that the interaction is multiplicative due primarily to (1) the
conclusions of Hammond et al. (1979), which was the largest of the early studies and was thus
most heavily weighted, and (2) lack of viable alternatives among the options considered. As
Liddell (2001) explains, commonly, reviewers examined fits of a simple multiplicative model or
a simple additive model and found that (of these two) the multiplicative model tended to provide
a better fit,  although the fits of either model were sometimes relatively poor.  Moreover, only
limited sets of mortality data from asbestos exposed cohorts could be used to evaluate these
model  fits because only limited smoking histories had been collected among such cohorts and,
among the data sets studied, the evaluation was typically limited to distinguishing effects among
dichotomous groups (i.e., non-smokers and smokers, the latter of which were treated as a single,
collective group).

In a more recent analysis of cigarette smoking among the cohort of Quebec miners and millers
exposed to chrysotile (which had been previously described in numerous studies, see Appendix
A), Liddell and Armstrong (2002) found that the interaction between smoking and asbestos
exposure was complex:  certainly less than multiplicative, but not quite linear either.

To get some idea of the possible impact of a non-multiplicative asbestos-smoking interaction
upon our work to reconcile lung cancer exposure-response coefficients calculated from different
environments, consider the generalized model for RR,

                           RR = 6 + ps.S + pA.A + Y«A.S                    (Eq. 7-3)

where  S is a measure of the amount smoked, A is a measure of asbestos exposure, and 6, ps, PA
and Y are parameters. If y=0, the model predicts an additive smoking/asbestos interaction, and if
Y=ps * PA / S> the model becomes

                        RR= 6 * [1 + (Ps/6). S] * [1 + (pA/6), A]                 (Eq. 7-4)

which  is a multiplicative smoking/asbestos interaction. Thus, this model generalizes both
additive and multiplicative interaction, and has been used by a number of researchers, including
Liddell and Armstrong (2002), to study smoking/asbestos interactions.
                                          7.12

-------
Note that the generalized model Equation 7-4 can be written in the form
                                                             ps , S)]         ;,   (Eq. 7-5)
                                                                             !•
Further, note that Equation 7-5 is of the form: const, * (1+constj* [asbestos exposure]), which is
the same form as the U.S. EPA lung cancer model (Equation 7-2), except that in Equation 7-5
const, (=6+ps,S) and const2 (=(pA+Y.S)/(6+ps.S)) depend upon the amount smoked, while in the
U.S. EPA lung cancer model (Equation 7-2) const, (=a) likewise depends upon the amount
smoked, but the constz (=KL) does not.  Thus even with this generalized model, the U.S. EPA
model would still apply, although the KL estimated would depend upon the smoking habits of the
underlying cohort.  That the generalized model Equation 7-3 is at least approximately correct is
supported by the fact that the U.S. EPA model (Equation 7-2), which we have noted is
comparable to Equation 7-3 except that the KL value may depend upon smoking, provided a
valid fit to data from all of the 20 studies to which it was applied.

This at least suggests that, even if the multiplicative interaction assumed by the U.S.,EPA model
(Equation 7-2) is not correct, as long as the smoking habits in different cohorts don't differ
extremely, the KL value from different cohorts will be estimates of nearly the same quantity. In
particular, the  differences in the estimated KL produced by differences in smoking habits in
different cohorts is likely to be a relatively small component of the very large variation in the KL
computed from different studies (see, e.g., Table A-l in Appendix A). In that case, KL computed
from different environments using the U.S. EPA model (Equation 7-2) would still be comparable
and it would be meaningful to search for measures of asbestos exposure ("exposure indices")
that better rationalize KL values (calculated using Equation 7-2) from different studies.3

Although the above argument suggests that efforts to reconcile KL values from different cohorts
is still a valid and useful exercise, even if the smoking/asbestos interaction is not multiplicative,
as assumed by the U.S. EPA model, further evaluation of the interaction between smoking and
asbestos exposure,  including evaluation of different exposure-response models, is clearly
warranted. However,  it must be recognized that the ability to conduct such evaluation will be
limited by the  small number of asbestos-related epidemiology studies in which smoking data are
available, and  possibly further limited by the ability to gain access to the raw data from these
studies. It should also be recognized that smoking data are not available for most of the studies
currently available (see, e.g., Table A-l in Appendix A), and, for most of those studies, the
published data are probably not sufficient to allow fitting of a model that incorporates any
smoking/asbestos interaction other than multiplicative.
       3However, if the interaction is not multiplicative as assumed by the EPA model, it could be problematic to
use a KL value estimated from the model to quantify absolute asbestos risk (e.g., from lifetime exposure) while
taking into account smoking habits.
                                          7.13

-------
7.2.1.4              Conclusions Concerning the Adequacy of the U.S. EPA Lung Cancer
                    Model

Three principle assumptions inherent to the U.S. EPA lung cancer model were considered in this
study to evaluate the overall adequacy of the model for describing the manner in which asbestos
exposure induces lung cancer. These are:

     •       that lung cancer RR is proportional to cumulative exposure (lagged 10 years);

     •       that risk remains constant following 10 years after the cessation of exposure; and

     •       that the interaction between smoking and asbestos exposure is multiplicative.

The findings from this evaluation are summarized briefly below. For convenience, timed-
dependence is addressed first.

In Section 7.2.1.2, the time response predicted by the lung cancer model (Equation 7-2) was
evaluated using the Wittenoom and South Carolina data, along with the data from a New Jersey
amosite factory (Seidman et al. 1986).  Follow-up in the Wittenoom and South Carolina was
sufficient to permit evaluation of the adequacy of the model 40-45 years past the end of
exposure, and follow-up in the New Jersey factory permitted the model to be evaluated up.
through 3 5+years past the end of exposure.                  ^

The data from all three studies were consistent with the assumption inherent in the lung cancer
model that RR of lung cancer mortality remains constant after 10 years past the end of exposure.
That this result holds equally for a cohort exposed to chrysotile (South Carolina) as for cohorts
exposed to crocidolite (Wittenoom) and amosite (New Jersey factory) is particularly noteworthy,
because chrysotile fibers have been reported to be much less persistent than amphibole fibers
in vivo (see Section 6.2.4). The fact that the lung cancer risk in South Carolina remained
elevated 40-45 years following the cessation of exposure, along with the fact that the KL from
the South Carolina cohort is the largest KL value obtained from our review of 20 studies (see
Table A-l in Appendix A) suggests that carcinogenic potency of asbestos is not strongly related
to durability, at least for lung cancer (see Section 6.2).

Based on the results indicated in Section 7.2.1.1, the linear, RR exposure-response model for
lung cancer (Equation 7-2) provides an adequate description of the exposure-response
relationship from 20 studies. There is little evidence that exposure-response is sub-linear or
"threshold-like". However, in a number of these cases the fitted model predicts a background
response that is higher than that in the control population.  If the background rate in the control
population was actually appropriate in  these studies, the exposure-response relationship in these
studies would appear in some cases to be supra-linear. These results further reinforce the
recommendation of the expert panel (Appendix B) that evaluation of a broader range of
exposure-response models (including those that incorporate smoking-asbestos interactions that
are other than multiplicative) for lung cancer is appropriate.
                                          7.14

-------
Although further evaluation of alternate models is warranted, the specific studies for which
«=1.0 does not provide an adequate fit to the data do not appear to be related by mineral type,
industry, or size of study cohort and such studies total fewer than half of the studies|evaluated.
Coupled with the additional evidence that the linear, RR model for lung cancer (Equation 7-2)
provides a good description of the time-dependence of disease mortality during and following
exposure and that it is not likely to grossly underestimate exposure despite limitation in the
manner that smoking is addressed (Section 7.2.1.3), this suggests that use of this model may be
adequate for  estimating exposure-response factors for the existing studies, at least unless and
until a model that provides a superior fit to the data is identified.
                                                                            r
Further support for the U.S. EPA lung cancer model  is also provided by Stayner et a'l. (1997).
Stayner et al. evaluated the relative fit of a variety of additive, multiplicative, and more complex,
empirical models to the same raw data from the South Carolina textile plant that we'evaluated
and conclusions drawn by these researchers generally parallel and reinforce those reported here.

Stayner et al. (1997) found a highly significant exposure-response relationship for lung cancer
and that a linear, RR model (similar to that described by Equation 7-1, except that th'ey initially
assumed a lag of 15 years) provided the best overall fit to the data (among the additive and
multiplicative models evaluated).  Moreover, the fit to the data was not significantly; improved
by adding additional parameters, nor was there any indication of a significant interaction with
any of the covariates evaluated (age, race, sex, or year).  These researchers did find an
interaction with time such that by applying different slopes for different latencies (15-29 years,
30-39 years,  and >40  years), they were able to obtain a significantly better fit.  However,
Stayner et al. (1997) did not evaluate a RR model with a single slope using a lag of 10 years.
Therefore, their analysis cannot be used to evaluate the effect of assuming different slopes for
different latencies with respect to the U.S. EPA model, which assumes a lag of 10 years.

Because we found that mortality among the South Carolina cohort is adequately described with
o=1.0, the Stayner et al. (1997) study does not directly address questions concerning supra
linearity that  have been suggested in our findings. Nevertheless, taken as a whole, the evidence
evaluated here suggests that the model described in Equation 7-2 may be adequate  for evaluating
the relationship between asbestos exposure and lung cancer mortality among the existing
epidemiology studies. Certainly, no clearly superior model has yet been identified. Therefore,
the U.S. EPA lung cancer model was employed in this study for evaluating lung cancer risk
while we also recognize that further evaluation of alternate models is warranted. The primary
obstacle to  more fully evaluating alternate models has been and continues to be lack/of access to
the raw data for a greater number of cohorts.                   ' '   "

7.2.2  Estimating KL values from the Published Epidemiology Studies

The U.S. EPA model for lung cancer (Equation 7-2) was applied to each of the  available
epidemiology data sets to obtain study-specific estimates for the lung cancer exposure-response
coefficient, KL. Based on the results presented in Section 7.2.1.4, while there is some evidence
that models in addition to the U.S. EPA lung cancer model should be explored,  there is little
indication that the current model does not provide an adequate description of lung cancer
mortality that is sufficient to support general risk assessment. The set of KL values derived from
                                          7.15

-------
available epidemiology studies are presented in Table 7-6, which is a reproduction of Table A-l
in Appendix A.

In Table 7-6, Column 1 lists the fiber types for the various.studies, Column 2 lists the exposure
settings (industry type), and Column 3 indicates the specific locations studied.  Column 4
presents the best estimate of each KL value derived for studies, as reported in the original 1986
Health Effects Update and Column 5  presents the reference for each respective study.  Columns
6,7, and 8 present, respectively: the best-estimates for the KL values derived in our evaluation
for all of the studies currently available (including studies corresponding to those in the 1986
Health Effects Update); a statistical confidence interval for each KL value (derived as described
in Appendix A); and an uncertainty interval for each KL value (also derived as described in
Appendix A).  The reference for the respective study from which the data were derived for each
KL estimate is provided in the last column of the table. To assure comparability across studies,
values for all studies (even those that have not been updated since their inclusion in the 1986
Health Effects Update) were re-derived using the modified procedures described in Appendix A.

As explained in Appendix A, the uncertainty intervals for KL values (and corresponding intervals
for KM values, the exposure-response coefficients for mesothelioma) are intended to reflect, in
addition to statistical variation, other  forms of uncertainty that are difficult to quantify, such as
model uncertainty and uncertainty in  exposure estimates.  We interpret these informally as
providing a range of KL values that are reasonable, based on the data available for a given study.
Accordingly, if uncertainty intervals for two KL values do not overlap, these two underlying sets
of data are considered to be incompatible. Potential reasons for such incompatibility include the
possibility that the KL values are based on an exposure measure that does not correlate well with
biological activity of asbestos. One of the goals of this report is to determine an exposure index
that correlates better with biological activity, and consequently brings the KL values into closer
agreement. The degree of overlap of the corresponding uncertainty intervals provides  an
indication of the extent to which a groups of KL values are in agreement.

The KL values derived in this study and the corresponding values derived in the original 1986
Health Effects Update generally agree at least to within a factor of 3; a couple vary by a factor of
4; one varies by a factor of 5; and one varies by a factor of about 15.  However, in every case the
KL from the 1986 update lies within the uncertainty interval for the KL derived in the current
update.

Perhaps the most interesting of the changes between the 1986 KL value estimates and the current
KL value estimates involves the friction products plant in Connecticut (McDonald et al. 1984).
Although a relatively small, positive exposure-response was estimated from this study in the
1986 Health Effects Update, the best current estimate is that this is essentially a negative study
(no excess risk attributable to asbestos). The difference derives primarily  from allowing a to
vary in the current analysis. The exposure groups in this cohort do not exhibit a monotonically
increasing exposure-response relationship and, in fact, the highest response is observed among
the group with the lowest overall exposure. As indicated in Appendix A, lack of a
monotonically increasing exposure-response relationship is a problem observed in several  of the
studies evaluated.
                                           7.16

-------
.1
•3
9
lemiological S
'S.
u
Various
S
o
•e
S
>
S
^_
^
oefficients (1
U
3!
e
o
a
S
w
9
O
a,
m
n
a
y
Ml
o

3
±
R
H


Reference
Uncertainty
Interval*
4>
^ n
ol £
0 S
U


2
•3 *
en ^
e


Reference
^
QO o
^ft C!
•>—' **
(d

O
•g-
u

S
o
• MM
*P»
2
V
D.
0


a
i
tt.
13
•*«*
"o
oo ^*
§OS
0
2-°
C
q
d
o
S


(N
0
d


"rt
3§ i
C wi ^
o — "3
9 -3 .-gg


vn r»»
d d

,S
|1




oo op
c .e
'c HS
is

JU
"o
1
6
Piolatto et al.
1990
McDonald
etal. 1984
O <^
• *°.
S o"

r— r-
d d
S S


V)
o o
d



<-. 2 ^
1 1^
^3 o Q — ;


^^
§ q
d <=>

a 3
.S .0
2 S c •£
ll li




.il
ts "g
•c 2
b CL,




W Kl
t3 t>
4eh
3 <^ .^ os
^ ^H Q ^^
o" 5
^ S

VO "*
d j^-
d — '


«o
^* i
0 ^



13
ts
-t_>
g*
§s
02



oo
CN

1
C3 CQ
S "3,
•c *
? 1 *l
« § g a
?: & &o u




ill
U S H




McDonald
1983a
de Klerk et al.
-1994C
$ i
s- e
^^^
"O" oo
^ d !
4 rf .
s^ 5


r-
^M ^^
0 '
1
i
i
McDonald
etal. 1983a



>/•> i
 o\ e ^ S 13
*> o\ 5 > « **
J — • •< 5^ i< «
^ r
-^. ^ oo
«« •* '(Si
" S' U
S o g
o 2/
^
x-s S^ n
d - K
e 5 ^
S- 0,

1
^ - OS
^H "*) rn
000
.r
!'
!.
1



r

a
!r,l
0> 0 XI C5
•^ o ^ O
CA
S :'


ll .
•3 tn ''
i s :
^is •
..
u
"o
H
Berry and
Newhouse
1983
CO
oj

^-^
oo
d
d^


00
o
d



Berry and
Newhouse
1983


oo
0
d

|
^
•S
m




§1
t> •§
•c £
[i. CU


"S
_x
2

-------
(continued)
en
4)
•3
9
•**
W
"i
f
o

V
3
Q-
w

e^ T3 C
o e 
n
^
O

O)
t
H
t-
<
£
S | -
to « tS
Is f£ IS
S ON 3 ON "3 ON
il^ «—» »JH ^* *% *•*

^Tt* irs ^T^
—> (^ ^^
^ vo ^
^ o M
s s ^



|x-
ON m !>,
 ON
fc S ? iQ

* R
^ d

i-t
I i
illi!
« y 5 ^ ^
O **,» C3
•
JU
'S
{2




~5
tS
2£
U ON
O< ^^

r
1
.S
'i
>,
•C*
.S
1
3
^

^ §§
"rt E (S
go-S>
•sis
i> 3 t
O *^ "N ^^
g ^ 0.
-8 w =>
« O s
imbining 90% con
Terri Schnorr (NI
iw data with folio1
S E e
O ^3
-&*!
T3 03 "R
4> 'g !S
E TJ •§
111
C4 «-,*-,
> flj {rt
C *•* -S
1 1 1
"^ S 8
.s al
533
O je* _f-
O *5 :£3
?FF
CO

-------
Among the KL values derived in the current study, the lowest and highest of the best-estimate
values differ by a factor of 300 (excluding the negative study of the Connecticut friction
products plant, which would make the spread even larger) and several pairs of uncertainty
intervals have no overlap. For example, the KL uncertainty interval for the chrysotile miners in
Quebec lies entirely below the corresponding intervals for chrysotile textile workers' (in either of
the two studies for South Carolina, which are of highly redundant cohorts  in the same plant), for
textile workers in the Pennsylvania plant, and for arriosite insulation manufacturers (in the
Seidman study (1986).

The KL values and the associated uncertainty intervals are plotted in Figure 7-1. Each exposure
environment is plotted along the X-axis of the figure and is labeled with a 4-digit code that
indicates fiber type (chrysotile, mixed,  crocidolite, or tremolite), industry (mining, friction
products, asbestos-cement pipe, textiles, insulation manufacturing, or insulation application); and
a 2-digit code indicating the study from which the data were derived. A key is also provided. In
Figure 7-1, the chrysotile studies are grouped on the left, amphibole studies are grouped on the
right, and mixed studies are in the middle.

For studies conducted at the same facility (generally among highly overlapping cohorts), such as
the Dement et al. (1994) and the McDonald et al. (1983a) studies of the same South|Carolina
textile facility, a single study was selected for presentation in Figure 7-1. Thus, for South
Carolina, the Dement et al. (1994) study is presented because we had access to the r^w data for
this study. It is also a newer study.  Similarly, the Amandus and Wheeler (1987) study was
selected to represent the Libby Vermiculite site over the other study at this facility (McDonald et
al.  1986). The effects of such selection is expected to be small in any case because the KL values
estimated for the individual studies in each pair vary only by a factor of 2.         '

Comparisons of KL values across the available studies are instructive. Within chrysotile studies
alone (and excluding the negative friction products study), lowest and highest KL values vary by
approximately a factor of 70.  Moreover, as previously indicated, the uncertainty intervals for the
lowest (non-zero) value (for Quebec miners) and the highest value (textile workers);have no
overlap. Uncertainty intervals for the negative friction product study and the  other estimates for
chrysotile do overlap, primarily due to the wide confidence interval associated withjthe negative
study.

Among the apparent variations, differences in lung cancer potency observed among'Quebec
miners versus that observed among South Carolina textile workers has been the subject of much
discussion and evaluation, which is worthy of review (Appendix D). The  inability to reconcile
these differences, appears to be among  the biggest obstacles to reliably estimating an overall
chrysotile dose-response coefficient for lung cancer.

As  indicated in Appendix D, the leading hypothesis for the apparent differences in lung cancer
risk per unit of exposure observed between chrysotile mining and textile manufacturing is the
relative distribution of fiber sizes found in dusts in these industries. Evidence from several
studies indicates that textile workers were exposed to dusts containing substantially'greater
concentrations of long structures than dusts to which miners were exposed. Thus, the effects of
fiber size is considered further in Section 7.4.
                                           7.19

-------
                                 Figure 7-1:
Plot of Estimated KL Values and Associated Uncertainty Intervals by Study Environment

     1000-g
      100,
       10-

      0.1 1
     0.01-
     1E-3
                                          T  T
            S
 .
Sfclfclfc
             lOCDh-    COOIOT-
             T-,— ^-    T— T- CJ CM
                                    7.20

-------
                         Key for Figures 7-1 through 7-6 Code
FiberTypes
(First Digit of Code)

A = amosite
C = chrysotile
M = mixed fibers
R = crocidolite
T = tremolite (in vermiculite)
Study Environment
(Second Digit of Code)
A = insulation application
F = friction products manufacturing.
I = insulation manufacturing
M = mining  .
P ='ac pipe manufacturing
T = textile manufacturing
X = misc. products manufacturing :'
Study Cohorts
(Last 2 digits)

1 = Quebec miners (Liddell et al. 1997)
2 = Quebec miners (Liddell et al. 1997, raw data)
3 = Italian miners (Piolatto et al. 1990)                   '.
4 = Connecticut friction product workers (McDonald et al. 1984)
5 = New Orleans ac pipe manufacturers (Hughes et al. 1987)
6 = South Carolina textile manufacturers (Dement et al. 1994, raw data)
7 = British friction product manufacturers (Berry and Newhouse 1983)
8 = Ontario ac pipe manufacturers (Finkelstein 1984)
9 = New Orleans ac pipe manufacturers (Hughes et al. 1987)
10 = Swedish ac pipe manufacturers (Albin et al. 1990)
11 = Belgium ac pipe manufacturers (Laquet et al. 1980)
13 = Retired factory workers (Enterline et al. 1986)
14 = Factory workers (Liddell et al. 1997)                 ;
15 = Insulation appliers (Selikoff and Seidman 1991)
16 = Pennsylvania textile workers (Mcdonald et al. 1983b)  '
17 = British textile workers (Peto et al. 1985)
18 = Australian crocidolite miners (de Klerk, unpublished, raw data)
19 = New Jersey insulation manufacturers (Seidman et al. 1986)
20 = Texas insulation manufacturers (Levin et al. 1998)    !
21 - Libby vermiculite miners (Amandus and Wheeler 1987)
                                         7.21

-------
Ignoring the negative Connecticut friction products study, the range of KL values observed
across chrysotile studies (70) appears to be substantially narrower than the range observed across
all studies (300). However, if the nearly negative study of the Belgium asbestos-cement pipe
manufacturers is also removed, the range observed for chrysotile studies is almost identical to
the range observed across all studies. This is because the KL values for Quebec represents the
low extreme of both ranges and South Carolina represents the high  extreme of both ranges.

Among "pure" amphibole studies, the lowest and highest of the best-estimate KL values vary by
a factor of approximately 9 and the two extremes both derive from within the same industry
(amosite insulation).  The two studies in question are of amosite insulation manufacturing plants
(one in Patterson, New Jersey, and one in Tyler, Texas) that utilized the same equipment
(literally) and apparently had similar sources of asbestos (South Africa). Despite the 9-fold
difference in KL values, the uncertainty intervals for these two estimates have substantial
overlap. If the arithmetic mean of the values for the two amosite insulation studies is used, KL
values estimated, respectively,  for crocidolite mining, amosite insulation manufacture, and
mining of vermiculite contaminated  with tremolite vary by less than a factor of 2.  However, this
is possibly fortuitous, given the magnitude of the associated uncertainty intervals (Figure 7-1).

As indicated in Appendix D, for example, it is possible that mining studies tend to exhibit low
KL values relative to studies of asbestos products industries. This is due to the presence of large
numbers of cleavage fragments in the dusts that may not contribute to biological activity
(because the majority of these may not exhibit the requisite size to be biologically active) but
which, nevertheless, are included in  estimates of asbestos concentrations in the original
epidemiology studies. If the KL estimates for Libby and Wittenoom could be adjusted for this
effect, they might be closer in value  to that obtained from the Seidman study.

It is also instructive to compare variation within and between industries. Within industries
(especially for a single fiber type), the data are limited. The studies from the two chrysotile
mines (Quebec and Italy) show remarkably close agreement, varying by less than a factor of 2.
The three studies involving mining of amphibole (Wittenoom and the two studies of Libby) also
vary by less than a factor of 2.  However, the mean of the amphibole mining group is
approximately 10 times the mean of the chrysotile mining group. Moreover, based on inspection
of their respective uncertainty intervals, the KL values for chrysotile mining in Quebec and
Crocidolite mining in Wittenoom appear to be incompatible.

Across all asbestos types (including  mixed), the asbestos-cement pipe industry shows the
greatest variation, including a nearly negative study (best estimate KL=0.000068) and four more
studies with KL values that range up to 0.0029 producing a variation within this industry of a
factor of 40.  The friction products industry includes one negative and one positive study.  Better
agreement is observed among textiles. The two mixed textile plants show KL values that vary by
no more than a factor of 5  from each other and  from the KL for the South Carolina chrysotile
textile plant.
                                           7.22

-------
7.3   MESOTHELIOMA                                                  ]4

The model proposed in the Airborne Health Assessment Update (U.S. EPA 1986) to describe the
mortality rate from mesothelioma in relation to asbestos exposure assumes that the mortality rate
from asbestos-induced mesothelioma is independent of age at first exposure and increases
according to a power of time from onset of exposure, as described in the following relationship:
              I
              M
= KMf [(T-10)3-(T-10-d)3]

= Kw-f-CT-10)3

= 0
forT>10 + d

forlO + d>T>10

forlO>T
                                                          (Eq.7-6)
     where:
             K
               M
is the mesothelioma mortality rate at T years from onset of exposure to
asbestos for duration d and concentration f;                :

is the proportionality constant between exposure and mesothelioma
response and represents the potency of asbestos;
A more general expression that holds for variable exposure is given by
                        IM »
                                                          (Eq. 7-7)
where f(x) is the concentration of fibers at time x following the beginning of exposure. This
expression reduces to Equation 7-6 when exposure is constant (see Appendix A).

7.3.1 The Adequacy of the Current U.S. EPA Model for Mesothelioma

Access to the raw epidemiology data from a few key studies allowed us to evaluate the adequacy
of the U.S. EPA model (Equation 7-6) for describing the timerdependence for mesothelioma in
asbestos-exposed cohorts. For this analysis, the raw data from a cohort of chrysotile miners in
Quebec was graciously provided by Drs. Douglass Liddell and Corbett McDonald (described in
Liddell et al. 1997), the raw data for the cohort of crocidolite miners in Whittenoom; Australia
was graciously provided by Dr. Nick de Klerk (unpublished) and the raw data for the cohort of
chrysotile textile workers (described by Dement et al. 1994) was graciously provided by Ms.
Terri Schnoor of NIOSH and Dr. John Dement of Duke University. The Whittenoom cohort was
originally described by Armstrong et al. (1988), but the data provided by Dr. de Klerk included
additional follow-up through 1999.

To identify potential effects due to varying statistical procedures, different methods ]for fitting
the U.S. EPA mesothelioma model to epidemiological data were evaluated.  In this evaluation,
three methods were used to fit the U.S. EPA mesothelioma model to data from Wittenoom.
                                         7.23

-------
In the first approach the data were categorized in a manner often available in published form, so
this method mimics the method generally used when raw data are not available. The observed
mesotheliomas and person-years of observation were categorized by time since first exposure,
and the mean exposure level and duration of exposure were calculated for each such category.
The U.S. EPA model was then applied to such data using the approach for the typical situation
(as described in Appendix A) and results for the Wittenoom cohort are presented in Table 7-7.
The KM value estimated for Wittenoom using this approach is 7.15xlO'8 (90% CI: 6.27x10'8,
8.1 IxlO"8).  The fit of this model to data categorized by time since first exposure is good
(p=0.65).

For most of the published epidemiology data sets, the average level and duration of exposure for
individual time-since-first-exposure categories are not provided and have to be estimated from
cruder data representing cohort-wide averages.  Thus for most of the epidemiology data sets, the
calculation of KM is based on cruder information than the calculation presented in Table 7-7.

 Table 7-7. Fit of EPA Mesothelioma Model to Observed Mesothelioma Mortality Among
 Wittenoom, Australia Miners (Deklerk 2001) Categorized  by Years Since First Exposure
Years Since
Range
0-5
5-10
10-15
15-20
20-25
25-30
30-40
40-100
Total
First Exposure
Average
0
0.19
0.69
1.6
3.3 •
6.2
11.8
21.5

Average
Duration
(years)
0.643
0.91
0.958
0.970
0.953
0.957
1.05
1.13

Average
Concentration
(ffml)
30.2
30.2
30.0
29.7
29.5
29.3
29.0
28.1

Observed
Deaths
0
0
1
5
20
25
90
23
164
Predicted Deaths
by Model
0.0
0.0
0.6
6.6
17.6
30.1
78.1
31.1
164.0
                                         Goodness of Fit P-value
0.65
 Estimates of KM
 KM=7.15xlO-s
 90% CI: (6.27xlO-8, 8.1 IxlO'8)

A second approach to fitting the U.S. EPA model to epidemiology data exploits the fact that the
mesothelioma model (Equation 7-7) expresses the mesothelioma mortality rate as the product of
KM and an integral involving the exposure pattern, the time of observation, and the 10-year time
lag, but not any parameters that require estimation.

The value of this integral was calculated for each year of follow-up of each subject.  Person-
years of follow-up and mesothelioma deaths were then categorized according to the values of the
                                          7.24

-------
integral, and the average value of the integral determined for each category. Results are
presented in Table 7-8.                                                      "•.

 Table 7-8. Fit of EPA Mesothelioma Model to Observed Mesothelioma Mortality Among
  Wittenoom, Australia Miners (Deklerk 2001) Categorized by Average Value of Integral
                                    (Equation 6-12)
Average Value of Integral
0
28.4
206.3
732.3
2,038.1
5,153.7
12,385.8
30,639.7
144,801
Total
Goodness of Fit P-value
Observed Deaths
0
0
1
4
10
23
32
27
67
164

. Predicted Deaths
0.0
0.0
1 . 0.3
1.0
2.9
6.8
; 13.8
25.8
113.3
164.0
O.0001
by Model
;
*





;.


•'i
 Estimates of KM
 KM=9.00xlO-8
 90% CI:  (7.89xlO'8, 10.2x10'8)
KM was estimated by maximum likelihood from Table 7-8 assuming that the observed numbers
of cancer deaths in the different categories of the integral were independently Poissori distributed
with a mean equal to the mean value of the integral for that category times KM. The value of Kw
obtained in this fashion was KM=9.00xlO'8 (90% CI: 7.89xlO-8, J0.2xlO'8). This estimate is
about 25% larger than obtained with the data categorized by time since first exposure; although
confidence intervals obtained from the two procedures overlaps The fit of the model to data
categorized by the integral is poor (pO.OOOOl), as the model predicts too few mesotheliomas for
small values of the integral and too many mesotheliomas for large values of the integral.

A third method of estimating KM (termed the "exact method") employs a likelihood that does not
involve any categorization of data.  With this method, the hazard function, h(t)=IM(t) and the
corresponding survival function (probability of surviving to age t without death from
mesothelioma in the absence of competing causes of death), S(t)=exp(-/0' h(s)ds), are'computed
for time at the end of follow-up.  The contribution to the likelihood of a subject who died of
mesothelioma t years after beginning of follow-up is h(t)*S(t), and the contribution of a subject
whose follow-up was not terminated by death from mesothelioma is S(t). The complete
likelihood is the product of such terms over all members of the cohort. The estimate of KM
                                         7.25

-------
obtained by maximizing the logarithm of this likelihood was 7.95x10'8 (90% CI: 7.0x10"8,
9.0x10'8).

We consider the exact method of computing KM to be the most accurate and results from this
method are reported in the summary table for mesothelioma (Table 7-9, which is a reproduction
of Table A-2). The Quebec and South Carolina data sets were thus evaluated using the exact
method. However, it is noteworthy that the two other methods described above, one of which is
often applied to published data, give similar estimates of KM, at least for this data set.

The Quebec cohort was subdivided into three subcohorts, believed to correspond to differing
amounts of amphibole exposure due to tremolite contamination of the ore at different mining
locations, and to use of some imported commercial amphibole at one factory location (Liddell et
al. 1997).  Location I consisted of workers at the mine at Asbestos where the ore reportedly had
less tremolite contamination. Locations 3 and 4 consisted of workers at the large central mine
and at smaller mines, respectively, near Thetford, where the ore was more heavily contaminated
with tremolite. Location 2 consisted of workers at an asbestos products factory at Asbestos,
which processed some commercial amphibole fibers in addition to chrysotile.  The exact method
of calculating KM produced the following estimates: Location 1 (8 cases): KM=1.3xlO"10 (90%
CI: 0.3x10-'°, 4.9X10'10); Location 2 (5 cases): KM=9.2xlO-10, (90% CI: 2.0x10-'°, SSxlO'10);
Locations 3 and 4 (22 cases): KM=2.1xlO-10 (95% CI: 0.65x10'10, 6.5x1O'10).

The relative magnitudes of these estimates track with the relative amounts of amphibole
exposure estimated for these locations (Liddell et al. 1997), which is consistent with the
hypothesis that the mesothelioma risk in this cohort is due, at least in large measure, to exposure
to amphiboles.

There were only two confirmed mesothelioma deaths in the South Carolina cohort and four
additional suspected deaths. These were too few to permit detailed analysis. Based on both
confirmed and suspected mesothelioma deaths, the exact method of analysis gave an estimate of
KM=0.43xlO'8, 90% CI: (0.20xlO'8, 0.79xlO'8). Using only the two confirmed mesotheliomas,
the same analysis yielded KM=0.14xlO-8,90% CI: (0.034xlO-8, OJSxlO'8). Very similar
estimates were obtained by estimating KM from data categorized by time since first exposure and
fitting a linear model to the categorized value of the integral in the definition of the U.S. EPA
model.  Thus, for this cohort comparable KM values are estimated no matter which of the three
methods described above are used for fitting the U.S. EPA mesothelioma model to the
epidemiology data.
                                          7.26

-------



J
•j
5
1
"S
e
.1
4
"E
fct
0
c
'C
>
1^^
|
1
1
r
I
y
^^
JS
.1
u
1
o
U
V
35
a
o


£
tn
i
£
S
1
09
1
cK
I--
—
lO
rot
H




!
\
1
tj
I
a

£.
f%
r i
5 <
2l
3

_
^ g 5
e^ -O C
S «5 "
§£
u

« 2
.s s 2
^!1
=> ^


41
W
D
a

_ o
< ^S
s« S *
M S s
— «

1
3
B
_S
^i
2
V
s

V
. £
u
JB
E
TT
00
f- r- 2 «•> h
ON ON " S 2J
ON ON -j ON ON
•3 T3 " "3 "3
w « 1 « «
? =5 1 1 1
i i i t §
J J S S Q
^-v ;•"••
5T S "- ^
S « c? ^ 2
6 «=> $ "
•s j/^1 o f*i r^i
S I o S S
« o ^ o d
So - -
S » sr
^"" *s] ON
S o ?? ^
0 ^ — 0
oo" _..- o 1 ^-•>
VO 2 _- C?,
S o S e>
II s


i § o S §
00 °











1 s 1
i i i" i 1
rf 13 ~ JJ. §
§ ,0 8 6 . 2
u -^ e i -f s <-•
JD D a 5 c 3 c
< {S 6 2-g. cSl
f
i I „

1 | li 1
e o E g £
S 'c r^ •! &
2. u. U 2 H

V
1
b
JS


S
O1
"S "
ts •=.
McDonald
1983a
de Klerk el

/— V
es
* /— \
""^ 00
<^>
S > "0
<=>. Q
B
«'

^
r*T



so

en
oo
Os
WN
.£
'5
3s
J^
c
£


(N


b
<«3
_o
a


•g
ll
II


X
s


i 1
SiJ ON
^— t-H
"73 •"•
2 *
« «
W5 ^_
•S S
ap 5
x 3

y"*«-,
^-S ^
wri ro
. ;
ON" ®
§ 2
d p
^ S

oo
d
1 *"
o
£
•

f^ ON
° §

I


l
i






New Orleans
plant
Asbestos, Quebec .
N
^
U.
O,
b
. I,
b1

i
!



2
'u
I/I
•M
Seilkoffani
1991


^o
VO
IA"
ts
S

/— V
•<*
rf
v^^


rn



"«
u
*c
^R
S2


<0


U.S. insulation
workers


c .2
1.1
11
£ <







13
S
McDonald
1983b


/— N
NO
NO
t~"
E

^— K
»rj
£
o
N_X


-











Pennsylvania
plant



Textiles




,


If)
oo
ON
"rt
«
a
£


^5
«•;

(2:
C!'
i!'
^* (
£?;:
•|.
u I
§i
j3 •
^3 -
I1'
"J3 %.

Jl

y
bining 90% confiden
iddeir :" ""•
E u
8'S
>>. P
.fftta
*o _s
^ s
1 ^
^*
£ fS
S §
£ E
ti
S 3
8«
§!?





















11
g 
&l
S 8
^*
'^>
E^ -a
CNj
!•>.'

-------
7.3.L1
Time Dependence
The U.S. EPA mesothelioma model (Equation 7-6 or 7-7) was next evaluated to determine
whether it adequately describes the time-dependence of mesothelioma mortality following
cessation of exposure in the Wittenoom and Quebec cohorts. The small number of
mesotheliomas observed among the South Carolina cohort precluded a meaningful evaluation of
this issue for that cohort.   .

For times since first exposure longer than 10 years past the end of exposure the mesothelioma
model (Equation 7-6) can be rewritten as

             IM = 3 * KM  * f * d * (t -10)2 * {1 -  3 *  [d / (t -10)] + [d / (t -10)]2}    (Eq. 7-8)

From this expression we see that, when  time since first exposure lagged 10 years, (t-10), is large
compared to duration of exposure, (d), the model predicts that the mesothelioma mortality rate is
approximately proportional to the product of cumulative exposure (the exposure level, f, (f/ml)
times the duration of exposure, (d) and the square of time since first exposure lagged 10 years.
Thus, the model predicts that the mesothelioma mortality will increase indefinitely with age as
the square of time since first exposure lagged 10 years. The availability of raw data from the
Wittenoom and Quebec cohorts provides an opportunity to evaluate this assumption.

Table 7-10 shows the fit of the U.S. EPA mesothelioma model to Wittenoom data characterized
by time since last exposure, based on the KM estimated from the exact analysis. There is no
indication from this table that the mesothelioma mortality rate declines after the cessation of
exposure, or that the model over-predicts the mesothelioma risk at long times after the cessation
of exposure. In fact, the model under-predicts the number of deaths at the longest times, as it
predicts 76.8 deaths after 30 years from the end of exposure, whereas 96 were observed.

Table 7-11 shows the observed number of mesothelioma deaths in the three separate locations at
Quebec, categorized by time since last exposure, and compared with the predicted numbers
obtained using the KM values obtained using the exact  fitting method. From all three locations
combined the model predicts 10.6 deaths from mesothelioma after more than 30 years following
the cessation of exposure, whereas 10 were observed.  Although the small numbers of
mesotheliomas make it difficult to draw definite conclusions about the adequacy of the model,
there is little evidence that the model under or over-predicts the numbers of mesotheliomas at
long periods after the end of exposure.
7.3.L2
Exposure Dependence
The lack of fit of the mesothelioma model (Equations 7-7 and 7-8) to the Wittenoom data
categorized by the value of the integral in Equation 7-8 (Table 7-8) suggests that the Wittenoom
data may not be consistent with the assumption that the mortality rate is linear in the intensity of
exposure (f/ml). Specifically, the mesothelioma model predicts that for fixed time since first
exposure and duration of exposure the mesothelioma mortality rate varies linearly with f, the
asbestos air concentration. To test this prediction, expected numbers of mesothelioma deaths
were calculated for each of four categories of asbestos air concentrations while controlling for
                                          7.28

-------
both time since first exposure and duration of exposure." Person-years in the first 10 years
following the beginning of exposure were ignored since no mesothelioma deaths occurred in this
time interval, as predicted by the model. The relatively few workers who were employed for
longer than 5 years were also excluded from this analysis (exposure durations were generally
quite short in this cohort, with the average employment duration being <1  year), which means
that all follow-up in this analysis occurred after exposure had ended. Results of this analysis are
shown in Table 7-12.                                         !
     Table 7-10. Fit of EPA Mesothelioma Model to Observed Mesothelioma Mortality
   Among Wittenoom, Australia Miners (Deklerk 2001) Categorized by Years Since Last
                                           Exposure         ;
Years Since
Range
0-1
1-5
5-10
10-20
20-30
30-40
40+
30+
Total
Last Exposure
Average
0.3
3
7.5
14.9
24.7
34
43.6


Average
Value
of Integral
35
72
327
4025
18058
39802
57952


Observed
Deaths
0
0
0
10
58;
79
17 •
96
164.
Predicted
Deaths
by Model
0
0.1
0.6
14.4
52.9
61.3
!5.5
76.8
144.9
Observed/
Predicted
0
0
• o
', °-7
• 1.1
1.3
; 1.1
1.3
;•
                                                        Goodness of Fit P-value
0.21
       4In this analysis the Wittenoom data were categorized by time since first exposure {10-20, 20^-30, 30-40,
and 40+ years), exposure intensity (0-15, 15-30, 30-60, and 60+ f/ml), and duration of exposure (0-1 and 1+ years).
This categorization was facilitated by the facts that in the subcohort being analyzed, exposure had ended prior to the
beginning of follow-up, and in the Wittenoom data base exposure intensity was assumed to be constant throughout
employment. Within each of the eight [time since first exposure] x [duration of exposure] categories tHe total
number of mesothelioma deaths were allocated to the various exposure intensity sub-categories in proportion to the
product of the average exposure intensity times the person-years of observation in each sub-category, thereby
producing the expected number of deaths in each sub-category under the assumption that the response was linear in
exposure intensity within each [time since first exposure] x [duration of exposure] category, as predicted by the
mesothelioma model. These expected deaths and the corresponding numbers of observed deaths were summed
across time-since-first-exposure and duration categories to yield observed and expected deaths categorized only by
exposure intensity                                                                     •'
                                                                                               V
                                              7.29                                  :

-------
 Table 7-11. Fit of EPA Mesothelioma Model to Observed Mesothelioma Mortality
Among Quebec Miners (Liddell 2001) in Each of Three Mining Areas, Categorized by
                         Years Since Last Exposure
Years Average
Since Last Value
Exposure of Integral
Person- Years
Observed
Deaths
Predicted
Deaths
by Model
KM
Location 1
0-10
10-20
20-30
30-40
40-50
50+
30+
Total
94,267.5
119,190
119,971
181,470
278,1 18
339,160


77,008
27,225
22,831 .,
18,321
11,846
6,673


3
2
0
1
2
0
3
8
3.1
1.4
1-2
1.4
1.4
1.0
3.7 .
9.3
1.34x10-'°







Location 2
0-10
10-20
20-30
30-40
40-50
50+
30+
Total
56,377.5
56,445.7
58,307.7
78,685.8
88,541.2
59,590.4


15,065
4,835
3,991
3,129
1,844
795


0
2
0
2
0
1
3
5
2.5
0.8
•0.7
0.7
0.5
0.1
1.4
5.4
9.55x1 0-'°







Locations 3 and 4
0-10
10-20
20-30
30-40
40-50
50+
30+
Total
194,913
262,898
179,052
251,766
348,264
298,743


95,299
23,885
18,777
14,311
8,800
4,451


13
5
0
0
i
3
4
22
12.7
4.3
2.3
2.5
2.1
0.9
5.5
24.8
2.18x10-'°







                                    7.30

-------
  Table 7.12.  Comparison of Wittenoom, Australia (DeKlerk 2001) Mesothelioma Deaths
     to Predicted Deaths Assuming Risk Varies Linearly with Exposure Intensity After
          Controlling for Years Since First Exposure and Duration of Exposure
            Intensity (ffml)
      Range
   Average
                                  Mesothelioma Deaths
Person-Years
Observed
                                        Goodness of Fit P-value
Predicted
0-15
15-30

30-60

>60
Total
9.7
17.0

50.3

100.2

50,736
23,881

18,166

13,353

I 32
51

27
1
40
• 150
;: 19.5
22.4
i
, 41.9
i
,. 66.3
' 150.0
Table 7-12 shows that the assumption that the mesothelioma risk varies linearly with exposure
intensity leads to a 2-fold under-prediction of risks for exposure intensities below 60 f/ml (83
deaths compared to only 41.8 predicted) and a corresponding over-prediction for exposure
intensities above 60 f/ml (only 67 deaths compared to 108.2 predicted). Thus, instead of risk
varying linearly with exposure intensity, Table 7-12 indicates that the exposure response is
supra-linear, with lower fiber intensities being more potent per f/ml.      .         '
7.3.1.3
Discussion of Adequacy of Mesothelioma Model
The mesothelioma model (Equations 7-6 and 7-7) provides adequate fits to each of the three data
sets evaluated (Wittenoom, Quebec and South Carolina) when the data are categorized by time
since first exposure. The value of KM estimated from the cohort of crocidolite miners
(Wittenoom) was largest and was about 60-fold larger than the KM estimated from the South
Carolina chrysotile textile workers (based only on confirmed mesotheliomas in South Carolina)
and more than 100-fold larger than the estimates obtained from the Quebec chrysotile miners
who did not work in the factory that utilized crocidolite. This is consistent with numerous
indications from the literature that crocidolite is more potent than chrysotile in causing
mesothelioma. The relative magnitudes of the KM for the Quebec data estimated from three
locations track with the relative amounts of amphibole exposure estimated for these (locations,
which is also consistent with the hypothesis that the mesothelioma risk is greater from amphibole
exposure than from chrysolite exposure.  Despite the very few mesotheliomas in the South
Carolina cohort, the KM estimated from these data is larger than those from Quebec,' although the
discrepancy is not as large as estimated for lung cancer.

The Wittenoom and Quebec data were evaluated further to see if they were consistent with the
prediction of the mesothelioma model that risk continues to increase indefinitely after exposure
has ceased. By comparing the observed number of mesothelioma deaths to the number predicted
by the mesothelioma model at various times since the cessation of exposure, no evidence was
found that mesothelioma risk dropped off below that predicted by the model, at least up to 40 to
50 years after the cessation of exposure.  To the contrary, in Wittenoom there was some evidence
that the model under-predicted at the longest times since the end of exposure, as past 30 years
                                          7.31

-------
from the cessation of exposure there were 96 observed mesothelioma deaths compared to only
76.8 predicted by the model.

Chrysotile is much more soluble than crocidolite and consequently a chrysotile fiber exhibits a
much shorter residence time in the body than a comparable-sized crocidolite fiber.  Based on
in vitro studies a chrysotile fiber with a diameter of 1 nm will dissolve in body fluid in
approximately 1  year whereas a 1-um crocidolite fiber will take 60 years to dissolve (see Section
6.2.4). It is noteworthy that despite the short residence time of chrysotile fibers, mesotheliomas
deaths have occurred in the Quebec cohort more than 50 years following the cessation of
exposure.  This suggests that either these mesotheliomas are the result of amphibole
contamination of the ore, or else long residence times for inhaled fibers are not necessary for the
production of mesotheliomas.

The mesothelioma model predicts that risk is proportional to the intensity of exposure (Equation
7-6) and, at long times past the end of exposure, to cumulative exposure (Equation  7-8).  Two
analyses of the Wittenoom mesothelioma data suggest that the assumption of a linear exposure-
response may not be valid. First, whereas the model is linear in the value of the integral  in
Equation 7-7, a very poor fit was obtained when the data were categorized according to the value
of the integral (Table 7-8). Second, an analysis that categorized the data by intensity of
exposure, while controlling  for both duration of exposure and time since last exposure, also
provided a poor fit (Table 7-12).  Both of these  analyses exhibit a supra-linear exposure-response
in which less intense lower exposures are more  potent  per f/ml than more intense ones.

In light of these findings, it is interesting that supra-linearity in the exposure-response
relationship for lung cancer has also been suggested by fits of data in several studies (see Section
7.2.1.1). Although the effects for lung cancer are difficult to separate from the confounding
effects from smoking, common suggestions of supra linearity for both disease endpoints
certainly indicate a need to evaluate the nature of the exposure-response models for asbestos in
greater detail. This is consistent with the recommendations of the expert panel (Appendix B)
that evaluation of a broader range of exposure-response models for mesothelioma is appropriate.

This analysis of the Wittenoom data appear to be one of very few that provide information on the
shape of the exposure-response relationship for mesothelioma.  It is possible that systematic
errors in the exposure data for Wittenoom could have resulted in an apparent supra-linear
exposure response.  Like most asbestos-exposed cohorts, the estimated exposures for Wittenoom
are uncertain. If higher estimated intensities are overestimates, a linear exposure response would
appear to be supra-linear, and a linear fit to such data would underestimate the true KM.

In addition, errors in exposure measurement even if unbiased, can tend to make a linear dose-
response appear  supra-linear (Crump 2003). Even  if the supra-linear exposure response in the
Wittenoom data  is real, the exposure-response at lower doses is likely to be linear, but with a
larger KM value than was obtained in the fit to the complete data set. If this is the case, our
analysis (Table 7-12) suggests that the KM for the lower exposures is about a factor of two larger
than the value estimated from the complete cohort. A factor of 2 is not extremely large
compared to the  other sources of uncertainty in  the analysis.  Provisionally, the value of KM
estimated  from the complete Wittnoom cohort will be applied in our analysis. In revisions to this
                                          7.32

-------
document, it will be important to attempt to evaluate the exposure-response for mesothelioma
using data from additional other cohorts.                                        ;
                                                                             i'
Given the importance of these two issues: (1) the relative potencies of chrysotile andjthe
amphiboles and (2) the adequacy of U.S. EPA models for predicting the time and exposure
dependence of disease, limited analysis of raw data from a small number of additional cohorts is
warranted. Recommendations for further research along these lines are discussed in the
conclusions to this Chapter (Section 7.6) and parallel the recommendations of the expert panel
(Appendix B).

7.3.2 Estimating KM Values from Published Epidemiology Studies

At the time that the Health Effects Update was published (U.S.: EPA 1986), four studies were
found to provide suitable quantitative data for estimating a value for KM and six additional
studies provided corroborative support for the mesothelioma model applied. Currently, there are
14 published studies with adequate data for deriving  an estimate of KM (including updates to all
four of the quantitative studies evaluated in 1986).                               ..
                                                                             i
The U.S. EPA mesothelioma model (Equations 7-6 and 7-7) was applied to each of these data
sets to obtain study-specific estimates for the mesothelioma dose-response coefficient, KM.  The
resulting set of KM values are presented in Table 7-9. The format for this table  is identical to that
described in Section 7.2.2 for Table 7-6. As with the KL values in Table 7-6, the KM values for
all studies presented in Table 7-9 (including those studies that have not been updated since their
inclusion in the 1986 Health Effects Update) were re-derived using the modified procedures
described in Appendix A. Uncertainty intervals in Table 7-9 for each estimated KM were derived
using the method described in Appendix A.                 •                    '
                                                                             t
As Table 7-9 indicates, the KM values derived in this study and the corresponding values derived
in the original  1986 Health Effects Update are in close agreement, none vary by more than a
factor of 1.5. Among the KM values derived in the current study, the lowest and highest of the
best-estimate values differ by a factor of approximately 1,400 (excluding the one negative study
of Connecticut friction product manufacturers) and many of the pair-wise sets of uncertainty
intervals do not overlap.  For example, none of the uncertainty intervals for the  KM values
derived for any of the environments involving exposure to chrysotile overlap the uncertainty
intervals associated with the KM values derived for either crocidolite mining or  asbestos-cement
manufacture using mixed fibers at the Ontario plant (Finkelstein 1984). Furthermore^ neither of
the uncertainty intervals for the KM values derived for chrysotile mines in Quebec (Asbestos and
Thetford) overlap the intervals around KM values for  any of the amphibole environments or any
of the mixed environments, except the Quebec factory that is associated with the Asbestos mine
(Liddell et  al. 1997).                                       [

The KM values and the associated uncertainty bounds derived in the current study are plotted in
Figure 7-2. Each exposure environment is plotted along the X-axis of the figure and is labeled
with a 4-digit code that indicates fiber type (chrysotile, mixed, crocidolite, or tremolite), industry
(mining,  friction products, asbestos-cement pipe, textiles, insulation manufacturing, or insulation
application); and a 2-digit numeric code indicating the study from  which the data were derived.
                                          7.33

-------
The key for Figure 7-1 also applies to this figure.  In Figure 7-2, the chrysotile studies are
grouped on the left, amphibole studies are grouped on the right, and mixed studies are in the
middle. As in Figure 7-1, data from the Dement et al. (1994) study are used to represent the
South Carolina textile cohort and data from the Amandus and Wheeler (1987) are used to
represent the Libby mine cohort in Figure 7-2. Also, the estimated KM values for the Quebec
miners (Liddell raw data) from Asbestos and Thetford Mines, respectively, are averaged in this
figure.
                                         Figure 7-2:
          Rot of Estimated K^, Values and Associated Uncertainty Intervals by Study Environment

              1000^
               100.
          00
          o
                0.1,
               0.01 r
               1E-3
                         5    u. 0. I-
                         O    000
00 O)
CL Q.
                                                           10 (o Is-    oo
                                                              55    2
                                           7.34

-------
Figure 7-2 indicates that, within'chrysotile studies alone, lowest and highest KM values vary by
approximately a factor of 15 (excluding the negative friction products study), which is minimal
variation relative to the spread across the values from all of the studies in the table. Also, the
corresponding uncertainty intervals have considerable overlap.                    l:
                                                                             s
The KM values for the two "pure" amphibole studies (crocidolite mining and amosite,
manufacturing) agree to within a factor of 2 and, with the exception of the value from the
Ontario asbestos-cement plant (MP8 in Figure 7-2 and the ninth study listed in Table 7-9) and
the factory plant associated with the Asbestos, Quebec mines (MF14 in Figure 7-2 arid the 11th
study listed in Table 7-9), the mixed exposure KM values lie between the high values;for the
"pure" amphibole exposures and the lower values reported for all of the chrysotile sites.
Although the K.M value for the asbestos-cement plant (MP8) appears high, its corresponding
uncertainty interval overlaps those of the other "pure" amphibole studies as well as those of a
number of the other mixed exposures. The KM value reported for the factory associated with the
Asbestos, Quebec mine (MF14) is the lowest of those reported for mixed exposures, put its
uncertainty interval overlaps those for several of the other mixed exposures, as well as all of
those involving only chrysotile exposures.                  .                    \

As with KL values, the asbestos-cement pipe industry shows the greatest variation in'KM values
across all asbestos types, with a range of more than a factor of 90. Moreover, the uncertainty
intervals for the largest and smallest values in this industry do not overlap. Within the textile
industry, the KM value for the South Carolina chrysotile plant is only an eighth of the values
reported for the two textile plants using mixed fibers, although there is considerable overlap in
their uncertainty intervals. The potential sources of variation across all of the KM  values are
likely attributable to the same sources of variation identified  for the KL values and previously
described in detail (Section 7.22).                          '                   ''.

7.4   EVALUATION OF ASBESTOS EXPOSURE INDICES

As indicated in Chapter 3, for exposure-response coefficients derived from one environment to
be applicable in a different environment requires that both of the following two conditions be
satisfied:                                                ;                   !'
                                                        I

      •      that asbestos be measured in both environments in an identical manner; and

      •      that such measurements reflect (or at least remain proportional to) the
             characteristics of asbestos exposure that determine biological activity:

When these two conditions are met, it is possible to define an "exposure index" that accurately
reflects biological activity, and consequently an exposure-response coefficient based upon such
an exposure index derived in one environment can confidently be applied in a different
environment. Such an index can be defined as a weighted sum of concentrations of categories of
structures of different asbestos types and sizes, where the weights reflect the relative
carcinogenic potencies of the different type and size categories.  For example, as described in
Section 6.4.3, Herman et al. (1995) derived an optimal exposure index from an analysis of rat
inhalation studies involving exposures to different types of asbestos and fibrous structures  of
differing dimensions. The optimum index (defined in Equation 6-11) consists of a weighted  sum

                                          7.35

-------
of the air concentrations of structures between 5 and 40 um in length and >40 urn in length (all
thinner than 0.4 urn).

There is considerable evidence that the manner in which asbestos is quantified in the available
epidemiology studies (i.e., PCM) may not adequately reflect the characteristics that relate to
biological activity (Sections 6.4 and 6.5).  Therefore, the second of the above two criteria may
not be satisfied when exposure-response coefficients (i.e., KL and KM values) derived from these
studies are used to predict risk in new environments.  We therefore investigated the possibility of
adjusting the KL and KM values so as to apply to a measure of exposure that better reflects
biological activity. Such an adjustment requires both (1) data from each studied environment on
fiber size and asbestos type needed to adjust the corresponding KL and KM and (2) evidence
regarding what measure of exposure better reflects biological activity.

7.4.1  Fiber Type and Size Distribution Data Available for Deriving Exposure Indices

Since the range of possible adjustments to the KL and KM values is constrained by the data on
fiber size and type available from each studied environment, we first consider the characteristics
of the data currently available for making such adjustments. The data considered to be pertinent
consisted of TEM analyses of samples conducted in the same environment in which an
epidetniological study was conducted or from an environment involving a similar operation (e.g.,
mining, textile manufacture, etc.). Table 7-13  lists available fiber size distributions obtained
from a search of the literature and categorized  by fiber type and type of operation.  The
epidemiological studies from which KL or KM have been calculated are categorized accordingly.

Assuming that the available TEM size distributions are representative of dust characteristics for
the exposure settings (industries) studied, these were paired with corresponding epidemiology
studies. The TEM size distributions were then used to convert the exposure measurements used
in the epidemiology studies to the new exposure index that potentially  better reflects biological
activity.  Studies were paired as indicated in Table 7-13.

Only a subset of the TEM size distributions listed in Table 7-13 were actually employed in the
effort to normalize KL and KM values. To minimize variability resulting from differences in
TEM analysis methodology employed by different authors,  it was decided to employ
distributions from common studies conducted by common groups of researchers, to the extent
that this could be accomplished without reducing the number of "size-distribution-
epidemiological study" pairs available for inclusion in the analysis. Also, studies containing the
best documented procedures were favored. Ultimately, with one exception the size distributions
selected for use came from only two studies, which were reported in three publications: Dement
and Harris (1979), Gibbs and Hwang (1980), and Hwang and Gibbs (1981). In the case of the
one exception, size distributions for Libby (tremolite asbestos in vermiculite) were derived from
unpublished TEM data recently acquired directly from the site.
                                           7.36

-------
    Table 7-13. Correlation Between Published Quantitative Epidemiology Studies and
                         Available Tern Fiber Size Distributions
Fiber Type  Exposure Setting
                      Distribution Reference
                             Epidemiology Reference
Chrysotile  , Textiles
Chrysotile
and
Crocidolite
Crocidolite

Amosite
            Friction Products
            Mining and Milling
            Asbestos Cement
            Manufacturing
Asbestos Cement
Manufacturing
                      Dement and Harris 1979
                      Cherrieetal. 1979

                      Dement and Harris 1979
                      Marconi etal. 1984
                      Winer and Cossett 1979
                      Roberts and Zumwalde 1982
                      Rood and Scott 1989
                      Gibbs and Hwang 1980
                      Winer and Cossett 1979    '
Dement and Harris 1979

Snyderetal. 1987
Winer and Cossett 1979
Hwang and Gibbs 1981
Mining and Milling      Gibbs and Hwang 1980
                      Hwang and Gibbs 1981
Insulation Manufacturing Dement and Harris 1979
            Insulation Application
            Insulation Clearance

Tremolite    Vermiculite Mining
                      Snyderetal. 1987
                      Cherrieetal. 1979
                      U.S. EPA, unpublished
                             Dement etal. 1994, 1983b
                             McDonald et al. 1983a,b
                             Peto 1980a; Petoetal. 1985
                             Berry and Newhouse 1983
                             McDonald et al. 1984
Liddelletai. 1997
McDonald et al. 1980b
Nicholson et al. 1979
Piolatto et al. 1990
Hughs etal. 1987;'
Finkelstein 1984 ,
               i
               \
Finkelstein 1983 '
Hughs etal. 1987
Weilletal. 1979
Weill etal. 1994
Albin etal. 1990
Armstrong et al. 1988
de Klerk, unpublished data
Levin etal. 1998 ;
Seidmanetal. 1986
Seidman 1984
Selikoffetal. 1979
                             McDonald et al. 1986
                             Amandus and Wheeler 1987
                                          7.37

-------
Table 7-14 presents the resulting bivariate fiber size distributions derived from the published
TEM data that are paired with representative KL and KM values from the corresponding
epidemiology studies. This table shows 17 KL values and 11 KM values matched with a fiber size
distribution from the literature. In this table some length and width categories from some  •
published distributions have been combined, so that, for the most part, only those categories
available for all of the epidemiological studies are presented. The column labeled "PCME"
("PCM-equivalent") provides the relative concentrations of asbestos fibers that would have been
identified by PCM (*5 nm in length and ^0.2 urn in width).
The fiber size distributions in Table 7-14 are not all of equal relevance to the respective
epidemiological studies to which they were paired. As indicated in Table 7-15, some of the
distributions are based on data collected at the same facility, others are based on data collected at
a similar facility, still others are based on a combination of data from similar facilities, etc. The
uncertainty factors listed in Table 7-15 were developed to quantify the relevance of each fiber
size distribution to its paired epidemiological study, where larger factors indicate a less certain
relevance. How these factors were used is explained below. It should also be kept in mind that,
whereas these fiber distributions were based on air samples collected over a fairly narrow time
range, they are used to represent the fiber size distributions throughout the exposure period,
which in most of the epidemiological studies covers many years.

As indicated in Tables 7-14 and 7-15, for the two environments for which multiple studies were
available (i.e., the South Carolina textile plant and the Libby, Montana vermiculite mine), a
single study was selected to represent each environment. For the South Carolina textile plant the
Dement et al. (1994) study was selected and for Libby, the Amandus and Wheeler (1987) study
was selected.

7.4.2 Modification of Existing KL and KM to Conform to a New Exposure Index

The fiber size distribution data in Table 7-14 are used to transform the existing KL and KM values
(which are defined in terms of PCM measurements) so they conform to a different exposure
index based on TEM. To see how this is accomplished, consider a KL value pertaining to a
specific environment, let CPCM be an air concentration from that environment measured by PCM,
let Cns.w be the concentration in the same air measured by a new exposure index using TEM (e.g.,
perhaps defined as a weighted sum of TEM concentrations in various length, width and asbestos
type categories), and let KL* be the adjusted exposure-response coefficient corresponding to the
new exposure index. It is clear that for the U.S. EPA lung cancer model (Equation 7-1 or 7-2) to
estimate the same risk from the given air concentration using either exposure index, it is
necessary that

                               (K. ) (C    ) — (K *) (C   )                        (Eq 7-9)

Using CpcME (the air concentration of PCM-equivalent fibers -  fibers measured by TEM that
would be identified  by PCM) as a replacement for CPCM, we get

                                                '
(Eq.7-10)
                                          7.38

-------
 a

 E

I
 _e


 x

 M
 S

 8
 v.
 S


 I

 S
 S
 "S
TS
 a
A

H
       !


            I
            o

            I
          fs

            I
            e
            S
          rfl
                                    x   ^
                                                        2    S
                       i§  gs
                       §2  §2
                                             o - a i
                   ?   8
                       3
                       3
                  I   I
                            s   S
                           a
S
.s   I   s
3   S   *
I   1   s
I   -g   s
•<   •>   o


!   i   I
d,   d,   S

s   "   £
sis
o   o   ^
                      S.    8
                      o    d

                      a    a
TV    Ml
                      9.   S
                                **i   —
                           o   o

                           2   P
                           p   a
         o
         o"
S   I
5   S
                                        i
                                        i
                                r-   c*1    r»
                                2   2    R
                                    S    3
                                    g    s
                                    I    I
                                    d    d
                                    !   I
                           §    §
                                             §   §
                                             d   d
                                             f*1    NO

                                         S   3    P
                           8
                           d
                                             C%    ^|

                                             P:    S
                                             m    o
                                             1   1
                  £
                  3
                                             O   m   n
                                                      s
                                                      d

                                                      >o
                                                      P
                                                          ^ 3 f § -
                                                                 i
                                                                 d
                                                    S.sf 00 AI 0*
                                                    s ^ s *



                                                I    m   s
                                                                         g
                                                                         d
                                                                                             t '  oo   —
                                                                     §•
                                                                              I   I
                                                          §OK

                                                          «H
                                                                                            fi   I
        I!
                                                                                              ' V °
                                                                                              ;;' s
                                                        £•  I
                                                        r!  3
:,' s
                                                                                             11
                                                        g
                                                        3
                                                        o


                                                        R
                                                        S
                                                        o
                                                                                             s   s
                                                                                             d  ' d
                                                                                            S   8
                                                                                            e  • d
                                                                                            S   g
                                                                                            i  •
                                                                               3
                                                                               d
                                                                      o    d   d
                                                                      ?*•  '  v   v
                                                                               §
                                                                                                a   a
                                                                                                                  0)
                                                                                                                  CO

-------
I
                                 «8
.1
w
 w



I


I
 09
 s

f
.1
 Efl
^
*
 a

I
 8
a

i
 A

I
w

in
                        v

                       1
                       H
1
B
_e

A

1
Q.
X
                               &

                               -S  U
                               "§"§
 e


 i

•o
2
                                             «S
                                             -a
                                                           oo
                                                           ON
                                                           tS
                                                           •Q
                                                                  r~
                                                                  oo
                                                                          T 0
                         &     •&
                                                                   ON

                                                                   C
                                                                                          •§
                                                                             r-

                                                                             ON
                                                                             "S


                                                    8
                                                    1
                                                    OH
                                    5      S
                                    «      -
                                    1
                                    2

                                        I
                                        Q

                                               I
                                               I
                                               CQ
                                                    !
                                                    PU
                                                                                 a.§      >f
                                                                                   1
                                                                                             *e
                                                                                          J 1
                                                                                           o5
                                                                                          «§i

 IS
 I

1
 a
•§ ja
a'i
                                             l
                                             1




                                                                                                                   >,


                                                         I
                                                         42
                                                                                CQ
                                                                   2
                                                                   1
                                                                    n
                                                                   o
                                                                   l



-------

g
•+*
QQ



t
gned
*P4
S

K

a
««
r^
       2
Explanation
   •a
   W
     a o
       c
       e
       •s
       S
       o



       I
            !i

            II
            os  S
            •« ^
            3 <8
,£>
(•<•>
00
o\
                    V

                    •o
                         15
                         a,
          i

          •a
          «
          .1

          1
                    to
                    00
                    os


                              "8
          _0

          a,
          ^
          (L>
          a
                              I
                              •O
                              .0    —


                   ao
                   o\


                                         .5 <*5
                    CM
             rN  •— <

             S  2
                                        S


             n>
          w  -
                           i

                «
                            1
                                            -

-------
In the actual calculation of KL*, this equation is applied with the ratio of air concentrations
appearing on the right side of this expression replaced by the equivalent ratio of fiber proportions
from Table 7-14.

As an example of the calculation of a KL*, an earlier draft of this report proposed use of an
exposure index defined as the weighted sum,
0.997
                                            0.003 C
                                                   5
-------
 In view of the limitations of the fiber distribution data, it was decided, as an interim measure, to
 adopt a maximum fiber width of 0.4 urn in the proposed exposure index for both lung cancer and
 mesothelioma. This is the width indicated by the animal data (Herman et al. 1995). Subject to
 that decision, we then derived separate exposure index indices for lung cancer and   ,
 mesothelioma, respectively, that are each optimal (as defined below) with respect to fiber type
 (chrysotile and amphibole) and fiber length (5-10 um and >10 um).  Based on results' of Berman
 et al.  (1995) and lack of compelling evidence elsewhere in the literature (assuming that size
 effects are adequately addressed, see Chapter 6), it was assumed that all similarly-sized
 amphibole fibers are equipotent.                           ,                     '

 To develop separate potency estimates for chrysotile and amphibole fibers (adjusted for fiber
 size)  it was necessary to estimate the relative amounts of chrysotile and amphibole in;each
 environment.  Table 7-16 presents our estimates of the fraction of exposure in each study
 environment contributed by amphibole asbestos, based on information on  each environment
 available in the literature.  The source of the information used to develop each estimate as well
 as a brief description of how the estimate was developed is also provided.
 7.4.3.1
Optimizing the Exposure Index for Lung Cancer
A statistical model was fit to the KL values in Table 7-14 and results from the fitting were used to
estimate separate potencies for amphibole and chrysotile, to estimate relative potencies of fibers
of different sizes, and to test certain hypotheses. In this model Ln(KL) (the log transform of a KL
value in Table 7-14) was assumed to be normally distributed with mean equal to
Ln{KLa*
        + rpc . (1 - f^)]. [q . C3.10 + (1 - q) . C>10]
                                                                               (Eq. 7-12)
In this expression C5.,0, and C>]0 are the fractions of fibers among those thinner than OA jim that
are between 5 and 10 pm in length or longer than 10 urn, respectively (from Table 7-T4), CPCME
is the fraction of PCME fibers (also from Table 7-14), and fmpll is the fraction of amphiboles
estimated for each environment (from Table 7-16). In addition there are four parameters that are
estimated by fitting the model to the KL values:

  q - the relative potency of fibers thinner than 0.4 p,m and between 5  and 10 \im in length,
  relative to fibers thinner than 0.4 \im and >10 jim in length;   •

  KLa* - the potency (KL value) of pure amphibole (based upon the exposure index defined by q);

  rpc - the relative potency of chrysotile, relative to amphibole                     !

The fourth parameter, a, is described below.

Note that the part of Equation 7-12 that  is inside the curly brackets is like Equation 7-9 solved
for KL but with an additional term to account for the relative amounts of chrysotile and
amphibole.                                                '
                                          7.43

-------










•8
CB
iS
1
S
.A
on
•3
.s
J
"o
.0
Fraction of Amphil
•o
4»
*•
w
.§
i

SO
^•H
t^
_«
3
rt
H





























1
'S
D.
E
<
> «
•0 -S
s ®
s <•>

B
_O
C3
g
hJ
>,
"O
2
!»
£
sS
t**

Os
2 
"«3
3
J


^ (4
rt rt
T3 -d
•«.g VD-3
oo 
3 ns 3
a E ex










o
S
13
«
o
1
o
Cu.







u-i
2
o



CO
o




ff\
s
U

T3
1
§
1
e
•*
00
Os

13
ts
•^
13
i
u
s
•T
OO
Os
"rt
McDonald et
b
2
t/1
2
•*" §
McDonald etal. 198
extrapolated from pi:



2




ID
O




•
1
"a.
1
'€
^
1 S


r-
oo
Os
13
u
1
g>
X

r-
oo
OS
"3
•s
33
f



«_t
"5.
^ E
si
• -a
13 «
« S
tit
3 tt to
X G .5



2





*"^




ID
OH
U


(U
s.
5 e
U «S
7, "3,
1
o
S2
S
-?
I'i
«N -g
i!
« 5
D S

S
Os
Dement et al
(raw data)
o
U
•g-8
*U A
S3 o'
•2 -
^ e
Sebastien et al. 1989
on), extrapolated fro
source material



2





O




P
O


1
6
•s **
11
Cfl (X









U
tn
3
O

b fi
g ^
a 2
b
.a
<-TI OT
S 2
ai
rti TO
Berry and Newhous*
extrapolated from pi



2




ID
O




r-
u-
2


b
2
1
^+H
CA
S
m




•
f
a
b
2
.12
15
•4_»
i
Hughes etal. 1987,
extrapolated from pi



*D
»— <
r!.




v>





0\
OH
^H
2


crt
i
§«
Is
^ "3.










o
o\
Os
13
•s
c
'£
<
"O
S
a
I
Albin etal. 1990, ex
from plant history



2




fTJ




O

2


1'
'a-
*^5
.2
1
£

-------
. I









DS Dusts (continued)
-*^
2
^P
Amphiboles in ;
•s
|

£
•o
1
E
•«
w
*?
n
H







y
B
4)
£
S




KL Reference




rp

^


o
••

f*\ ^&
«—« ,(-H V%
"*b^ id. |~^
s^ s . s
o
- '| c
§ I
g CX 3
ta »T 3
«I! O vi
a i .s
«i -s "»
D < -D






ON
2


ON
ON



t
crt














workers
00
ON
*"^
•**
13
3
c
Q
2

_o
McDonald et al. 1983
£>
Q
IS
•£>" "3
it
i!
u -o
2 i
13 .5
§ a
1 1

VN
T



oo


1Q
,_—
LM
•S
1
"5.
Pennsylvania ]




oo
ON
2
"S
(X


oo
ts
|
Q.

"«
5
etal. 1985, extrapc
plant history
| |

2
i
1/1
(S


VI


[""•*
1_—
. CJn
S

Rochedale,
England plant
T3
V
.2

3
1
•fcf
JU
Q -S'


DeKlerk, unpublished
data



ral estimate8

i^
1
•3
1
£
13
1
5,
1
o
t*-c
1 o

ji*
|
1
t_i
to

-------
As previously indicated, the current approach for estimating asbestos-related risk (U.S. EPA
1986) uses (effectively) PCME as the exposure index. To allow comparison with this approach,
analyses were also conducted using PCME as the exposure index, rather than the size range of
fibers considered heretofore (i.e., Equation 7-12 was simplified to Ln{KLa* , [f^ + rpc, (1 -
famph)])-  Results of this analysis are shown on the right side of Table 7-17. With this exposure
index, the best estimate is that chrysotile is about one-half as potent as amphibole, the hypothesis
that chrysotile is non-potent can be rejected (p=0.001), and the hypothesis that chrysotile and
amphibole are equally potent cannot be rejected (p=0.20).

Based upon a comparison of either the residual variance, o, or the likelihood, it appears that the
exposure index defined by fibers longer than 10 urn and thinner than 0.4 urn (corresponding to
q=0 in Table  7-17) provides at most a very marginal improvement in fit over use of PCME as the
exposure index for lung cancer when rpc is held at 1. Similarly, adjusting for fiber type but not
size (i.e., optimizing rpc to reflect separate potencies  for chrysotile and the amphiboles) also
provides at best a marginal improvement over the current approach (PCME with rpc=l).
However, when the exposure index is adjusted for both fiber size and type, (comparing the
optimized values for Equation 7-12 to PCME with rpc=l), a small improvement is apparent. The
log-likelihood increases by half a unit and the spread in the estimated KL values (represented by
o) decreases by about 7%. Moreover, as discussed in the following section, the improvement for
mesothelioma is substantial..

In addition to the analyses reported in Table 7-17, other analyses were conducted in which an
additional term was added to the linear combination of fiber lengths appearing in Equation 7-12
to represent fibers shorter than 5 urn (still considering only fibers thinner than 0.4 fim).  In this
analysis, the best estimate of both the potencies of fibers shorter than 5 ^m, and between 5 and
10 p,m in length, was zero.

Discussion of the results from the analysis of the mesothelioma values (also presented in Table
7-17) is provided in Section 7.4.3.2. Based on this analysis and the rest of the evaluation
described in this chapter, a recommended set of lung  cancer and mesothelioma exposure-
response coefficients is presented in Section 7.5.
 7.4.3.2
Optimizing the Exposure Index for Mesothelioma
Concomitant with the analysis reported in Section 7.4.3.1 regarding development and evaluation
of an improved exposure index for lung cancer exposure-response coefficients (KL values) that
correlates better with biological activity, a parallel analysis was performed for the mesothelioma
exposure-response coefficients (KM values) presented in Table 7-14. This table presents data
from 11 environments in which KM values are paired with fiber size distribution data. The
corresponding uncertainty intervals for these 11 KM values are provided in Table 7-9. The same
relationship defined by Equation 7-9 was used to adjust these KM values to a different exposure
index and the same statistical model (Equation 7-12) was applied both to evaluate different
adjustments and to develop an adjustment that was optimal for the available data.  The results of
this analysis of KM values are presented in the bottom half of Table 7-17, which also contains the
results of the comparable analysis of KL values.
                                           7.48

-------
The results of fitting the model defined by Equation 7-12 to the KM values in Table 7-14 are
shown in the columns of Table 7-17 labeled "Equation 7-12".  The first of these columns, labeled
"optimized values" contains the resulting parameter estimates and log-likelihood with all four
parameters estimated. Just as was the case with the analysis of KL values, the best estimate of q
is q=0.  Since q represents the potency of fibers between 5 and 10 jim in length, relative to fibers
>10 um in length (considering only fibers thinner than 0.4 u.m), just as was the case for lung
cancer, the model predicts that fibers between 5 and 10 um are non-potent in causing
mesothelioma.                                                               i.
                                                        i                    i'
For mesothelioma, the best estimate of rpc is rpc=0.0013, which predicts that chrysotile is only
0.13% as potent as amphibole in causing mesothelioma (after adjusting for fiber size).  This
small estimate for rpc is not significantly different from rpc=0 (p=0.79). Consequently, in this
analysis, the data are consistent with the hypothesis that all of the mesotheliomas occurring in
cohorts exposed primarily to chrysotile are due to small amounts of amphibole contamination
within the chrysotile. Moreover, the hypothesis that chrysotile and amphibole are equally potent
in causing mesothelioma (the assumption inherent in the U.S. EPA (1986) asbestos health effect
document) is clearly rejected (p=0.0001).

Results using PCME as the exposure index for mesothelioma (the bottom right side of Table
7-17) are similar. The best estimate is that chrysotile is only 0.0033 as potent as amphibole, the
hypothesis that chrysotile is non-potent cannot be rejected (p=0.61), and the hypothesis that
chrysotile and amphibole are equally potent is definitely rejected (p=0.0007).

Based upon a comparison of either the residual variance, a, or the likelihood, it appears that the
exposure index defined by fibers longer than 10 urn and thinner than 0.4 |im (corresponding to
q=0 in Table 7-17) provides an improvement in fit over use of PCME as the exposure index for
mesothelioma. Even after adjusting for the effects of fiber type (i.e., comparing the 0 values
estimated for the optimized  values with PCME and Equation 7-12 as the exposure index,
respectively), the variation across KM values appears to decrease by more than 20% when values
are adjusted to the index of fibers that are longer than 10 um and thinner than 0.4 \im;
Moreover, comparing a values between the index in current use (PCME with rpc=l) (JU.S. EPA
1986) and the optimized index of longer and thinner fibers (with rpc=0.0013), use  of the latter
exposure index results in a 67% reduction in variation across K^ values.            :

As was also the case for the lung cancer analysis, the fit based on the.exposure index defined by
Equation 7-11, which was proposed in an earlier version of this report (with q,  the relative
potency of fibers between 5  and 10  \an compared to fibers longer than 10 urn fixed at q=0.003),
is virtually identical to the fully optimized fit (which predicts q=K)).

In addition to the analyses reported  in Table 7-17, another analysis for mesothelioma were
conducted in which an additional term was added to the linear combination of fiber lengths
appearing in Equation 7-12 to represent fibers shorter than 5 urn (still considering only fibers
thinner than 0.4 am). In this analysis, the best estimate of the potency of fibers shorter than 5
jim was also zero.                                          •         ,          ;.
                                          7.49

-------
Based on results in this section and the corresponding evaluation of lung cancer (Section 6.4.3.1)
a recommended set of lung cancer (and mesothelioma) exposure-response coefficients are
developed and presented in Section 7.5.

7.5  THE OPTIMAL EXPOSURE INDEX

7.5.1 Definition of the Optimal Index and the Corresponding Exposure-Response Factors

Table 7-17 presents the results of fitting a statistical model to asbestos exposure-response
coefficients to estimate the relative potencies for fibers of various sizes and types that define an
optimized exposure index for asbestos.  Although the coefficients for lung cancer (the KL values)
and mesothelioma (the K.M values) were separately evaluated, the optimal index for each
incorporates the same size range of fibers (at least based on the limited range of options
evaluated). These are fibers longer than 10 urn and thinner than 0.4 um.

Results in Table 7-17 also differentiate between the potency of chrysotile and amphiboje for both
lung cancer and mesothelioma.  Amphibole is estimated as being about four times as potent as
chrysotile for lung cancer (although the difference is not significant) and about 800 times as
potent as chrysotile for mesothelioma (a highly significant difference). Moreover, the data are
consistent with the hypothesis that chrysotile has zero potency toward the induction of
mesothelioma.

The optimized dose-response coefficients (rounded up) from Table 7-17 for pure fiber types
(chrysotile or amphibole) are summarized in Table 7-18. These coefficients apply to exposures.
quantified in terms of concentrations (in f/ml) of fibers longer than 10 [im and thinner than
0.4 |im.
                         Table 7-18. Optimized Dose-Response
                           Coefficients for Pure Fiber Types
Fiber Type
Chrysotile
Amphiboles
KL, x 100
0.6
3
KM.xl08
0.04
30
7.5.2 Evaluation of the Optimal Exposure Index

The optimal coefficients presented in Table 7-18, result from adjustments for both fiber type and
fiber size to the KL and KM values obtained directly from the literature (Tables 7-6 and 7-9),
which are linked to PCM measurements.  To get some idea of the relative effects of adjustments
for fiber size (from PCM to fibers longer than 10 um and thinner than 0.4 urn) and fiber type
toward the goal of rationalizing the KL and KM values obtained from different environments, the
effect of the two adjustments are considered sequentially—first the effect of adjusting for fiber
size is considered and next the added effect of adjusting for fiber type is evaluated.
                                          7.50

-------
o
              To compare the effects of adjusting KL and KM values for fiber size, the KL, and KM, values,
              which are adjusted to the optimal exposure index (using Equation 7-9) are plotted (along with
              their associated uncertainty intervals) in Figure 7-3, and 7-4. These figures are in a format
              identical to that of Figures 7-1 and 7-2 of the untransformed values. The "key" provided for
              Figure 7-1 is also directly transferable to Figures 7-2, 7-3, and 7-74.

              Note that one of the points plotted in Figures 7-1 and 7-2 was omitted from Figures 7-3 and 7-4
              (for the Enterline study [1986] of retired factory workers: study MX13) because it was felt that
              none of the available size distributions were  suitably applicable for this study site, so no
              conversions were possible. Also, the confidence intervals are larger in Figures 7-3 and 7-4
              because, as previously indicated, we have attempted throughout our analysis of the epidemiology
              data to account for major sources of uncertainty. Thus, the confidence intervals depicted in
              Figure 7-3 and 7-4 are modified from those depicted in Figures 7-1 and 7-2 to account for the
              uncertainty of making the adjustment for fiber size using paired data from published size
              distributions.  The intervals were adjusted by multiplying the upper bound and dividing the
              lower bound by a factor thought to represent the relative contribution to uncertainty contributed
              by the need to match data from separate studies to perform the conversions. The factors
              employed (along with the rationale for selecting the value of each factor) are provided in
              Table 7-15.     '

              The visual impressions from a comparison between Figures 7-1 and 7-3 for lung cancer and
              between Figures 7-2 and 7-4 for mesothelioma are that the changes resulting from adjusting for
              fiber size are subtle. This visual impression  is reinforced by the relative similarity of the fits
              (based on comparison of o and log-likelihoods) reported in Table 7-17 for PCME and the
              optimal exposure index, particularly for lung cancer.

              Figures 7-5 and 7-6 show the effects of adjusting exposure-response coefficients for both fiber
              size and type. To develop Figure  7-5 for lung cancer, a size-adjusted KL, as plotted in Figure
              7-3, was further adjusted to correspond to pure amphibole by dividing it by the factor [f^ + rpc
              . (1 - famph)], where fmf is the proportion of amphibole fibers estimated for a given environment
              (as listed in Table 7-16) and rpc=0.267, the optimal value from Table 7-17. The corresponding
              adjusted factors corresponding to pure chrysotile can be calculated simply by multiplying the
              amphibole value by rpc=0.267. Since the study-specific values for pure amphibole all differ
              from the corresponding value for pure chrysotile by a multiplicative constant, values for both
              types of asbestos are plotted simultaneously in Figure 7-5 by using a different scale for
              amphibole and chrysotile. The same method was used to develop Figure 7-6 for mesothelioma
              exposure-response coefficients, from the size-adjusted KM values'plotted in Figure 7-4, the only
              difference being that the mesothelioma rpc=0.0013 (Table 7-17) was used for this latter
              conversion. Comparing Figure 7-5 to 7-1 suggests that the adjustments for size and type resulted
              in as somewhat better reconciliation of the dose-response coefficients for lung cancer, although
              the improvement is still somewhat subtle. However, a comparison of Figures 7-6 to 7-2
              indicates a much more dramatic improvement in the case of mesothelioma.  Comparisons of
              Figures 7-2, 7-4, and 7-6  indicate this is primarily due to the adjustment for fiber type, although
              the subsequent adjustment for fiber size provides further, noticeable improvement.
                                                        7.51

-------
                               Rgire7-6:
Rotrf EsGmfedK^and K^V/auesa^
iuuu-
100
§
X 10
 O
          S   ^LLQ-I—   LL Q_ 0- *-, T- T-   \-T-T-   5-T-CM
          O   oOOO   5SS||^   ig^   |<<
                                 7.54

-------
Rot cf Estimated K^ and K^VaLes and
                              Rgue7€:
     10000 1
      1000-
 b
 _g>
  o

 Ic
  Q.


  03
       Q1
      001
                                                              1
                                                    oo
                                                    O


                                                Q'i  S
                                                 I   ^^^


                                                    5

                                                QQ1 8
                                                  :  T3
                                                1E-4  ®
                                                  ,  to
                                                  I  .3,


                                                1E5<
CM


O
                    00
                             oo en
                             CL Q.
                                7.55

-------
In addition to the graphical comparisons, it is also instructive to consider numerical comparisons.
Table 7-19 reproduces values for all of the original (study-specific) KL and KM (from Tables 7-6
and 7-9, respectively) along with all the corresponding values for the adjusted KL, and KM,.
Study-specific estimates for the all the corresponding KLA, KLC, KMA, and KMC are also presented.

Table 7-20 presents the magnitude of the spread in the range of original and adjusted KL and KM
values estimated as the quotient of the maximum and minimum values of each range. Note that,
of necessity, such an analysis requires that the zero values obtained for the Connecticut friction
products plant (CF4) be omitted.  Note further that the data sets evaluated  in Table 7-20 for the
original and adjusted values are identical (i.e., the one study for which no suitable fiber size
distribution could be found was excluded).

In Table 7-20, for mesothelioma, the spread in unadjusted values among "pure" fiber types (i.e.,
chrysotile only or amphibole only) are both substantially smaller than those for mixed data sets
(containing both fiber types). This provides further evidence of differences in the potencies of
each fiber type toward mesothelioma. It is also apparent that adjusting for fiber size decreases
the range within pure fiber types. Moreover, by simultaneously adjusting  for both fiber size and
type (as illustrated by the column labeled, "K^"), the range of variability across the entire data
set of 10 mesothelioma studies is reduced from almost 1,100 to a factor  of 30, which is a clear
arid substantial improvement.

For lung cancer, the results presented in Table 7-20 are a bit more complicated. While there is a
substantial reduction in the spread of unadjusted values among "pure" amphibole environments
when mixed environments are excluded, the spread among "pure" chrysotile environments is no
different than that observed for the entire data set.  This is because, as previously described
(Section 7.2.2) the difference between the KL values observed among chrysotile miners in
Quebec and chrysotile textile workers in South Carolina represent the extremes of the entire
range of reported  KL values. Adjusting these exposure-response coefficients for size provides
some improvement in agreement across these two environments.  This reinforces the finding
from Appendix D that, if size adjustments incorporating cutoffs for longer fibers could be
incorporated into  a new exposure index for asbestos, the apparent discrepancy between the
exposure-response observed among Quebec miners and South Carolina  textile workers (and,
thus, among KL values as a whole) is likely to be further reconciled.  The impressions from the
figures discussed  above and from Table 7-20 confirm the conclusions concerning the
quantitative improvement in agreement across exposure-response coefficients highlighted in
Sections 7.4.3.1 and 7.4.3.2 and reinforce the conclusion that, especially with regard to
mesothelioma, adjusting exposure-response coefficients for fiber size and  type leads to
substantially improved agreement across studies.
                                           7.56

-------








$
9
1
<*
K>
*
•o
fl
«
J
*
1
JV
A
V.
s/
jdG
«
V
W
«
kJT
;*

J
«
e,
M
O
'3
w
a
I/}
>>
•B
S
5
ON
f
f-
Jj
3
£





^^
y»
g ®
1 *E
o
9)
t
5 ac
*E

Jo^
s
1
*H
! ,e
^i


,•*»!,
"s*
i/r<
m X
\*«^


|i--fc>
fc o
s>2
E
^^
so"
S/3
E


o
Jo
*§
N^P^
a>
o
i
«s
£
il

KL Reference
£• «
H
35U
Environment





VO
O
o




2:
ON
fi
O







oo
o
o






ON
ts
o
o






Liddell et al.
1997

2
O
Quebec mines
and mills

VO
*~*
p*
8
o


irt
vi







oo
fn


£
r**
ON


 C -5
O 3 T3

DeKlerk,
unpublished
data
oo
•g
2
O at
Q ••«
> <
VO
r-
ON
•*
fl
O
d
^ r2
§ S
§ S
^ d
5
o
ON
VD
«N

•* —
O c^i
oo v)
vo ON
f^l OO
in oo
f"" d

^
p
VO
fS
•* f-
r- fn
o o
vo oo
ro ^O
t-: °°
- d
ON

f*N


r*N

— o

,

73
0
§
is
u ON
t/5 —

73
.« 73
i t>
Is Is
•^2^2
ON O
— (N
N>M I-H
< <
:z a
-- |
•it »"t
;|j ^^





5-
^
ON
Tf

^
d






Amandus and
Wheeler 1987
^•sl
Ol
H
Libby,
Montana
oo
ON
•*
fl

t<^ rn
— •*
d ;. d
ON
S
rj
vo
< '""'

S i. p
1 , §
d ! -
ro
i| m
" 00
: os
vq
06
I ^
vo , vi
t=- :• trt


F^


-------












f
1
a
s



a

j>.
J

•H
B

^
J
s
r
s

J
!p
*
£
u
tu
Q«
on
•e
a?

•
1






u£
S Jjji?
•""• liC M
*-C ^w*
o
1
5 ^C

^j T"j
n2

i jf
O sJi
^g
a
s -o
*rf , 5* ^^
ptf *"2
s*^r

-4-

3 w
* o

^
d8
* *,

if
>-'
d!
a*
u
i
*5?
tf
i
01
w
1
£
£
^
-&«
85
—
a
i
B
B
OO
OO
o

s
t^
VO
«N
cs
as
3

i»*
oo
d
oo
CM
00
p
'"H
d

m
d


~3
15
en
•f £
x2


"|SS
x 2
Os
1

1
Sco
g "I
IZ "H.






OJ
oo
oo
Os
VO
«— «
d


in
m
oo
tf)
VO
d





fN
OO
rt
OS
oo
d


i i s







„ _;
« s 1
d n
r-
Os
oo
p

?!
r-
o
^
o 2
d
oo
d

S
T3 "^ "e
** & »
w 5 t*-
« ^- tw
•§ r- is1 ^
3 gj g TJ
J — 5 00


Selikoffand
- 12
S §
c
o
I
ll 1
•S S <«
< a s
































^^
OS
•"^
§
'53
on

^_,
Os
'5




workers

oo
VO
^J-
o
d

g

^^
Os
M
OO
Os
in '

rs)

-------
  Table 7-20. Comparison of Spread in Range of Original and Adjusted K, and Km Values
                               for Specific Fiber Types
Fiber Type Ranges of Values
Number in KL,
KK ^ptf K K
L"»*»LX aels *H, **L*
All fiber types 15 72 67
combined
Chrysotile only .4 72 51
(excluding
mixed)
Chrysotile only 3 8.6 5.0
(also excluiding
textiles)

Chrysotile and 11 72 51
mixed settings
Textiles only 3 5.1 5.1
Amphibole and 11 31 55
mixed settings
Amphiboles 4 8.5 8.5
only (excluding
mixed)
Number in
KLX KM, KM.,KMX
Sets
52 10
•
52 • 3


5.0 ; 2

1

52 8

5.7 i 3
30 , 7

8.5 2
I

KM
1,089

15


12



1,089

5.2
60

2.0


*„.
795
1*
11


7.0
1

i
794
i
5.2
60

1.7
!
i
K.
30

19


7.0



30

2.9
11

1.7


7.6  GENERAL CONCLUSIONS FROM QUANTITATIVE ANALYSIS OF HUMAN
     EPIDEMIOLOGY STUDIES

The following conclusions result from our evaluation of the available epidemiology studies.

 (1) To study the characteristics of asbestos that relate to risk, it is necessary to combine results
     (i.e., in a meta analysis) from studies of environments having asbestos dusts of differing
     characteristics. More robust conclusions regarding risk can be drawn from an analysis of
     the set of epidemiology studies taken as a whole than results  derived from individual
     studies.                                                                i

 (2) By adjusting for fiber size and fiber type, the existing database of studies can beireconciled
     adequately to reasonably support risk assessment.        :

 (3) The U.S. EPA models for lung cancer and mesothelioma both appear to track the time-
     dependence of disease at long times following cessation of exposure, however, the
     relationship between exposure concentration and response may not be adequately
     described by the current models for either disease. There is some evidence that these
     relationships are supra-linear.                         '                  !
                                         7.59

-------
(4) Whereas the U.S. EPA model for lung cancer assumes a multiplicative relationship
    between smoking and asbestos, the current evidence suggests that the relationship is less
    than multiplicative, but possibly, more complex than additive. However, even if the
    smoking-asbestos interaction is not multiplicative as predicted by the U.S. EPA model,
    exposure-response coefficients estimated from the model are still likely to relate to risk
    approximately proportionally and, consequently, may be used to determine an exposure
    index that reconciles asbestos potencies in different environments.  However, adjustments
    to the coefficients may be required in order to use them to estimate absolute lung cancer
    risk for differing amounts of smoking. This issue needs to be investigated further in the
    next draft of this document.

(5) The optimal adjustment found for fiber size and type that best reconciles the published
    literature assigns equal potency to fibers longer than 10 \an and thinner than 0.4 urn and
    assigns no potency to fibers of other dimensions.  Different exposure-response coefficients
    for chrysotile and amphibole are assigned both for lung cancer and mesothelioma. For
    lung cancer the best estimate of the coefficient (potency) for chrysotile is 0.27 times that
    for amphibole, and for mesothelioma the best estimate of the coefficient (potency) for
    chrysotile is only 0.0013 times that for amphibole.

(6) Without adjustments, the lung cancer exposure-response coefficients (KL values) estimated
    from 15 studies vary by a factor of 72 and these values are mutually inconsistent (based on
    non-overlap of uncertainty intervals). By simultaneously adjusting these values for fiber
    size and type, the overall variation in KL values across these studies is reduced to a factor
    of 50.

(7) Without adjustments, the mesothelioma exposure-response coefficients (KM values)     -
    estimated from 10 studies vary by a factor of 1,089, and these values are likewise mutually
    inconsistent. By simultaneously adjusting these values for fiber size and type, the overall
    variation in KM values across these studies is reduced to a factor of 30.

(8) The exposure index and exposure-response coefficients  embodied in the risk assessment
    approach developed  in this report are more consistent with the literature than the current
    U.S. EPA approach.  In particular, the current approach  appears highly likely to seriously
    underestimate risk from amphiboles, while possibly overstating risk from chrysotile.
    Furthermore, most the remaining uncertainties regarding the new proposed approach also
    apply to the current approach. Consequently, we recommend that the proposed approach
    begin to be applied in assessment of asbestos risk on an  interim basis, while further work,
    as recommended below, is being conducted to further refine the approach.

(9) The residual inconsistency in both the lung cancer and mesothelioma potency values is
    primarily driven by those calculated from Quebec chrysotile miners and from South
    Carolina chrysotile textile workers. The difference in the lung cancer potency estimated
    between these studies has long been the subject of much attention.  A detailed evaluation
    of the studies addressing  this issue, the results of our analysis of the overall epidemiology
    literature, and implications from the broader literature, indicate that the most likely cause
    of the difference between these studies is the relative distribution of fiber sizes in the two
    environments.  It is therefore likely that the variation between these studies can be further

                                         7.60

-------
      reduced by developing improved characterizations of the dusts that were present in each of
      these environments (relying on either archived samples, or newly generated samples using
      technologies similar to those used originally).
                                                          1                    i
Recommendations for Limited, Further Study

The two major objectives identified above for further study are: •
                                                                              t
  (1)  to evaluate a broader range of exposure-response models in fitting the observed
      relationship between asbestos exposure and lung cancer or mesothelioma, respectively.
      For lung cancer models, this would also include an attempt to better account for the
      interaction between asbestos exposure and smoking; and

  (2)  to develop the supporting data needed to define adjustments for exposure-response
      coefficients that will allow them to be used with an exposure index that more closely
      captures the criteria that determine biological activity (see Section 6.5). Amongiother
      things, this work should focus on obtaining data that would permit more complete
      reconciliation of the exposure-response coefficients derived  for Quebec miners and South
      Carolina textile workers.
                                                                              i
                                                                              i
The first of the above objectives requires access to raw data from a small number of selected,
additional epidemiology studies.  The best candidate studies include: (for chrysotile) the lung
cancer data from Quebec (best) or, potentially, from the New Orleans asbestos-cement|pipe plant
studied by Hughes et al. For amphiboles, the best candidate studies include: the lung cancer and
newest mesothelioma data from Libby, or, potentially the lung cancer and mesothelioma data
from the Paterson, New Jersey insulation manufacturing plant studied by Seidman et al. (1986).
The possibility of obtaining some  or all of these data sets needs to  be further explored.'

The second of the above objectives requires more detailed  size characterization data for the
environments of interest.  Although archived air samples do not appear to be available from any
of the study locations of interest (except South Carolina), we believe that suitable data .can be
developed from appropriate bulk samples. Thus, for example, it would be useful to obtain
samples of the raw ore from Libby, Wittenoom, and Quebec andithe textile, asbestos-cement
pipe, and friction-product grade products from Quebec.

Results from our review of the supporting literature suggest that the optimum cutoff for
increased potency occurs at a length that is closer to 20 u,m than 10 um, (the latter of which is
the  cutoff in the exposure index provided in this study).  Data do.not currently exist to improve
on this latter cutoff. However, provided that study-specific size distribution data could be
obtained as indicated above, with the appropriate analyses, it will be possible to develop the size
distributions necessary to evaluate a range of considerations including:

  (1) delineation of size fractions among individual length categories out to lengths as long as
     30 or 40 u,m;

  (2) determination of the relative presence and importance of cleavage fragments (of non-
     biologically relevant sizes) in mine ores vs. finished fibers; and
                                                           I
                                          7.61

-------
(3) the relative fraction of fibrous material vs. non-fibrous particles in the various exposure
    dusts of interest.
                                           7.62

-------
             8.0 DISCUSSION, CONCLUSIONS, AND

                         RECOMMENDATIONS

Although gaps in knowledge remain, a review of the literature addressing the health-related
effects of asbestos (and related materials) provides a generally .consistent picture of the
relationship between asbestos exposure and the induction of disease (lung cancer and
mesothelioma). Therefore, the general characteristics of asbestos exposure that drive the
induction of cancer can be inferred from the existing studies and can be applied to define
appropriate procedures for evaluating asbestos-related risk.  Moreover, such procedures provide
substantial improvement in the confidence that can be placed in predicting risks in exposure
environments of interest compared to risk predictions based on procedures in current iuse.

Following a general discussions of the findings of this study, specific recommendations for
finalizing a protocol for assessing asbestos-related risks using the procedures identified in this
document are provided in Section 8.2, Recommendations are described in Section 8.3 for
limited, focused, additional studies to:

       (I)    settle a small number of outstanding issues (concerning whether better models
             exist than the current U.S. EPA models for tracking the time and concentration
             dependence of the exposure-response relationships for asbestos-induced lung
             cancer and mesothelioma);                                      ;

       (2)    improve the manner in which smoking is addressed in the modeling of asbestos-
             induced lung cancer; and

       (3)    provide the data required to fully optimize the approach recommended'in this
             document (i.e., reconciling the published epidemiology studies by addressing the
             effects of fiber size and type).

Moreover, these recommendations parallel those of the expert panel convened to peer-review the
previous version of this report (Appendix B).                '                   ;!
                                                                          i.
8.1    DISCUSSION AND CONCLUSIONS

8.1.1   Addressing Issues

The issues identified in the introduction (Chapter 2) as part of the focus of this study can now be
addressed. These are:

       •     adequacy of existing models: whether the exposure-response models currently in
             use by the U.S. EPA for describing the incidence,of asbestos-related diseases
             adequately reflect the time- and exposure-dependence for the development of
             these diseases;
                                         8.1

-------
       •      relative potency for different mineral types: whether different potencies need to
              be assigned to the different asbestos mineral types to adequately predict risk for
              the disease endpoints of interest;

       •      biodurability: to the extent that different asbestos mineral types are assigned
              distinct potencies, whether the relative in vivo durability of different asbestos
              mineral types determines their relative potency;

       •      minerals of concern: whether the set of minerals included in the current definition
              of asbestos adequately covers the range of minerals that potentially contribute to
              asbestos-related diseases;

       •      analytical methods: whether the analytical techniques and methods currently used
              for determining asbestos concentrations adequately capture the biologically-
              relevant characteristics of asbestos (particularly with regard to structure sizes), so
              that they can be used to support risk assessment; and

       •      extrapolation of risk coefficients: whether reasonable confidence can be placed in
              the cross-study extrapolation of exposure-response relationships that are required
              to assess asbestos-related risks  in new environments of interest.

The Adequacy of Existing Models. Regarding the first of the above-listed issues,  both the U.S.
EPA lung cancer model and mesothelioma model appear to adequately reflect the time-
development of asbestos-induced lung cancer. For lung cancer, the assumption  in the model that
risk remains constant with time following the end of exposure was confirmed for cohorts
exposed, respectively, to chrysotile, to crocidolite, and to amosite (Section 7.2.1). A similar
analysis for mesothelioma suggests that, as that model predicts, risk for mesothelioma continues
to increase with the square of time since the end of exposure (Section 7.3.1).

Regarding exposure concentration, we did find some evidence suggesting that, for both lung
cancer and mesothelioma, the relationship between exposure concentration and risk may not be
linear, but rather supra-linear (Sections 7.2.1.1 and 7.3.1.2). If confirmed, this would contradict
the assumed strictly linear relationship in both the current lung cancer and mesothelioma models.

For 15 of the 18 studies that were fit using the lung cancer model (Equation 7-2) in  our analysis
(Appendix A and Section 7.2.2), the parameter,«, (which indicates differences in background
lung cancer incidence between cohorts and controls) was greater than  1.0 and significantly so in
six cases. In these cases, if it is assumed instead that the background lung cancer mortality rate
applied to the cohort is appropriate,  the correct fitting would be with 
-------
 Importantly, because any such evaluation would be greatly enhanced by broadening the number
 of data sets utilized, it is further recommended that all means be explored for acquiring raw data
 sets for cohorts from additional epidemiology studies.

 At the same time, although there are suggestions from our analysis that models other than the
 current models might better describe the relationship between exposure and risk for both
 mesothelioma and lung cancer any limitation associated with use of the current models for lung
 cancer and mesothelioma would be common to the procedures recommended in this report and
 the existing U.S. EPA approach for assessing asbestos-related risks. Therefore, given the degree
 of improvement demonstrated for the proposed approach over the current U.S. EPA approach,
 there appears to be little reason not to adopt the proposed approach  as an interim measure, while
 further research is conducted to address the remaining outstanding issues highlighted by the
 expert panel (Appendix B) and highlighted in this report.                         i'
                                                                      i       '
 Relative Potency for Different Mineral Types.  As indicated in Sections 7.4.3.2 and 7.5.2, our
 analysis indicates a substantial difference in the relative potency of amphiboles and chrysotile
 toward the induction of mesothelioma, with amphiboles estimated as almost 1,000 times more
 potent than chrysotile (fiber-for-fiber). Moreover, this difference was shown to be highly
 statistically significant. When fiber size and type  are simultaneous  addressed,  variation across
 the 10 published epidemiology studies included in our analysis drops from a factor of more than
 1,000 to a factor of 30 (Table 7-19). This, coupled with the growing evidence  in the;literature
 supporting this difference in potency among fiber types, provides strong support for defining
 distinct risk coefficients for chrysotile and the amphiboles to assess the risk of mesothelioma.
                                                                             i
 The situation with lung cancer is less clear. Although our analysis suggests that the best estimate
 is that (fiber-for-fiber) amphiboles are about 4 times more potent than chrysotile toward the
 induction of lung cancer, this difference was not found to be statistically significant (Sections
 7.4.3.1  and 7.5.2).  This issue also remains unresolved in the wider  literature. It is also possible
 that the confounding effects of smoking, coupled with the lack of adequate  data for properly
 assessing the effects of smoking may be limiting our ability to.address this question.! Thus, this
 is one of the issues that would likely benefit from  additional research.             !,

 At the same time, when a small difference in potencies is incorporated into  our meta analysis of
 the epidemiology studies we evaluated, variation across the data set is reduced by about 30%
 (Table 7-20). This suggests that incorporating a small difference in lung cancer riskipoefficients
 for chrysotile and the amphiboles is reasonable.

 Biodurability. Because the in vitro dissolution rate for chrysotile in biological fluids is
 substantially greater than for crocidolite and, likely, other amphiboles (Section 6.2.4), effects
 potentially attributable to differences in the relative biodurability of these asbestos types are
 addressed  in several sections of this document.  It is possible that such differential biodurability
.at least partially explains the clear difference in potency between chrysotile and the amphiboles
 toward mesothelioma (along with the possible, albeit smaller, difference toward lung cancer).
 However, that no difference was observed in the time-development  of either lung cancer or
 mesothelioma following exposure to chrysotile or amphibole asbestos (respectively), suggests
 that any relationship that exists between potency and biodurability must be  more subtle and
 complicated than the obvious effect on internal dose. At the same time, there is ample literature

                                            8.3

-------
evidence that some kind of relationship in fact exists (Section 6.2.4). While more research into
this relationship may prove interesting (and may be useful for assessing effects of less durable
fibers), it is unlikely to have a direct impact on procedures for assessing asbestos-related risk.

Minerals of Concern. Regarding the range of fibrous minerals that potentially contribute to
lung cancer and mesothelioma, available evidence (Sections 6.2 and 6.4)  suggests that:

       •      several minerals and the most biodurable among synthetic fibers (such as erionite
              or refractory ceramic fibers) in addition to those included strictly within the
              definition for asbestos  have been shown capable of inducing lung cancer and/or
              mesothelioma (as long as the corresponding fibers fall within the appropriate size
              range);                                                       .

       •      fibrous minerals that differ radically in chemical composition or crystal structure
              (such as erionite, chrysotile, and the amphibole asbestos types) appear to exhibit
              substantially different potencies (even after adjusting for size); and

       •      fibrous minerals that exhibit closely related chemical compositions and crystal
              structures (such as the  family of amphibole asbestos types) appear to exhibit
              relatively similar potencies (once effects are adjusted for size). To illustrate this
              point for mesothelioma, consider that the range of variation in estimated KM's
              over 10 studies (which include studies of exposures to tremolite, amosite, and
              crocidolite in addition  to chrysotile) is reduced to a factor of 30 (from almost
              1,100), once the effects of size and differences between the amphiboles and
              chrysotile are accounted for (Table 7-20). Moreover, among the 7 of these
              studies reflecting exposure exclusively to amphiboles (which still includes
              exposures to tremolite', crocidolite, and amosite), the range of variation is
              reduced to a factor of 11 (Table 7-20). Regarding lung cancer, the four studies of
              "pure" amphibole environments (which includes exposures to tremolite,
              crocidolite, and amosite) vary by only a factor of 8.5 (Table 7-20) and this range
              is bounded by studies of the same  mineral: amosite, which certainly suggests that
              mineralogical differences do not drive the observed variation (Section 7.2.2).

Given the above-described observations, it is clear that fibrous minerals beyond those included
in the definition for asbestos can contribute to lung cancer and mesothelioma. It is also likely
that potencies for minerals that exhibit similar chemistry and crystal structure (and which
therefore likely exhibit similar physical-chemical properties) also exhibit corresponding potency
(for similarly sized fibers). However, the carcinogenicity of fibers exhibiting radically different
chemical compositions and crystal structures than those already identified as carcinogenic should
be evaluated on a case-by-case basis.

Two additional  considerations may be useful for  focusing such evaluations. First, the size
distribution for  fibers composed of the mineral of concern should be shown to include substantial
        'Formally, the exposures at Libby are to the amphibole mineral richterite. However, this mineral is closely
 related to tremolite and has sometimes been called "sodium tremolite" because it contains a greater fraction of
 sodium than the composition range that is commonly termed tremolite.

                                             8.4

-------
 numbers within the range of structures that potentially contribute to biological activity (e.g., that
 fall within the size range defined by the improved index recommended in this study, i.e.,
 structures longer than 10 urn and thinner than 0.4 [am, see Section 7.5).  Second, such fibers
 should also be shown to be relatively biodurable (i.e., that they exhibit dissolution rates less than
 approximately  100 ng/cmz-hour, Lippmann 1999).

 Analytical Methods. Given the need to detect the thinnest fibers and the need for reliable
 measurements in outdoor settings, the only analytical technique that appears to be capable of
 providing quantitative data useful for supporting risk assessment  is TEM (Sections 4.3 and 7.6).
 Further,"given the specific size range  of structures that need to be evaluated and the specific
 manner in which they need to be counted (to assure both cross-study comparability and
 compatibility with the recommended  dose-response coefficients), analyses should be performed
 using the specific analytical methods  recommended in this document. These are ISO 10312
 (modified to focus on interim index structures) for air and the Modified Elutriator method
 (Berman and Kolk 2000) for soils or bulk materials. However, on a study-specific basis, other
 methods may be shown to provide comparable results so that they can also be used as part of a
 properly integrated investigation.                                              :

 8.1.2  Comparison with  Other, Recent Risk Reviews

 Although several other reviews have also recently been published that nominally address risk-
 related issues for asbestos  (including questions concerning the identification of an appropriate
 exposure index and the relative potency of varying fiber types), these studies are either
 qualitative or involve analysis of data in a manner that does not allow formal evaluation of the
 nature of specific exposure-response relationships for the various diseases. Therefore, they are
 not well suited to support development of a protocol for conducting formal assessment of
 asbestos-related risks. Nevertheless, the general conclusions from these reviews arelnot
 inconsistent with our findings.
                                                                            t-
 Hodgson and Darnton (2000). Hodgson and Darnton (2000) conducted a comprehensive
 quantitative review of potency of asbestos for causing lung cancer and mesothelioma in relation
to fiber type. They concluded that amosite and crocidolite were, respectively, on the order of
 100 and 500 times more  potent for causing mesothelioma than chrysotile. They regarded the
evidence for lung cancer to be less clear cut, but concluded nevertheless that amphiboles
(amosite  and crocidolite) were between 10 and 50 times more potent for causing lung cancer
than chrysotile. In reaching this latter conclusion they discounted the high estimate of chrysotile
potency obtained from the South Carolina cohort. Hodgson and Darnton concluded that inter-
study comparisons for amphibole fibers suggested non-linear exposure-response relationships for
lung cancer and mesothelioma, although a linear relationship was possible for pleural
mesothelioma and lung tumors, but not for peritoneal mesothelioma.

The Hodgson and Darnton study was  based on 17 cohorts, 14 of which were among the 20
included  in the present evaluation. This study had different goals from the present evaluation
and used  different methods of analysis. Hodgson and Darnton>did not use the exposure-response
information within a study. Instead, lung cancer potency was expressed as a cohort-wide excess
mortality divided by the cohort mean exposure.  Likewise, mesothelioma potency was expressed
as the number of mesothelioma deaths divided by the expected total number of deaths,

                                          8.5

-------
normalized to an age of first exposure of 30, and by the mean exposure for the cohort. These
measures have the advantage of .being generally calculatable from the summarized data available
from a study. However, since they are not model-based, it is not clear how they could be used to
assess lifetime risk from a specified exposure pattern, which is an important goal of the current
project.  Use of average cohort exposure could cause biases in the estimates, if, e.g., a large
number of subjects were minimally exposed. Also, the recognized differences between studies
in factors, in addition to level of exposure, that may affect risk will also affect the reliability of
conclusions concerning the dose response shape based on comparisons of results across studies.

Lippmann (1994,1999).  In the most recent of these reviews, Lippmann (1999) reinforces our
general findings that it is longer fibers (those longer than a minimum of approximately 5 urn)
that contribute to lung cancer and mesothelioma. He further indicates that, based primarily on
the limits observed for fibers that can be phagocytized, fibers that contribute most to lung cancer
are likely longer  than a minimum of 10 jim.  In his review, based on a series of comparisons of
mean and  median dimensions reported for the relevant exposures across a broad range of studies,
Lippmann draws several fairly specific conclusions on the ranges of fiber sizes that may
contribute to various diseases (i.e., that the minimum length fibers that contribute to asbestosis,
lung cancer, and  mesothelioma are 2, 5, and 10 u.m, respectively).  He also suggests that fibers
that contribute to mesothelioma may need to be thinner than 0.1 \im while those that contribute
to lung cancer may need to be thicker than 0.15 um. While it is not clear that drawing such
specific conclusions can be firmly supported by the kinds of qualitative comparisons across
reported mean and median dimensions for exposures in various studies that are described in this
paper, the author indicates that further, more formal study of the dose-response relationships that
he posits is warranted. It is noted that many of the studies reviewed by Lippmann (1999) are
also  incorporated in our analysis.

In the earlier review, Lippmann (1994) plots lung tumor incidence as a function of inhaled
animal dose for data from a series of broadly varying studies based, respectively, on fibers
longer than 5,  10, and 20 urn (no widths considered) and suggests that the quality of the fits  are
comparable. The author further suggests, based on this evaluation, that PCM seems to provide a
reasonable index of exposure. However, no formal goodness-of-fit tests were performed in this
analysis and, based on visual inspection, none of the plots would likely show an adequate fit.
Moreover, the plot of the tumor response vs. dose as a function of fibers longer than 5 |im
appears to be substantially worse than the other two plots; if one removes the single highest  point
in this plot, it appears that any correlation will largely disappear.

Stayner et al. 1996. In the context of evaluating the "amphibole hypothesis", Stayner et al.
(1996) computed the excess relative risk of lung cancer per fiber/ml/year from  10 studies
categorized by the fiber types to which the cohort was exposed.  Each of these studies was also
included in the present evaluation. Both the lowest and highest excess relative risks came from
cohorts exposed  exclusively to chrysotile. Based on their evaluation, they concluded that the
epidemiologic evidence did not support the hypothesis that chrysotile asbestos is less potent than
amphibole for inducing lung cancer. However, based on a review of the percentage of deaths in
various cohorts from mesothelioma, they concluded that amphiboles were likely to be more
potent than chrysotile in the induction of mesothelioma.  They also noted that comparison of the
potency of different forms of asbestos are severely limited by uncontrolled differences in fiber
sizes. None of these conclusions are inconsistent with our general findings.

                                           8.6

-------
8.2    RECOMMENDATIONS FOR ASSESSING ASBESTOS-RELATED RISKS
                                                        1                    '(
The optimum values for the risk coefficients for lung cancer and mesothelioma (the adjusted KL
and KM values) derived in our analysis are provided in Table 7.-18. Although these values are
optimized (within the constraints of the current analysis) and use of these values reduces the
apparent variation across published studies substantially (Section 7.5.2), the need to pianage and
minimize risk when developing a general approach for assessing risk, is also recognized. Thus,
to reduce the chance of under-estimating risks, a conservative set of potency estimates were
developed (Table 8-1) by adjusting upward the best-estimate potency coefficients listed in
Table 7-18 to provide additional health protectiveness.  The manner in which this was
accomplished is described in the following paragraphs.                           '
                     Table 8-1. Conservative Risk Coefficients for
                                   Pure Fiber Types
Fiber Type
Chrysotile
Amphiboles
K;.xlOO
5.5
20
K^xlO8
: 0.15
100
Reviewing the KLc and KLa dose response coefficients values listed in Table 7-19, it is apparent
that the maximum values for lung cancer derive from the Dement et al. (1994) study;of the South
Carolina textile plant. As indicated in Appendix A, the South Carolina study is a high quality
study. Moreover, we were able to obtain the raw data from this study and have analyzed these
data in detail.  We found, among other things, a well-behaved (i.e., monotonic) exposure-
response trend in this study for lung cancer. Therefore, because this study is of high quality and
provides the largest values of Ku and KLa, the values from this study were selected for our
conservative estimates of the corresponding potency coefficients.                 ,
                                                                            I;
Reviewing the KMa and KMc values listed in Table 7-19, it is apparent that the maximum values
derive from the Finkelstein study (1984) of the Ontario asbestos-cement factory.  As indicated in
Appendix A, the data from this study appear problematic. Among other things, the exposure-
response relationships observed in this study are not well-behaved (i.e., not monotonic).
Moreover, the value for a estimated for this study is the highest of any study we evaluated.
Possible reasons for such a large a include large discrepancies between the background
incidence of lung cancer between cohort and controls in this study and/or serious  errors in
exposure estimates. Given the potential problems associated with this study, (which suggests
that the potency coefficients estimated from this study may be less reliable than for many of the
other available studies), we decided to bypass this study and select the next highest values in
Table 7-19 for KMa and KMc as the conservative estimates to be recommended in this study.
Interestingly, these also turn out to be from the South Carolina textile study. Note that the
difference in the estimates from these two studies vary by less than a factor of two in any case.

Based on the above evaluation, conservative estimates  for the various potency coefficients
recommended in this  report are presented in Table 8-1.
                                          8.7

-------
Importantly, a measure of the degree of reconciliation among the results of the published
epidemiology studies that has been accomplished by the analysis presented in this report is
indicated by the ratios of the values presented in Tables 7-18 and 8-1, respectively.  The ratios
between the corresponding coefficients in Tables 7-18 and 8-1 are no more.than 10 for the lung
cancer potency coefficients and no more than 4 for the mesothelioma potency coefficients.
Moreover, the bounding study for the values presented in Table 8-1 is once again the South
Carolina textile study, which further reinforces earlier discussions identifying the particular need
to reconcile this study with the other chrysotile studies (particularly the Quebec mining study).

To assess risk, depending on the specific application, either the best-estimate risk coefficients
presented in Table 7-18 or the conservative estimates presented in Table 8-1 can be incorporated
into procedures described below for assessing asbestos-related risks.

Tables 8-2 and 8-3 present estimates of the additional risk of death from lung cancer, from
mesothelioma, and from the two diseases combined that are attributable to lifetime, continuous
exposure at an asbestos concentration of 0.0001 f/cm3.(for fibrous structures longer than 10 um
and thinner than 0.4 jim) as determined using TEM methods recommended herein. Table 8-2
was developed using the best-estimate values for risk coefficients defined in Table 7-18, and
Table 8-3 was developed using the conservative estimates defined in Table 8-1. Separate risk
estimates are provided for males and females and for smokers and non-smokers. The method
used to construct these tables is described in detail in Appendix E.

Separate estimates are presented for smokers and nonsmokers because the lifetime asbestos-
induced risk of both lung cancer and mesothelioma differ between smokers and non-smokers.
The asbestos-induced risk of lung cancer is higher among smokers because the lung cancer
model (Equation 7-2) assumes that the increased mortality rate from lung  cancer risk due to
asbestos exposure is proportional to background lung cancer mortality, which is higher among
smokers. Note that, while this is consistent with a multiplicative effect between smoking and
asbestos exposure that has been reported by several researchers (see, for example, Hammond et
al. 1979), some of the latest studies of the interaction between smoking and asbestos exposure
suggest a more complicated relationship (Section 7.2.1.3). This issue needs to be addressed
more fully in future analyses of these data. However, we believe the effects of such
considerations on the overall accuracy of asbestos-related risk estimates is, likely to be small
relative to other sources of error.
                                           8.8

-------
  Table 8-2. Estimated Additional Deaths from Lung Cancer or Mesothelioma per
100,000 Persons from Constant Lifetime Exposure to 0.0001 TEM f/cc Longer than 10
  urn and Thinner than 0.4 um - Based on Optimum Risk Coefficients (Table17-18)
Chrysotile
NonSmokers

Lung Cancer
Mesothelioma
Combined
Males
0.185
0.0836
0.269
Females
0.207
0.096
0.303
Smokers
, Males
1.6
0.0482
1.65
Females
1;5
0:0702
1.57
Amphibole
Lung Cancer
Mesothelioma
Combined
0.2
62.7
62.9
0.286
72.3
72.5
2.22
36.1
38.3
2.47
52.7
55.1
  Table 8-3. Estimated Additional Deaths from Lung Cancer or Mesothelioma per
100,000 persons from Constant Lifetime Exposure to 0.0001 TEM f/cc Longer than 10
 um and Thinner than 0.4 um - Based on Conservative Risk Coefficients (Table 8-1)
Chrysotile


Non-Smokers

Lung Cancer

Mesothelioma
Combined
Amphibole
Lung Cancer
Mesothelioma

Combined
Males
1.7

0.314
2.02

3.77
209

213
Females
1.9

0.361
2.26

4.41
241

245
i
!
Smokers
• Males
14.7
'
0.181
' 14.9

34.1
120
1
.154
Females
13.8
l
0.263
14
.t
33.2
175
<
209
                                    8.9

-------
If there is a desire to generate population averaged risks, this can also be accomplished using the
data in Tables 8-2 or 8-3. For such a case, simply choose the appropriate row (for lung cancer,
mesothelioma, or combined risk) from either Table 8-2 or 8-3 and substitute the four values
given in the row into the following equation to derive a single, population averaged risk:
                          0.5*[0.214*(MS + FS) + 0.786*(MNS + FNS)            (Eq.8-1)

       Where:
                    R^g   is the population averaged risk for the chosen disease endpoint;
                    MS   is the corresponding risk for male smokers;
                    FS    is the corresponding risk for female smokers;
                    MNS  is the corresponding risk for male non-smokers; and
                    FNS   is the corresponding risk for female non-smokers.

Note that Equation 8-1 is derived based on the assumption that 21 .4% of the general population
smokes (see Appendix E).

The asbestos-induced risk of mesothelioma is smaller among smokers because the mesothelioma
model (Equation 7-6) assumes that risk from constant exposure increases rapidly with age, with
the result that the predicted mortality rate is highest among the elderly. Thus, since smokers
have a shorter life span than non-smokers, their risk of dying from mesothelioma is also
predicted to be smaller.

Risks from lifetime exposures to asbestos levels other than 0.000 1 may be estimated from the
appropriate entry in Tables 8-2 or 8-3 by multiplying the value in the selected cell from the table
by the airborne asbestos concentration of interest and dividing by 0.0001 (i.e., by assuming that
the additional risk is proportional to the asbestos exposure level). Airborne asbestos
concentrations to be used in this manner must be estimates of lifetime average exposure and must
be expressed as structures longer than 10 (am and thinner than 0.4 (am derived as described
below.

Importantly, the risks provided in Tables 8-2 and 8-3 relate to exposure estimates expressed in
terms of the interim exposure index (i.e., estimates including only asbestos structures longer than
10 (im and thinner than 0.4 um).  Only exposures expressed in terms of the same exposure index
can be used to adjust the risk estimates to other levels of exposure (in the manner described
above). Use of exposures expressed in terms of any other index of exposure (such as the PCME
index in current use) will result in invalid estimates of risk.

The procedure described above for estimating risks using Tables 8-2 or 8-3 should provide good
approximations as long as the projected risk is no greater than 1,000 per 100,000.  Risks greater
than 1,000 per 100,000 (i.e., I in 100) that are derived from the tables are likely to be over-
estimated.  However, for risks associated with  short-duration exposures or exposures that differ
radically over time, it may be better to use a lifetable analysis or a modified version of Tables
8-2 or 8-3 that reflect the differences in exposure duration and frequency. This  is to avoid
substantially under- or over-estimating risk (depending on how the table might otherwise be
applied).
Tables 8-2 and 8-3 were derived using the approach described in Appendix E by incorporating

                                           8.10

-------
 the age-, sex-, and smoking-specific death rates reported for the general U.S. population and
 assuming that exposure is constant and continuous at the level indicated in the table, the
 underlying models are provided in Chapter 7 for cases in which'exposure is not constant
 throughout life and for which sufficient data exist to characterize the time-dependence of such
 exposure.  If available, there may also be cases in which it is advantageous to employ mortality
 data from a control population that better matches the exposed population of interest than the
 U.S. population as a whole.                                                     '

 For the interim, it is recommended that asbestos-related risks from constant low-leveliexposures
 be estimated using Tables 8-2 and 8-3. Although it is possible also to use the tables to estimate
 risk from short-term exposures by applying the corresponding long-term average exposure
 (derived from appropriate measurements, as described below),- this method can result in
 significant errors in some cases. It is anticipated that a flexible and user-friendly software
 package for evaluating risk will eventually be developed to supplement this document.  Such a
 package will be capable of accurately implementing the calculation method presented in
 Appendix E to calculate risks from general exposure patterns, rather than from constant
 exposures only.

 Requirements for Asbestos Measurements. One additional advantage of the approach for
 evaluating asbestos-related risks recommended in this document (in comparison to the current
 approach) is that the procedure for assessing risks is tied unambiguously to a specificjndex for
 measuring and expressing exposure (i.e., the index of all fibrous structures longer than 10 urn
 and thinner than 0.4 urn, as defined in Section 7.5) and this, in turn, is tied unambiguously to
 requirements for analyzing asbestos samples.                                    '

 Estimates of airborne asbestos concentrations that are required to support risk assessment can be
 derived either by extrapolation from airborne measurements or by modeling release and
 dispersion of asbestos from sources (soils or other bulk materials). In either case, exposure
 estimates must be representative of actual (time-dependent) exposure and must provide
 measurements that include only fibrous structures satisfying the dimensional criteriailisted in the
 last paragraph.  Additional considerations that need to be addressed to assure the validity of risk
 estimates derived using this protocol include:               '
                                                         <
       •      the array of samples collected for estimating airborne asbestos concentrations
              must be representative of the exposure environment;

       •      the time variation of airborne asbestos concentrations must be properly addressed;

       •      airborne samples must be collected on membrane filters that are suitable for
              preparation for analysis by transmission electron microscopy (TEM).
              Appropriate rocedures for sample collection are described in Chatfield and
              Herman (1990) or the ISO Method (ISO 10312)2;                  ':
       2Note that the ISO Method (ISO 10312) is a refinement of the method originally published as the Interim
Superfund Method (Chatfield and Berman 1990).  It incorporates improved rules for evaluating fiber morphology.
Both methods derive from a common development effort headed by Eric Chatfield.
                                           8.11

-------
       •     sample filters must be prepared for analysis using a direct transfer procedure (e.g.,
             ISO 10312).. Should indirect preparation be required (due, for example, to
             problems with overloading of sample filters), a sufficient number of paired
             samples will need to be collected and analyzed to establish a site-specific
             correlation between directly and indirectly prepared samples; .

       •     samples must be analyzed by TEM;

       •     samples must be analyzed using the counting and characterization rules defined in
             ISO 10312 and the structures used to determine exposure concentrations for use
             with this protocol need to satisfy the dimensional criteria defined in Section 7.5
             (i.e., structures longer than 10 jim and thinner than 0.4 nm).  Importantly, ISO
             Method rules require separate enumeration and characterization of component
             fibers and bundles that are observed within more complex clusters and matrices.
             Such components, if they meet the.dimensional criteria defined above must be
             included in the structure count;

       •     when risks are estimated using the risk tables (Tables 8.-2 or 8-3) the risk values
             selected from the-tables must be appropriate for the fiber type (i.e., chrysotile or
             amphibole) and the disease endpoint (i.e., lung cancer or mesothelioma) relevant
             to the situation of interest; and

       •     to use Tables 8-2 or 8-3, the concentration of total asbestos structures longer than
             10 jam and thinner than 0.4 u.m must be derived, divided by 0.0001, and
             multiplied by the risk estimate listed in the appropriate cell of the selected table to
             generate the risk estimate of interest.

Considerations that need to be addressed to assure the validity of risk estimates derived from soil
or bulk measurements combined with release and transport modeling include:

       •     the array of samples collected for  estimating source concentrations must be
             representative of the surface area or volume of source material from which
             asbestos is expected to be released and contribute to exposure;

       •     samples must be prepared and analyzed using the Modified Elutriator Method for
             soils and bulk materials (Berman and Kolk 1997, 2000), which is the only method
             capable of providing bulk measurements that can be related to risk;

       •     membrane filter samples prepared using the tumbler and vertical elutriator per the
             Superfund method must themselves be prepared for TEM analysis using a direct
             transfer procedure;

       •     TEM analysis must be conducted  using the counting and characterization rules
             defined in the ISO Method (ISO 10312) in precisely the same manner that is
             described above for air measurements. Also, the same size categories need to be
             evaluated (in the same manner described above) to estimate exposures for use
             with this protocol to assess risk; and

                                          8.12

-------
       •      release and dispersion models that are selected for assessing risks must be
              appropriate to the exposure scenario and environmental conditions of interest.
              Such models must also be adapted properly so that they accept input estimates
              expressed in terms of fiber number concentrations. Procedures suggested for
              adapting such models are illustrated in a recent publication (Berman 2000).

Note, if new analytical procedures can be designed to focus on long structures, risks can be
evaluated more cost-effectively. The alternate approach of spending a large proportion of
available resources counting many (potentially non-potent or marginally potent) short structures,
while not characterizing longer structures with adequate sensitivity or precision, leaves open the
possibility of missing potentially serious hazards because a small population of extremely potent,
long fibers were missed  in a particular environment.  Moreover, any potential contribution to risk
by shorter structures will be incorporated to some extent by default, i.e., to the extentjthat similar
proportions of such structures were also present in the environments from which the exposure-
response coefficients were derived and such structures are known to have been ubiquitous in
these environments (see, for example, Dement and Harris 1979; Gibbs and Hwang  1980; Hwang
and Gibbs 1981).                                          '                    ;

8,3    RECOMMENDATIONS FOR FURTHER STUDY                     :

A small number of limited and focused studies (described previously) are recommended in this
document because they are likely to provide very cost-effective improvements to the quality of
this document and may support substantial improvement to the recommended procedures for
assessing asbestos-related risks. The recommended studies are:                    ij
       (1)
              a focused study to expand our evaluation of the current U.S. EPA models to
              include other candidate models that might better track the exposure dependence
              of asbestos-related disease (Sections 7.2.2 and 7.3.2). Such models snould also
              be used to explore and better represent the relationship between smoking and
              asbestos exposure (to the extent that data suitable for supporting suchjan analysis
              can be acquired); and
       (2)
              a focused study to develop the supporting data needed to define adjustments for
              potency factors that will allow them to be used with an exposure index that even
              more closely captures asbestos characteristics that determine biological activity
              than the currently proposed index (Section 7.5).
Note that, by properly designing the second of the above-listed studies, it may also be possible to
further address another outstanding issue that was previously identified: the question of whether
exposure-response coefficients derived from mining studies are under-estimated relative to
studies involving asbestos products because exposures in the mining studies may contain large
numbers of non-asbestos particles contributed by the disturbance of host rock (Appendix D).
                                          8.13

-------
                            9.0  REFERENCES

Aalto M; Heppleston AG. Fibrogenesis by Mineral Fibres: An In-Vitro Study of the Roles of the
Macrophage and Fibre Length. British Journal of Experimental Pathology. 65:91-99. 1984.

Adamson IYR. Early Mesothelial Cell Proliferation After Asbestos Exposure: In Vivo and
In Vitro Studies. Environmental Health Perspectives.  I05(Suppl 5): 1205-1208. September.
1997.                         .                         '

Addison J. Consultant Mineralogist, John Addison Consultancy, Cottingham, East Yorkshire,
UK. 2001.  (private communication)

Afaq F; Abidi P; Matin R; Rahman Q. Activation of Alveolar Macrophages and Peripheral Red
Blood Cells in Rats exposed to Fibers/Particles. Toxicology Letters. 99:175-182. 1998.

Albin M; Jakobsson K; Attewell R; Johansson L; Welinder H.  Mortality and Cancer Morbidity
in Cohorts of Asbestos Cement Workers and Referents. British Journal of Industrial Medicine.
79(9):602-610. September. 1990.

Albin M; Pooley FD; Stromberg U; Attewell R; Mitha R; Johansson L; Welinder H. Retention
Patterns of Asbestos Fibres in  Lung Tissue Among Asbestos Cement Workers. Occupational
and Environmental Medicine.  51:205-211. 1994.

Albin M; Magnani C; Krstev S; Rapiti E; Shefer I.  Asbestos and Cancer: An Overview of
Current Trends in Europe. Environmental Health Perspectives. 107(2):289-298. May. 1999.,

Amandus HE; Wheeler R. The Morbidity and Mortality of Vermiculite Miners and Millers
Exposed to Tremolite-Actinolite: Part II. Mortality. American Journal of Industrial Medicine
11:15-26. 1987.

Amandus HE; Wheeler R; Jankovic J; Tucker J. The Morbidity and Mortality of Vermiculite
Miners and Millers Exposed to Tremolite-Actinolite: Part I. Exposure Estimates. American
Journal of Industrial Medicine. 11:1-14.  1987.

Armstrong BK; de Klerk NH;  Musk AW; Hobbs MST. Mortality in Miners and Millers of
Crocidolite in Western Australia.  British Journal of Industrial Medicine. 45:5-13. 1988.

Asgharian B; Yu CP. Deposition of Inhaled Fibrous Particles in the Human Lung. Journal of
Aerosol Medicine. 1:37-50.  1988.

Barchowsky A; Lannon BM; Elmore LC; Treadwell MD.  Increased Focal Adhesion Kinase-
and Urokinase-Type Plasminogen Activator Receptor-Associated Cell Signaling in Endothelial
Cells Exposed to Asbestos. Environmental Health Perspectives.  105:(Suppl )5:1131-1137.
September.  1997.
                                         9.1

-------
Barchowsky A; Roussel RR; Krieser RJ; Mossman BT; Treadwell MD. Expression and Activity
of Urokinase and its Receptor in Endothelial and Pulmonary Epithelial Cells Exposed to
Asbestos.  Toxicology and Applied Pharmacology.  152(2):388-396. 1998.        |

Baris YI; Simonato L; Artvinli M; Pooley F; Saracci R; Skidmpre J; Wagner C.    •
Epidemiological and Environmental Evidence of the Health Effects of Exposure to Erionite
Fibres: A Four-Year Study in the Cappadocian Region of Turkey. International Journal of
Cancer. 39:10-17. 1987.                                                  •
                                                                        i
Bauman MD; Jetten AM; Bonner JC; Kumar RK; Bennett RA; Brody AR.  Secretion,of a
Platelet-Derived Growth Factor Homologue by Rat Alveolar Macrophages Exposed to
Particulates/n Vitro. European Journal of Cell Biology. 51:327-334. 1990.      ;
                                                      1                  i
Beckett ST. The Generation and Evaluation of UICC Asbestos Clouds in Animal Exposure
Chambers. Annals of Occupational Hygiene. 18:187-198.  1975.                ;

Bellman B; Konig  H; Muhle H; Pott F.  Chemical Durability of Asbestos and of Man-Made
Mineral Fibres In Vivo. Journal of Aerosol Science. 17(3):341-345. 1986.       ,'

Bellman B; Muhle H; Pott F; Konig H; Kloppeel H; Spurny K. Persistence of Man-Made Fibers
(MMF) and Asbestos in Rat Lungs. Annals of Occupational Hygiene. 31:693-709. 1987.

Berman DW.  Asbestos Measurement in Soils and Bulk Materials: Sensitivity, Precision, and
Interpretation - You Can Have It All. In Advances in Environmental Measurement Methods for
Asbestos. ASTM STP 1342. Beard ME; Rook HL (eds.). American Society for Testing and
Materials.  2000.
                                                                        i.
                                                      ;                  i
Berman DW.  President, Aeolus, Inc., Albany, CA. (unpublished data)

Berman DW Chatfield EJ.  Interim Superfund Method for the Determination of Asbestos in
Ambient Air/'Part 2: Technical Background Document. Office of Solid Waste and Remedial
Response.  U.S. Environmental Protection Agency, Washington, D.C, EPA/540/2-90/005b.
May. 1990.                                                              ;

Berman DW and Crump KS. Methodology for Conducting Risk Assessments at  Asb'estos
Superfund Sites. Part 2: Technical Background Document.  Prepared for: Kent KitcHingman,
U.S. Environmental Protection Agency, Region 9. Work Assignment No. 59-06-D800 under
Contract No. 68-W9-0059.  1999.                                           j
                                                      ;                  i
Berman DW; Crump KS. Technical Support Document for a Protocol to Assess  Asbestos-
Related Risk.  Final Draft.  Prepared for the Volpe Center, U.S. Dept. of Transportation,
Cambridge, MA and the U.S. EPA, Region 8, Denver, CO.  Prepared under Contract No.
DTRS57-01-C-10044. September. 2001.                  '                  |:

Berman DW; Kolk AJ.  Superfund Method for the Determination of Asbestos in  Soils and Bulk
Materials.  Office of Solid Waste and Emergency Response. U.S. Environmental Protection
Agency, Washington, D.C., EPA 540-R-97-028.  1997.       '                  |'

                                        9.2

-------
Berman DW; Kolk AJ. Draft: Modified Elutriator Method for the Determination of Asbestos in
Soils and Bulk Materials. Aeolus, Inc. 751 Taft St., Albany, CA 94706.  2000.  (unpublished)

Berman DW; Crump KS; Chatfield EJ; Davis JMG; Jones AD. The Sizes, Shapes,' and
Mineralogy of Asbestos Structures that Induce Lung Tumors or Mesothelioma in AF/HAN Rats
Following Inhalation. Risk Analysis.  15(2): 181-195. 1995.

Berman DW; Crump KS; Chatfield EJ; Davis JMG; Jones AD. The Size  Distribution of Fibers
and Particles in Airborne Dust Generated for Selected Animal Inhalation  Studies, (in
preparation)

Bernstein DM; Morscheidt C; Grimm H-G; Thevenaz P; Teichert U. Evaluation of Soluble
Fibers Using the Inhalation Biopersistence, a Nine-Fiber Comparison. Inhalation Toxicology.
8:345-385. 1996.                         .          .         .

Berry G;  Newhouse ML. Mortality of Workers Manufacturing Friction Materials Using
Asbestos. British Journal of Industrial Medicine. 40:1-7.  1983.

Bertrand  R; Pezerat H. Fibrous Glass: Carcinogeriicity and Dimensional  Characteristics. In
Biological Effects of Mineral Fibres.  Wagner JC (ed.). IARC Scientific Publications.
pp. 901-911.  1980.                    '

Bignon J; Monchaux G; Sebastien P; Hirsch A; Lafuma J. Human and Experimental Data on
Translocation of Asbestos Fibers Through the Respiratory System. Annals of New York
Academy of Sciences. 330:745-750.  1979.                    "  .

Blake T;  Castranova V; Schwegler-Berry D; Baron P; Deye GL; Li C; Jones W. Effect of Fiber
Length on Glass Microfiber Cytotoxity. Journal of Toxicology and Environmental Health.
54(Part A):243-259.  1998.

Bolton RE; Davis J; Donaldson K; Wright A.  Variation in the Carcinogenicity of Mineral
Fibres. Annals of Occupational Hygiene. 26(l-4):569-582. 1982.

Bolton RE; Davis JMG; Miller B;  Donaldson K; Wright A. The Effect of Dose of Asbestos on
Mesothelioma Production in the Laboratory Rat.  Proceedings of the VIth International
Pneumonoconiosis Conference. 2:1028-1035.  1984.

Bolton RE; Addison J; Davis JMG; Donaldson K; Jones AD; Miller BG;  Wright A. Effects of
the Inhalation of Dusts from Calcium Silicate Insulation Materials in Laboratory Rates.
Environmental Research. 39:26—43.  1986.

Bonneau L; Malard C; Pezerat H.  Studies on Surface Properties of Asbestos.  Environmental
Research. 41:268-275. 1986.

Boutin C; Dumortier P; Rey F; Viallat JR; De Vuyst P. Black Spots Concentrate Oncogenic
Asbestos Fibers in the Parietal Pleura: Thoracoscopic and Mineralogic Study. American Journal
of Respiratory and Critical Care Medicine. 153:444-449.  1996.

                                         9.3

-------
 Brass DM; Hoyle GW; Poovey HG; Liu J-Y; Brody AR. Reduced Tumor Necrosis Factor-
 Alpha and Transforming Growth Factor-Beta-1 Expression in the Lungs of Inbred Mi
-------
CHAP (Chronic Hazard Advisory Panel on Asbestos). Report to the U.S. Consumer Product
Safety Commission. July. 1983.

Chatfield EJ. Ambient Air: Determination of Asbestos Fibres, Direct Transfer Transmission
Electron Microscopy Procedure. Submitted to the International Standards Organization: ISO/TC
10312. 1995.

Chatfield EJ; Berman DW. Interim Superfund Method for the Determination of Asbestos in
Ambient Air. Part 1: Method. Office of Solid Waste and Remedial Response. U.S.
Environmental Protection Agency, Washington, D.C. EPA/540/2-90/005a. May. 1990.

Chen YK. Ph.D. Dissertation. Mechanical and Aerospace Engineering, State University of New
York at Buffalo. 1992.

Chen YK; Yu CP.  Deposition of Charged Fiber in the Human Lung. Journal of Aerosol
Science. 24(Suppl 1):5459-5460.  1993.

Cherrie JW; Dodgson J; Groat S; Carson M.  Comparison of Optical and Electron Microscopy
for Evaluating Airborne Asbestos. Institute of Occupational Medicine, Edinburgh, (U.S.
Department of Commerce). 1979.

Cherrie JW; Dodgson J; Groat S; Carson M.  Comparison of Optical and Electron Microscopy
for Evaluating Airborne Asbestos. Institute of Occupational Medicine, Edinburgh, U.S.
Department of Commerce. 1987.

Chesson J; Rench JD; Schultz BD; Milne KL. Interpretation of Airborne Asbestos
Measurements. Risk Analysis.  10(3):437-47. 1990.

Choe N; Tanaka S; Xia; W; Hemenway DR; Roggli VI; Kagan E. Pleural Macrophage
Recruitment and Activation in Asbestos-Induced Pleural Injury. Environmental Health
Perspectives.  105(Suppl 5): 1257-1260. September. 1997.

Choe N; Tanaka S; Kagan E. Asbestos Fibers and Interleukin-1 Upregulate the Formation of
Reactive Nitrogen Species in Rat Pleural Mesothelial Cells.  American Journal of Respiratory
Cell and Molecular Biology.  19(2):226-236.  1998.

Choe N; Zhang J; Iwagaki A; Tanaka S; Hemenway DR; Kagan E. Asbestos Exposure
Upregulates the Adhesion of Pleural Leukocites to Pleural Mesothelial Cells via VCAM-1.
American Journal of Physiology.  277(2:Part 1):L292-L300. 1999.

Churg A. Deposition and clearance of chrysotile asbestos. Annals of Occupational Hygiene.
38(4):625-633.  1994.

Churg A; Wright JL; Stevens B. Exogenous mineral particles in the human bronchial mucosa
and lung parenchyma. I. Nonsmokers in the general population. Experimental Lung Research.
16:159-175. 1990.
                                         9.5

-------
Churg A; Wiggs B; Depaoli L; Kampe B; Stevens B. Lung Asbestos Content in Chrysotile
Workers with Mesothelioma.  American Review of Respiratory Disease.  130:1042-1045.  1984.

Coin PG; Roggli VL; Brody AR. Deposition, Clearance, and Translocation of Chrysotile
Asbestos from Peripheral and Central Regions of the Rat Lung. Environmental Research.
58:970-116.  1992.                                                         /

Coin PG; Roggli VL; Brody AR. Persistence of Long, Thin Chrysotile Asbestos Fibers in the
Lungs of Rats. Environmental Health Perspectives. 102(Suppl 5):197-199. 1994.  j
                                                                          i
Costa DL; Dreher KL.  Bioavailable Transition Metals in Particulate Matter Mediate
Cardiopulmonary Injury in Healthy and Compromised Animal Models. Environmental Health
Perspectives. 105(Suppl 5):1053-1060. September. 1997.

Cox D; Lindley D.  Theoretical Statistics. Chapman and Hall, London.  1974.       i

Cox D; Oakes DV.  Analysis of Survival Data.  Cox DR; Hinkley DV (eds.). Chapman and Hall,
London.  1984.

Crump KS. The Effect of Random Error in Exposure Measurement upon the Shape of the  Dose
Response. Nonlinearity in Biology, Toxicology and Medicine. 2003. (pending publications)

Crump KS.  Asbestos Potency Assessment for EPA Hearing. Prepared for Asbestos Information
Association/North America.  116pp. 1986.                 •                   '
                                                                         i
Davis JMG; Cowie  HA. The Relationship Between Fibrosis and Cancer in Experimental
Animals Exposed to Asbestos and Other Fibers. Environmental Health Perspectives.
88:305-309.  1990.                                      '                   ,

Davis JMG; Jones AD. Comparisons of the Pathogenicity of Long and Short Fibres'of
Chrysotile Asbestos in Rats. British Journal  of Exposure Pathology.  69(5): 171-73 8'. 1988b.

Davis JMG; Beckett ST; Bolton RE; Col lings P; Middleton AP. Mass and Number 6f Fibres in
the Pathogenesis of Asbestos-Related Lung Disease in Rats. British Journal of Cancer.
37:673-688.  1978.                                      ;                   ['

Davis JMG; Beckett ST; Bolton RE; Donaldson K. A Comparison of the Pathological Effects in
Rats of the UICC Reference Samples of Amosite and Chrysotile with Those of Amosite and
Chrysotile Collected from the  Factory Environment. In Biological Effects of Mineral Fibres.
Wagner JC (ed.). IARC Scientific Publications, pp. 288-292.' 1980.

Davis JMG; Addison J; Bolton RE; Donaldson K; Jones AD; Miller BG.  Inhalation Studies on
the Effects of Tremolite and Brucite Dust in Rats. Carcinogenesis. 6(5):667-674.  1985.

Davis JMG; Addison J; Bolton R; Donaldson K; Jones AD; Smith T. The Pathogenicity of Long
Versus Short Fibre Samples of Amosite Asbestos Administered to Rats by Inhalation and
Intraperitoneal Injection. British Journal of Experimental Pathology. 67:415-430, 1986a.

                                         9.6                             r

-------
Davis JMG; Gylseth G; Morgan A. Assessment of Mineral Fibres from Human Lung Tissue.
Thorax, 41:167-175.  1986b.

Davis JMG; Bolton RE; Brown D; Tully HE. Experimental Lesions in Rats Corresponding to
Advanced Human Asbestosis.  Exposure Molecular Pathology. 44(2):207-221. 1986c.

Davis JMG; Addison J; Bolton RE; Donaldson K; Jones AD.  Inhalation and Injection Studies in
Rats Using Dust Samples from Chrysotile Asbestos Prepared by a Wet Dispersion Process.
British Journal of Pathology.  67:113-129.  1986d.

Davis JMG; Jones AD; Smith  T. Comparisons of the Pathogenicity of Long and Short Fibres of
Chrysotile Asbestos in Rats. Institute for Research and Development of Asbestos, Montreal
(ed.).  Institute of Occupational Medicine. Report No. TM-87/08. 1987.

Davis JMG; Bolton RE; Douglas AN; Jones AD; Smith T. Effects of Electrostatic Charge on the
Pathogenicity of Chrysotile Asbestos. British Journal of Industrial Medicine. 45(5):292~309.
1988a.

Davis JMG; Addison J; Mclntosh C;  Miller BG; Niven K. Variations in the Carcinogenicity of
Tremolite Dust Samples of Differing  Morphology. Annals New York Academy oj'Sciences.
473^90. 1991.

de Klerk N. Unpublished Raw Data Provided to Dr. Wayne Berman by Dr. Nick de Klerk from
Study of Crocidolite Miners in Wittenoom, Australia, originally described in Armstrong et al.
(1988) but with followup extended through 1999.  2001.

de Klerk NH; Musk AW; Armstrong  BK; Hobbs MST. Diseases in Miners and Millers of
Crocidolite from Wittenoom, Western Australia: A Further Followup to December 1986. Annals
of Occupational Hygiene.  38(Suppl l):647-655.  1994.

Dement JM.  Estimation of Dose and Evaluation of Dose-Response in a Retrospective Cohort
Mortality Study of Chrysotile  Asbestos Textile Workers. Ph.D. Thesis. The University of North
Carolina at Chapel Hill. 1980.

Dement JM; Brown DP. Lung Cancer Mortality Among Asbestos Textile Workers: A Review
and Update. Annals of Occupational Hygiene.  3 8(4): 525-532.  1994.

Dement JM; Brown DP. Cohort Mortality and Case-Control Studies of White Male Chrysotile
Asbestos Textile Workers. Journal of Clean Technology, Environmental Toxicology, and
Occupational Medicine. 7:1052-1062.  1998.

Dement JM; Harris RL. Estimates of Pulmonary and Gastrointestinal Deposition for
Occupational  Fiber Exposure.  NTIS  PB80-149644.  U.S. HEW Contract #78-2438.  1979.

Dement JM; Harris RL; Symons MJ; Shy CM.  Estimates of Dose-Response for Respiratory
Cancer Among Chrysotile Asbestos Textile  Workers.  Annals Occupational Hygiene.
26(l-4):869-887.  1982.

                                         9.7

-------
Dement JM; Harris RL; Symons MJ; Shy CM.  Exposures and Mortality Among Chrysotile
Workers. Part I: Exposure Estimates. American Journal of Industrial Medicine. 4:399-419.
1983a.

Dement JM; Harris RL; Symons MJ; Shy CM.  Exposures and Mortality Among Chrysotile
Workers. Part II: Mortality. American Journal of Industrial Medicine.  4:421-433.  1983b.

Dement JM; Brown DP; Okun A. Follow-up Study of Chrysotile Asbestos Textile Workers:
Cohort Mortality and Case-Control Analysis. American Journal of Industrial Medicine.
26:431-447.  1994.

Doll R; Peto R.  Cigarette Smoking and Bronchial Carcinoma: Dose and Time Relationships
Among Regular Smokers and Lifelong Non-Smokers. Journal of Epidemiology and Community
Health. 32:303-313. 1978.                             '                  |.
                                                                         '.

Doll R; Peto J. Asbestos: Effects on Health of Exposure to Asbestos. Health and Safety
Commission, London, United Kingdom.  1985.

Dopp E; Schiffmann D. Analysis of Chromosomal Alterations Induced by Asbestos and
Ceramic Fibers. Toxicology Letters.  96-97:155-162. 1998.

Driscoll KE; Carter JM; Hassenbein DG; Howard B.  Cytokines and Particle-Induced
Inflammatory Cell Recruitment. Environmental Health Perspectives. 105(Suppl 5):1159-1164.
September. 1997.

Bastes W; Hadley JG. A Mathematical Model of Fiber Carcinogenicity and Fibrous in
Inhalation and Intraperitoneal Experiments in Rats. Inhalation Toxicology. 8:323-343. 1994.

Eastes W; Hadley JG. Dissolution of Fibers Inhaled by Rats.  'Inhalation Toxicology.
7:179-196. 1995.                                                         ''

Eastes W; Hadley JG. A Mathematical Model of Fiber Carcinogenicity and Fibrous ;in
Inhalation and Intraperitoneal Experiments in Rats. Inhalation Toxicology. 8:323-343. 1996.

Economou P; Samet JM; Lechner JF.  Familial and Genetic Factors in the Pathogenesis of Lung
Cancer. In: Epidemiology of Lung Cancer.  Chapter 14. Samet JM (ed.).  Marcel Dekker, Inc.,
New York.  1994.

Enterline PE; Harley J; Henderson V.  Asbestos and Cancer --(A Cohort Followed to Death.
Graduate School of Public Health, University of Pittsburgh. 1986.

Everitt JI; Gelzleichter TR; Bermudez E; Mangum JB; Wong BA; Janszen DB; Moss OR.
Comparison of Pleural Responses of Rats and Hamsters to Subchronic Inhalation of'Refractory
Ceramic Fibers.  Environmental Health Perspectives. 105(Suppl 5):1209-1213. September.
1997.
                                         9.8

-------
Finkelstein MM.  Mortality Among Long-Term Employees of an Ontario Asbestos-Cement
Factory. British Journal of Industrial Medicine. 40:138-144. 1983.

Finkelstein MM.  Mortality Among Employees of an Ontario Asbestos-Cement Factory.
American Review of Respiratory Disease.  129:754-761. 1984.

Finkelstein MM; Dufresne A. Inferences on the Kinetics of Asbestos Deposition and Clearance
Among Chrysotile Miners and Millers. American Journal of Industrial Medicine.
35(4):401-412. April 1999.

Finkelstein JN; Johnston C; Barrett T; Oberdorster G. Particulate-Cell Interactions and
Pulmonary Cytokine Expression. Environmental Health Perspectives. 105(Suppl
5): 1179-1182. September.  1997.

Floyd RA. Role of Oxygen Free Radicals in Carcinogenesis and Brain Ischemia. The FASEB
Journal. 4:2587-2597.  June. 1990.

Fubini B.  Surface Reactivity in the Pathogenic Response to Particulates. Environmental Health
Perspectives. 105(Suppi 5):1013-1020. September. 1997.

Gehr P; Geiser M; Stone KC; Crapo JD. Morphometric Analysis of the Gas Exchange Region of
the Lung.  In Toxicology of the Lung, 2nd edition.  Gardner DE; Crapo JD; McClellan RO (eds.).
Raven Press, New York. 1993.

Ghio AJ; Kadiiska MB; Xiang Q-H; Mason RP. In Vivo Evidence of Free Radical Formation
After Asbestos Instillation: An ESR Spin Trapping Investigation. Free Radical Biology and
Medicine.  24(1): 11-17. 1998.

Gibbs GW; Hwang CY. Physical Parameters of Airborne Asbestos  Fibres in Various Work
Environments - Preliminary Findings. American Industrial Hygiene Association Journal.
36(6):459-466.  1975.

Gibbs GW; Hwang CY. Dimensions of Airborne Asbestos Fibers.  In Biological Effects of
Mineral Fibers. Wagner JC  (ed.). IARC Scientific Publication,  pp. 69-78.  1980.

Gilbert O.  Statistical Method for Environmental Pollution Monitoring. Van Nostrand Reinhold,
New York. 1987.

Gold J; Amandusson H; Krozer A; Kasemo B; Ericsson T; Zanetti G; Fubini B. Chemical
Characterization and Reactivity of Iron Chelator-Treated Amphibole Asbestos. Environmental
Health Perspectives. 105(Supple 5):1021-1030.  1997.

Goldstein B; Rendall R; Webster I. A Comparison of the Effects of Exposure of Baboons to
Crocidolite and Fibrous-Glass Dusts. Environmental Research.  32:334-359. 1983.
                                         9.9

-------
Golladay SA; Park S-H; Aust AE.  Efflux of Reduced Glutathione after Exposure of Human
Lung epithelial Cells to Crocidolite Asbestos.  Environmental Health Perspectives. )05(Suppl
5): 1273-1278. 1997.
                                                                         i
Goodglick LA; Kane AB. Cytotoxicity of Long and Short Crojcidolite Asbestos Fibers In Vitro
and Hn Vivo. Cancer Research. 50:5153-5163.  1990.      '                   i

Governa M; Camilucci L; Amati M; Visona I; Valentino M; Botta GC; Campopiano A; Fanizza
C. Wollastonite Fibers In Vitro Generate Reactive Oxygen Species Able to Lyse Erythrocytes
and Activate the Complement Alternate Pathway. Toxicological Sciences. 44(l):32-38.  1998.

Gross TJ; Cobb SM; Peterson MW. Asbestos Exposure Increases Paracellular Transport of
Fibrin Degradation Products Across Human Airway Epithelium. American Journal of
Physiology. 266(3):L287-295. March.  1994.                                 '•
                                                                         t
                                                                         i
Hammond EC; Selikoff IJ; Seidman H. Asbestos Exposure, Cigarette Smoking and Death Rates.
Annals New York Academy of Sciences.  330:473^90. 1979.                    '

Harris RL; Timbrell V. The Influence of Fibre Shape in Lung Deposition - Mathematical
Estimates. Inhaled Particles IV. Walton WH (ed.).  Pergamon Press.  1977.       !
                                                                         i
Hart GA; Kathman LM; Hesterberg TW. In Vitro Cytotoxicity of Asbestos and Man-Made
Vitreous Fibers: Roles of Fiber Length, Diameter and Composition.  Carcinogenesis.
15(5):971-977.  May. 1994.                             !

Health Effects Institute - Asbestos Research (HEI-AR). Asbestos  in Public and Commercial
Buildings: A Literature Review and Synthesis of Current Knowledge. HEI-AR, 141 ;Portland
St., Suite 7100, Cambridge, MA.  1991.                    '                   '

Health Effects Institute (HEI). Asbestos in Public and Commercial Buildings: A Literature
Review and Synthesis of Current Knowledge.  1991.

Hei TK; Wu LJ; Piao CQ. Malignant Transformation of Immortalized Human Bronchial
Epithelial Cells by Asbestos Fibers. Environmental Health Perspectives. 105(Suppl
5): 1085-1088. September.  1997.

Heidenreich WF; Luebeck EG; Moolgavkar SH. .Some Properties of the Hazard Function of the
Two-Mutation Clonal Expansion Model. Risk Analysis.  17:391-399.  1997.

Henderson VL; Enterline PE. Asbestos Exposure: Factors Associated with Excess Cancer and
Respiratory Disease Mortality. Annals New York Academy of Sciences. 330:117-126. 1979.

Hesterberg TW; Miller WC; McConnell EE; Chevalier J; Hadley JG; Bernstein DM;;Thevenaz
P; Anderson R. Chronic Inhalation Toxicity of Size-Separated Glass Fibers in Fischer 344 Rats.
Fundamental and Applied Toxicology. 20:464—476. 1993.
                                        9.10

-------
Hesterberg TW; Miiller WC; Thevenaz P; Anderson R.  Chronic Inhalation Studies of Man-
Made Vitreous Fibres: Characterization of Fibres in the Exposure Aerosol and Lungs. Annals of
Occupational Hygiene. 39(5):637-653.  1995.

Hesterberg TW; Miiller WC; Musselkman RP; Kamstrup O; Hamilton RD; Thevenaz P.
Biopersistence of Man-Made Vitreous Fibers and Crocidolite Asbestos in Rat Lung Following
Inhalation. Fundamentals and Applied Toxicology.  29:267-279.  1996.

Hesterberg TW; Axten C; McConnell EE; Oberdorster G; Everitt J; Miller WC; Chevalier J;
Chase GR; Thevenaz P.  Chronic Inhalation Study of Fiber Glass and Amosite Asbestos in
Hamsters: Twelve-Month Preliminary Results. Environmental Health Perspectives.  105(Suppl
5): 1223-1230. September.  1997.

Hesterberg TW; Chase G; Axten C; Miller WC; Musselman RP; Kamstrup O; Hadley J;
Morscheidt C; Bernstein DM;  Thevenaz P. Biopersistence of Synthetic Vitreous Fibers and
Amosite Asbestos in the Rat Lung Following Inhalation. Toxicology and Applied
Pharmacology. 151:262-275.  1998a.                               '

Hesterberg TW; Hart GA; Chevalier J; Miller WC; Hamilton RD; Bauer J;.Thevenaz P. The
Importance of Fiber Biopersistence and Lung Dose in Determining the Chronic Inhalation
Effects of X607, RCF1, and Chrysolite Asbestos in Rats. Toxicology and Applied
Pharmacology. 153(l):68-82. 1998b.

Hodgson AA. Fibrous Silicates. Lecture Series No. 4.' The Royal Institute of Chemistry,
London, United Kingdom. 1965.

Hodgson J; Darnton A. The Quantitative Risk of Mesothelioma and Lung Cancer in Relation to
Asbestos Exposure. Annals of Occupational Hygiene. 44(8):565-601. 2000.

Holian A; Uthman MO; Goltsova T; Brown SD; Hamilton RF.  Asbestos and Silica-Induced
Changes in Human Alveolar Macrophage Phenotype. Environmental Health Perspectives.
105(Suppl5):1139-1142. September. 1997.

Hughes JM; Weill H.  Asbestos Exposure: Quantitative Assessment of Risk. American Review
of Respiratory Disease.  133:5-13. 1986.

Hughes JM; Weill H; Hammad YY. Mortality of Workers Employed at Two Asbestos Cement
Plants. British Journal of Industrial Medicine.  44:161-174. 1987.

Hume LA; Rimstidt. The Biodurability of Chrysotile Asbestos. American Mineralogist.
77:1125-1128. 1992.

Hwang CY; Gibbs GW.  The Dimensions of Airborne Asbestos Fibres —I. Crocidolite from
Kuruman Area, Cape Province,.South Africa. Annals Occupational Hygiene. 24(1):23-41.
1981.
                                         9.11

-------
 Ilgren E; Chatfield E. Coalinga Fibre: A Short, Amphibole-Free Chrysotile.  Part 3: Lack of
 Biopersistence. Indoor Built Environment. 7:98-100.  1998.                     '

 International Agency for Research on Cancer (IARC).  Monographs on the Evaluation of '
 Carcinogenic Risks to Man. Volume 14. IARC Scientific Publications. Lyon, France. 1977.

 Integrated Risk Information System (IRIS). Toxicological Review of Asbestos.  U.S.;
 Environmental Protection Agency. Office of Research and Development, National Center for
 Environmental Assessment. Washington, D.C. http://www.epa.gov/iris/subsl/Q37Lhtm.  1998.
                                                                          !:
 Ishizaki T; Yano E; Evans PH. Cellular Mechanisms of Reactive Oxygen Metabolite Generation
 from Human Polymorphonuclear Leukocytes Induced by Crocidolite Asbestos. Environmental
 Research. 75(2): 135-140.  1997.                           '                  |;

 International Organization for Standardization (ISO).  Ambient Air-Determination of Asbestos
 Fibres - Direct-Transfer Transmission Electron Microscopy Method. ISO 10312. 1995.

 Jagirdar J; Lee TC; Reibman J; Gold LI; Aston C; Begin R; Rom WN.  Immunohistochemical
 Localization of Transforming Growth Factor Beta Isoforms in Asbestos-Related Diseases.
 Environmental Health Perspectives.  105(Suppl 5):1197-1203. September. 1997.  j

 Jaurand MC. Observations on the Carcinogenicity of Asbestos Fibers.  Annals New York
 Academy of Science. 643:258-70. 1991.                    '                  •

 Jaurand MC. Mechanisms of Fiber-Induced Genotoxicity. Environmental Health Perspectives.
 105(Suppl 5): 1073-1084. September.  1997.

 Jesch NK; Dorger M; Enders G; Rieder G; Vogelmeier C; Messmer K; Krombach F. Expression
 of the Inducible Nitric Oxide Synthase and Formation  of Nitric Oxide by Alveolar Macrophages.
 Environmental Health Perspectives.  105(Suppl 5): 1297-13 00. September. 1997.

 Johnson NF. Asbestos-Induced Changes in Rat Lung Parenchyma. Journal of Toxicology and
 Environmental Health. 21:193-203. 1987.                                     •

 Johnson NF; Jaramillo RJ.  P53, Cip 1, and Gadd 153 Expression Following Treatment of A549
 Cells with Natural and Man-Made Vitreous Fibers. Environmental Health Perspectives.
 105(Suppl 5): 1143-1145. September.  1997.                 .                   '.

 Jones AD; McMillan CH; Johnston AM; Mclntosh C; Cowie H; Bolton RE; Borzuki G; Vincent
 JH. Pulmonary Clearance of UICC Amosite Fibres Inhaled by Rats During Chronic Exposure at
 Low Concentrations. British Journal of Industrial Medicine. 45:300-304. 1988.

 Kaiglova A; Hurbankova M; Kovacikova Z. Impact of Acute and Subchronic Asbestos
 Exposure on Some Parameters of Antioxidant Defense System and Lung Tissue Injuryl
Industrial Health. 37(3):348-351. July. 1999.
                                        9.12

-------
Kamp DW; Graceffa P; Pryor WA; Weitzman SA. The Role of Free Radicals in Asbestos-
Induced Diseases.  Free Radical Biology & Medicine. 12:293-315.1992.

Kamp DW; Dunne M; Dykewicz MS; Sbalchiero JS; Weitzman SA; Dunn MM. Asbestos-
Induced Injury to Cultured Human Pulmonary Epithelial-Like Cells: Role of Neutrophil
Elastase. Journal of Leukocyte Biology, 54:73-80. July. 1993.

Kamp DW; Greenberger MJ; Sbalchierro JS; Preusen SE; Weitzman SA. Cigarette Smoke
Augments Asbestos-Induced Alveolar Epithelial Cell Injury: Role of Free Radicals.  Free
Radical Biology & Medicine. 25(6):728-739. 1998.

Kane AB; MacDonald JL. Mechanisms of Mesothelial Cell Injury, Proliferation, and Neoplasia
Induced by Asbestos Fibers.  Chapter 14. Fiber Toxicology, pp. 323-347.  1993.

Kauffer E; Vigneron JC; Hesbot A; Lemonnier M. A Study of the Length and Diameter of
Fibres, in Lung and in Broncho-Alveolar Lavage Fluid, Following Exposure of Rats to
Chrysotile Asbestos. Annals of Occupational Hygiene. 31(2):233-240.  1987.

Keane MJ; Miller WE; Ong T;  Stephens JW; Wallace WE; Zhong B-Z. A Study of the Effect of
Chrysotile Fiber Surface Composition on Genotoxicity In Vitro.  Journal of Toxicology and
Environmental Health Part A.  57(8):529-541. August. 1999.

Kimizuka G; Wang N; Hayashi Y. Physical and Microchemical Alterations of Chrysotile and
Amosite Asbestos in the Hamster Lung. Journal of Toxicology and Environmental Health.
21:251-264. 1987.

Kodama Y; Boreiko CJ; Maness SC; Hesterberg TW. Cytotoxic and Cytogenetic Effects of
Asbestos on Human Bronchial  Epithelial Cells in Culture.  Carcinogenesis.  14(4):691-697.
1993.

Kostyuk VA; Potapovich AI. Antiradical and Chelating Effects in Flavonoid Protection Against
Silica-Induced Cell Injury. Archives of Biochemistry and Biophysics.  355(l):43-48. 1998.

Kravchenko IV; Furalyov VA;  Vasylieva LA; Pylev LN. Spontaneous and Asbestos-Induced
Transformation of Mesothelial  Cells In Vitro. Teratogenesis Carcinogenesis and Mutagenesis.
18(3):141-151.  1998.

Krombach F; Miinzing S; Allmeling AM; Gerlach JT; Behr J; DOrger M. Cell Size of Alveolar
Macrophages: An Interspecies  Comparison. Environmental Health Perspectives.  105(Suppl
5):1261-1263. September. 1997.

Lacquet LM; VanderLinden L; Lepoutre J. Roentgenographic Lung Changes, Asbestosis and
Mortality in a Belgian Asbestos-Cement Factory. In Biological Effects of Mineral Fibres,
Wagner JC (ed.). lARCSciPubl. pp. 783-793. 1980.

Law BD; Bunn WB; Hesterberg TW. Solubility of Polymeric Organic Fibers and Manmade
Vitreous Fibers in Gambles Solution. Inhalation Toxicology. 2:321-339.  1990.

                                         9.13

-------
Law BD; Bunn WB; Hesterberg TW.  Dissolution of Natural Mineral and Man-Made'Vitreous
Fibers in Karnovsky's and Formalin Fixatives. Inhalation Toxicology.  3:309-321.  1991.

Leanderson P; Soderkvist P; Tagesson C; Axelson O.  Formation of 8-hydroxydeoxyguanosine
by Asbestos and Man-Made Mineral Fibres.  British Journal of Industrial Medicine. •'•
45:309-311.  1988.

Le Bouffant L. Physics and Chemistry of Asbestos Dust. In Biological Effects of Mineral
Fibres.  Wagner JC (ed.). IARC Scientific Publications, pp. 15-34.  1980.

Le Bouffant L; Daniel H; Heninn JP; Martin JC; Normand C; Tichoux G; Trolard F. '',,
Experimental Study on Long-Term Effects of Inhaled MMF on the Lungs of Rats. Annals of
Occupational Hygiene. 31(4B):765-790. 1987.                                ':

Lechner JF; Tesfaigzi J; Gerwin BI. Oncogenes and Tumor-Suppressor Genes in Mesbthelioma
- A Synopsis. Environmental Health Perspectives. 105(Suppl 5): 1061 -1067. September.
1997.                                                                     I'

Lee KP; Barras CE; Griffith RS; Waritz RS; Lapiri CA. Comparative Pulmonary Responses to
Inhaled  Inorganic Fibers with Asbestos and Fiberglass. Environmental Research. 24:167—191.
1981.                                                                     i
                                                                           I
Leigh J; Wang H; Bonin A; Peters M; Ruan X. Silica-Induced Apoptosis in Alveolar and
Granulomatous Cells In Vivo. Environmental Health Perspectives. 105(Suppl 5):1241-1246.
September. 1997.

Leikoff G; Driscoll K. Cellular Approaches in Respiratory Toxicology. In Toxicology of the
Lung, 2nd edition. Gardner DE; Crapo JD; McClellan RO (eds.).  Raven Press, New York.  1993.

Levin JL; McLarty JW; Hurst GA; Smith AN; Frank AL. Tyler Asbestos Workers: Mortality
Experience in a Cohort Exposed to Amosite. Occupational and Environmental Medicine.
55:155-160.  1998.                   .                                       ''

Levresse V; Reiner A; Fleury-Feith J; Levy F; Moritz S; Vivo C; Pilatte Y; Jaurand M-C.
Analysis of Cell Cycle Disruptions in Cultures of Rat Pleural Mesothelial Cells Exposed to
Asbestos Fibers. American Journal of Respiratory Cell and Molecular Biology.  17(6):660-671.
1997.

Li XY; Gilmour PS; Donaldson K; MacNee W. In Vivo and In Vitro Proinflammatory Effects of
Particulate Air Pollution (PM10). Environmental Health Perspectives.  105(SuppI
5): 1279-1284. September.  1997.

Liddell FDK.  The Interaction of Asbestos and Smoking in Lung Cancer.  Annals of
Occupational Hygiene. 45:341-56. 200la.
                                        9.14

-------
Liddell FDK. Unpublished raw mesothelioma data provided to Dr. Wayne Herman by Dr. FDK
Liddell from multiple studies of the 1891-1920 Birth Cohort of Quebec Chrysotile Miners and
Millers most recently described in Liddell et al. 1997. 2001b.

Liddell FDK; Armstrong BG. The Combination of Effects on Lung Cancer of Cigarette
Smoking and Exposure in Quebec Chrysotile Miners and Millers.  Annals of Occupational
Hygiene. 46(1):5-13. 2002.

Liddell FDK; McDonald AD; McDonald JC. The 1891-1920 Birth Cohort of Quebec Chrysotile
Miners and Millers: Development From 1904 and Mortality to 1992. Annals of Occupational
Hygiene. 41:13-36. 1997.

Lira Y; Kim  S-H; Kim K-A; Oh M-W; Lee K-H. Involvement of Protein Kinase C,
Phospholipase C, and Protein Tyrosine Kinase Pathways in Oxygen Radical Generation by
Asbestos-Stimulated Alveolar Macrophage Environmental Health Perspectives.  105(Suppl
5): 1325-1328.  September.  1997.

Lippmann M. Deposition and Retention of Inhaled Fibres: Effects on Incidence of Lung Cancer
and Mesothelioma.  Occupational and Environmental Medicine. 51:793-798. 1994.

Lippmann M. Asbestos and Other Mineral and Vitreous Fibers. In Environmental Toxicants:
Human Exposures and Their Health Effects. Lippman M (ed). Wiley-Interscience; 2nd edition.
December. 1999.

Lippmann M; Schlesinger RB. Interspecies Comparisons of Particle Deposition and Mucociliary
Clearance in  Tracheobronchial Airways. Journal of Toxicology and Environmental Health.
3:441[261H69[289]. 1984.

Luster MI; Simeonova PP. Asbestos Induces Inflammatory Cytokines in the Lung Through
Redox Sensitive Transcription Factors.  Toxicology Letters (Shannon). 102-103:271-275. 1998.

Lynch JR; Ayer HE; Johnson DJ. The Interrelationships of Selected Asbestos Exposure Indices.
American Industrial Hygiene Association Journal.  31 (5):598-604. 1970.

Marconi A; Menichini E; Paoletti L. A Comparison of Light Microscopy and Transmission
Electron Microscopy Results in the Evaluation of the Occupational Exposure to Airborne
Chrysotile Fibres. Annals of Occupational Hygiene. 28(3):321-331. 1984.

Marsella JM; Liu BL; Vaslet CA; Kane AB. Susceptibility of p53-Deficient Mice to Induction
of Mesothelioma by Crocidolite Asbestos Fibers. Environmental Health Perspectives.
105(Suppl 5):1069-1072. September. 1997.

Martin LD; Krunkosky TM; Dye JA; Fischer BM; Jiang NF; Rochelle LG; Akley NJ; Dreher
KL; Adler KB. The Role of Reactive Oxygen and Nitrogen Species in the Response of Airway
Epithelium to Particulates. Environmental Health Perspectives. 105(Suppl 5): 11301-1308.
September.  1997.
                                         9.15

-------
Mattson SM.  Glass Fiber Dissolution in Simulated Lung Fluid and Measures Needed to
Improve Consistency and Correspondence in In-Vitro Studies. Presented at the IARG Conf.
Biopersistence of Respirable Synthetic Fibres and Minerals. Lyon, France, September 7-9.
1992, Environmental Health Perspectives. 1994.

McConnell E; Wagner J; Skidmore J; Moore J. A Comparative Study of the Fibrogenic and
Carcinogenic Effects of UICC Canadian Chrysotile B Asbestos and Glass Microfibre!(JM 100).
Biological Effects of Man-Made Fibres. Proceedings of a WHO/IARC Conference, pp.
234-252.  1982.                                        '                   '

McConnell EE; Rutter HA; Ulland BM; Moore JA. Chronic Effects of Dietary Exposure to
Amosite Asbestos and Tremolite in F344 Rats. Environmental Health Perspectives. 53:27-44.
1983.                                                                     :

McConnell EE; Adkins B. Studies on the Chronic Toxicity (Inhalation) of Wollastonite in
Fischer 344 Rats. Inhalation Toxicology.  3:323-337. 1991.                     lj

McConnell EE; Mast RW; Hesterberg TW; Chevalier J; Kotin P; Bernstein DM; Thevenez P;
Glass LR; Anderson R. Chronic Inhalation Toxicity of a Kaolin-Based Refractory Ceramic
Fiber in Syrian Golden Hamsters.  Inhalation Toxicity: 7:503-532.  1995.          !

McDonald AD; Fry JS; Wooley AJ; McDonald JC. Dust Exposure and Mortality  in ah
American Chrysotile Textile Plant British Journal of Industrial Medicine. 39:361-367.  1983a.

McDonald AD; Fry JS; Wbolley AJ; McDonald JC. Dust Exposure and Mortality in an
American Factory Using Chrysotile, Amosite, and  Crocidolite in Mainly Textile Manufacture.
British Journal of Industrial Medicine.  40:368-374. 1983b.                     !

McDonald AD; Fry JS; Woolley AJ; McDonald JC. Dust Exposure and Mortality in an
American Chrysotile Asbestos Friction Products Plant. British Journal of Industrial Medicine.
41:151-157.  1984.

McDonald JC. Mineral Fibre Persistence and Carcinogenicity. Industrial Health.
36(4):372-375. October. 1998a.

McDonald JC. Invited Editorial: Unfinished Business - The Asbestos Textiles Mystery. Annals
of Occupational Hygiene. 42(l):3-5. 1998b.

McDonald JC; Gibbs GW; Liddell FDK. Chrysotile Fibre Concentration and Lung Cancer
Mortality: A Preliminary Report.  In Biological Effects of Mineral Fibres. Wagner JC |(ed).
IARC Scientific Publications, pp.  811-817.  1980a.

McDonald JC; Liddell FDK; Gibbs GW; Eyssen GE; McDonald'AD. Dust Exposure and
Mortality in Chrysotile Mining, 1910-1975.  British Journal of Industrial Medicine. 37:11-24.
1980b.
                                        9.16

-------
McDonald JC; McDonald AD; Armstrong B; Sebastian P. Cohort Study of Mortality of
Vermiculite Miners Exposed to Tremolite. British Journal of Industrial Medicine. 43:436-444.
1986.

McDonald JC; Liddell FDK; Dufresne A; McDonald AD. The 1891-1920 Birth Cohort of
Quebec Chrysotile Miners and Millers: Mortality 1976-1988. British Journal of Industrial
Medicine.  50:1073-1081.  1993.

McGavran PD; Butterick CJ; Brody AR. Tritiated Thymidine Incorporation and the
Development of an Interstitial Lesion in the Bronchiolar-Alveolar Regions of the Lungs of
Normal and Complement Deficient Mice After Inhalation of Chrysotile Asbestos. Journal of
Environmental Pathology, Toxicology, and Oncology. 9(5-6):377-391. December. 1989.

McGavran PD; Moore LB; Brody AR. Inhalation of Chrysotile Asbestos Induces Rapid Cellular
Proliferation in Small Pulmonary Vessels of Mice and Rats. American Journal of Pathology.
136(3):695-705.  1990,

Mercer RR and Crapo JD. Three-Dimensional Analysis of Lung Structure and its Application to
Pulmonary Dosimetry Models. In Toxicology of the Lung, 2nd edition, Gardner DE; Crapo.JD;
McClellan RO (eds.). Raven Press, New York. 1993.

Middleton AP; Beckett ST; Davis JMG.  Further Observations on the Short-Term Retention and
Clearance of Asbestos by Rats, Using UICC Reference Samples.  Annals of Occupational
Hygiene. 22:141-152.  1979.

Miller FJ; Overton JH; Kimbell JS; Russell ML. Regional Respiratory Tract Absorption of
Inhaled Reactive Gases.  In Toxicology of the Lung, 2nd edition, Gardner DE; Crapo JD;
McClellan RO (eds.). Raven Press, New York. 1993.

Moolgavkar SH; Dewanji A; Venzon DJ. A Stochastic Two-Stage Model for Cancer Risk
Assessment. 1: The Hazard Function and the Probability of Tumor.  Risk Analysis.
8(3):383-392.  1988.

Moolgavkar SH; Luebeck EG; De Gunst M; Port RE; Schwarz M. Quantitative Analysis of
Enzyme-Altered Foci in Rat Hepatocarcinogenesis Experiments.  Single Agent Regimen.
Carcinogenesis. 11 (8): 1271-1278. 1990.

Moolgavkar SH; Luebeck EG; Krewski D; Zielinski JM. Radon, Cigarette Smoke, and Lung
Cancer: A Re-Analysis of the Colorado Plateau Uranium Miners' Data. Epidemiology.
4:204-217. 1993.

Morgan A; Holmes A. Concentrations and Dimensions of Coated and Uncoated Asbestos Fibres
in the Human Lung.  British Journal of Industrial Medicine.  37:25-32. 1980.

Morgan A; Talbot RJ; Holmes A. Significance of Fibre Length in the Clearance of Asbestos
Fibres from the Lung. British Journal of Industrial Medicine. 35:146-153. 1978.
                                        9.17

-------
Morgan A; Black A; Evans N; Holmes A; Pritchard JN. Deposition of Sized Glass Fibres in the
Respiratory Tract of the Rat. Annals of Occupational Hygiene-. 23:353-366.  1980.!'

Morrow PE. Issues: Possible Mechanisms to Explain Dust Overloading of the Lungs.
Fundamental and Applies Toxicology. 10:369-384. 1988.                       \

Mossman BT. Mechanisms of Asbestos Carcinogenesis and Toxicity: The Amphibole
Hypothesis Revisited. British Journal of Industrial Medicine. 50:673-676.  1993.
                                                       '                   |t
Mossman BT; Churg A. Mechanisms in the Pathogenesis of Asbestosis and Silicosis. American
Journal of Critical Care Medicine.  157:1666-1680.  1998.                       :•
                                                                          i
                                                                          i
Mossman BT; Marsh JP. Role of Active Oxygen Species in Asbestos-Induced Cytotpxicity, Cell
Proliferation, and Carcinogenesis.  In Cellular and Molecular Aspects of Fiber Carcinogenesis,
Cold Spring Harbor Laboratory Press. 0-87969-361-4/91.  pp. 159-168. 1991.      ;

Mossman BT; Bignon J; Corn M; Seaton A; Gee JBL. Asbestos: Scientific Developments and
Implications for Public Policy. Science. 247:294-301. 1990.                    '

Mossman BT; Kamp DW; Sigmund A; Weitzman SA. Mechanisms of Carcinogenesis and
Clinical Features of Asbestos-Associated Cancers. Cancer Investigation.  14(5):466-480. 1996.

Mossman BT; Faux S; Janssen Y; Jimenex LA; Timblin C; Zanella C; Goldberg J; Walsh E;
Barchowsky A; Driscoll K. Cell Signaling Pathways Elicited by Asbestos. Environmental
Health Perspectives. 105(Suppl 5):1121-1125. September.  1997.

Muhle H; Bellman B; Takenata S; Ziem Y. Inhalation and Injection in Rats to Test the
Carcinogenicity of MMMF. Annals Occupational Hygiene. 31(4B):755-764. 1987. ;

Nario RC; Hubbard AK. Localization of Intercellular Adhesion Molecule-1 (ICAM-A) in the
Lungs of Silica-Exposed Mice. Environmental Health Perspectives.  105(Suppl 5): 1183-1190.
September. 1997.

Nehls P; Seller F; Rehn B; Greferath R; Bruch J. Formation and Persistence of 8-Oxoguanine
Rat Lung Cells as an Important Determinant for Tumor Formation following Particle Exposure.
Environmental Health Perspectives. 105(Suppl 5):1231-1240. September.  1997.

Nicholson WJ. Part III. Recent Approaches to the Control of Carcinogenic Exposures. Case
Study 1: Asbestos - The TLV Approach. Annals New York Academy of Science.  271:152-169.
1976.

Nicholson WJ; Selikoff IJ; Seidman H; Lilis R; Formby P. Long-Term Mortality Experience of
Chrysotile Miners and Millers in Thetford Mines, Quebec. Annals New York Academy of
Sciences. 330:11-21. 1979.
                                        9.18

-------
Nikula KJ; Avila KJ; Griffith WC; Mauderly JL.  Sites of Particle Retention and Lung Tissue
Responses to Chronically Inhaled Diesel Exhaust and Coal Dust in Rats and Cynomolgus
Monkeys. Environmental Health Perspectives. 105(Suppl 5):1231-1240. September. 1997.

National Institute for Occupational Safety and Health (NIOSH). Method for Determination of
Asbestos in Air Using Positive Phase Contrast Microscopy. NIOSH Method 7400. NIOSH,
Cincinnati, Ohio, U.S.A.  1985.

National Institute for Occupational Safety and Health (NIOSH). Method for Determination of
Asbestos in Air Using Transmission Electron Microscopy. NIOSH Method 7402. NIOSH,
Cincinnati, Ohio, U.S.A.  1986.

Oberdorster G. Macrophage-Associated Responses to Chrysotile.  Annals of Occupational
Hygiene. 38(4):601-615.  1994.

Oberdorster G; Morrow PE; Spurny K. Size Dependent Lymphatic Short Term Clearance of
Amosite Fibres in the Lung.  Annals of Occupational Hygiene. 32(Suppl  1):149-156. 1988.

Ollikainen T; Linnainmaa K; Kinnula VL. DNA  Single Strand Breaks Induced by Asbestos
Fibers in Human Pleural Mesothelial Cells In Vitro.  Environmental and Molecular Mutagenesis.
33(2): 153-160. 1999.

Ontario Royal Commission.  Report of the Royal Commission on Matters of Health and Safety
Arising form the Use of Asbestos in Ontario. Volume 3.  1984.

Osier M; Baggs RB; Oberd5rster G. 1997.  Intratracheal Instillation vs. Intratracheal Inhalation:
Influence of Cytokines on Inflammatory Response. Proceedings of the Sixth International
Meeting on the Toxicology of Natural and Man-Made Fibrous and Non-Fibrous Particles.
Environmental Health Perspective. 105(Suppl 5):1265-1271.

Osornio-Vargas AR; Kalter VG; Badgett A; Hernandez-Rodriguez N; Aguilar-Delfm I; Brody
AR. Rapid Communication: Early-Passage Rat Lung Fibroblasts do not Migrate In Vitro to
Transforming Growth Factor-Beta.  American Journal of Respiratory Cell Molecular Biology.
8(5):468-471.  May. 1993.

Palekar LD; Spooner CM; Coffin DL. Influence of Crystallization Habit of Minerals on In Vitro
Cytotoxicity. Annals New York Academy of Sciences, pp. 673-687. 1979.

Park S-H; Aust AE. Regulation of Nitric Oxide Synthase Induction by Iron and Glutathione  in
Asbestos-Treated Human Lung Epithelial Cells. Archives of Biochemistry and Biophysics.
360(l):47-52.  1998.

Peto J. Lung Cancer Mortality in Relation to Measured Dust Levels in an Asbestos Textile
Factory.  In Biological Effects of Mineral Fibres.  Wagner JC  (ed.). 1ARC Scientific
Publications, pp. 829-836.  1980a.
                                         9.19

-------
 Peto J. The Incidence of Pleural Mesothelioma in Chrysotile Asbestos Textile Workers.  In
 Biological Effects of Mineral Fibres. Wagner JC (ed.). IARC Scientific Publications1; pp.
 703-711. 1980b.

 Peto J; Doll R; Howard SV; Kinlen LJ; Lewinsohn, HC. A Mortality Study Among Workers in
 an English Asbestos Factory. British Journal of Industrial Medicine. 34:169-173. 1977.

 Peto J; Seidman H; Selikoff IJ.  Mesothelioma Mortality in Asbestos Workers: Implications for
 Models of Carcinogenesis and Risk Assessment. British Journal of Cancer. 45:124-135. 1982.

 Peto J; Doll R; Hermon C; Binns W; Clayton R; Goffe T. Relationship of Mortality to Measures
 of Environmental Asbestos Pollution in an Asbestos Textile Factory. Annals of Occupational
 Hygiene.  29(3):305-355. 1985.

 Pinkerton KE; Brody AR; McLauren DA; Adkins B Jr; O'Connor RW; Pratt PC; Crapo JD.
 Characterization of the Three Types of Chrysotile Asbestos After Aerosolization.
 Environmental Research.  31:32-35.  1983.

 Pinkerton KE; Pratt PC; Brody AR; Crapo JD. Fiber Localization and its Relationship to Lung
 Reaction in Rats After Chronic Inhalation of Chrysotile Asbestos. American Journal of
 Pathology. 117:484-498.  1984.                                               '

 Pinkerton KE; Plopper CG; Mercer RR; Rogli VL; Patra AL; Brody AR; Crap JD.  Airway
 Branching Patterns Influence Asbestos Fiber Location and the Extent of Tissue Injury; in the
 Pulmonary Parenchyma. Laboratory Investigation.  55(6):688-695.  1986.

 Piolatto G; Negri E; LaVecchia C; Pira E; Decarli A; Peto J. An Update of Cancer Mortality
 Among Chrysotile Asbestos Miners in Balangero, Northern Italy.  British Journal of Industrial
 Medicine. 47:810-814. 1990.

 Platek SF; Groth DH; Ulrich CE; Stettler LE; Finnell MS; Stoll M. Chronic Inhalation of Short
 Asbestos Fibers.  Fundamental and Applied Toxicology.  5:327-340. 1985.

 Pooley FD. An Examination of the Fibrous Mineral Content of Asbestos Lung Tissue from the
 Canadian Chrysotile Mining Industry.  Environmental Research. 12:281-298. 1976.
                                                                            l
 Pooley FD. Tissue Burden Studies.  In Short and Thin Mineral Fibres: Identification, Exposure,
 and Health Effects. Chatfield EJ (ed).  pp. 96-129.  1982.

 Pott F. Some Aspects on the Dosimetry of the Carcinogenic Potency of Asbestos and Other
 Fibrous Dusts. Staub-Reinhalt.  38(12):486-490.  1978.                           •   '

 Pott F. Animal Experiments with Mineral Fibers. In Short and Thin Mineral Fibers:  '
Identification, Exposure, and Health Effects. Chatfield EJ (ed.). pp. 133-161. 1982.

 Pott F; Huth F; Friedrichs KH. Tumorigenic Effect of Fibrous  Dust in Experimental Animals.
Environmental Health Perspectives.  9:313-315. 1974.

                                         9.20

-------
Pott F; Huth F; Friedrichs KH.  Results of Animal Carcinogenesis Studies After Application of
Fibrous Glass and Their Implications Regarding Human Exposure. Occupational Exposure to
Fibrous Glass. U.S. HEW Publication No. 76-151. pp. 183-191. 1976.

Pott F; Ziem U; Reiffer RJ; Huth F; Ernst H; Mohr U.  Carcinogenicity Studies on Fibres, Metal
Compounds, and Some Other Dusts in Rats. Experimental Pathology.  32:129-152.  1987.

Raabe O. Deposition and Clearance of Inhaled Particles. In Occupational Lung Diseases. Gee
JB; et al. (eds.). Raven Press, pp.1047. 1984.

Quinlan TR; Berube KA; Hacker MP; Taatjes DJ; Timblin CR; Goldberg J; Kimberley P;
O'Shaughnessy P; Hemenway D; Torino J; Jimenez LA; Mossman BT. Mechanisms of
Asbestos-Induced Nitric Oxide Production by Rat Alveolar Macrophages in Inhalation and
In Vitro Models.  Free Radical Biology and Medicine. 24(5):778-788. 1998.

Rahman Q; Mahmood N; Khan SG; Arif JM; Athar M. Mechanism of Asbestos-Mediated DNA
Damage: Role of Heme and Heme Proteins. Environmental Health Perspectives. 105(Suppl
5):1109-1112. September.  1997.

Roberts DR; Zumwalde RD. Industrial Hygiene Summary Report of Asbestos Exposure
Assessment for Brake Mechanics.  NIOSH Reports No. IWS-32-4A, Industrial Hygiene Section,
NTIS# PB 87-105433.  1982.

Robledo R; Mossman B. Cellular and Molecular Mechanisms of Asbestos-Induced Fibrosis.
Journal of Cellular Physiology. 180:158-166.  1999.

.Roggli VL; Brody AR. Changes in Numbers and Dimensions of Chrysotile Asbestos Fibers in
Lungs of Rats Following Short-Term Exposure. Experimental Lung Research. 7:133-147.
1984.

RoggH VL; George MH; Brody AR.  Clearance and Dimensional Changes of Crocidolite
Asbestos Fibers Isolated from Lungs of Rats Following Short-Term Exposure. Environmental
Research. 42:94-105.  1987.

Rood AP; Scott RM. Size Distributions of Chrysotile Asbestos in a Friction Products Factory As
Determined by Transmission Electron Microscopy. Annals of Occupational Hygiene.
33(4):5 83-590. 1989.

Rubino GF; Piolatto GW; Newhouse ML; Scansetti G; Aresini GA; Murray R. Mortality of
Chrysotile Asbestos Workers at the Balangero Mine, Northern Italy. British Journal of
Industrial Medicine. 36:187-194. 1979.

St. George JA; Harkema JR; Hyde DM; Plopper CG.  Cell Populations and Structure/Function
Relationship of Cells in the Airways.  In Toxicology of the Lung, 2nd edition. Gardner DE; Crapo
JD; McClellan RO (eds.). Raven Press, New York. 1993.
                                         9.21

-------
Sanden A; Jarvholm B; Larsson S; Thiringer G. The Risk of Lung Cancer and Mesothelioma
After Cessation of Asbestos Exposure: A Prospective Cohort Study of Shipyard Workers.
European Respiratory Journal. 5:281-285. 1992.                              '
                                                                         i
Schnoor T. Unpublished Raw Data Provided to Dr. Wayne Berman by Ms. Terri Schhoor of
NIOSH from Study of South Carolina Textile Workers (Dement et al. 1994). 2001. \

Sebastien P; Plourde M; Robb R; Ross M.  Ambient Air Asbestos Survey in Quebec Mining
Towns - Part 1, Methodological Study.  Environmental Protection Service, Quebec Region,
3/AP/RQ/1E.  pp. 1-41. 1984.                                              j

Sebastien P; Plourde M; Robb R; Ross M; Nadon B; Wypruk T. Ambient Air Asbestos Survey
in Quebec Mining Towns.  Part II: Main Study. Environmental Protection Service, Environment
Canada.  EPS 5/AP/RQ/2E. July. 1986.

Sebastien P; McDonald JC; McDonald AD; Case B; Harley R.  Respiratory Cancer in;Chrysotile
Textile and Mining  Industries: Exposure Inferences from Lung Analysis.  British Journal of
Industrial Medicine. 46:180-187.  1989.

Seidman  H. Short-Term Asbestos Work Exposure and Long-Term Observation — July 1984
Update. Department of Epidemiology, American Cancer Society.  1984.           •

Seidman  H; Selikoff IJ; Hammond EC.  Short-Term Asbestos Work Exposure and Lohg-Term
Observation. Annals New YorkAcademy of Sciences. 330:61-89.  1979.           .
                                                                         !
Seidman  H; Selikoff IJ; Gelb SK. Mortality Experience of Amosite Asbestos Factory^ Workers:
Dose-Response Relationships 5 to 40 Years After Onset of Short-Term Work Exposure.
American Journal of Industrial Medicine.  10(5/6) :479-514.1986.

Selikoff IJ; Seidman H. Asbestos-Associated Deaths among Insulation Workers in the United
States and Canada,  1967-1987. Annalsof the New YorkAcademy of Sciences. 643:1-14. 1991.

Selikoff IJ; Hammond EC; Seidman H.  Mortality Experience of Insulation Workers in the
United States and Canada 1943-1976. Annals New YorkAcademy of Sciences.  330:91-116.
1979.
                                                                         it
Smith DM; Oritiz LW; Archuleta RF; Johnson NF.  Long-Term Health Effects in Hamsters and
Rats Exposed Chronically to Man-Made Vitreous Fibres. Annals of Occupational Hygiene.
34(4B):731-754.  1987.                 -                 \                   ;

Snyder J; Virta R; Segreti J. Evaluation of the Phase Contrast Microscopy Method forthe
Detection of Fibrous and Other Elongated Mineral Particulates  by Comparison with a STEM
Technique. American Industrial Hygiene Association Journal.  48(5):471-477. 1987.

Stanton M; Wrench  C. Mechanisms of Mesothelioma Induction with Asbestos and Fibrous
Glass. Journal of the National Cancer Institute. 48:797-821.  1972.
                                        9.22

-------
Stantpn M; Layard M; Tegeris A; Miller E; May M; Kent E. Carcinogenicity of Fibrous Glass:
Pleural Response in the Rat in Relation to Fiber Dimension. Journal of the National Cancer
Institute.  58(3):587-597. 1977.

Stanton M; Layard M; Tegeris A; Miller E; May M; Morgan E.  Relation of Particle Dimension
to Carcinogenicity in Amphibole Asbestos and Other Fibrous Minerals. Journal of the National
Cancer Institute. 67(5):965-975. 1981.

Stayner L; Smith R; Bailer J; Gilbert S; Steenland K; Dement J; Brown D; Lemen R. Exposure-
Response Analysis of Risk of Respiratory Disease Associated with Occupational Exposure to
Chrysotile Asbestos. Occupational and Environmental Medicine. 54:646-652. 1997.

Stayner LT; Dankovic DA; Lemen RA. Occupational Exposure to Chrysotile Asbestos and
Cancer Risk: A Review of the Amphibole Hypothesis.  American Journal of Public Health.
86(2): 176-186. February. 1996.

Stober W; McClellen RO; Morrow PE. Approaches to Modeling Disposition of Inhaled
Particles and Fibers in the Lung. In Toxicology of the Lung, 2nd edition. Gardner DE; Crapo JD;
McClellan RO (eds.). Raven Press, New York. 1993.

Strom KA; Yu CP.  Mathematical Modeling of Silicon-Carbide Whisker Deposition in the Lung-
Comparison Between Rats and Humans. Aerosol Science and Technology. 24(3): 193-209.
1994.

Sussman RG; Cohen BS; Lippmann M. Asbestos Fiber Deposition in a Human
Tracheobronchial Cast. I. Experimental. Inhalation Toxicology. 3:145-160. 199la.

Sussman RG; Cohen BS; Lippmann M. Asbestos Fiber Deposition in a Human
Tracheobronchial Cast. II. Empirical Model. Inhalation Toxicology.  3:161-179.  1991b.

Takeuchi T; Nakajima M; Morimoto K. A Human Cell System for Detecting Asbestos
Cytogenotoxicity In Vitro. Mutation Research. 438(1):63-70.  1999.

Tanaka S; Choe N; Hemenway DR; Zhu S; Matalon S; Kagan E. Asbestos Inhalation Induces
Reactive Nitrogen Species and Nitrotyrosine Formation in the Lungs and Pleura of the Rat.
Journal of Clinical Investigation. 102(2):445^54. 1998.

Timblin CR; Guthrie GD; Janssen YWM; Walsh ES; Vacek P; Mossman BT. Patterns of C-Fos
and C-Jun Proto-Oncogene Expression, Apoptosis, and Proliferation in Rat Pleural Mesothelial
Cells Exposed to Erionite or Asbestos Fibers.  Toxicology and Applied Pharmacology.
151(l):88-97. 1998a.

Timblin CR; Janssen YMW;  Goldberg JL; Mossman BT. GRP78, HSP72/73, and CJUN Stress
Protein Levels in Lung Epithelial Cells Exposed to Asbestos, Cadmium or H2O2. Free Radical
Biology & Medicine. 24(4):632-642.  1998b.
                                         9.23

-------
                                                        I                   1
 Timbrell V. Deposition and Retention of Fibres in the Human Lung. Annals of Occupational
 Hygiene.  26(I-4):347-369. 1982,

 Timbrell V; Hyett AW; Skidmore JW. A Simple Dispenser forGenerating Dust Clouds From
 Standard Reference Samples of Asbestos. Annals of Occupational Hygiene.  11:273-281. 1968.

 Unfried K; Kociok N; Roller M; Pott F; Dehnen W.  P53 mutations in tumours induced by
 intraperitoneal injection of crocidolite asbestos and benzo[a]pyrene in rats. Experimental
 Toxicololgoy and Pathology. 49:181-187. 1997.                               !,
                                                                           t
 U.S. EPA (U.S. Environmental Protection Agency).  Airborne Asbestos Health Assessment
 Update. Report 600/8-84-003F,  U.S. Environmental Protection Agency. 1986.      •;

 U.S. Environmental Protection Agency (EPA). Asbestos Hazard Emergency Response Act:
 Asbestos-Containing  Materials in Schools. Final Rule and Notice (Appendix A: AHERA
 Method).  Federal Register, 40 CFR 763, Vol. 52, No. 2, pp. 41826-41903. October. :'1987.

 Venzon D; Moolgavkar S.  A Method for Computing Profile-likelihood-based Confidence
 Intervals.  Applied Statistics. 37:87-94.  1988.
                                                                           !
 Vincent JH. On the Practical Significance of Electrostatic Lung Deposition of Isometric and
 Fibrous Aerosols. Journal Aerosol Science. 16(6):511-519. 1985.                '

 Vincent JH; Johnston AM; Jones AD; Bolton RE; Addison J. Kinetics of Deposition and
 Clearance of Inhaled Mineral Dusts During Chronic Exposure. British Journal of Industrial
 Medicine. 42:707-715. 1985.                                                [
                                                        '                   |
 Wagner JC. Opening Discussion — Environmental and Occupational Exposure to Natural
 Mineral Fibres. In Biological Effects of Mineral Fibres. Wagner JC (ed.). IARC Scientific
 Publications,  pp. 995-998.  1980.

 Wagner JC; Berry G;  Skidmore JW; Timbrell V.  The Effects of the Inhalation of Asbestos in
 Rats.  British Journal of Cancer.  29:252-269.  1974.

 Wagner JC; Berry G;  Skidmore JW. Studies of the Carcinogenic Effects of Fiber Glass of
 Different Diameters Following Intrapleural Innoculation in Experimental Animals. NIOSH
 76-151. pp.193-197.  1976.                               '.

 Wagner JC; Berry G;  Hill R; Munday D; Skidmore J. Animal Experiments with MMM(V)F -
 Effects of Inhalation and Intrapleural Inoculation  in Rats. In Biological Effects of Man-Made
Fibres - Proceedings of a WHO/IARC Conference, Copenhagen, pp. 209-233.  1982]
                                                                           t
 Wagner JC; Skidmore JW; Hill RJ; Griffith DM.  Erionite Exposure and Mesotheliomas in Rats.
British Journal of Cancer. 51:727-730.  1985.

 Wagner JC; Griffiths DM; Munday DE. Experimental Studies with Palygorskite Dusts. British
Journal of Industrial Medicine. 44(ll):749-763.  1987.       '

                                        9.24

-------
Walker AM. Declining Relative Risks for Lung Cancer After Cessation of Asbestos Exposure.
Journal of Occupational Medicine.  26(2):422-426. 1984.

Walton WH. The Nature, Hazards, and Assessment of Occupational Exposure to Airborne
Asbestos Dust: A Review. Annals of Occupational Hygiene. 25:117-247. 1982.

Warheit DB; Snajdr SI; Hartsky MA; Frame SR. Lung Proliferative and Clearance Responses to
Inhaled para-Aramid RFP in Exposed Hamsters and Rats: Comparisons with Chrysotiles
Asbestos Fibers.  Environmental Health Perspectives. 105(Suppl 5):1219-1222. September.
1997.

Weill H.  1994. Cancer Mortality in Chrysotile Mining and Milling: Exposure-Response.
Asbestos-Cement. Annals of Occupational Hygiene. 38(4):412.  1994.

Weill H; Hughes J; Waggenspack C. Influence of Dose and Fibre Type oh Respiratory
Malignancy Risk in Asbestos Cement Manufacturing. American Review of Respiratory Disease.
120:345-354.  1979.

WilksSS. Mathematical Statistics.  2nd edition. Wiley Publication, New York. 1963.

Winer AA; Cossette M. The Effect of Aspect Ration on Fiber Counts: A Preliminary Study.
Annals New York Academy of Sciences. 330:661-672. 1979.

Weitzman SA; Graceffa P. Communication: Asbestos Catalyzes Hydroxyl and Superoxide
Radical Generation from Hydrogen Peroxide.  Archives of Biochemistry and Biophysics.
228(1 ):373-376. 1984.

World Health Organization (WHO). Reference Methods for Measuring Airborne Man-Made
Mineral Fibers (MMMF).  1985.

Wright GW; Kuschner M. The Influence of Varying Lengths of Glass and Asbestos Fibres on
Tissue Response in Guinea Pigs.  In Inhaled Particles: Part 2.  Walton WH; McGovern B (eds.).
Pergamon Press Oxford,  pp. 455^74. 1975.

Wylie AG; Virta RL; Segretti JM. Characterization of Mineral Population by Index Particle:
Implications for the Stanton Hypothesis.  Environmental Research. 43:427-439.  1987.

Wylie AG; Bailey KF; Kelse JW; Lee RJ. The Importance of Width in Asbestos Fiber
Carcinogenicity and its Implications for Public Policy. American Industrial Hygiene Association
Journal.  54:239-252.  1993.

Wylie AG; Skinner HCW; Marsh J; Snyder H; Garzione C; Hodkinson D; Winters R; Mossman
BT. Mineralogical Features Associated with Cytotoxic and Proliferative Effects of Fibrous Talc
and Asbestos on Rodent Tracheal Epithelial and Pleural Mesothelial Cells. Toxicology and
Applied Pharmacology. 147:143-150. 1997.
                                         9.25

-------
Yamaguchi R; Hirano T; Ootsuyama Y; Asami S; Tsurudome Y; Fukada S; Yamato|H; Tsuda T;
Tanaka I; Kasai H.  Increased 8-Hydroxyguanine in DNA and Its Repair Activity in Hamster and
Rat Lung After Intratracheal Instillation of Crocidolite Asbestos. Japanese Journal of Cancer
Research. 90(5):505-509.  1999.                                           j

Yamate G; Agarwal SC; Gibbons RD. Methodology for the Measurement of Airborne Asbestos
by Electron Microscopy. U.S. EPA Report No. 68-02-3266. U.S. Environmental Protection
Agency, Washington, D.C., U.S.A. 1984.                                     !
                                                                         i:
                                                                         |!
Yeh H-C; Harkema JR.  Gross Morphometry of Airways. In Toxicology of the Lung, 2nd edition.
Gardner DE; Crapo JD; McClellan RO (eds.). Raven Press, New York.  1993.      ''

Yu CP; Asgharian B. A Kinetic Model of Alveolar Clearance of Amosite Asbestos Fibers from
the Rat Lung at High Lung Burdens. Journal of Aerosol Science. 21:21-27.  1990b. ;••

Yu CP; Yoon KJ. Investigator's Report: Retention Modeling of Diesel Exhaust Particles in Rats
and Humans.  1991.                                                       i*
                                                                         I

Yu CP; Asgharian B; Abraham JL. Mathematical Modeling of Alveolar Clearance of Chrysotile
Asbestos Fibers from the Rat Lungs. Journal of Aerosol Science. 21:587-594. 1990a.

Yu CP; Asgharian B; Pinkerton KE. Intrapulmonary Deposition and Retention Modeling of
Chrysotile Asbestos Fibers in Rats. Journal of Aerosol Science. 22(6): 75 7-761. 1991.

Yu CP; Zhang L; Oberdoster G; Mast RW; Glass LR; Utell Ml; Deposition Modeling  of
Refractory Ceramic Fibers in the Rat Lung. Journal of Aerosol Science. 25(2):407-417. 1994.
                                                                         !,
Yu CP; Zhang L; Oberdoster G; Mast RW; Maxim D; Utell MJ. Deposition of Refractory
Ceramic Fibers (RCF) in the Human Respiratory Tract and Comparison with Rodent Studies.
Aerosol Science and Technology.  23(3):291-300. 1995a.     ,                  '•'

Yu CP; Zhang L; Oberdoster G; Mast RW; Glass LR; Utell MJ. -Clearance of Refractory
Ceramic Fibers (RCF) from the Rat Lung - Development of a Model. Environmental Research.
65(2):243-253.  1995b.

Zalma R; Bonneau L; Guignard J; Pezerat H; Jaurand M-C. Formation of Oxy Radicals by
Oxygen Reduction Arising from the Surface Activity of Asbestos. Canadian Journal of
Chemistry. 65:2338-2341.  1987.                                              i

Zanella CL; Timblin CR; Cummins A; Jung M; Goldberg J; Raabe R; Tritton TR; Mossman BT.
Asbestos-Induced Phosphorylation of Epidermal Growth Factor Receptor is Linked to c-fos and
apoptosis.  American Journal of Physiology. 277(4:Part 1):L684-L693.  1999.

Zhang Y; Lee TC; Guillemin B; Yu M-C; Rom WN.  Enhanced IL-lBeta and Tumor Necrosis
Factor-Alpha Release and Messenger RNA Expression in Macrophages from Idiopathic
Pulmonary Fibrosis  or After Asbestos Exposure. Journal of Immunology.  150(9):4188-^4196.
May.  1993.

                                        9.26

-------
Zhu S; Manuel M; Tanaka S; Choe N; Kagan E; Matalon S.  Contribution of Reactive Oxygen
and Nitrogen Species to Particulate-Induced Lung Injury.  Environmental Health Perspectives.
106(5):1157-1163.  1998.

Zoitus BK; De Meringl A; Rouyer E; Thelohan S; Bauer J; Law B; Boymel PM; Olson JR;
Christensen VR; Guldberg M; Koenig AR; Perander M. In Vitro Measurement of Fiber
Dissolution Rate Relevant to Biopersistence at Neutral pH: An Interlaboratory Round Robin.
Inhalation Toxicology. 9:525-540.  1997.
                                         9.27

-------
                              APPENDIX A:

            UPDATE OF POTENCY FACTORS FOR
     LUNG CANCER (KL) AND MESOTHELIOMA

Estimates of risk of dying of lung cancer or mesothelioma from asbestos exposure are quantified
by means of mathematical, models that express risk as a function of exposure. The models
utilized in the 1986 U.S. EPA Airborne Asbestos Health Assessment Update (U.S. EPA 1986)
contain parameters (KL for lung cancer and KM for mesothelioma) that gauge the potency of
asbestos for causing these health effects.  USEPA calculated KL and KM values from a number of
studies.  In this section these KL and KM calculations are revised.using the same models as in the
U.S. EPA (1986) update, but incorporating newer data from more recent publications. ;,Since the
1986 update, additional cohorts have been studied from several new exposure settings and the
followup periods have been extended for several of the previously studied cohorts.   '

In the 1986 update, KM values were not calculated from all of the available studies, perhaps
owing to the limited number of mesotheliomas observed in some of these studies. In this update,
an attempt has been made to utilize any study with suitable health and exposure data, regardless
of the number of mesotheliomas reported, and to quantify the statistical uncertainty  attributable
to small numbers using statistical confidence limits.  Since the present work utilizes somewhat
different methods from the 1986 update, for consistency, all of the KL and KM values v^ere
recalculated, even from studies for which no new data were available. Table A-l contains a
summary of the new values  for KL and Table A-2 contains the new values for KM. The original
values from the 1986 update are also provided for comparison. These tables also contain
statistical confidence limits  and ad hoc "uncertainty limits" for K^ and KM. The derivation of
these limits will be described in detail in subsequent sections.

A.1   LUNG CANCER MODEL                        '

The 1986 U.S. EPA lung cancer model (U.S. EPA 1986) assumes that the relative risk, RR, of
mortality from lung cancer at any given age is a linear function of cumulative asbestos exposure
(fiber-years/ml, or f-y/ml, as measured by PCM), omitting any exposure in the most recent
10 years. This exposure variable is denoted by CE10.  The 10-year lag embodies the assumption
that exposures during the most recent 10 years do not affect current lung cancer mortality risk.
The mathematical expression for this model is
                                                                           : A-l)
RR = 1+ 1C * CEi,
where the linear slope, KL, is the "lung cancer potency factor." To make allowance for the
possibility mat the background lung cancer risk in the exposed population differs from that of the
comparison population, the model is expanded to the form,
                                 = a*(l+KL*CE10).
                                         (Eq. A-2)
                                      A.I

-------
With this form of the model the relative risk at zero exposure is a rather than 1.0. Both K^ and a
are estimated by fitting the model to data.  The type of data usually available for applying this
model are from cohort studies in which observed and expected (based on an appropriate
comparison population, e.g., U.S. males) numbers of lung cancers are categorized by cumulative
exposure incorporating a 10 year lag.  To explore the adequacy of the model, it is useful to have
the data cross-classified by one or more other variables, such as latency.

Frequently the cumulative exposure variable available from the published report of a study does
not incorporate a lag (or, less frequently, incorporates a lag of less than 10 years). In this report,
rather than attempting an ad hoc correction, no correction for lag has been made.- Although this
tends to cause KL values to be slightly underestimated, this is unlikely to be a serious problem.
For most cohorts, exposures decreased significantly over time. Also,  in many studies, followup
didn't begin until several years after the start of exposure and the bulk of the lung cancers
occurred at older ages. All of these factors tend to mitigate the error created from use of data
with no lag.  Moreover, use of an ad hoc correction for lag could hinder comparisons of KL
values among studies  that do not employ  a lag (which includes the majority of studies).

A.2    MESOTHELIOMA MODEL

The 1986 U.S. EPA mesothelioma model (U.S. EPA 1986) can be derived by assuming that the
mortality rate at time t after the beginning of exposure can be calculated by summing the
contributions from exposure at each increment of time, du, in the past. The contribution to the
mortality rate at time t from exposure to E(u) f/ml (as measured by PCM) at time u is assumed to
be proportional to the product of the exposure rate, E(u), and (t-u-10)2, the square of the elapsed
time minus a lag of 10 years.  Thus, as with the lung cancer model, the mesothelioma model
assumes a 10-year lag before exposure has any effect upon risk. With the additional assumption
that the background rate of mesothelioma is zero, the mesothelioma mortality rate at time t since
the beginning of exposure is given by
                IM(t) =  3*KM*rl0E(u)*(t-u-10)2du,
                        (Eq. A-3)
where t and u are in years, and IM(t) is the mortality rate per year at year t after the beginning of
exposure. The proportionality factor, KM, is called the "mesothelioma potency factor." The
factor of "3" is needed to retain the same meaning of KM as defined by U.S. EPA (1986).

If exposure is at a constant level, E, for a fixed duration, DUR, this model can be written as

             0
                  = KM*E*(t-10)3
                 * E * [(t -10)3 - (t - 10 - DUR)3]
0<;t< 10
10 <; t< 10 +DUR
DUR  <; t
(Eq. A-4)
The genesis of this model and its agreement with data were discussed in U.S. EPA (1986).

Through the courtesy of Dr. Corbett McDonald, Professor Douglass Liddell, Dr. Nicholas
de Klerk, Dr. John Dement, and Hie National Institute for Safety and Health (NIOSH), raw data
on mesothelioma mortality were obtained from a cohort of Quebec chrysotile miners and millers
                                          A.2

-------
 (Liddell et al. 1997; McDonald et al. 1980a), a cohort of Wittenoom, Australia crocidblite miners
 and millers (Armstrong et al. 1988; de Klerk et al. 1994), and a cohort of workers from a plant in
 Charleston, South Carolina that manufactured textiles from chrysotile (Dement et al.' 1983a,b,
 1994; Dement and Brown 1998).  These data were used to calculate KM values in a more
 accurate manner for these cohorts (using the "exact" approach described below) and to explore
 the potential magnitude of the errors incurred by the crude application of cohort-wide! averages
 when fitting the mesothelioma model.
                                                                            i

 A.3   STATISTICAL FITTING METHODS

 The method of maximum likelihood (Cox and Oakes 1984; Venzon and Moolgavkar 1988) was
 used herein to fit the lung cancer and mesothelioma models to data and to estimate KL and KM.
 The profile likelihood method was used to calculate statistical confidence intervals and
 likelihood ratio tests were used to assess goodness-of-fit and test hypotheses.        I'
                                                                            '»
 Typically the data for calculating a lung cancer potency factor, KL, consist of observed and
 expected (based on an external control group, such as U.S. males) numbers of cancer deaths
 categorized by cumulative exposure. The likelihood of these data is determined  by assuming
 that the deaths in different exposure categories are independent and that the number of deaths in
 a particular category has a Poisson distribution with expected number given by the expected
 number predicted by the external control group times the relative risk given by either expression
 (Eq.A-lorA-2).                                                            •
                                                                            i

 In the typical situation, the published data most useful for calculating the mesothelioma potency
 factor, KM, consist of the number of mesothelioma deaths and person-years of observation
 categorized by time since first exposure.  The likelihood of these data is determined by assuming
 statistical independence of the number of mesothelioma deaths in different categories and that
 the number of mesothelioma deaths in a category has a Poisson distribution with mean equal the
 number of person-years in the category times expression (Eq. A-4), using average values for E,
 DUR, and t appropriate for that category.                     '                   ;

 The fitting of the mesothelioma model (Eq. A-3) to raw (unsummarized) mesothelioma data is
 accomplished using an "exact" maximum likelihood method. The cumulative mesothelioma
 hazard is defined as
                                  = f'lM(u)du.
. A-5)
The contribution to the likelihood of a person whose followup terminated at t is
S(t) = exp[ -H(t) ]  if the followup did not terminate in death from mesothelioma, and 1^0) * S(t)
if the person died of mesothelioma The complete likelihood was defined as the product of these
individual contributions. The integrals in expressions (Eq. A-3 and A-5) were evaluated
numerically.
                                         A.3

-------
A.4    SELECTION OF A "BEST ESTIMATE" OF KL AND KM

For each study for which a KL or KM is estimated, a "best estimate" is provided. For lung cancer,
the best estimate of K^ (Table A-l) was generally assumed to be the maximum likelihood
estimate (MLE) obtained with a estimated. For mesothelioma, the best estimate of KM (Table
A-2) is generally the maximum likelihood estimate derived from the best-fitting model in the
form (Eq. A-3) for raw data and (Eq. A-4) for published data. As described in the descriptions of
the individual studies, in a few cases these general rales had to be adapted to fit the particular
form of the data available.

A.5    UNCERTAINTY IN KL AND KM

Statistical uncertainty in KL and KM estimates is expressed using 95% upper and lower statistical
confidence limits.  These limits (summarized in Table A-l for lung cancer and Table A-2 for
mesothelioma) were computed using the profile likelihood method and (for K^) with a estimated.

However, non-statistical sources of uncertainty, such as model uncertainty and uncertainty in
exposure, are also  likely to  be very important.  Although these uncertainties are difficult to
quantify, it is important to attempt quantification, since presentation of statistical uncertainty
alone may provide a misleading picture of the reliability of the estimates.  Consequently, an ad
hoc approach to quantifying non-statistical uncertainty was adopted in this report. In this
approach, the primary sources of uncertainty are identified. Then, for each study, a factor was
selected for each uncertainty source using guidelines that will be described in this appendix.  The
individual factors were combined with the statistical confidence bounds to arrive at an
"uncertainty range" for KL or K^ for each particular cohort.  These ranges are described in detail
in following sections and are summarized in Table A-l for lung cancer and Table A-2 for
mesothelioma.

Because the most serious uncertainties among published epidemiology studies are often
attributable to the  estimation of exposure, three factors (Fl, F2, and F3) were defined to address
distinct sources of uncertainty associated with exposure. Two additional factors (F4L and F4M)
were defined to account for uncertainty due to special limitations that had to be addressed to
facilitate estimation of exposure-response factors from specific studies for lung cancer and
mesothelioma, respectively.

To define the factors we used to address uncertainty associated with exposure, we first
considered that, ideally, cumulative exposure would be estimated in an epidemiology study by:

       •      continuously monitoring the concentrations to which the worker is exposed over
              their entire working life;

       •      measuring such concentrations using personal monitors (samplers worn by
              workers with sampling ports placed within a few inches of the breathing zone of
              the worker); and
                                          A.4

-------
        •      analyzing samples in a manner appropriate for determining the concentration of
               the specific range of structures of interest1.                           '

 In practice, however, measurements are collected only periodically at fixed locations considered
 representative of worker exposures for jobs performed at that location (local operations).
 Moreover, measurements were frequently derived using analytical methods that report results in
 units different from those of interest, so that some type of conversion is required. Then,
 typically, cumulative exposures are estimated for individual workers as the sum (over:the set of
 jobs held by that worker) of the product of the mean exposure concentration for each job and the
 duration over which that job is performed. Thus:                                  ;
                                    -PCM,
-QZc
      j
LO1
(Eq. A-6)
 where:
        CPCM  is the cumulative exposure experienced by a worker to PCM fibers (f-years/ml);

       Q     is a factor used to convert concentration measurements in a particular study to
              PCM fiber concentrations whenever the measurements in the study were collected
              using a different method (usually dust concentrations determined by midget
              impinger, in which case the units of Q are f7ml/mppcf);

       CLO   is the concentration estimated for a particular "local operation" typically derived
              by a combination of measurement and extrapolation; and

       Dj     is the duration of time that the worker spent working in local operation '"j".

Note that, because exposure concentrations at specific locations have generally been observed to
decrease over time due to changes in process, introduction of dust control equipment, and other
factors, cumulative annual exposures are generally estimated for workers in the manner
described above and the annual exposures are then summed. However, this does not change the
general applicability of Equation A-6.

Based on Equation A-6, a factor, Fl, is defined to account for uncertainty introduced iii the
manner that the CLO are determined in specific epidemiology studies; a factor, F2, is used to
address uncertainty associated with the determination of the conversion factors, Q, for!specific
studies; and F3 is defined to represent uncertainty in the manner that job matrices are developed
       ^ost comparisons of epidemiology studies involve converting estimates of cumulative exposures to fiber
concentrations as determined by phase contrast microscopy (PCM) using the "membrane filter method". Thus, for
the discussion above, the range of structures (exposure index) of interest would be those determined using the
membrane filter method. Importantly, however, discussions in other portions of this document deal with determining
asbestos concentrations using an exposure index representing the specific range of structures that contribute directly
to biological activity, which should not be confused with the exposure index reported using the membrane filter
method.

                                           A.5

-------
in specific studies to assign workers to specific local operations over specific durations.  The
manner in which values were assigned for each uncertainty factor is described more fully below.

A.5.1  The Factor Fl

As indicated above, the factor, Fl, represents the uncertainty in concentration estimates to which
workers are exposed (in whatever units of exposure that are reported in a particular study).  In
addition to analytical uncertainty, considerations addressed when assigning values for Fl for
specific epidemiology studies include:

       •     to what extent exposure concentrations were directly determined from
             measurements collected at the locations and times that worker exposures  actually
             occurred; and

       •     whether measurements were derived from personal monitoring or from area
             monitoring (sampling a general area that is assumed representative of exposure
             conditions associated with jobs performed within the local area).

Regarding the latter consideration, exposure concentrations estimated in the published
epidemiology studies were almost universally determined by area, rather than personal
monitoring.  As has been reported in several of these studies (see, for example, McDonald et al.
1983b), area monitoring can miss short-term, high-level exposures contributed by the personal
actions being performed by a worker. Moreover, certain periodic activities potentially
associated with extremely high exposure (typically involving cleanup) were not performed
during time periods when work areas were routinely monitored.

Regarding the first bullet above, published epidemiology studies differ in the frequency  and time
period over which sampling was conducted. With few exceptions, little or no sampling was
conducted prior to the 1950's when exposure concentrations are thought  generally to be higher
than those monitored more recently, due to  lack of use of dust control equipment and procedures
that were introduced only later. For many studies, therefore, early exposures had to be estimated
by extrapolation from later measurements and the care with which such extrapolations were
performed also  varies from study to study.

Studies vary in  the degree to which the range  of local operations associated with a particular
facility were individually sampled.  Exposure conditions attendant to jobs performed in
association with local operations not sampled directly would then be extrapolated from
measurements collected for other local operations assumed to be associated with "comparable
exposures." As with extrapolations in time, the care with which such spatial extrapolations were
performed varies from study to study.

Values assigned for Fl  vary between 1.5 and 4 (all to the nearest 0.5). The most typical value
assigned is 2.0 for studies in which additional uncertainty is introduced due to use of area
samplers rather than personal samplers,  lack of measurements representative of episodic but
high-exposure jobs (usually associated with cleanup), and lack of direct measurements from the
earliest periods of exposure (when dust  control equipment and procedures were absent). To be
assigned a value of 2.0, however, authors must have had access to substantial numbers of

                                          A.6

-------
 samples representative of the majority of the local operations of interest, must have described a
 systematic procedure for extrapolating exposure estimates to less well studied local operations,
 and must have described a systematic procedure for extrapolating exposure estimates to earlier
 times when measurements were lacking. The logic used to assign Fl values (and values for the
 other uncertainty factors) for individual studies is described for each study in Section A. 6 of this
 appendix.
                                                                            i

 A.5.2  The Factor F2                                                       !

 F2 is a factor used to characterize the uncertainty introduced in deriving conversion factors to
 convert from the exposure indices measured in a particular study to the exposure index typically
 reported using the membrane filter method (as determined by PCM). In about half of ihe studies,
 concentrations are estimated in millions of dust particles per cubic foot (mppcf) as determined by
 midget impinger (see Section 4.3). The uncertainty introduced by such conversions varies from
 study to study because:                                                       '•'
                                                                            t

       •      for a small number of studies, the majority of measurements were performed by
              the membrane filter method so that conversion was unnecessary;
                                                                            i.
       •      for some studies, conversion factors were derived from a statistical analysis of a
              set of side-by-side measurements determined, respectively, using the membrane
              filter method and the other method from which measurements need to be
              converted (typically the midget impinger method);                  :•'
                                                                            i
       •      for some studies, lack of side-by-side measurements required expert judgement
              for comparing across samples collected at different times and locations; and

       •      for some studies, conversion factors were not derived at all, but were adapted
              from other studies of similar processes.                            '

Moreover, as has been demonstrated in several  studies, the factors used to convert other
measurements (primarily midget impinger) to the exposure index determined by PCM'vary as a
function of study environment, local operation, and time.  For example, the ratio of PGM to
midget impinger derived from side-by-side measurements in a single study reportedly varied
between 0.3 and 30 (McDonald et al. 1980a).

Note that, given the above, the factors used to convert measured concentrations to exposure
concentrations in units of interest (Q in Equation A-6) ideally should be brought into the sum on
the right and determined individually for each local operation.  However, with the exception of
the South Carolina study by Dement and cowbrkers (Dement et al. 1994; Dement and Brown  '
1998), only average (study-wide) conversion factors are typically estimated in any particular
study.                                                                       !'
                                                         i
Values for F2 assigned to particular studies vary between 1.0 and 3.0.  Studies in which
conversions were not required (due to routine use of PCM) or studies in which conversion
factors were determined for specific operations were assigned an F2 value of 1.0. Studies in
which a study-wide conversion factor was determined from paired measurements are assigned a
                                                         i
                                         A. 7

-------
value of 1.5. Studies in which conversion factors were adapted from other studies or for which
authors did not define a conversion factor were assigned larger values for F2.

A.5.3 The Factor F3

The factor, F3, is used in this study to represent the uncertainty attributable to the manner in
which job-exposure matrices were constructed in the various published epidemiology studies.
Authors for some studies had detailed work histories that could be used to identify the complete
set of specific jobs that each worker performed over their working life and the duration of time
spent on each job. Authors from other studies did not have access to individual work histories so
that crude estimates of average duration was applied to all members of the cohort. The factor,
F3, is used to account for conditions in which less than optimal job histories were used to .
identify the set of jobs performed by each worker and the duration that each worker spent
performing each such job.

A.5.4  The Factor F4L for Lung Cancer and F4M for mesothelioma

An additional factor is included (F4L for lung cancer) and (F4M for mesothelioma) to account
for uncertainties in mortality data (e.g., when diagnosis is uncertain for a substantial fraction of
potential mesothelioma cases) or when approximations or assumptions are required because the
data are not presented in the form needed for fitting the exposure-response models. Two
assigned F4L values are greater than 1.0 (1.5 and 2.0)} and six F4M values are greater than 1.0;
these six values range from 2.0 to 5.0.

A.5.5  Combining Individual Uncertainty Factors into an Overall "Uncertainty Range"

As indicated above, in addition to statistical confidence intervals, four uncertainty factors have
been proposed: Fl: exposure, general; F2: exposure conversion factor; F3: lack of individual
work histories; and F4L (lung cancer) and F4M (mesothelioma): non-exposure related. Since it
is unlikely that all of the uncertainty sources would cause errors in the same direction in the
same study, rather than multiplying the uncertainty factors, an overall uncertainty factor, F, was
calculated as:

                   F = exp{ [  Ln2(Fl) + Ln2(F2) + Ln2(F3) + Ln2(F4) ]* },

where 1.0 is the default value for any factor not explicitly provided. The overall "uncertainty
range" for KL or KM was calculated by dividing the statistical 95% lower bound by F and
multiplying the 95% upper bound by F.

A.6    ANALYSIS OF INDIVIDUAL EPIDEMIOLOGY STUDIES

Predominately Chrysotile Exposure

Quebec Mines and Mills. Liddell et al. 1997 extended the followup into 1992 of a cohort of
about eleven thousand workers at two  chrysotile asbestos mines and related mills in Quebec that
had been studied earlier by McDonald et al. 1980b (follow-up through 1975) and McDonald
et al. 1993 (follow-up through 1988).  Production at the mines began before 1900.  The cohort

                                         A.8

-------
 consisted of workers who worked ^ 1 month and who were bom between the years of 1891 and
 1920.  Follow-up began for each individual after 20 years from first employment. Trie most
 recent follow-up (Liddell et al. 1997) traced 9,780 men through May 1992, whereas i,138 (10%)
 were lost to view, most of whom worked for only a few months prior to 1935. Of those traced,
 8,009 (82%) were deceased as of 1992.                      '                  '>

 Estimates of dust levels in specific jobs were made from some 4,000 midget impingef
 measurements collected systematically starting in 1948 and periodically in the factory beginning
 in 1944. Estimates for the period prior to 1949 utilized interviews with long-term employees and
 comparison with more recent conditions.  These dust-level estimates were matched to individual
job histories to produce estimates of cumulative exposure for each worker (mppcf-years).
 Conversions between dust levels and PCM concentrations were derived from side-by^side
 samples. On the basis of over 600 side-by-side midget impinger and optical microscopy
 measurements, it was estimated that 3.14 fibers/ml was, on the average, equivalent to' 1.0 mppcf
 (McDonald et al. 1980b).                                 .                   :
                                                                           (

 Liddell et al. (1997) categorized cancer deaths after age 55 from of lung, trachea, and;  bronchus
 by cumulative asbestos  exposure to that age (Liddell et al. 1997, Table 8). Standardized
 mortality ratios (SMRs) were calculated based on Quebec rates from 1950 onward, arid
 Canadian, or a combination of Canadian and Quebec rates, for earlier years. Table A-4 shows
 the fit of the lung cancer model to these data. Although the models both with ct=l and a variable
 provided reasonably adequate fits to the data, the hypothesis a=l can be rejected (p=0.014). The
 model  with a estimated yields a best estimate of K^ of 0.00029  (f-y/ml)-1, 90% CI: (0.'00019,
 0.00041). With o=l, the estimate was Kj,=0.00041 (f-y/ml)'1, 90% CI: (0.00032,0.00051).

 Smoking history was obtained in 1970 by a questionnaire administered to current workers, and
to proxies of those who had died after 1950. Although no analyses of lung cancer and asbestos
 exposure were presented for the 1992 follow-up (Liddell et al. 1997) that controlled for smoking,
such an analysis was conducted for the follow-up that continued through 1975 (McDonald et al.
 1980a). Table 9 of McDonald et al. (1980a) contained data on lung cancer categorized jointly by
cumulative exposure to  asbestos and by smoking habit. Two models were fit to thesejdata: the
multiplicative model for relative risk
                            RR = a * (1 + b * d) * 0 + c * x),
and the additive model
where d is cumulative exposure to asbestos to age 45, x is number of cigarettes smoked per day,
and a,b, and c are parameters estimated from the data. The multiplicative model fit the data
well, but the fit of the additive model was inadequate.  This corroborates the multiplicative
interaction between smoking and asbestos exposure in causing lung cancer (Hammond et al.
1979).  The estimate of potency using the multiplicative model was 0.00051 (f-y/ml)'^ which
was very close to that of 0.00045 (f-y/ml)-1 estimated from Table 5 of McDonald et al. (1980a),
which did  not utilize smoking data. This suggests that the association between lung cancer and
asbestos exposure is not strongly confounded with smoking in this cohort.

                                         A.9

-------
By 1993,38 deaths from mesothelioma had occurred in this cohort (Liddell et al. 1997).
Through the courtesy of Dr. Corbett McDonald and Professor Douglass Liddell, the underlying
mesothelioma data from this study were provided for additional analysis (Liddell 2001). These
data contained the following information on each worker: the date of birth, asbestos exposure
history, last date of follow-up, whether follow-up ended as a result of death from mesothelioma,
location of first employment, and whether a worker had been employed at more than one
location.

Nine distinct locations for first employment were coded. Locations 5-9 referred to small
operations, some having very heterogeneous exposures, and were omitted from the analysis.
Also, workers who worked at more than one location were omitted. After these exclusions, there
remained 9,244 workers who worked at Locations 1-4, and among whom 35 deaths from
mesothelioma occurred. Location 1 (4,195 men, 8 deaths from mesothelioma) was the mine and
mill at the town of Asbestos. Location 2 (758 men, 5 deaths) was a factory at the town of
Asbestos that, in addition to processing chrysotile, had also processed some crocidolite.
Location 3 (4,032 men, 20 deaths) comprised a major mining and milling company complex near
Thetford Mines. Location 4 (259 men, 2 deaths) comprised a number of smaller mines and mills
also in the vicinity of Thetford Mines. Because of the small number of workers at Location 4,
the fact that both locations were near Thetford Mines, and the fact that the separate KM values
obtained from Locations 3 and 4 were similar, data from.these locations were combined. The
remaining groups were analyzed separately, because of the crocidolite used at Location 2, and
because of evidence of greater amounts of tremolite in the ore at Thetford Mines that at Asbestos
(Liddell etal. 1997).

The availability of the raw data from this study made it possible calculate K^ from this study
using an "exact" likelihood approach based on expression (Eq. A-3) that did not involve any
grouping of data, or use of average values. For Location 1 (Asbestos mine and mill),
KM=0,013xlO-8, 90% CI: (0.0068x10-*, 0.022xlQ-8). For Location 2 (Asbestos factory),
KM=0.092xlO-8, 90% CI: (0.040x10'8, O.lSxlO'8). For Locations 3 and 4, KM= 0.021X10'8, 90%
CI: (0.014x10'8, 0.029x10-*). The KM estimate from Location 1 (whose ore was reported to have
a lower tremolite content) was about one-half that from Locations 3 and 4, although this
difference was not significant (p=0.22). The KM estimated from Location 2, the mill where
substantial crocidolite was used, was 4-7 times higher than the KL  estimated from Location 1
and Locations 3 and 4.

For comparison purposes, KM were also calculated using grouped data and applying expression
(4), since mis is the method that must be used with most studies. For Location 1 (3 and 4) the
KM estimate based on the "exact" analysis was 34% (25%) higher than that based upon grouped
data  This suggests that reliance upon published data for calculating KM may introduce some
significant errors in some  cases.  Such errors may be further compounded by the failure of some
studies to report the needed data on levels and durations of exposure in different categories of
time since first exposure.

For this study Fl is set equal to 2.0. This study is the paradigm used to define the typical case
(see Section A.5.1) in which increased uncertainty  can be attributed to use of area rather than
personal samplers, lack of measurements early in the study, and lack of direct measurements
from certain episodic but high-exposure  operations. At the same time, the authors of this study

                                         A.1O

-------
appear to have used the available data in a systematic and objective manner to address the issues
raised by the lack of sampling.                                                ,

The uncertainty factor F2 is set equal 1.5 for this study to reflect use of a conversion factor that
is derived from paired samples, but that is based on a project wide average, rather than
addressing variation for specific, local operations.                               *

All other uncertainty factors are set equal to 1.0 for this study due to lack of remarkable
distinctions. Thus:                                        '                   '•:
       Fl = 2.0                                                              •
       F2 = 1.5
       F3 = 1.0
       F4L = 1.0                                                             ''
       F4M = 1.0                                                            ;

These uncertainty factors, when coupled with the statistical confidence limits, resulted in the
uncertainty ranges for KL and KM shown in Tables A-l and A-2.

Italian Mine and Mill.  Piolatto et al. (1990) conducted additional follow-up of workers at a
chrysotile mine and mill in Italy that was earlier studied by Rubino et al. (1979). The; cohort ^
consisted of 1058 workers with at least 1 year of employment between 1946 and 1987.  Follow-
up extended from 1946 through  1987, which is 12 more years of follow-up than in Rubino et al.
(1979).  Lung cancer mortality was compared to that of Italian men.                ,,

As described in Rubino et al. (1979), fiber levels were measured by PCM in 1969.  Inlbrder to
estimate earlier exposures, information on daily production, equipment changes, number of
hours worked per day, etc. were used to create conditions at the plant during earlier years. PCM
samples were obtained under these simulated conditions and combined with work histories to
create individual exposure histories.                          '                   :

Piolatto et al. (1990) observed 22 lung cancers compared to 11 in the earlier study (Rubino et al.
1979). Lung cancer was neither significantly in excess nor significantly  related to  cumulative
asbestos exposure. Piolatto et al. (1990, Table 1) presented observed and expected lung cancers
(based on age- and calendar-year-specific rates for Italian men) categorized by cumulative
exposure in f-y/ml. The lung cancer model with fixed a provided a good fit to these data (Table
A-5, p=0.75) and allowing a to vary did not significantly improve the fit The KL estimate with
a=l was 0.00035 (f-y/ml)4, with 90% CI: (0, 0.0015). With a allowed to vary the  estimate was
KL=0.00051 (f-y/ml)-1 with 90% CI: (0, 0.0057).

Two mesotheliomas were observed by Piolatto et al. (1990), compared to one found by Rubino
et al. (1979).  However, data were not presented in a form from which K^ could be estimated. "

Regarding uncertainty factors, Fl is assigned a value 2.0 for this study for reasons  similar to
those described for Quebec. F2 is assigned a value of 1.0 because measurements werejconducted
using PCM so that conversion is unnecessary for this study.  All other factors are also assigned a
value of 1.0 because there are no other unique limitations.  Thus:

                                         A.11            '•

-------
       Fl=2.0
       F2=1.0  "
       F3 = 1.0
       F4L = 1.0

These uncertainty factors, when coupled with the statistical confidence limits, resulted in the
uncertainty range for KL shown in Table A-l.

Connecticut Friction Product Plant McDonald et al. (1984) evaluated the mortality of
workers employed in a Connecticut plant that manufactured ashestos friction products.  The
plant began operation in 1913 and used only chrysotile until 1957, when a little anthophyllite
was used.  Also, a small amount of crocidolite (about 400 pounds) was handled experimentally
between 1964 and 1972.  Brake linings and clutch facings were made beginning in the 1930s,
and production of automatic transmission friction materials, friction disks and bands was begun
in the 1940s.

The cohort was defined to include any man who had been employed at the plant for at least
1 month before 1959, omitting all that had worked at a nearby asbestos textile plant that closed
in 1939. This cohort consisted of 3,515 men, of whom 36% had died by the end of follow-up
(December 31,1977). Follow-up of each worker was only begun past 20 years from first
employment.

Information on dust levels from impinger measurements were available for the years 1930,1935,
1936, and 1939.  There was little other exposure information available until the 1970s.  An.
industrial hygienist used these measurements and information on processes and jobs,
environmental conditions and dust controls to estimate exposures by process and by period in
units of mppcf. No conversion from mppcf to r7ml value was suggested by the authors, a
conversion factor or between 1.4 and 10 is suggested by other studies. The most common value
seems to be around 3 f/ml per mppcf, which has been observed in diverse environments such as
mining and textile manufacture. This value was provisionally applied to this cohort, although
this conversion has considerable uncertainty associated with it.

Total deaths and  deaths from most individual causes investigated were elevated; these elevations
were due primarily to increased deaths in the group working for <1 year. This pattern holds for
lung cancer in particular; the SMR for lung cancer was highest (180) for persons exposed for
<1 year. A similar pattern holds when the analysis was carried out by cumulative exposure
(Table A-6), the SMR in the lowest exposure category is higher than in any other category. The
linear relative risk lung cancer model provided a poor fit (p=0.01) to these data when the
Connecticut rates were assumed to be appropriate for this cohort (fixing the parameter a=l); use
of U.S. rates gave similar results.  However, the fit was adequate (p=0.28) if the background
response is allowed to rise above that of Connecticut men (allowing the parameter a to vary).
Although the reason for this increased response in persons that worked for a short period or have
low exposures is  not clear, the analysis in which the background response is allowed to vary
appears to be the most appropriate. This analysis yields an estimate of K^O.O (f-y/ml)"1, 90%
CI: (0,0.0017).  The analysis with cc=l yielded KL=0.0019 (f-y/ml)4, 90% Cl: (0,0.0061).
                                         A.12

-------
McDonald et al. did not find any mesotheliomas in this cohort.  It is useful to determine the
range of mesothelioma risk that is consistent with this negative finding. Although McDonald et
al. do not furnish data in the form needed for this calculation, these data can be approximated
from Table 1 of McDonald et al. (1984). In this table they list 511 deaths occurring after age 65.
Assuming that the overall SMR of 108.5 held for persons over 65 years of age, the expected
number of deaths is 511/1.085 = 471. The death rate in U.S. white males between 65 and 75
years of age is approximately 0.050 per year (from 1971 vital statistics). Therefore the number
of person years observed in persons post 65 years of age is estimated as 471/0,050=9^420.

A lower bound on the person-years of follow-up between ages 45 and 65 can be estimated by
assuming that follow-up was complete for this age group. First we estimate the number of
persons that would have had to have been in the cohort to-experiehce the observed deaths.
Assuming that x persons in the cohort are alive at age 45, we have the following estimates of the
number entering each successive five-year age interval and the corresponding number of deaths
(based on death rates  in 1,971 white males).
Age
45-50
50^55
55-60
60-65
65+
TOTALS
Number Entering Interval
X
x(l -0.0063 8) 5=0.97x
0.97x(l -0. 0 1 072) 5=*0. 92x
0.92x(l-0.01718)5=0.84x
0.84x(l-0.02681) 5=0.73x

Number of Deaths
in Interval
0.032x
0.052x
0.076x
O.llx

0.27x
i
Person-Years in
Interval
4.9x
4<7x
4.4x
3.9x
,,
18. Ox
Since there were 616 deaths in men between the ages of 45 and 65, the expected number of
deaths is estimated as 616/1.085=567.7 expected deaths between ages of 45 and 60, the number
of persons entering this age interval is estimated as x=567.7/0.27=2,100. The person-years is
then estimated as (2, 100)(I7.964)=38,000.      .             ,                   ;
                                                                            I,
Using the average age of beginning work of 30.95 years (McDonald et al. [1984], Table 3) yields
the data in Table A-7.  Moreover, the average duration of exposure in this cohort was 8.04 years
and the average exposure level was 1 .84 mppcf (McDonald et al. [1984], Table 3), which is
equivalentto 1. 84x3=5.52 fibers/ml.  These data yields an estimate of KM=0.0 and a 90% upper
bound of KM=1.2xlO'9.
The best estimate of KM was assumed to be zero. For uncertainty factors, Fl is assigned a value
of 2.0 for reasons similar to those described for Quebec.  F2 is assigned a value of 3.0 for this
study because there is no conversion factor reported by the authors so that an average value of 3
for tiie range of conversion factors observed among the available studies (U.S. EPA 1986) was
selected for use with this study. To derive an exposure-response, factor for mesothelioma from
this study, an upper bound had to be estimated by reconstructing the data because the authors do
not provide the data in a form suitable for performing the required calculation. Therefore, F4M
is assigned a value of 3 for this study. Thus:                  '
                                         A.13

-------
       Fl = 2.0
       F2 = 3.0
       F3 = 1.0
       F4L = 1.0
       F4M = 3.0

These values, when coupled with the statistical confidence limits, resulted in the uncertainty
ranges for KL and KM shown in Table A-l and A-2, respectively.

New Orleans Asbestos-Cement Plants. Hughes et al. (1987) report on follow-up through 1981
of a cohort of Louisiana workers from two asbestos cement plants studied previously by Weill et
al. (1979). Although chrysotile, amosite and crocidolite were used at these plants, a group of
workers at one of the plants were only exposed to chrysotile.  The cohort contained 6,931
workers, of whom 95% were traced, compared to a 75% success in tracing by Weill et al. (1979).
This improved trace was the result both of greater access to Social Security Administration
records and greater availability of computerized secondary information sources (Dr. Hughes,
personal communication).

Both of the plants have operated since the 1920s.  Chrysotile was used predominantly in both
plants. Some amosite was used in Plant 1 from the early 1940s until the late 1960s, constituting
about 1% of some products,  and crocidolite was used occasionally for approximately 10 years
beginning in 1962.  Plant 2 utilized only chrysotile, except that pipe production, which began in
1946 and was housed in a separate building, produced a final product that contained about 3%
crocidolite. Since the total percentage of asbestos fiber in most asbestos cement products ranges
from 15 to 28%, it is estimated that crocidolite constituted between 10 and 20% of the asbestos
used to make cement pipe (Ontario Royal Commission 1984).  Workers from Plant 2 that did not
work in pipe production were exposed only to chrysotile.

Estimates of airborne dust levels were made for each job by month and year from midget
impinger measurements initiated in the early 1950s.  Levels estimated from initial samples in the
1950s were also assumed to  hold for all earlier periods because no major dust control measures
had been introduced prior to that time. New exposure data from Plant 2 became become
available after the earlier study (Weill et al.  1979) was completed, and these, along with a
complete review of all the exposure data, were used to revise the previous estimates of exposure.
In Plant 1 the earlier and revised estimates were reasonably similar, but in Plant 2, the revised
estimates tended to be about one-third of the previous estimates through the 1940s and about
one-half Hie previous estimates thereafter. Based on 102 side-by-side measurements by midget
impinger and PCM in various areas of one of the plants, Hammond et al. (1979) estimated an
overall conversion factor of 1.4 fibers/ml per mppcf. There were substantial variations in this
factor among different areas of the plant.

The principal cohort studied consisted of all workers who, according to company records, were
employed for at least one month prior to 1970, had a valid Social Security number, and were first
employed in 1942 or later (Plant 1), or in 1937 or later (Plant 2). Mortality experience was
compared with that expected based on Louisiana rates.
                                         A.14

-------
Hughes et al. found no significant difference between the exposure responses for lung cancer in
Plant 2 among workers exposed to chrysotile only and those who were also exposed to
crocidolite in pipe production. A single lung cancer exposure response model adequately
describes the lung cancer data from Plants 1 and 2 combined (pi 0.42, Table A-8).  The fit of this
model is good when Louisiana men are assumed to be an appropriate control group (fixing the
parameter a=l).  This fit provides an estimate of 1^=0.0040 (fiber-y/ml)0, 90% CI: (6.0015,
0.0070). With a allowed to vary, the estimate is 0.0025 (fiber-y/ml)-1, 90% CI: (0,0.6066 ).

Six mesotheliomas were identified in the primary cohort studied by Hughes et al., two in Plant 1
and four in Plant 2. Four other mesotheliomas are known to have occurred, one among those
initially employed in Plant 2 before 1937 and three among Plant 2 workers shortly after follow-
up ended in  1981. A case control analysis conducted among Plant 2 workers found a,
relationship  between mesothelioma risk and length of employment and proportion of !time spent
in the pipe area after controlling for length of exposure, which is consistent with a greater risk of
mesothelioma from crocidolite exposure.                                       .

Data were not presented in the paper in the form required for estimating KM. However, Hughes
and Weill (1986) present estimates of mesotheliomas potency from several data sets, including
the cohort studied in Hughes et al. and containing six mesotheliomas, but using a model slightly
different from the 1986 EPA model (3). Estimating KM by multiplying the potency estimated by
the Hughes and Weill (1986) model by the ratio of the potency values estimated for another
study using the 1986  U.S. EPA model and the Hughes-Weill (1986) model yielded the following
estimates of KM for the Hughes et al. (1987) data: 0.25x10'8 (Selikoff et al. 1979); 0.21x1O'8
(Dement et al. 1983b); 0.27xlO-8(Seidman et al. 1979); and 0.43xlO-8 (Finkelstein 1983).  Based
on these calculations, 1^=0.30x10"8 seems to be a reasonable estimate for the Hughes' et al.
cohort.
                                                                           i

It would be worthwhile to estimate mesothelioma risk using additional follow-up that included
the three cases that occurred shortly after follow-up ended.  However, such an estimate should be
no larger than about 1^=0.45x10"8. This is because, since there were six mesotheliomas in the
cohort studied by Hughes et al., even if the additional person years of follow-up post-1981 is not
taken into account, the three additional mesotheliomas would increase the estimate of tKM by
only about 50%.                                                              '

The finding by Hughes et al. (1987) of an association with crocidolite exposure implies that a
smaller KM would correspond to the chrysotile-only exposed group in Plant 2. Although Hughes
et al. didn't furnish the data needed for precise estimation of KM from this cohort, it is possible to
make some reasonable approximations to this KM. Since none of the six mesotheliomas occurred
among workers exposed only to chrysotile, KM=0 would be the point estimate derived from the
data used by Hughes et al.

However, one mesothelioma was discovered in  a person whose employment began in 1927 and
thus was not eligible for inclusion in the cohort.  This person was employed continuously for 43
years in the shingle production area, where only chrysotile was used.  In an attempt to compute
an alternative KM using this one case, it was  noted that the duration of observation of the Hughes
et al. cohort was roughly equivalent to that of the Dement et al. (1983b) cohort. If the person-
years from this cohort, categorized by years  since first exposure, are adjusted by the ratio  of the
                                                         I
                                        A.15

-------
sizes of Dement et al. and the Hughes et al. non-crocidolite-exposed cohort from Plant 2, one
mesothelioma is assumed to occur (in 30+ years from first exposure category) and the average
duration of exposure (2.5 years) and fiber level (11.2 fibers/ml) appropriate for the Hughes et al.
cohort are applied to these data, a KM=0.2xlO'8 is obtained.

The best estimate of KM was assumed to be 0.2xlO"8 for workers exposed only to chrysotile and
0.3x10"8 for workers exposed to both chrysotile and amphibole. For uncertainty factors, Fl is
assigned a value of 2.0 for reasons similar to those described for Quebec. F2 is assigned a value
of 1.5 because most early measurements were collected by midget impinger and the authors
report using a conversion factor of 1.4 derived from paired measurements.  Due to the lack of
adequate data for estimating both the overall mesothelioma rate and a confidence interval for
such rates and the consequent need to reconstruct the data (incorporating numerous assumptions)
to be able to obtain the needed estimates, a value of 5.0 was assigned to the factor F4M for
chrysotile exposures and 2.5 for mixed exposures. Thus:

      Fl = 2.0
      F2 = 1.5
      F3 = 1.0
      F4L = 1.0
      F4M = 5.0

These values, when coupled with the statistical confidence limits, resulted in the uncertainty
ranges for KL and KM shown in Tables A-l and A-2.

South Carolina Textile Factory. Dement and coworkers (Dement et al. 1994; Dement and
Brown 1998) conducted a retrospective cohort study of employees of a chrysotile textile plant in
South Carolina. In an earlier study of this plant (Dement et al. 1982,1983a,b), the cohort was
defined as all white male workers who worked for one or more months between 1940 and 1965,
and follow-up was through 1975. Dement et al. (1994) expanded the cohort to include black
male and white female workers who met the entrance requirements, and  extended follow-up
through 1990, an additional 15 years. This expanded cohort included 1,247 white males (2.8%
lost to follow-up), 1,229 white females (22.8% lost to follow-up) and 546 black males (7.8% lost
to follow-up).  A total of 1,259 deaths were identified, and a death certificate was located for all
but 79 (6.2%) of the deaths.

Based on data from 5,952 air samples taken at the plant between 1930 and  1975, linear statistical
models were used to reconstruct exposure levels, while taking into account textile processes,
dust control methods, and job assignments (Dement et al.  1983a). For each worker, time spent in
each job was multiplied by the estimated exposure level for that job to estimate cumulative
exposure (f/ml-days). Based on regression analyses applied to 120 side-by-side particle and
fiber counts, Dement (1980) estimated a f/ml to mppcf ratio of 2.9, 95%  CI: (2.4, 3.5). Also,
between 1968 and 1971 both impinger and PCM samples were collected (a total of 986 samples).
Based upon a regression analysis of these data, Dement (1980) determined that a common
conversion factor could be used for jobs except fiber preparation.  For fiber preparation, a
conversion factor of 7.8 was found, 95% CI: (4.7-9.1). For all other operations, a value of 2.5,
95% CI: (2.1-3.0) was calculated. Based on this information, Dement et al. (1983a) concluded
                                         A.16

-------
that a conversion factor of 3 was appropriate for all operations except preparation, for which a
factor of 8 was adopted.

The underlying data for this cohort were obtained from the National Institute for Safety and
Health (NIOSH).  These data consisted of a work history file and a file with exposure levels by
job category and time period. The work history file contained codes for race, sex, month and
year of birth, vital status, month and year of death, and the department, operation, start date, and
stop date for each job worked. The exposure level file contained the exposure start arid stop
dates and the exposure level (fiber/ml) by the plant code, the department code, and the operation
code.
                                                                            !•
The cohort was defined as the white and black males and the white females who met the
employment requirements described above. This cohort included 1,244 white males (1.5% lost
to follow-up), 550 black males (7.5% lost to follow-up), and 1,228 white females (22/1% lost to
follow-up).        .                                                          •'

Table A-9 shows observed and expected deaths for lung cancer among white males, black males
and white females, categorized by cumulative exposure.  This table shows an excess of lung
cancers that exhibited an exposure response relationship.  U.S. rates were used for calculating
expected deaths, whereas South Carolina lung cancer rates are higher for white men, but slightly
lower for white women and black men. Whereas twelve categories of cumulative exposure were
used for fitting the model, these were been combined into seven categories for display, in Table
A-9. The model with a=l and a variable fit the data well (p^O.8), and the hypothesis that cc=l
cannot be rejected (p=0.19). The estimate of KL with a=l was 0.028 (f-y/ml)-1, 90% CI: (0.021,
0.037), and the estimate with a variable was KL=0.021 (f-y/ml)'', 90% CI: (0.012,0.034). An
analysis applied to white men alone gave somewhat higher estimates (KL=0.040 (f-y/ml)*1 with
o=l, and KL=0.026 (f-y/ml)-1 with a variable).                                    ;

Two deaths were certified as due to mesothelioma on the death certificates.  In addition, Dement
et al. (1994) considered four other deaths as likely due to mesothelioma. The availability of the
raw data from this study made it possible calculate KM from this'study using an "exact"
likelihood approach based on Equation A-3 that did not involve any grouping of data, or use of
average values.  Using the six confirmed and suspected mesotheliomas, KM= 0.43xlO'8, 90% CI:
(0.20xlO-8, 0.79xlO'8).  Using the two confirmed mesotheliomas; KM=0.14xlO-8, 90% CI:
(0.034x10-8, 0.38x10'8).

For comparison purposes, KM were also calculated using grouped data and applying Equation
A-4, since this is the method that must be used with most studies'. The data were divided into 10
categories by the tabulated values of Equation A-4. The KM estimate based on the "exact"
analysis was 2% greater than that based upon grouped data.

The best estimate of KM was assumed to be the geometric mean of the MLE estimates computed
using either confirmed or both confirmed and suspected mesotheliomas (0.25x10"*).  The
statistical lower bound used for this estimate was the one based on confirmed cases and the
upper bound used was the one based on confirmed and suspected cases.
                                        A. 17

-------
Regarding uncertainty factors, Fl is assigned a value of 1.5 for this study to give credit for the
reasonably complete sampling coverage of exposures by a combination of midget impinger and
extensive PCM, and the formal statistical evaluation conducted to derive job-specific exposure
estimates.  However, the exposure estimates are still based on analyses of area rather than
personal samples. Because multiple factors were used to convert midget impinger measurements
to PCM based on side-by-side samples collected from specific areas (associated with specific
operations) within the plant, a value of 1 .0 is assigned for F2 for this study.  The treatment of
statistical confidence limits described above was considered adequate to account for the
uncertainty in the number of mesotheliomas, and a value of K.M=1 was assigned. In summary:
       F2 = 1.0
       F3 = 1.0
       F4L = 1.0
       F4M=1.0

These uncertainty factors, when coupled with the statistical confidence limits, resulted in the
uncertainty ranges for KL and KM shown in Tables A-l and A-2.

McDonald et al. (1983a) conducted a cohort mortality study in the same South Carolina textile
plant that was studied by Dement et al. (1994).  Their cohort consisted of all men employed for
at least 1 month before 1959 and for whom a valid social security record existed. This cohort
consisted of 2,410 men, of whom 36% had died by the end of follow-up (December 31, 1977).
Follow-up of each worker was begun past 20 years from first employment.

McDonald et al. (1983a) had available the same exposure measurements as Dement et al.
(1983b) and used these to estimate cumulative exposures for each man in mppcf-y.  In their
review of the environmental measurements in which both dust and fiber concentrations were
assessed, they found a particle to fiber conversion range of from 1 .3 to 10.0 with an average of
about 6 fibers/ml per mppcf. This value, which is intermediate between the values of 3 and 8
found by Dement et al. (1983b) for different areas of the same plant, will be used in the
calculations involving the McDonald et al. (1983 a) study.

McDonald et al. describe two practices at the plant that entailed very high exposures and which
were not reflected in either their's or Dement and coworkers estimates: cleaning of burlap bags
used in the air filtration system by beating them with buggy whips during the years 1937-1953,
and the mixing of fibers, which was carried out between 1945 and 1964 by men with pitch forks
and no dust suppression equipment.

A strong exposure response for lung cancer was observed (Table A-10), which parallels the
results of Dement et al.  (1994). Unlike Dement et al., McDonald et al. used South Carolina men
as the control group rather than U.S. men. Use of this control group provided an adequate
description of the data and lung cancer potency values estimated both  with a=l and allowing a
to vary provided excellent descriptions of the data (p^O.88) and the  hypothesis a=l could not be
rejected (p=0.80).  Assuming a=l resulted in 1^=0.012 (f-y/ml)-1, 90% CI: (0.0075, 0.016), and
when a was allowed to vary, 1^=0.010 (f-y/miy1, 90% Cl: (0.0044, 0.025).  These results are
                                         A. 1.8

-------
 reasonably consistent with the potency estimated from Dement et al. (1994), and the differences
 can be largely accounted for by the different assumptions regarding the fiber/particle ratio.
                                                                            if
 McDonald et al. (1983a) found one case of mesothelioma in this cohort, apparently the same one
 discovered by Dement et al. (1983b): a rrian born in 1904 who died in 1967 and worked at the
 plant for over 30 years.  Since this study was conducted exactly as McDonald et al. (1-984), the
 same method used there to reconstruct person-years by years from first exposure can be applied
 to this cohort as well. The reconstructed data are listed in Table A-l 1. The estimated potency
 MLE is KM=0.088 xlO'8, with a 90% CI: (0.0093x10-*, 0.32xlO'8).                  j1

 For uncertainty factors, Fl is assigned a value of 2.0 for reasons' similar to those described for
 Quebec. F2 is assigned  a value of 1.0 because McDonald essentially used the same data that
 Dement and coworkers  used to estimate conversion factors (see above), although they favored a
 slightly higher mean value. We used the values favored by Dement when evaluating this study.

 Unlike the study by Dement and coworkers (for which we received the raw data so that we
 could calculate the exposure-response factor and the attendant confidence interval for
 mesothelioma directly), the mesothelioma data published in the McDonald study of this facility
 was not suitable to estimating confidence bounds.  Thus the data had to be reconstructed, which
 required incorporation of numerous assumptions. To account for the uncertainty associated with
 the reconstruction, F4M is assigned a value of 3 for this study. Thus:              \
                                                                            \'
       Fl = 2.0                                                             ;
       F2 = 1.0                                                             !'
       F3 = 1.0   •
       F4L = 1.0
       F4M = 3.0

 These uncertainty factors, when coupled with the statistical confidence limits, resultediin the
 uncertainty ranges for KL and KM shown in Tables A-l and  A-2.                    '

 Predominant Crocidolite Exposure

 Wittenoom, Australia Mine and Mill, de Klerk et al. (1994) followed a cohort of 6,904 men
 and women employed at a crocidolite mine and mill in Wittenoom, Australia. This cohort was
 followed through 1999 and the raw data were obtained through the courtesy of Dr. de Klerk.
 The data consisted of a record number, date of birth, sex, employment start date, total days  of
 employment, average exposure level (f/ml), cumulative exposure (f-y ear/ml), date of last contact,
 ICD code for cause of death, indicator variable for mesothelioma death, and date of death if
 applicable.                                           .                       |

 A number of subjects from the full cohort were removed from the analysis reported herein:  412
 because the sex was not designated as  male; one because the date of last contact was missing;
.1,275 subjects because the follow-up period was <5 years; 41  because the number of days
 worked was 0 or missing. After these  subjects were removed, the cohort consisted of 5,173 men
 who were employed at Wittenoom Gorge between 1943 and 1966.
                                        'A.I 9

-------
The concentrations of fibers greater than 5 [im in length as measured by PCM were measured at
various work sites in a survey conducted in 1966.  Job category data were obtained from
employment records and supplemented by records from the Perth Chest Clinic and the Western
Australian Mineworkers Relief Fund. The concentration measurements and job category
information were used to estimate the exposure level for each subject in the cohort (de Klerk
et al. 1994). The exposure levels were high with a median of 17.8 (fiber/ml). The durations of
employment were low with a median of 1 28 days.

There were 251 lung cancer deaths in the cohort. Table A- 12 shows the observed, expected, and
predicted lung cancer deaths among the males categorized by cumulative exposure (fiber-
year/ml). The number of expected lung cancer deaths are based on Australian lung cancer
mortality rates. With no allowance for difference between the background lung cancer death
rates among Australia and the members of this cohort (a=l), the fit of the model is poor
(p<0.01). Allowing for difference in the background lung cancer death rates (a variable), the
model provides a reasonably good fit to the data (p=0. 10) and estimates KL=0.0047 (fiber-
year/ml)-1, 90% CI: (0.0017, 0.0087). The hypothesis cc=l can be rejected with high confidence
There were 165 mesotheliomas in the cohort. The availability of the raw data from this study
made it possible calculate KM from this study using an "exact" likelihood approach based on
Equation A-3 that did not involve any grouping of data, or use of average values. With this
approach, KM=7.95xlO-8, 90% CI: (6.97xlQ-8, 9.01x1 0'8).

For comparison purposes, KM were also calculated using grouped data and applying Equation
A-43 since this is the method that must be used with most studies. The KM estimate based on the
"exact" analysis was 12% lower than the estimate based upon grouped data.

Regarding uncertainty factors, Fl is assigned a value 2.0 for this study for reasons similar to
those described for Quebec. F2 is assigned a value of 1.0 because measurements were conducted
using PCM so that conversion is unnecessary for this study. All other factors are also assigned a
value of 1.0 because there are no other unique limitations. Thus:

       Fl = 2.0
       F2 = 1.0
       F3 = 1.0
       F4L = 1.0
       F4M=1.0

These uncertainty factors described earlier, when coupled with the statistical confidence limits,
resulted in the uncertainty ranges for KL and KM and shown in Table A-l and A-2.

Predominant Amosite Exposure

Patterson, N.J. Insulation Factory. Seidman et al. (1986) studied  a cohort of 820 men (mostly
white) who worked at an amosite asbestos factory that operated in Patterson, New Jersey from
1941 through 1954. The men began work between 1941 and 1945 and follow-up was through
1982.  The follow-up of a worker began 5 years following the beginning of employment.

                                         A. 20

-------
Workers who had prior asbestos exposure .were not included in the cohort, and followfup was
stopped when a worker was known to have begun asbestos work elsewhere (6 men). Exposures
were generally brief, as 76% were exposed for <,2 years, although a few were exposed for as
long as 10 years.
                                                                            i
No asbestos exposure measurements are available for this plant. Estimates of exposures in
particular jobs were made based on air measurements made between 1967 and 1970 at plants in
Tyler, Texas and Port Allegheny, Pennsylvania that were operated by the same company and
made the same products using some of the same machinery as the Patterson facility. The
estimated median exposure level was 50 C'ml. Amosite was the only type of asbestos used at the
plant.

Seidman et al. cross-categorized lung cancer deaths by cumulative exposure (eight categories of
f-y/ml) and length of time worked (seven categories, Seidman et al.  1986, Table XXXIV).
Although this table apparently was created by categorizing workers by their final cumulative
exposure (rather than categorizing person-years of follow-up by the cumulative exposure to that
point in time, which is more appropriate for calculating a KL), because exposures were brief this
likely made little difference. Expected number of lung cancer deaths were based on age- and
year-specific rates for New Jersey white males.

Table A-13 shows the results of applying the lung cancer model to these data, after collapsing
the table by summing over length-of-time worked.  Results were highly dependent upon whether
or not the background lung cancer mortality rate was assumed to be equal to that predicted by
the comparison population of New Jersey white males (equivalent to a=l). The test for
departure from the null hypothesis, ct=l, was highly significant, and the maximum likelihood
estimate  was a=3.3. Similarly, the model gave a poor overall fit to the data with a=l (p<0.01),
but the fit was quite good when a was allowed to vary (p=0.90). The estimated potency
parameter, KL, also was highly dependent upon the assumption regarding the parameter, a.  The
estimate  of FQ, was 0.062 (f-y/ml)*1, 90% CI: (0.050, 0.076), when a was fixed at a=l, and
0.011 (f-y/ml)-J, 90% CI: (0.0058, 0.019), when a was allowed to vary, a 6-fold difference.  The
lung cancer model was also  fit to the data cross-classified by both cumulative exposure/and
length of time worked, allowing a to assume a different value in each category of time worked.
Although the estimated values of a tended to increase with increasing duration of exposure,
allowing different values of a did not significantly  improve the fit (p=0.64).

The reason for this behavior is not clear. There is no indication that workers with shorter
durations experienced disproportionately high mortality, since, as noted above, a tended to
increase with increasing duration of exposure. Although it is possible that cumulative exposure
is not the appropriate exposure metric, it is difficult to envision what metric would predict this
response, so long as a linear model is assumed. It is also possible that a linear model for relative
risk is not correct and a supralinear model is more appropriate, or. that the increased risk is not
proportional to the background risk, as assumed by this simple relative risk model.  Finally, it is
possible that the background rate in this population is significantly greater than that in the
comparison population, although it seems unlikely  that it could be 3 times greater as suggested
by the model.
                                         A.21

-------
Seidman et al. (1986) discovered 17 deaths from mesothelioma in this population.  Table III of
Seidman et al. categorized mesothelioma deaths and person-years of observation by years since
onset of work.  In order to apply the 1986 U.S. EPA mesothelioma model it is necessary to have
estimates of the duration of exposure and level of exposure for each category. Using the
categorization of the members of the cohort by duration of work in Table XXIII of Seidman
et al., it was estimated that the mean duration of work was 1.5 years. Using data from Seidman
et al. Table XIV, an average cumulative exposure was for each category of time from onset of
exposure by weighting exposures according to the expected total number of deaths. These
averages were divided by 1.5 years to obtain the average fiber concentrations in Table A-14.
The estimated exposure levels decrease with time since onset, which is consistent with higher
mortality among more heavily exposed workers.

The 1986 mesothelioma model provided an adequate fit to these data (p=0.35), although it over-
predicted somewhat the number of cases in the highest latency category (>35 years). The
estimate of KM was 3.9x10-*, 90% CI: (2.6X10'8, 5.7xlQ-8).

Regarding uncertainty factors, Fl is assigned a value of 3.5 for this study because exposure
concentrations were not measured at this facility  at all. Rather'exposures were estimated (as
described in Lemon et al. [1980]) based on measurements collected at another facility in Tyler,
Texas (see below) that manufactured the same products from the same source of raw materials
using some of the same equipment, which was moved from the Patterson plant to the plant in
Tyler.  Because the measurements collected in Tyler were analyzed by PCM, no conversion
factor is required. Thus, F2 is assigned a value of 1.0 for this study. All other factors are also
assigned a factor of 1.0 due to lack of other remarkable limitations.  Thus:

       Fl = 3.5
       F2=1.0
       F3=LO
       F4L = 1.0
       F4M=1.0

These uncertainty factors, when coupled with the statistical confidence limits, resulted in the
uncertainty ranges for KL and KM shown in Table A-l and A-2.

Tyler, Texas Insulation Factory. Levin et al. (1998) studied the mortality experience of 1,121
men who formerly worked at a plant in Tyler, Texas that manufactured asbestos pipe insulation.
The plant operated from 1954 through February 1972. The plant used the same raw materials
and some of the same equipment that was used in the Patterson, New Jersey plant that was
studied by Seidman et al. (1986).  The asbestos used was amosite from the Transvaal region of
South Africa. The insulation was manufactured from a mixture that contained 90% amosite
asbestos.

Environmental surveys were conducted at the plant in 1967,1970, and 1971, with average fiber
concentrations ranging from 15.9 through 91.4 f/ml.  An average exposure of 45 0ml is assumed
for this plant, which is near the middle of this range obtained in the three surveys.  It is also
consistent with average levels assumed for the Patterson, New Jersey plant, which operated
under very similar conditions.

                                         A.22

-------
 The cohort consisted of 744 whites, 305 non-white (mostly black), and 72 with missing race
 (assumed to be white, based on hiring practices at that time). For the entire cohort, the'median
 age of first employment was 25 years, and the mean duration of employment was 12.7 months
 (range of one day to 17.3 years). Follow-up was through 1993. Death certificates were obtained
 for 304 of the 315 men known to be dead. In the mortality analysis only white men were
 evaluated and follow-up started 10 years after first employment. After additional exclusions of
 men with missing birth dates or missing employment information,, the cohort analyzed in the
 mortality analysis consisted of 753 former workers, among whom 222 deaths were recorded.
 These deaths were compared with those expected based on age, race and sex-specific U.S. rates.

 There was an excess of deaths from respiratory cancer (SMR=277, based on 36 deaths; not
 including four deaths from mesothelioma). Table A-15 contains observed and expected numbers
 of deaths from respiratory cancer, categorized by duration of exposure. Cumulative exposure in
 f-y/ml was  estimated by multiplying the duration of exposure times the assumed average fiber
 level of 45  f/ml.  There was an excess of lung cancer deaths  in the lowest exposure group
 (23 observed, 8.9 expected), and consequently the model with a=l did not fit these data
 (p<0.01), and the hypothesis a=l could be rejected (p<0.01). The KL with a variable was
 KL=0.0013, 90% CI: (0, 0.0060). With o=l, K^O.013 (f-y/ml)-1, 90% CI: (0.0055, 0.022).

 Four mesotheliomas were reported in this study.  However, the data are not presented in a form
 that would permit application of the U. S. EPA 1986 mesothelioma model.

 Regarding uncertainty factors, Fl is assigned a value of 3.0 for this study because, although
 exposure concentrations were measured at this facility, the data are sparse so that only'.an overall
 average concentration for the entire plant could be derived.  Because the measurements collected
 were analyzed by PCM, no conversion factor is required. Thus, F2 is assigned a valu6 of 1.0 for
 this study.  All other factors are also assigned a factor of 1.0 due to lack of other remarkable
 limitations. Thus:                                                            >•
                                                                            i',.
      Fl=3.0
      F2=1.0
      F3 = 1.0
      F4L=1.0                                                           r

These uncertainty factors, when coupled with the statistical confidence limits, resulted in the
uncertainty  range for KL shown in Table A-l.

Predominant Tremolite-Actinolite Exposure                                   :

Libby, Montana Vermiculite Mine. Amandus and Wheeler (1987) conducted a retrospective
cohort study of 575 men who were exposed to tremolite-actinolite while working at a'Vermiculite
mine and mill in Libby, Montana.  A dry mill began operation in 1935 and a wet millibegan
operating in the same building as the dry mill in 1950 (Amandiis et al. 1987).
                                                                           .i
A total of 376 impinger samples were available that had been collected during 1950-1969,
although only 40 of these were collected prior to 1965. In addition 4,118 PCM samples were
available from the period 1967-1982. Exposure estimates for years later than 1968 were based

                                        A.23

-------
on historical measures of fiber concentrations (f/ml), and those for earlier years were based on
concentrations measured by midget impinger (mppcf) and converted to f/ml assuming a
conversion ratio of 4 f/ml per mppcf. This conversion factor was derived from 336 impinger
samples collected during 1965-1969 and 81 filter samples collected during 1967-1971.
individual cumulative fiber exposure estimates (f-y/ml) were computed from job-specific
exposure estimates and work histories (Amandus et al. 1987).

The cohort consisted of all men hired prior to 1970 and employed for at least 1 year in either the
mine or the mill. Follow-up was through December 31,1981.  The vital statuses of 569 of the
men (99%) were determined and death certificates were obtained for 159 of the 161 who were
deceased.

Smoking information was available for 161 men employed between 1975 and 1982 and with at
least 5 years of tenure. The proportion of these workers who smoked (current or former) was
84% compared to 67% among U.S. white males during the same time period.

A total of 20 deaths from lung cancer were observed (9 expected, SMR=223.2, using U.S. white
males as the comparison population).  Table A-16  (based on Amandus  and Wheeler 1987, Table
II) shows that the excess occurred mainly in workers whose cumulative exposure exceeded 400
f-y/ml (10 observed, 1.7 expected). The 1986 U.S. EPA lung cancer model fit these data
adequately (p^O.25) both with «=1  and a variable, and the hypothesis a=l could not be rejected
(p=0.8). With o=l, KL was estimated as 0.0061 (f-y/ml)-1, 90% CI: (0.0029,0.010), and with a
variable, KL= 0.0051 (f-y/ml)-1, 90% CI: (0.0011, 0.020).

Amandus and Wheeler (1987) observed 2 deaths from mesothelioma in this cohort. However,
information on these cases was not sufficient to permit application of the 1986 U.S. EPA
mesothelioma mo4el.

For uncertainty factors, Fl is assigned a value of 2.0 for reasons similar to those described for
Quebec. F2 is assigned a value of 1.5 because most early measurements were collected by
midget impinger and the authors report using a conversion factor of 4 derived from temporally
overlapping (but not paired) measurements.  All other uncertainty factors were assigned a value
of 1.0. Thus:

       Fl  = 2.0
       F2 = 1.5
       F3 = 1.0
       F4L = 1.0

These uncertainty  factors, when coupled with the statistical confidence limits, resulted in the
uncertainty range for KL shown in Table A-l.

McDonald et al. (1986) also conducted a cohort study of workers at the Libby, Montana
vermiculite mine and mill. Their cohort was composed of 406 workers employed prior to  1963
for at least 1 year.  Follow-up was until July 1983. Vital status was determined for all but one
man and death certificates were obtained for 163 of the 165 men who had died.  Cumulative
exposures (f-y/ml)-were estimated for each worker using work histories based on 42 job

                                         A.24

-------
 categories, and 1,363 environmental measurements, including samples analyzed by PGM (f/ml)
 and by midget impinger (mppcf).

 A total of 23 deaths from lung cancer were observed (SMR=303, based on Montana rates).
 Table A-17 shows these deaths categorized by cumulative exposure (based on Table 4 of
 McDonald et al. 1986). Both the models with a=l and a variable fit these data adequately
 (p^0.16) although the test of ot=l was marginally significant (p=0.11). The estimate of KL with
 a=l was 0.011, (f-y/ml)-1, 90% CI: (0.0055, 0.017), and with a variable, KL= 0.0039 (f-y/ml)-1,
 90% CI: (0.00067, 0.012).

 McDonald et al. (1986) observed 2 deaths from mesothelioma. 'However, information1 on these
 cases was not sufficient to permit application of the 1986 U.S. EPA mesothelioma model.
                                                                            r '
 Because this study and the Amandus study used virtually the same data and very similar
 approaches to analysis, the same values are assigned to uncertainty factors for this study that are
 assigned for the Amandus and Wheeler study. These factors, when coupled with the statistical
 confidence limits, resulted in the uncertainty range for KL shown in Table A-l.
                                                                            f
 Exposure to Mixed Fiber Types
                                                                            i1
 British Friction Products Factory. Berry and Newhouse (1983) conducted a mortality study of
 13,460 workers in a factory in Britain that manufactured brake blocks, brake and clutch linings,
 and other friction materials.  Only chrysotile was used at the plant except for two relatively short
 periods before 1945 when crocidolite was used in the production of railway blocks.   ,
                                                                            M
 The cohort studied consisted of all men or women employed at the plant between 1941 and 1977.
 Follow-up was to the end of 1979 and the mortality experience was examined  after 10 years
 from first exposure.  Airborne dust measurements were only available from 1967 onward and
 these were made using the PCM method. Fiber concentrations in earlier years were estimated by
 reproducing earlier working conditions using knowledge of when processes were changed .and
 exhaust ventilation introduced.
                                                                            •i
 Deaths from all causes were less than expected both prior to 10;years from first employment
 (185 observed versus 195.7 expected) and afterward (432 observed versus 450.8 expected).
 There was no indication of an effect of employment at the plant upon lung cancer; there were 51
 lung cancers >10 years from first employment compared to 47.4 expected. A significant deficit
 of gastrointestinal cancers was observed after 10 years from first employment  (25 observed
'versus 35.8 expected, p-0.04).

 A linear exposure response model relating cumulative exposure and lung cancer was fit to case-
 control data presented by Berry and Newhouse. The resulting KL was 0.00058 (f-y/ml)"1 and the
 95% upper limit was 0.0080 (f-y/ml)*1. This estimate was used as the best estimate of KL, and
 the lower confidence bound was assumed to be zero.

 A case control study on mesothelioma deaths showed that 8 of the 11 cases had been exposed to
 crocidolite and another possibly had intermittent exposure to crocidolite. The other two had
 been employed mostly outside the factory and possibly had other occupational exposures to

                                         A. 25

-------
asbestos. The case control analysis showed that the distribution of cases and controls in respect
to exposure to crocidolite was quite unlikely assuming no association with crocidolite.  This
indicates that some, and possibly all, of the eight mesotheliomas with crocidolite exposure were
related to this exposure. The data were not presented in a form that permitted a quantitative
estimate of mesothelioma risk.

Regarding uncertainty factors, Fl is assigned a value of 2.0 for this study because, although the
manner in which unmeasured exposure was estimated in this study is different than for that
reported for the majority of other studies (see, for example, Quebec), it is unlikely to introduce
greater uncertainty. Rather than extrapolating measured estimates to earlier times based on
expert judgements, judgements were used to simulate earlier conditions at the plant and
exposures were measured directly. Because the measurements collected were analyzed by PCM,
no conversion factor is required. Thus, F2 is assigned a value of 1.0 for this study. An
uncertainty factor F4L-1.5.was included to account for the fact that a was not estimated. F3 was
assigned a factor of 1.0 due to lack of other remarkable limitations. Thus:

       Fl = 2.0
       F2 = 1.0
       F3 = 1.0
       F4L=1.5

These factors, when coupled with the statistical confidence limits, resulted in the uncertainty
range for KL shown in Table A-l.

Ontario Asbestos-Cement Plant. Finkelstein (1984) studied mortality among a group of 535
exposed and 205 unexposed employees of an Ontario asbestos-cement factory who had been
hired before 1960 and who had been employed for at least 1 year. This cohort contained the
cohort studied by Finkelstein (1983) and which required at least 9 years of employment for
membership. Follow-up continued until  1977 or 1981.

The plant produced asbestos cement pipe from 1948, asbestos cement board from 1955-1970,
and manufacture of asbestos insulation materials was added in 1960.  Both chrysotile and
crocidolite were used in each batch processed in the pipe process, but only chrysotile was used in
the cement board operation. Crocidolite constituted approximately 20% of the asbestos used in
the pipe process (Ontario Royal Commission 1984).

Fiber concentrations in various work areas and for various epochs were estimated from
membrane filter samples taken after 1969, impinger measurements taken during 1949,1954,
1956,1957, and semiannually during the 1960s, and information on changes in dust control
methods. Finkelstein judged that the resulting exposure estimates were "probably accurate to
within a factor of three or five."  Exposures of maintenance workers were not estimated, and the
exposure response analysis consequently involved only the unexposed workers (N=205) and the
production workers (N=428).

Only 21 deaths from lung cancer were observed among production workers.  Based on these
deaths, Finkelstein compared age-standardized lung cancer mortality rates in production workers
after a 20-year latency, categorized into five groups according to their cumulative exposure

                                          A.26

-------
 through 18 years from date of first employment (Finkelstein 1984, Table 7). Mortality rates
 were standardized with respect to age and latency using the man-years distribution in the cohort
 as a whole as the standard. Using similarly standardized mortality rates in Ontario males as the
 comparison population, lung cancer rates were elevated in all five categories, and Finkelstein
 found a significant exposure-response trend. However, the trend was not monotone, as rates
 increased up to the middle exposure category and decreased thereafter (Table A-18).

 These data may be put into a form roughly equivalent to the more conventional age-adjusted
 comparison of observed and expected lung cancer deaths by dividing the rates in the exposed
 group by that of Ontario men. (The rate for unexposed workers was not used because^! was
 based on only 3 deaths.) The results of this are shown in Table A-l 8, which also shows the
 results of fitting the 1986 U.S. EPA lung cancer model both  assuming the Ontario rates were
 appropriate for this cohort (fixing the parameter a=l) and not making this assumption!(allowing
 the parameter a to vary).  Neither approach provided an adequate fit to these data (p^O.05) and
 the test of a=l was marginally significant (p=0.07). The maximum likelihood estimate of a was
 4.26, which seems too large to be due to differences in smoking habits. The KL estimate with
 ct=l was 0.048 [f-y/ral]*1,  90% CI: (0.028, 0.074). With a allowed to vary the estimate was
 KL=0.0029 [f-y/ml]'1, 90% CI: (0, 0.037). The fact that the lower limit was zero indicates that
 the exposure-response trend was not significant when the background was allowed to vary.

 Based on a "best evidence" classification of cause of death, Finkelstein identified 17 deaths from
 mesothelioma among production workers. Table 3 of Finkelstein (1984) gives these  ,
 mesotheliomas categorized by years since first exposure. This table also provides the mortality
 rate, from which can be calculated the person-years of observation. Finkelstein states ithat the
 average cumulative exposure for production workers was about 60 f-y/ml, but does not provide
 information for determining duration and level of exposure separately. CHAP (1983)>used an
 average exposure  of 9 f/ml for a subcohort of production workers, although they provided no
 support for this assumption. If this value is assumed to be appropriate for the expanded cohort,
 the average duration is estimated as about 60/9=6.7 years.  However these values are uncertain.
 Table A-19 presents the result of applying the 1986 U.S. EPA niesothelioma model tojthe
 Finkelstein (1984) data based on these assumptions. The mesothelioma model describes these
 data adequately (p=0.26) and provides an estimate of KM=18xlO'8, 90% CI: (13x10-*, 24xlO'8).
Regarding uncertainty, Fl is assigned a value of 4 because Finkelstein indicates that exposure
estimates derived for this study are probably good to within a factor of 3 or 4. Findlestein also
notes that many of the assumptions employed to extrapolate exposures were only weakly
supported by limited, earlier impinger measurements. The source of the conversion factor
employed to link impinger measurements and PCM measurements in this study is unclear.
Therefore a value of 3.0 is assigned to F2. Because data for evaluating mesothelioma incidence
was not provided in a format suitable for deriving confidence intervals, so that some
reconstruction was required, a value of 2.0 is assigned for F4M. All other factors are also
assigned a factor of 1.0 due to lack of other remarkable limitations.  Thus:
                                         A.27

-------
      Fl = 4.0
      F2 = 3.0
      F3 = 1.0
      F4L = 1.0
      F4M = 2.0

These factors, when coupled with the statistical confidence limits, resulted in the uncertainty
ranges for KL and KM shown in Tables A-l and A-2.

Swedish Asbestos-Cement Plant. Albin et al, (1990) studied workers at a Swedish plant that
operated from 1907 to 1978 and produced various asbestos cement products, including sheets,
shingles, and ventilation pipes. The asbestos handled was mainly chrysotile (>95%).
Crocidolite was used before 1966, but never exceeded 3-4% of the total asbestos. Amosite was
used for a few years in the 1950s but never exceeded 18% of the total asbestos used.  Fiber
length classes were the commercial grades 3-7, and all asbestos was milled prior to
incorporation into products.

Impinger and gravimetric dust measurements were available for 1956-1969,  and PCM
measurements after 1969. These data, along with information on production and dust control,
were used to estimate exposures for different jobs and periods of time.

The cohort contained 2,898 men and was defined as all male employees who worked for at least
3 months between 1907 and 1977.  A reference cohort was composed of 1,233 men who worked
in other industries in the region and who were not known to have worked with asbestos.  Vital
status of both groups was determined through 1986. Follow-up of both began after 20 years
from first employment.

Excluding mesothelioma, other respiratory cancers were not significantly increased.  Albin et al.
present relative risks of these respiratory cancers and corresponding 95% CIs for three categories
of cumulative exposure (Table A-20), based on Poisson regression with control for age and
calendar year.  In order to obtain crude estimates of the range of KL that are consistent with these
data, the 1986 U.S. EPA lung cancer relative risk model  was fit, assuming that the Ln (RR) were
normally distributed with fixed variances computed from the reported confidence intervals for
the RR. Although elevated, the RR did not exhibit an exposure response, and the hypothesis a=l
was not rejected (p=0.13).  In this analysis K^  was not significantly different from zero,
regardless of whether a was fixed at 1.0 or estimated.  With ct=l the estimate of KL was 0.019
(f-y/ml)-1, 90% CI:  (0, 0.065), and KL=0.00067 (f-y/ml)-1, 90% CI: (0, 0.036) with a estimated.

Thirteen mesotheliomas were identified among exposed  workers and one in the referent
population, and a significant exposure response was observed with increasing cumulative
exposure. Unfortunately, the mesothelioma data were not presented in a format that would
permit application of the 1986 U. S. EPA mesothelioma model.

Regarding uncertainty, Fl is assigned a value of 4 due to the sparsity of data and the need to
extrapolate. Several assumptions were incorporated into the extrapolations performed that were
                                         A.28

-------
based, among other things, on the scarcity of raw-material asbestos during World War II. All
other factors are also assigned a factor of 1 .0 due to lack of other remarkable limitations.  Thus:

       Fl = 4.0                                                              !;
       F2=1.0                                                              i'
       F3 = 1.0                           .                                   I
       F4L=1.0                                         '                    ;'

These factors, when coupled with the statistical confidence limits, resulted in the uncertainty
range for KL shown in Table A-l .                                               ;

Belgium Asbestos-Cement Plant  Lacquet et a). (1980) conducted a roentgenologic, asbestosis,
and mortality study in a Belgium asbestos cement factory employing about 2,400 employees that
annually processed about 39,000 tons of asbestos, of which 90% was chrysotile, 8% crocidolite,
and 2% amosite. The mortality study considered male workers who worked in the factory for at
least 12 months during the 15-year period 1963-1977. Apparently no minimal latency was
required before follow-up began.

Fiber counts were available for the years 1970-1976; fiber levels were estimated for as far back
as 1928, but these estimates were considered to be "only good guesses at best." Individual
exposures were estimated in fiber-years from work histories and estimated yearly concentrations
in four work areas.                                                            ]

The incidence of respiratory cancer was very close to that which was expected in a Belgium
population of matched age and sex (Table A-21). The models with a=l (p=0.51) and a variable
(p=0.39) gave similar results and the hypothesis 
-------
Retirees from U.S. Asbestos Products Company.  Enterline et al. (1986) extended follow-up
through 1980 for a cohort of U.S. retirees from a large asbestos products company that had been
the subject of an earlier report (Henderson and Enterline 1979).  Products manufactured by the
company included textiles, cement shingles, sheets, insulation and cement pipe.  Exposure was
predominately  to chrysotile in most operations, although amosite predominated in insulation
production, and crocidolite in manufacture of cement pipe. Each worker's exposure was
estimated from dust measurements in mppcf obtained from environmental surveys that started in
the mid-1950's and were extrapolated back in time by the company industrial hygienist. No data
are provided for conversion from mppcf to PCM in f/ml. Given the wide range of products
manufactured,  this conversion likely varied according to operation. Conversions calculated in
different environments have ranged from 1.4 to 10, the most common value being around 3 f/ml
per mppcf, which has been observed in diverse environments such as mining and textile
manufacture. This value was provisionally applied to this cohort.

The cohort consisted of 1,074 white males who retired from the company during 1941-1967, and
who were exposed to asbestos in production or maintenance jobs. The average duration of
employment was 25 years.  Follow-up started at age 65 or at retirement if work continued past
age 65.  By the end of follow-up in 1980, 88% were deceased.

Overall, respiratory cancer was significantly increased (SMR=258 in comparison to U.S. rates,
based on 79 observed deaths). Enterline et al. (1986) categorized lung cancer deaths by
cumulative exposure (their Table 4).  Results of applying the 1986 U.S. EPA lung cancer model
to these data are shown in Table A-22. Although both the model with a=l and'a variable fit the
data adequately (p^0.75), the test of a=l was not rejected (p=0.24).  With a=l the estimate of K^
was 0.0021 (f-y/ml)-1, 90% CI: (0.0015, 0.0027).  With a variable, KL=0.0011  (f-y/ml)'1; 90% CI:
(0.00041,  0.0028).

From the death certificates Enterline et al. identified eight deaths from mesothelioma. These
data were not presented in a form that permitted application of the 1986 U.S. EPA mesothelioma
model.

Regarding uncertainty., Fl is assigned a value of 2.0 for this study for reasons  similar to those
described for Quebec. Because the manner employed for deriving the conversion factor used to
convert impinger counts to fiber concentrations is not documented, a value of 3.0 is.assigned to
F2 for this study. All other factors are also assigned a factor of 1.0 due to lack of other
remarkable limitations. Thus:

       Fl=2.0
       F2 = 3.0
       F3 = 1.0
       F4L = 1.0

These factors,  when coupled with the statistical confidence limits, resulted in the uncertainty
range for KL shown in Table A-l.
                                         A.3O

-------
 U.S. Insulation Applicators. Selikoff and Seidman (1991) reported on follow-up through 1986
 of a cohort of 17,800 asbestos insulation applicators that had been followed through 1976 by
 Selikoff et al. (1979).  The cohort consisted of men enrolled as members of the insulator's union
 in the United States and Canada.  Deaths were classified both based on the information the death
 certificate, and using "best evidence," in which death certificate information was augmented by
 clinical data, histopathological material and X-rays.                              ''

 Based on the composition of insulation material, it seems likely that these workers were exposed
 to substantial amounts of chrysotile and amosite.  Data on insulator's exposures were reviewed
 by Nicholson (1976), who concluded that average exposures of insulation workers in past years
 could have ranged 10-15 f/ml and could have been 15-20 f/ml in marine construction! U.S.
 EPA (1986) assumed a value of 15 f/ml as an overall average, with an associated 3-fojd
 uncertainty. This estimate of 15 f/ml will be used provisionally here as well.

 The form of the data provided in Selikoff and Seidman (1991) is not particularly suitable for
 calculating KL.  Table 4 of Selikoff and Seidman (1991) contain observed and expected deaths
 from lung cancer (determined from either death certificates or best information) categorized by
 years from first exposure  (<15,15-19,20-24,..., 50+). Death certificate informationVas
 utilized herein to facilitate comparisons with expected deaths (based on the mortality experience
 of U.S. white males), which were also based on death certificates.  Lung cancer was significantly
 increased over expected, except for the category of <15 years from onset of exposure. Selikoff
 and Seidman did not provide information on the duration of exposure. The U.S. EPA (1986,
 page 90) assumed an average exposure duration of 25 years. Assuming that all workers worked
 exactly 25 years and were exposed to 15 f/ml, the data in Table 4 of Selikoff and Seidman can be
 used to categorize lung cancer deaths by cumulative exposure lagged 10 years. The result is
 shown in Table A-23.  The 1986 U.S. EPA lung cancer model provided a reasonable fit to these
 data with a variable (p=0.12),  but not with a=l (p<0.01). Also, the hypothesis that a=l could be
 rejected (p<0.01). The estimate of KL with a variable was 0.00.18 (f-y/ml)-1, 90% €1:1(0.00065,
 0.0038).  With o=l, 1^=0.0087 (f-y/ml)'1, 90% CI: (0.0081, 0.0093).               \

 Based on best evidence, Selikoff and Seidman (1991) found 458 mesotheliomas in this cohort.
 Table A-24 shows these deaths categorized by years from onset (based on Selikoff arid Seidman
 1991, Tables 5 and 6). Table A-24 also shows the results of fitting the 1986 U.S. EPA
 mesolhelioma model to these data, assuming, as above, that workers worked for 25 years and
 were exposed to 15 f/ml.  The  1986 U.S. EPA mesothelioma model provided a poor fit to these
 data (p<0.01), as it overestimates by more than a factor of 2 the number of mesothelioma deaths
 after 50+ years from first exposure.  The estimate of KM was 1.3x10'8, 90% CI: (1.2x10'8,
 1.4x10-*).

 Regarding uncertainty, Fl is assigned a value of 4.0 for this study because data employed to
 estimate exposure is not facility-specific, but represents general, industry-wide exposure
 estimates derived from limited data. F3 is assigned a value  of 2 for this study because the study
 provides no information on worker histories. F4L is assigned a value of 2 for this study because
the data presented in the study were not provided in a form suitable for fitting the lung cancer
model. Thus, the data had to be partially reconstructed. Other factors are also assigned a factor
 of 1.0 due to lack of other remarkable limitations. Thus:
                                         A.31

-------
       Fl=4.0
       F2 = 1.0
       ¥3 = 2.0            -                                  '        .
       F4L = 2.0
       F4M = 1.0

These factors, when coupled with the statistical confidence limits, resulted in the uncertainty
ranges for KL and K.M shown in Tables A-l and A-2.                      ,                ;

Pennsylvania Textile Plant. McDonald et al. (1983b) report on mortality in an asbestos plant
located near Lancaster, Pennsylvania that produced mainly textiles, but also some friction
materials. About 3,000 to 6,000 tons of chrysotile were processed annually at the plant, which
began operation in the early 1900s. Crocidolite and amosite were used from 1924 onward; about
3-5 tons of raw crocidolite were processed annually and the use of amosite reached a peak of
600 tons during World War II.

The cohort consisted of all men employed for at least 1 month prior to 1959 and who had a valid
record with the Social Security Administration. This group consisted of 4,022 men, of whom
35% had died by the end of follow-up (December 31,1977). Follow-up of each worker was only
begun past 20 years from first employment.

To estimate exposures, McDonald et al. had available reports of surveys conducted by the
Metropolitan Life Insurance Company during the period 1930-1939, Public Health Service
surveys conducted during 1967 and 1970, and company measurements made routinely from
1956 onward. These data were used to estimate by department and year in units of mppcf.

The lung cancer mortality in this cohort exhibited a significant exposure response trend (Table
A-25), which was partially due to a deficit of cancers in the group exposed to <10 mppcf-y (21
with 31.4 expected). A survey of those employed in the plant in 1978 revealed a larger per cent
of nonsmokers (25%) man were found in the other plants studied by these researchers
(McDonald et al. 1983a, 1984), although this finding was based on a sample of only 36 workers.
Regardless of the reason for this shortfall in the number of lung cancers, it appears that the most
appropriate analysis is that in which the background is  allowed to vary; this analysis fits the data
well (p>0.7), whereas the analysis which assumes the Pennsylvania rates are appropriate
provides a marginal fit (p=0.08). The hypothesis a=l was rejected (p^O.Ol). Consequently, the
former analysis is judged to be the most appropriate (allowing the parameter a to vary).
McDonald et al. (1983b) did not provide  a factor for converting from mppcf to f7ml. Assuming
that 3 fml is equivalent to one mppcf, the resulting estimate of lung cancer potency with a
variable was 0.018 (f-y/ml)'1, 90% CI: (0.0075, 0.045). With «=1, KL=0.0057 (f-y/ml)'1, 90%
CI: (0.0027, 0.0094).

A diagnosis of mesothelioma was specified on 14 death certificates (ten pleura! and four
peritoneal). Thirty other deaths were given the ICD code 199 (malignant neoplasms of other and
unspecified sites) and the diagnosis given in many of these cases was said to be consistent with
an unrecognized mesothelioma. McDonald et al. (1983b) Table 3 lists the average age at
beginning of employment as 28.92 and the  average duration of employment as 9.18 years, and
their Table  1 lists 191, 667, and 534 deaths as occurring before age 45, between 45 and 65, and

                                         A.32

-------
after 65 years of age, respectively. Assuming that 1A of the deaths given the ICD code 199 might
have been due to mesotheliomas, the total number of mesotheliomas in this cohort is estimated to
be 23. Proceeding as in the mesothelioma analysis carried out for the McDonald et al.;1 (1984)
data, the data in Table A-26 were generated.  Noting that the age since first exposure categories
in which the mesotheliomas occurred is irrelevant as far as estimating KM is concerned, the
estimate of KM is l.lxKT8, 90% CI: (0.76x1 O'8, l.SxlO'8). These estimates are uncertain due to
the uncertainty regarding the number of mesotheliomas in the cohort.                '

Regarding uncertainty, Fl is assigned a value of 2.0 for mis study for reasons similar !to those
described for Quebec. Because the manner employed for deriving the conversion factor used to
convert impinger counts to fiber concentrations is not documented, a value of 3.0 is assigned to
F2 for this study.  A factor of 2.0 is assigned  for F4M because the number of mesotheliomas
observed in this study are reported to be estimates expected to be good to within a factor of 2.
Thus:                                                                        :

       Fl=2.0                                                               ,
       F2 = 3.0
       F3 = 1.0
       F4M-2.0                              '           '

These factors, when coupled with the statistical confidence limits, resulted in the uncertainty '
ranges for KL and KM shown in Tables A-l and A-2.                              i

Rochdale, England Textile Factory.  Peto et al. (1985) studied a textile factory in Rochdale,
England that has been the subject of a number of earlier reports (Peto et al. 1977; Peto 1980a,b).
Peto et al. (1985) has the most complete follow-up (through 1983) and emphasizes assessment of
risk. The factory, which began working with asbestos in 1879, used principally chrysotile, but
approximately 5% crocidolite was used between 1932 and 1968.                   ;

Quantitative estimates of risk were based on a subgroup of Peto et al. (1985) "principal cohort"
consisting of all men first employed in 1933 or later who had worked in scheduled areas or on
maintenance and had completed 5 years of service by the end of 1974. In the analyses of interest
relating to lung cancer, follow-up only begins 20 years after the beginning of employment and
exposure during the last 5 years of follow-up is not counted.

Routine sampling using a thermal precipitator began at 23 fixed sampling points in 1951.
Comparisons of particle counts and fiber counts taken in 1960 and 1961  were used tojconvert
between particles/ml and f/ml.  Dust levels prior to 1951 were assumed to be the same as those
observed during 1 95 1-1 955 for departments for which no major changes had been made.  In
departments in which conditions had improved, higher levels were assigned. These levels and
work histories were used to assign individual exposure estimates. A conversion factor of 34
particles/ml per r7ml was determined by comparing average results obtained by the Casella
thermal precipitator (particles/ml) with Ottway long running thermal precipitator (f/rril) at the
same sampling point during 1960 and 1961 . However, a conversion factor of 35.3 was used by
Peto et al. (1985) for the sake of consistency with earlier work, and this factor will be used here
as well.

                                         A. 33

-------
After 20 years from first employment, there were 93 lung.cancer deaths with only '64.6 expected.
Using a lung cancer model essentially the same as the 1986 U.S. EPA model, Peto et al.
estimated 1^=0,0054 (f-y/ml)"1 for the entire cohort, and KL=0.015 (f-y/ml)"1 when the analysis
was restricted to men first employed in 1951 or later. Peto et al. felt that the most plausible
explanation for this difference was that it was largely due to chance and also possibly to the
chance that exposure to the most carcinogenic fibers was not reduced as much as changes in
particle counts from 1951 to 1960 would suggest.

Table A-27 displays the exposure response data based on men first employed in 1933 or later for
lung cancer based on shows that the excess occurred mainly in workers whose cumulative
exposure exceeded 400 f-y/ml (10 observed, 1.7 expected). The 1986 U.S. EPA lung cancer
model fit these data adequately (p^O.63) both with a=l  and a variable, and the hypothesis a=l
could not be rejected (p=0.57).  With a=l, K^ was estimated as 0.0052 (f-y/ml)-1, 90% CI:
(0.0028, 0.0079), and with a variable, KL=0.0041 (f-y/ml)-1, 90% CI: (0.0012, 0.0087).

Ten mesotheliomas were observed in the cohort used by Peto et al. for quantitative analysis (an
11th case who was exposed for 4 months and died 4 years later was omitted because the short
latency made it unlikely that this case was related to exposure at the factory). Observed
mesotheliomas and corresponding person-years of observation by duration of service and years
since first employment (Peto et al. 1985, Table 8) are shown in Table A-28. An overall average
exposure was estimated by applying the Peto mesothelioma model to the data in Table A-28 with
a single exposure estimate selecting the value that gave the smallest least squares fit of this
model to the mesothelioma data. The fitting was carried out both unweighted and by weighting
by the person years, with resulting estimates of 360 and 322 particles/ml, respectively; the latter
value was the one selected. Using the conversion factor of 35,3 particles/ml per fi'ml, the
estimated average exposure is 322/35.2=9.1 f/ml. The 1986 U.S. EPA mesothelioma model fit
these data well, and the resulting estimate of mesothelioma potency (Table A-28) was
KM=1.3xlO-8, 90% CI: (0.74x10-8, 2.1xlO-8).

Regarding uncertainty, Fl is assigned a value of 2.0 for this study for reasons similar to those
described for Quebec.  Because a conversion factor was derived for measurements collected
using Otway long-running thermal precipitators and PCM measurements based on measurements
of each collected under similar conditions (but not side-by-side), a value of 2 is assigned to F2.
Thus:

       Fl = 2.0
       F2 = 2.0
       F3 =  1.0
       F4L=1.0
       F4M = 1.0

These factors, when coupled with the statistical confidence limits, resulted in the uncertainty
ranges for KL and KM shown in Tables A-l and A-2.
                                         A.34

-------
REFERENCES

Albin M; JaJcobsson K; Attewell R; Johansson L; Welinder H. Mortality and Cancer Morbidity
in Cohorts of Asbestos Cement Workers and Referents. British Journal of Industrial Medicine.
79(9):602-610.  September. 1990.                     '                      |.

Amandus HE; Wheeler R.  The Morbidity and Mortality of Vermiculite Miners and Millers
Exposed to Tremolite-Actinolite: Part II. Mortality. American Journal of Industrial Medicine
11:15-26. 1987.                                                          i:
                                                                         f
Amandus HE; Wheeler R; Jankovic J; Tucker J. The Morbidity and Mortality of Vermiculite
Miners and Millers Exposed to Tremolite-Actinolite: Part I. Exposure Estimates. American
Journal of Industrial Medicine.  11:1-14.  1987.
                                                                         i>
Armstrong BK; de Klerk NH; Musk AW; Hobbs MST. Mortality in Miners and Millers of
Crocidolite in Western Australia. British Journal of Industrial Medicine.  45:5-13.  1988.
                                                                         i

Berry G; Newhouse ML. Mortality of Workers Manufacturing Friction Materials Using
Asbestos. British Journal of Industrial Medicine.  40:1-7. 1983.                 [

CHAP (Chronic Hazard Advisory Panel on Asbestos). Report to the U.S. Consumer Product
Safety Commission. July.  1983.

Cox D; Oakes DV. Analysis of Survival Data. Cox DR; Hinkley DV (eds.). Chapman and Hall,
London.  1984.                                                            ;'.

de Klerk NH; Musk AW; Armstrong BK; Hobbs MST. Diseases in Miners and Millers of
Crocidolite from Wittenoom, Western Australia: A Further Follow-up to December 1986.
Annals of Occupational Hygiene. 38(Suppl 1): 647-655.  1994.                   !

Dement JM.  Estimation of Dose and Evaluation of Dose-Response in a Retrospective Cohort
Mortality Study of Chrysotile Asbestos  Textile Workers.  Ph.D. Thesis. The University of North
Carolina at Chapel Hill.  1980.

Dement JM; Brown DP.  Cohort Mortality and Case-Control Studies of White Male Chrysotile
Asbestos Textile Workers.  Journal of Clean Technology, Environmental Toxicologyt\and
Occupational Medicine.  7:1052-1062.  1998.

Dement JM; Harris RL; Symons  MJ; Shy CM. Estimates of Dose-Response for Respiratory
Cancer Among Chrysotile Asbestos Textile Workers. Annals Occupational Hygiene.
26(lr4):869-887. 1982.                                 ,                   "

Dement JM; Harris RL; Symons  MJ; Shy CM. Exposures and Mortality Among Chrysotile
Workers. Part I: Exposure Estimates. American Journal of Industrial Medicine.  4:399-419.
1983a.                                       '         !
                                       A.35

-------
   one   ' LaVecchia           •
ftolatto G; Negri P- r aV ,             " -^"rational
5^^a-^rs,^^

                       "™


               Sfflaa
       'T^00'^ Deafe -,


        -^^-^^^9-
          -^^
   ^-s^
»„! S— c^S-j,,™,,,

           A.38

-------
 McDonald JC; Liddell FDK; Gibbs GW; Eyssen GE; McDonald AD.  Dust Exposure and
 Mortality in Chrysotile Mining, 1910-1975. British Journal of Industrial Medicine. 37:11-24.
 1980a.                                                    i.
                                                          ' »        '       i
 McDonald JC; Gibbs GW; Liddell FDK. Chiysotile Fibre Concentration and Lung Cancer
 Mortality: A Preliminary Report.  In Biological Effects of Mineral Fibres.  Wagner JC (ed).
 IARC Scientific Publications, pp. 811-817. 1980b.

 McDonald AD; Fry JS; Wooley AJ; McDonald JC. Dust Exposure and Mortality in an
 American Chrysotile Textile Plant. British Journal of Industrial Medicine. 39:361-367. 1983a.

 McDonald AD; Fry JS; Woolley AJ; McDonald JC.  Dust Exposure and Mortality in an
 American Factory Using Chrysotile, Amosite, and Crocidolite in Mainly Textile Manufacture.
 British Journal of Industrial Medicine.  40:368-374. 1983b.

 McDonald AD; Fry JS; Woolley AJ; McDonald JC.  Dust Exposure and Mortality in an
 American Chrysotile Asbestos Friction Products Plant. British Journal of Industrial Medicine.
 41:151-157. 1984.

 McDonald JC; McDonald AD; Armstrong B; Sebastien P. Cohort Study of Mortality''of
 Vermiculite Miners Exposed to Tremolite.  British Journal of Industrial Medicine.  43:436-444.
 1986.
                                                                         i.

 McDonald JC; Liddell FDK; Dufresne A; McDonald AD.  The 1891-1920 Birth Cohort of
 Quebec Chrysotile Miners and Millers: Mortality 1976-1988. British Journal of Industrial
Medicine. 50:1073-1081.  1993.                                             '

Nicholson WJ. Part III.  Recent Approaches to the Control of Carcinogenic Exposures. Case
 Study 1: Asbestos - TheTLV Approach. Annals New York Academy of Science.  271:152-169.
 1976.                                                                   i,

Nicholson WJ; Selikoff IJ; Seidman H; Lilis R; Formby P. Long-Term Mortality Experience of
Chrysotile Miners and Millers in Thetford Mines, Quebec. Annals New York Academy of
Sciences. 330:11-21. 1979.          -                                    .

Ontario Royal Commission. Report of the Royal Commission on Matters of Health-and Safety ,
Arising form the Use of Asbestos  in Ontario. Volume 3.  1984.

Peto J. Lung Cancer Mortality in Relation to Measured Dust Levels in an Asbestos/Textile
Factory. In Biological Effects of Mineral Fibres. Wagner JC (ed.). IARC  Scientific
Publications, pp. 829-836. 1980a.

Peto J. The Incidence of Pleura! Mesothelioma in Chrysotile Asbestos Textile Workers. In
Biological Effects of Mineral Fibres. Wagner JC (ed.). IARC Scientific Publications,  pp.
703-711. 1980b.
                                       A.37

-------
Peto J; Doll R; Howard SV; Kinlen U; Lewinsohn, HC. A Mortality Study Among Workers in
an English Asbestos Factory. British Journal of Industrial Medicine. 34:169-173.  1977.

Peto J; Seidman H; SelikofFIJ.  Mesothelioma Mortality in Asbestos Workers: Implications for
Models of Carcinogenesis and Risk Assessment. British Journal of Cancer.  45:124-135.  1982.

Peto J; Doll R; Hermon C; Binns W; Clayton R; Goffe T.  Relationship of Mortality to Measures
of Environmental Asbestos Pollution in an Asbestos Textile Factory. Annals of Occupational
Hygiene. 29(3):305-355. 1985.

Piolatto G; Negri E; LaVecchia C; Pira E; Decarli A; Peto J. An Update of Cancer Mortality
Among Chrysotile Asbestos Miners in Balangero, Northern Italy. British Journal of Industrial
Medicine.  47:810-814.  1990.

Rubino GF; Piolatto GW; Newhouse ML; Scansetti G; Aresini GA; Murray R.  Mortality of
Chrysotile Asbestos Workers at the Balangero Mine, Northern Italy. British Journal of
Industrial Medicine.  36:187-194. 1979.

Seidman H. Short-Term Asbestos Work Exposure and Long-Term Observation - July 1984
Update. Department of Epidemiology, American Cancer Society.  1984.

Seidman H; Selikoff IJ; Hammond EC. Short-Term Asbestos Work Exposure and Long-Term
Observation. Annals New York Academy of Sciences.  330:61-89. 1979.

Seidman H; Selikoff IJ; Gelb SK. Mortality Experience of Amosite Asbestos Factory Workers:
Dose-Response Relationships 5 to 40 Years After Onset of Short-Term Work Exposure.
American Journal of Industrial Medicine. 10(5/6):479-514.  1986.

Selikoff IJ; Seidman H.  Asbestos-Associated Deaths among Insulation Workers in the United
States and Canada, 1967-1987. Annals of the New York Academy of Sciences.  643:1-14.  1991.

Selikoff IJ; Hammond EC; Seidman H. Mortality Experience of Insulation Workers in the
United States and Canada 1943-1976. Annals New York Academy of Sciences. 330:91-116.
1979.

U.S. EPA (U.S. Environmental Protection Agency). Airborne Asbestos Health Assessment
Update. Report 600/8-84-003F,  U.S. Environmental Protection Agency. 1986.

Venzon D; Moolgavkar S. A Method for Computing Profile-likelihood-based Confidence
Intervals. Applied Statistics. 37:87-94. 1988.

Weill H. 1994. Cancer Mortality in Chrysotile Mining and Milling: Exposure-Response.
Asbestos-Cement. Annals of Occupational Hygiene.  38(4):412. 1994.

Weill H; Hughes J; Waggenspack C. Influence of Dose and Fibre Type on Respiratory
Malignancy Risk in Asbestos Cement Manufacturing.  American Review of Respiratory Disease.
 120:345-354. 1979.
                                        A.3S

-------
I
•3

S
ce


1
OJ
o

.2

V
!S
'o
W
1

"i
p
•*»

S
**M
O


!
u
u
e

B.
K
6JD

S
Refe
   !l
   ii
90%

infldence

nterval
    I
    o
    1
    u
    a
    O
Liddell e

1997
      ii
      0 0
      o


      ON"
      f-H

      O
      o
      o

             « o
            o

            o"
                                    =   i"
p

o
                                 0  0   0
      o - -g

      Q ^J ^

      O 83 O
      >
tto e
                          2
                                           8
       <_. ^ ON
Pi

19

                                    05




                                    <
      g
      o
         ^  i
         o  ~
      11
            •a  o

            6== |




            ^1 J
            p—i ra w
                            -

                            1
                     §
                     X
                     ,u
               Hil
               il III
               •a    b

               g  § 1

               33 If
               .s s g §
               ss ^s


                           U
«s
11
1 s
II
                                    I
                                    o
                                          ffl
                                          i!
                                          il

-------
 Ml

"o
 S
a
o
I
I
en
e
a.
H
U
U
bi
C
Refere
     B •=
     ll
     y s
     u
   *il
   S'Bi
     as
     s
     W
      1
      I
      to
ald
                                       x>
                                       en   .
                                       8  "3
         -.a os  a av 5 a\
         UH 1™1  hM i—I ^i i-^
                      cs
                      o
JN
O


1
O
                            ?"S  §£
                            S ts  ol 2
                                 CM
                                 VO"
                                 S
                                 o
                                 r-T
             04
             O
B   0\
H'~   t>
    ON

Ixa S
         00
                      oo
             o
             o
                               oo
                               »-H
                               o
                               •a
                               •^^
                               0)
                                     o ^
                                     O rd
                                     U C9

         •a
                      '53)
U.S.
insulation
workers
          l«  3«
                                     CU G.
Roc
pla
en
uf
Ce
M
                      d .2
                      I*   3
                      lu   1
                      ^^   E^
                                                                      O
                                                                      <

-------
 i
•3
a
w
•a
 u

f
 Ed
 M

 o
I


1
C
 u
£
 u
 e
U
 v
 on
 e



f

i
e


I

is
•u
a
        8.
       o
        2L

           --   «  «    -H
           •a

           s
Liddell et


McDonalc
                3  .7
                o  —

                *•  °

                s  e
                C5
                o"
                              a
Hughe
                        0
                             o
                        tS

                        o
                             O
                            in
                            cs
le
Carol
                                        S  S
                                        ON  o\

                                  •a    ~  ~.

                                  •S3    "3  "3
                                  _M    *j  *^
                                  2    u  u



                                     am  j§  1
                                     f^  IN/  -H
                                  o oo  *  J3
                                  ^ ON  fli  flj
                                  ,< -H  -Ct  CO
                                  S
                                  0

                                  o\
                                  o
                                  o
                                  00
                                  00
                                  o
             f3  J^  O ™  W w  Q 3

             CXH  U'B.Z'ftOT'S,
Proc
Fricti
Texti
                                          S
                                          ON
                                          03
                                               OO
                                               0\
                                               U.
                                               1  a
                                               
-------
V3 1
&
"°\
V) 1
«- J

CO
1
•s
i
u
%
s
o
a
M
u
e«
i
3
m
2
M
U
u
•2
•i
e
a
C
N—
*
1
8
e
i — \
»J
o.
o
IS
£
Q
S
•o
«
en
-
£
O
i
Cb
>
£
js
5
1
^
ff
-i
jj
!

8
1
£



•e &
II
•° ~
E »
S u
35



M
1
w
(2

!«
'2
TO
'u
u
w
5
3
'u
V
£

9
«•

'«
w
iS u
*• IN
C '
U M
« 1
fi«
i>
1
8
^



a
o B
!»
3 »
* 5
U
M g
5 ^
S w
Jj cs


till,
A « TJ S S
t S-^ 5 *
S w ss s ^
l|l|
P <*- S 1-J
A
fa S a
5 'S a .2
££332
|" 91
5-233
tg> s
111 Is
a S" g
g< K

^> Bfi
.a ~s
S h U 5 3 u C3 ^S !is ^
uu r** o *^i
t-j r) cs u "^ v vl. ' . • •
NfsdZ^^-1^ ^ m

(S^^or^(S o «/">
(Ni^S^^1^*"^^ ^* '^



fO C^













•O "! "1 m "^
r-I t-J .-< •— '



*n ^^
tSfScSfNtses^n M ^



11 5
u -a a .3 -S e z
- -S 3 i °> °- 1 c
0 S o 1 a FS "-8 S sa 1 8
5 toUC5wdS32 ^ _c3^=!_e^5 S h
_D (f i .G n-iu raK C-*-* i^*J 9 rt *3 o •*!» **
•S ^uuq-j-5ae?cB§ss S c
5- la SI 11 at al «5 ^5 > ^
g •«
.iji ll ill
a 1 5 ££02^ -3S2|£
I M 2 fa
|S u <

-------

=3
i
I
t
^J,
*•
V)
b
^
1
Q
M
£
E
1
j
"t
<

tt
(rt
C
e
Cl
fl>
^
|

B
a
t>
&
k,

j:
*5
i
u
JB
»*
0i
£
"
o
0
5
"O
a>
M
&
c
O
-4^
U
M
to
-b
Uncertaii
*?
^n
^
«
2
CQ
H



<
i
i
t.
<{
C


•a ^
II
11
3|


_o
1
^
J
'«
'h
£

Jfi
C5
|3
"u
u
*



I
«J
(M
W

1 2
£ 0
b •>
5 o
"S ft
« N
r? Ed
ft ""I
i1
fl
I
**
|H
8
a
p


i i
ij

11
^
. a
I"S 1 a
.5 "C .2 e
*HH
s *- i a
s « (S 3
_ g M
&"» iii M
i» T2 ?; g
.5 •§ c §
I|ul2
8 "2 g>-=*
5ljl

i- he
I c S
r l|le
§ - i
J* < S
& ^
>. ,„.,
Jj> W)
|-ls
-c « V a
3 > fc fe
?? C M.
u » o
eo**
&r i
W

>. a
1" c1 £ -2
1 •-§ a «
1 1 S|E
8 1 & S
!2£ a
B w 5
U

•c
J
a

c
o
S 3
^s
f*
£
00 o\
&< TJ r~ ^-
^ S oo
— ' 3 o> -3
_! « -1 ™
" 5 S "
« ^1 a
.s 3 j §
nJ ^

1 1 i

O (N  tN fS
















S
in l/l
rt ^-4





r^ CS cs




s" 1 ^ 1 ^ 1
g-J 31 51
t^ *J~« K-J ^ 1— J xi
i-a
•3 §
o *
o 'S ^
*» E g ^g
"o o a -3
g > 2 2
H
II
     .3


i
fa "
   S

-------
t
s
 1.

 I
 w
 I
 s
 a
 t
  •O G
  §•9
   6ft
W
I.

•2 £
£ S

III
        i S
        8,£
           §
                  r-
                  rs
   t T i s
   II  1!
   m-r
   fa 8 S
   •1 '5 S .2
       ^> Wi

       •2 -^ ^5
       3 6 U
       1? a> o
       a S t. —
       3 -3 S 2 ^
       Illl"
            H i
            *• o
            u O
            A  I
            s  I
Textil
S « o
o a o
^- B ts


111
E C>*
t. K &
£§S
8^5,^
r b«a
3 §i
«-s f
== «> S
                    2 e §

                    O O Q>
                      §£ €
                      4> V
                    "&||
                    w ws . in
                    •s-s-s

                    m
pres
es
Qu
                              'S :
                                 ,
                               IE
09
(4
S
z
,§!«.
^ :c
£,£
•!•«
>
<
en
i
04
z
cfi
s
D4
X

-------
    tn
 3
 0)
 II
 O "K
      >
     =
   O
   ca
   "5J
.3
             IO
           '
 II
 O
           o>
           JS
           o

             C
             ra
          Q.
0
3.
                  h-oococoooogcoT-CD
                  CO
                               _       £^^

                            ^—*. CD  CD CD CD
                            O O  O O O O
                            O CM  CO ^ T- O
                            T-  __  _   _  _ O
                        _  _  .
                     -OOOOOOO
                                            CD
                                            h-'
                                           CM
                                    H$
                                   f*«
                                   V°8
                                    a   > o
                                                         .  o
                                                        JI
                                                        § §
= ET
«-s  d
' ~ •« .*"

-------
    o
   +•»
    0)
    0>
    o C
    CO 0>
           °*
<  g  2  •
    IT  W ffl
    5  m o
     •  c
     o  •-
     O)
     c
     3
   «o
   O)
   o
•o
 d>
 4-1
 O
 0)
 Q.
 X
 LLI
                   0)
 «
.a
O
                            0)
                                
                              •
                            o
                             II
                             o
                                       o>     £
                                                 >
>  >
 1    I
a.  0-
                                      c  u-  n
                                      0)4.
                                     •o  o  o
                                     ^  w  'i

                                 ooSf

                                 2°5  »
                                     -S  n  *^
                                 *  s>  o  u,
                                   _.P  o  2

-------








.2
0 3
(0 O
£* *^
_ o
O c
10 C
o
._ o
J2 .£ — *
o> '.Z ^
•** c S
o — *-
«> § °- IT
rf O) jS *"
c o ** •
* 0 = *
.a c ts -a
a. S
•- e °
sis
s
S&_
m
H
o to
o o
*^ ^^
«t ^
TO q>
0 X3
I/)
C <
3
j










a
15
"O II
* D
O
•o
V
a. "•"
ii
o


•D
Q>
^
0
to
A
0
•a
0)
o
0)
a.
X
LU

at:
(O

i c
IX  --X — v
O O O
CM TT CO N
0 " " "°
z: o o o co
T- CM TT II
V ^ ^- ^_, A

O
co




cq
CO
in

u?
j
S P
~ "— ' T- CO
^ o> o o ^
^ - o
T^ 2-
II
D
1
p
Oi


^>
4) p
^™ ^3 •• ^3 ^M^
CJ
^J ' ^^
D

3 3

~ lo ~a
co > >
£ a. a.
0)
*J
—
Q)
y ~ t-
c u- „
fl> <*-
•o o o
•*• in "
= S £
o o o •
J2 5 ° ^: o
B ~~? ** *J
I *_,« 2 «
h- ^ £• O 1—

-------



Ill
Si:
0>
"if
o
5
c
CO
Q.
"o
3
I
Q.
_g
'C. 00
LL. CD
** <«•
3
•— 15
t> *j
C -H
8-
o o
0)Q
C 0
1s
(Q
=
t
O
s
CO
E
o
^
o
V)


•o
D
J

3

ol

•o
0)
^
5
10
O

§£
JO (Q
*- 0>
£*



_
E
5=

*o g>
Duration
Exposui
2i
V)
o c
& 


o q

d d






0 0


3 °
CO °>



CM CN

m" in


q o
oo oo



CM Oi
CM CO





CO •*
. CO
Z^ A


0

d






0


CM
CO


-—V
CM
•r- O
o d °.
o" ^~
*"*"


;_.
> 3
m "ca
>
_C i
Q} ^-
ll
•o *o
'c w
eo O 0)

2 * a5 o
2 **£«
CO
<

-------



w
o

01
0>
jQ ™
0) ~
0)
o .-
5 3
1— °!
J
c -
•~ w
•o c
Q) 10
>> 0)
o "Z
a o ,_,
C Is*
5 eO
U 0) 0,
eo £ ^<
rf * •- —

w o ** **

& >
M ^ ^ Rf J0
1- 0)0. «
n .c
o ™ o>
c £ 3
= »- T-
« 3
>> o
•t: 
fc-
0
0)
.n
O

^^
E c
X Q
"- 0)

• S
^^^




c
(0
0)
t_ 5
><
o
a
E
01
O)
c
n
Q^
^* co OQ 10 r^*
CO OJ ^ IO N-







C3 CO ^f ''f CO
co oo TJ- m oo






O) OO h» OO ^~
CN CO CO ^C





CO O CM CO IO







CD CM O> CD CM
10 00 "*" CO CD
T- O 10
T- CM





^ CO lO ^ CO
«~ t*> 1*- OO
^_
m
 O
«• '"^ CM , , *-
= ^ ' in o ii
— V _x 	 v^
oo co r^ cj> ^-
^~ r^- r*^ Is* 05
CM T-






CM IO CM Is- O>
o> in r*- >*• o>
T- T-





a> in to m CM
oo •* 10 in
*~ T~




o en CM o CM
CM T- T- T- T-







CM 00 •* T|-  CM
CM





CO CM CO TT ^~
*- CO h- CO
^__
CO
0)
CD

o
Q.
E ^^
""^ O) CS O
CM "* "* O> O
5: *° ' 10 o ii
— V CO (N IO A
Q. >-x >— ' ^* v^ s_
a
in
a>



«


o
CD
00





(O
CM '
h-




10
O)






















0)
(0
o
l_








u?
-I ^
S Jo
"— in CD CM
"* ^ a i ^
^0 .0
i- O
•~*
II
g





_
0 h-
X o
Si * ' •* T-
o *"
T- ^ 0 0

II 0
g ^

0) 
3 3
~ CO CO
(B > >
> ' '
L; a. a.
0)
4-1
C
^™
di
o ± »-
c u- „
(V «-
•o o «
^^ ._ * •
= 0) X
o ° o>
o O c *£
Np <••'
^£.0 H
O)
<

-------
LU
w
O
«
|
o

c
O
' c
RI
0>
3
O
Ul
***
™ 
jQ

4-1
"ffl
4->
w
o

>-
0)

R>
^
0)
C
3

1C
C
"o
k.
o

4-1


c
1
"Z
M
£
O
—
w
i_
4)
^£
i_
O

CD
4J
X

o
^

*-
T5 £
5 CO
2o
c Z
Q ^***'
i-
0 O
to c
(0 £
• O
1 C/3
1 •••
<*'^^ fc»
^ ^
o> *
o> 1-
^. >»

. £
15
^J
0)

c
4)
C^
d>
Q




•o
0)
+••
0
•5

o
(I)
Q.
X
UJ
TJ
0)

i_
*d)
•





CM
II
o



^
II
a

















c

(0
0)
S



O)
£
 O CO 00 T-
T- t- T- CM



oo co CM T- o> m
CO 1 O) O) ^ T- CO
^~



h~ x— CM O) O) T—
1- T- T- T- CM






^ co o> co in co
t— co ^ m co r*^
O T- CD O) <*
^— in





rr -^ m m
""""* O CO 00
.00 r iX"
° CD CM" •*" T- co
O)
CO


m
CO
CO



CO
co




CO
CO






10
CO
CO
^






m
00
ii
A
T-
CM


T-
co
^~
'T~


CO
CO
co



CM
CM
^















10
(Q
^^
O
h-
-1
s
CM
CM
*-
II
O





0)
x

**w^
^
II
a





















^^
T- CO
CM CM"
•^







f^
00 «
f\J **
T
CM





«^*
re
C0
•**>
C
4t\
Of

c
0)
•a
•^^
*c
0 (J
T-
* e*"
^«

OJ









CO
•
o



0)
3
M*
ca
>
i
a.




•4-1

11.
*0

vt
c
•o
o
o
O











^
»
o



CO
3
^^
re
>
a.






ii
D

"o
H.
°
In
*
                                                      o
                                                      IT)

-------







JO)
•4-1
O
w
^ CQ
£ C
o:s
CO *-
«J
.Eo
4> 3 eo
.* O °°
o w *

° S .E T
i _ ^* ~~
^ c c
o> o — ^>
5 E °- -o
« «c 
D.
X
LU

QC
W

i c
X (O
^ 49
C S
^


C
re
0)
^S
i^^
j
a.
O)
10
a:

• fx. O ^1" U
° IO CO CO CD




. co ^~ co r*~
^ IO CO CO CO




^ 10 oo r*. co




^ h- co r- co
^™i • • • •
^ CM CN r- O



•r- h- c\j IO °^
oo CM ^ OT rr
?£g;S


Q ° S CD CO
" °> - co co




IO O O ^
"* •*- CO CD T~
1-



C7> O) C7)
i i i 0
° 0 0 0 °°
^ CN ^ II
V t_J^-^J A
O
a»




•r-
00
IO

u?
_l -~-
s w-
— CN 00
S K "~ V *
^^ ^^
^~ O
II "—
o

iq
f^\
CN


•o — «
Q) CO
C CN ^~ ^ °
*- ^ J2 o d
11 S
0 ^

— 3
^5 > >
£ QL Q.
0)
c

4)
o .ti ^
C ^Li
29°

^ W "o
o o a> ^
« 00=^
2 T 55 o ^
° ^2.0 H

-------



£
Q»
•£
o
5
*J
c
CO
Q.
j»
^^
X
0>
1-

^^,
C CO
>-— **%
— c*>
O 00
* °»
CO ^-
£ ° ^
rf £ «°
^ 3 *-
CD O *
.0 (ft -0
2 0) co
^ C E
O O
E Q
iS
^^

To
r
o
5
CO
E
o
"3
4->
o
(0
0)
5

D
0)
^1
0
6
£
Q.

•o
0)
£
0)
CO
A
O


5 »
O L.
(0 CO
»- 0)
Wl
s_
0. ^
i
t

c
o
"^
CO
3
0
CD
L.
3
W c
S re
Q. o>
x S
HI 2
*J
10
l_
IZ
^
0)
g.
<
£ g
ss
> a


h~ CO
o" d





0 -r-




O l-~
CD S
CM S°
CO £
CN ^
"* ^
IO IO


0 0




00 •<*
CM 1^-






^•^^
a>
CO ,-
^ ' o:
03 n
1 ^- A


o
t-^





,_




r>-
CO
o
a>
 3
L —
a> co
** •>
E -T
tt Q-
o.«
C LL
0) 4-
•D O
E «
C (o
« O <|)
^ 5 °e 5
ho * SS o
Q S2 0
| ^qH ^S,C3
CM
tf>

<

-------


«
k.
•o
0>
.2 -
(0 -Q
£ 3
|fl ^^m ^^
0) 3 = g
C < 3 2
O ȣ ^
£ E £ -e
Us?
>°-o o
+rf C « V.
S^Sc-S
i t £ « 0.
«£§ 1 ?
— 2 c .£ >
£ V. ™ Q. S
« 0) W o —
H ii * 3 o
S -S « *
« »_ _
o o : £
»5 ? "i
3 2 S «
J 4- T- (Q
« *—«

Obser



0)
7>
0)
a
X
UJ

0)
01
(6
V.
_ *
C ^r
^s,
^
C a>
'-' ra
c
10
UL


QjOCOOJOJ^COCOCOCO
^T-T-CNCNCOCNCNCN-sJ-




CDOCO^P^^:10. ^^
^coc6£!Trcoi^T;r;*;£







|>>.T-CNCOOOT-T-IOCO
UJCNT-CNCNCOCOCNCN'^r



CD05CNCD.0^C?CN!CN{0.CP
^^ Is*™* CX3 OD



CT)O5CT)h-O>IlSoCSJ
o^tqiocNt-^^R^
c)d-<-cocD^;^^:- ^^ ,_
,*5 "> o o
£, i^ co' q q
^ CN - O O
II — v v
D


ID O
2! 2
C" <0 «8
•» ' '
£ a. a.
a>
**
c
c il u
» >»-
•o o o
^ w *i
o O q>
S°So
r§!«
_l *•* O a)
* £.0.1-
in
<

-------
o  ->

I  5
-* c

Is-
Q. tL to
X 0) 00
m ** o>
   (0 T_
0)Q.—'

§JETS

1 £  «
£ ^x  C
3 1~  W
> 0)
 (0
o

 O)
 c
 3
«
<
               CM
               CO

               CO

            o n
            * ~
            «-• D
            O

            •5
            0)
            •o
            0)


            0)
            w
0)
4-1
o
0)
CL|
X
UJ
            (A
                0)
                ra|
                R>

                0)
             0)
             c
             (Q
            o:
                  co m co co
                              10
                  co o>
                              oo
                                        co
                           10
                                  10  o in
                       >

       £  OL  0.
       (I)
                                                           m
                                                           0)  •*-
                                                          •o  o  o
                                                              to
                                                              S
                                                                                    in

                                                                                    <
J2
 n
 o
i-
                                                       *
                                                        **  O aj
                                                        2.CD i-

-------



>
0)
(0
l_
0)
^
5

z
£
W
V
J£
o
ta^
§
c
o
^J ^^
« tf>
— 09
3 0>
W 1-
"* E *~
< ® «
 '3> tS
s 1 <=
CO C (0
•-< E
W^5
W
c '«
|2
(0
>»
•M
15
•c


CO
E
o

^
^_
4->
0
tfl
0)
S



•o
d>
4W
O
•5
(!)

£
•o
a>
>
£
o
(0
M
^
O
c .„
o £
W 10
t- o>
« si
Q. ^



1^^
*^
C
£
£!
(0
^
a
Q

0
^H
3
(0 C
OIB
2f
Q. 0)
X S
UJ ^
•4-1
2
z

Ito
0)
t; a)
<••>
D)
£ 5
(5 tf
CD
>-


T- -i- 00 CM TT CO
6 T- CN ^r xf ^






O O O CN m 00 CN


CN 00 00 CD T CD CD
if) C*4 Q> If) O) t*-  co •«- CN co m T-
CD co ^r co o co t-
t^ t^ ^f1 •vt1 v^ ro m
^^ ^^ ^^ ^^ ^^ v) c*j

ir> in to in in 10 m






^ in m «n m m in
K" CN (^ CN 1^- CN r-
•^ *- •«- CN CN CO CO


•



"~* T en ^r o) ^?r o>
O) •<- T- CN CN CO CO
i i i i i i i
LO O LO O tO O tO
v— •<- T- CN CN CO CO
<*~* *~* N^« S^» S.^ •>_»



CD
r^






h-


o
o
00

fs»
o> w I jo
• _ CO
CO m »-i
**-; (^

"~ Q.
S.t;
c u-
0) 
-------





1-
±


c
i
0>
JC
o
(A 0)
U> £ *
*7 


L.
4)
A
0
•o
o
0.
X
uu

I
1



I




c
n
c _
Q
*3
re
3 -
Q
C
re
tt


. CT> CO ^"
£J CN IO lO




^ CO CO 00

T~




^ co ^- to



O> ^— 00 IO
CO -f- •«- ^~

U> lO 1C
^ CO ^^ CO
r- CO CO



LO m 10 10




lO IO .->
CN r- co
o c> ^



y— S ^_^ ^^^
^O "*" if) "~*
c! "> i
V O T- ^


c>
<£)
CO




•t
^
CM
u?
_l
S-*~*

^ re re
re > >
t Ou CL
0)
_ ~
^
0 ~ T-
c. l^»
fl> *^
•o o o
•^ w "i
^ Q) T
O O ot
(A ^ ^ O
2 * s? o ^
V* MH^ ,». W^
O -i ° ° 0)
i- * 2,0 i-
in
<

-------





o> <"
C c
o 2

0) 5
3 >

CO _
0) flj VIM'
> Z i-
(O — •- ^>
"f"" JHJ3 ™ Q)
S a,
2 S c
«- 2 **
0)
0 «
c ~
o 3
—
c |
3 0)
-1 >



CO
*T
•o u
*- a
o
•5

i_
a. *~
u
- o


•o
0)

(A

O
TJ
(D
U
O
a
X
UJ

£
2
0)


0)
O)
(0
0)
1 •*

+- 0)
^ 0)
c
(0
tt


0 v- 00 0
m csi ^ co





 h~



If) IO •»- 00
T-' T-^ -r^ IO




10 10 g g
0x1 ^ CM CO

^^
'^> 0) 0 O
0 0) CO 0
IO i i 1^-
V O O ||
>_, >0 0 A
"^^ 3, "—•

o
0
CM

^^^
LiJ
S^^^
CM
IO *~~ T- .IT)
^ CO 10 "- CM
T- *•; o T o
*" 2^
II
o

o
CM


TO"
4} "***
^ .5 ,- "". T- b
«t> CO O) ^ 00
05 ^6^66
II °
II >— '
D
d) g)
— —
To "co
^ > >
5 CL CL
L.
a>
c

o .ti •«-
c il u
10*00
CV) o
(A T
o o a> ^
« r 3? o t;
h- ^ 2, (D H
IT)
<

-------




0>
c
o
E
{]} (Q
«- C
3 10
0 =
a o
x 5
UJ _<->
a> >» *°
.2 2 °*
T- "3 *- —
^ E w a
^ E CO ^j
CD ' 2 °
•° > 2 -
I"™ c C
>•- o
£ S Q

t .-§ s
° 3
S o
>

•o ii
•f O
o
•5
0)
a. ""
ii
o

•o
4)
S
CD
(0
O
•o
o
 m m
CM T- h- m



m co "*. "T
^ "^ co cr
^~ ^ CO 00

, 	 „ *" - ^^
«§!§
CM CM U,3 in
CM O A
"-^^
CD
CO
CM

uT
— 1 es?
g
r>- *- CT> ^~. co
C3i ^ CO ,» CM
,_ at 0 co o
T- °
ii 2,
. o
CO
CM


^3 ^<»«»s
0) 1s-
» S-*:»^
h« "*** •»— m
11 S
d "^


09 O
3 3
"(0 (0
O > >
5 a. a.
c
a>
I
« *- ^
o — *-
£ "• "
Q) *^
•o o o
1 m '*
O O 
CO
<

-------









0)
c
o
E
® («
3 X
U) w
x 5
UJ *-
> JS *"*•
33 °- to
„ re •*-< o

T"" 3 !J^
< E ES
.£ Q * ®
A <0
*^o 1
•w ** ^
2* 
^ 0>
c
re
Q:




00 00 CM CM O
IO ^T CM CO if)




CM ^ CM O OT
CM CO CM ^ h-




^ ^ 10
CO 00 \f) ^
*~ *~



CO CD O IO CM





co o ^r co r^
T- T- 0 0 0.



CM CD CM CO
O) ^~ O) O

 CM CM


"*^«T ° "
0 «0 ^ U5 o
co 'Y -^-^ m
{^0 ^^
P*^
>•*• >_ * ^~ •>— '
>-' ^^


.
CM

'


N-
0)
^
'
u?
_j
2 p
^J  o"
CM
II s_, _
O

0) 0)
(0 . (0
2" > >
W Q. (L
£

-------

^^
L.
O
"5
it
e
0)
E
0)
O
W
o
*•"
10
.Q
(0
o
IB
4-1
C
c oo
IB O
^W Til

**M
(I) ^"^
O *>
tl
UJ
D)
C
O
E
IB
X
™
IB
•c
o

IB
E
o
J^J
0>

0
(0
Q>
5

•D

3
(0 C
O- IB
^^ 09
x S
UJ
^J
fi
iZ

^
0)
$1 0)
< 0,
2 IB
Q>



g ^r CD °°. CM
^"^ * * ^^ '
0 T-




^ ^ 00 « CD



O O CO CO .n
O O CO CO Jij
in 10 o> o ;,;
CM CM CM CM ^


O) 
-------



c
•o
0>
^
(0


(fl
0>

o
§

0)
0) •—
O)
e w 
0>
o
c
to
0
o>
c

~*










^_
dl
o
c
(Q
O
0)
c
3
J
O
D)
C
*>
O
O
2"
01
<^™*'
(g
E
>
• ^M
"*•*
jc
or










M
00

•o n
0) M
*• O
0
^
l_
Q. ^~
II

D


>
Q
4J
in

«- TS
0) if"
°- 3
0. o
3m

v. ~o
$ 3
O O
-i m




o:
a:



_
E

•^*
1
CO CO O)
.
T— ^— ^~





•r- IO 1^
T^ W (N




O) CM h-
CO IO CO
o o d


O> CO T^
co in h-


oo h-; iq
O CD C)




00 Ol O)
T— T- T—





v- CO CM
co 10 oo
CM 00


CO

IO




uT
CM -1
« S .^ JT
"5 S S
CM 2 ^ ^
gy\ ^^ „ ^*
• ^^ ^^ ^^
II
D




•0
0> x*~,
>< IO ^
t °^ to co
V™
T- Q 0
II ^
D
CD
_g
^ZT' (0
(0 >
£ a.
0>
4-1
C
* *-
o .-
C LL
Q> 4—
!5 °

*C (A
o ® o>
f^ f\ ^
CO T- ^
5 * s? o
0 j 0 O
h- ^ 2.O




















CO

•

at
g
re
>
i;
a.




^
n
o
B €'
«=
^
^~
o
w
CO
o
1-
CD
<

-------

10
0
L_
0
5
>t
W



u.
"c
0)
E
0)
U
to
o *—•
** o
CO «a
fl> 8)
V* •** V*
i <^r •
< (B
.£ IB «

•S J5?4-
1— M 3
D) (B
C -1
O
E
IB
£
(B
t
O
S
V
u
IB
U
at
c
3
J


CM
O>
O
•o u
* 0
tl
•5
0)
u
0- T-

O



•o
0)

0)
(0
XI
o


•o

4-«
u
Q.
X
LLJ


0)
a
IB
a>
>
4
E
|
O)
c
IB
a:





oo co co o a>  10 CM
IO CM T N- T- O O






CD CO VO "t T- CM O






CM ^ CD IO O (O CM
10 CM ^r r-~ CM o o




.-, _ Q O O
ko m g § § o o
'CM ^ ^— p) (Q CM •'T
T— CM


'"""' *"^ *"^ rn O
^.-^ 01 oj CD S o
-T, O> Ol Oi O> Jx CM
^ C3> •«- co N- ;;: co
i ' ' ' ' i '
.-. o o o o ,-, o
a lO O O O S O
"-' ^^- •«- CM -^r S CD
^— ' -^^ ^~^ T~
^'1^-


q
CM





*~i
CM
CM






CM






^
CM
CM











V)
(B
4^
o
»-
ilT
-I CO P
S «D CM O)
S. o ,-: co
_^ X ° •
n 9 .o
CM 0 0
O> — '
d
n

D






T?
fl) X— s.
^ • ^H.
^T 0-d d
ii "^^
D




0) 0)
3 3
(B A
i i
a. a.

^«

c

o ** *-
S E •
0) <•- „
TS 0 B
5= (0 -o
5 10 I
o ° » !T
o 0 c "g
« 5"" O ^j
_i ® O d)
^ 2i.(D I-
(M
(0

-------




>»
c
CO
o.
E
o
O
(A
0
•*•>

0)

w
CO
3
to — ^
ES
o 0*
< 2 *
0) .± 0>
A 0> fl>
10 t^ .£
o) "^
^^ ^,
o g
Eiu
(0
CO
^
O
5
w
0>
u
c
(0
O
O)
c
3
— '



5
*"
•o n
d>
•••• D
O
^
0)

T- CM «O


s ; s 22 s
T— If) S
"*" "*"





-. CM CM CO CO
:,: oo o o h-
w T- CO CO O>

O) O> O>
•*" CM ^ r»- o
m i i i 10
^J If) O O ""*
CM 1O O II
V T- CM IO A
p
°?






CO
IO




CO
0
CO

0>



















in
(0
•*->
0
1-

'



^ ^ 	
_l 00
«VH
S ^
•»• ^- O CM
"t O ^ CD
T- O
" S
O "^




' p
a> CM
g - o « ^t
^ CM . r» CM
T- 0 "> C> C)
11 S
0 ^






0) 0)
_3 _3
^" CO (0
 >
t DL CL
0)
^
_
a)
S ±i *
S i "
•o o a
^ 
-------
  re
  c
  re
  O


  c
  re
  «o
  
  re
  4->
  V)

  •Q
  0)
   v P

   -s
   .- C.

   £ =
_  |r re
CO  W _
CM  .* E


«l|
•Q  = -O
*  .2 c
^  43 (H
   « J!
to
         CM

       ! °
      •o
       0)
       4-1
       u
       X
       Ul
0)
L.

w
fi
o
O)

o

E
re

(A
£
4-1
re
CO
Q

0)
u

re
U
O)
c
3
  CO
         10
          k. re
          0) Q)

          Q. >-


          C
          o
          !••
          4->
          re
       9
       Q.
       X
       Ul
          re
       0)
                  CM
             CT)
                         0£
             i
                         •
                         -.j~
                              o-^
              O)

              CO
                  r  (D i'a? O a?
                  CN -st m *$-. cr» T-
                     CM CM <*5 <"
                     ^ * «!.«-
                                     o
                                     o
                                     OJ

                                     "O
                               CO

                               Q
                               a>
                               CO
                                   O)
              CO
                T-OOCO
                     
-------
    (0
    0)

    o
   5
    c
    o
    as

    s-
    c  at
   —  o>
    w  T-
    o  —'
   **  c
    «  2
CV|
«  o'
»-  E  S
    a
    E
    o
   "w
   o
   (0
   0)
             •o
             V
             0)
             a!
             09  O
             (0 £
erson
Years
             3
             to c
             O n>
             QB ^

             £s
             **
             M
                       
                                                    00
                                                    to
lO
CO
CM
                                                       lO
            •  o
          ^  o
          CN" o
          T-  V
                                                       c
                                                       CD
                                                                                                        «
                                                       «>     _
    CO
    CD

   lO
   CN
    II
    C


n  ca

O  D
H Q
   Q.
 a>  ..
 o ~
 c u-
 Oi <*>•
p  °
*£  *"
 O  a>
O  c
g|

£.0

-------


k
^H
O
4^
o
10
5

£
Q)
J-
.2
c
<0
">
(A
C _
C A
CL 09
u> -. C.
oi .E _.
< £ «
— ^ *
«0 O —
•-§ 2
O) ®
g; Q
O °
ta
+j
(0

0
5

4>
O
C
re
u

0)
c
3




o»
10
o
^ M
IB
0)
CL *~
II
O


•o
O
O
Q.
X
UJ
•o
o>
>
l_
u
to
J3
O

s
CO



i
>.
«^



c
n
o
>, ^
i^
Q
Q.
Q.
E
o>
0)
c
n
n^

^ CO 00 CO CO
° in od oo C3)



T in r*- CD CD
^ ° __r .," .,"
CO




^ o ••* op co
^ CO CD CO CM




^ m ° CD ^



0) CO CD O ^
co" co «  oo . .^
to T- ^ in •
CO io h- O
O CD
II "^
„

CM
o
m


V ^
5 r- " oo ^
 >
£ DL CL
0)
c
^^
Q)
(.) .^ ^
S It "

!5 ° °

C W T*
o o o> „_
U) O O C Q
« r a? o ^
0 -J S ° <»
i- ^1 °> O \-
<0
<0

-------

%
^rf
o


Person
Years


E
&

c
_o
*3
52
"3
O
. 2
S: a
m ^
z; to
2? °
< Q.
(A X
S-
°2
L

CM CN ®
0 00 J


CD 0 ^




O> CO n
f*- CD 2
^~ °° 00
h- 0 S
T- Tr *»


CD CD CO
CT) 05 ro
CO CD CO


00 00 00
O O) O)



1 fN
^ TT T-
m eg ^>


o
CO-
CM


CO
CM




u^ ^
i -.;-s.
& "1?





^^
Is
1?
••• 1
« °-
8±s
c u.
0) i«-
•o o
*= (A
c (A
CO O CJ>
w ° 0 £
5 * 25 o
o =2 °
K- * 2le>
CD
<

-------



>
^
o

o
10
LL
™
X
O


(0
0)
(0
^If
u*
d) *~*
"Z ~
CM "O f-
< o - :
Q) O (0
to o> _
I— »- O
5 **
O (D
^Z cL
(0
>.
™
(0
^
o
2
0>
o
c
10
o
O)
c
3
-1




o
^J
•o n

^j B
.£

"O
Q)
V.
Q. T~
II
O


Q)

O
CD
Q.
X
LU

tJ
0>
0)
(0
O


—
E
••
i




(0
	 a>
ES
>
i
_J£
4-1
^
Q. *
C*
10
o:
CM 00 ^- 00 CM CO
CO O) O) OO |s- ^
CO CM





^f CM O lO CM CM
o o> O oo h- m
CO CM



m r*- co r- co oo
o r*- co m Tt o
CM T-





• •»— ^- o> m
JO O) •«- 00 00 tD
co t-- o) CM in
•<- CM


O> O> T— ^- T— h-
o o T- r«- 10 in
CM ^ m ^ m o
T- CM CO 't C3>

^ ^^ *~* ^-^
0000
o o o o
o o o o -^
•"•* CM co ^- m o
0 , , , , O
g 0 0 0 0 g
° 0 0 0 0
0000 II
V •«- CM CO ^T A

o
CO
Oi





CO
O)
00



CO
•<*•
co





CO
en




















j»
(0
o
H
— p
_i °9
s ^- ° S
— d ^ o
O t-

T 0
^™ "^^^^
II
o



"o oT
® r*-
.2 CM d c^ i^-
t, m , Is- in
^ 0 °° 0 0
CM

" O
O "^








q) g}
.2 .2
(0 (0
<^"« 1 1
"Jo o- o-
^
0)
4-1
c
s *- ^
« it "
•o *o o
*^ to "<=
" W T
o ® ®
° ° ^ *o
T ^ o ^
•1 ^^ ** Q)
*: £.0 i-
CO
(D

-------
                        Table A-28
Mesothelioma Mortality among Rochdale Asbestos Textile Factory
                      Petoetal. (1985)    ,                  ;•
Years After First Exposure
Range
(0-19)
( 20-24 )
( 25-29 )
( 30-34 )
( 35-39 )
(>=40)
(0-19)
( 20-24 )
( 25-29 )
( 30-34 )
( 35-39 )
(>=40)
(0-19)
( 20-24 )
( 25-29 )
( 30-34 )
( 35-39 )
( >=40 )
(0-19)
( 20-24 )
( 25-29 )
( 30-34 )
( 35-39 )
(>=40)
( 20-24 )
( 25-29 )
( 30-34 )
( 35-39 )
(>=40)
( 30-34 )
( 35-39 )
(>=40)
Totals
KM* 10"
(90% Confidence
Mean
11.5
22.5
27.5
32.5
37.5
42
11.5
22.5
27.5
32.5
37.5
42
11.5
22.5
27.5
32.5
37.5
42
11.5
22.5
27.5
32.5
37.5
42
22.5
27.5
32.5
37.5
42
32.5
37.5
42


Interval)
Duration
0.5
0.5
0.5
0.5
0.5
0.5
3
3
3
3
3
3
7.5
7.5
7.5
7.5
7.5
7.5
15
15
15
15
15
15
25
25
25
25
25
35
35
35



Goodness of Fit P-value
f/ml
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12
9.12

1.3
(0.74, 2.1)
0.80
Person
Years
28015
4668
3470
2041
840
402
4786
877
632
421
238
148
8521
1417
1104
707
383
249
4814
1423
870
470
204
102
848
935
600
257
122
86
107
103
69861




Observed
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
3
0
1
1
1
2
1
0
0
0
0
10




Predicted
; 0.01
0.2
' 0.3
: 0.3
0.2
0.1
'.0.003
; 0.2
0.3
0.3
i 0.3
0.2
0.01
0.5
' 0.9
1.1
0.9
.' 0.9
'0.003
0.5
'' 0.9
1.0
! 0.7
0.5
0.3
' 1.0
, 1.3
; 1.0
0.8
0.2
0.4
0.7
16.1

,

                          A.69

-------
               APPENDIX B:
REPORT ON THE PEER CONSULTATION WORKSHOP
 TO DISCUSS A PROPOSED PROTOCOL TO ASSESS
         ASBESTOS-RELATED RISK

-------
Report on the Peer Consultation Workshop to Discuss aj
  Proposed Protocol to Assess Asbestos-Related Risk   ,
                       Prepared for:

              U.S. Environmental Protection Agency
           Office of Solid Waste and Emergency Response
                   Washington, DC 20460

                EPA Contract No. 68-C-98-148
                  Work Assignment 2003-05
                       Prepared by:

                  Eastern Research Group, Inc.
                     110 Hartwell Avenue
                    Lexington, MA 02421
                      FINAL REPORT
                       May 30, 2003

-------
                                           NOTE
This report was prepared by Eastern Research Group, Inc. (ERG), an EPA contractor, as a general
record of discussion for the peer consultation workshop on a proposed protocol to assess asbestos-
related risk. This report captures the main points of scheduled presentations, highlights discussions
among the panelists, and documents the public comments provided at flie meeting. This report does not
contain a verbatim transcript of all issues discussed, and it does not embellish, interpret, or enlarge upon
matters Ihat were incomplete or unclear. EPA will use the information presented during ihe peer
consultation workshop to determine whether the proposed risk assessment methodology can be used to
support decisions at asbestos-contaminated sites. Except as specifically noted, no statements in this
report represent analyses by or positions of EPA or ERG.

-------
                                      CONTENTS


List of Abbreviations	iii

Executive Summary	;	 v

1.    Introduction	  1-1

      1.1     Background	  1-1
      1.2     Scope of the Peer Consultation Workshop	'....  1-2
              1.2.1    Activities Prior to the Peer Consultation Workshop		  1-2
              1.2.2    Activities at the Peer Consultation Workshop		  1-3
              1.2.3    Activities Following the Peer Consultation Workshop	  1-4
      1.3     Report Organization	'"....  1-5

2.    Background on the Proposed Protocol to Assess Asbestos-Related Risk	  2-1
                                                                             r
3.    Comments on Topic Area 1: Interpretations of the Epidemiology           \
      and Toxicology Literature	-.•••-,	  3-1.

      3.1     Lung Cancer  	\ . ..  3-1
              3.1.1    Lung Cancer and Fiber Type: Inferences from the Epidemiology Oteratufb-1
              3.1.2    Lung Cancer and Fiber Type: Inferences from Animal Toxicology.
                      and Mechanistic Studies 	'-....  3-6
              3.1.3    Lung Cancer and Fiber Dimension: Inferences from Ihe
                      Epidemiology Literature	  3-7
              3.1.4    Lung Cancer and Fiber Dimension: Inferences from Animal
                      Toxicology and Mechanistic Studies	  3-9
              3.1.5    Other Issues Related to Lung Cancer ...,	  3-10
      3.2     Mesolhelioma	;	p. ..  3-12
              3.2.1    Mesolhelioma and Fiber Type: Inferences from the Epidemiology!Litera1nte2
              3.2.2    Mesothelioma and Fiber Type: Inferences from Animal
                      Toxicology and Mechanistic Studies	  3-14
              3.2.3    Mesothelioma and Fiber Dimension: Inferences from the        ,
                      Epidemiology Literature	'	  3-16
              3.2.4    Mesolhelioma and Fiber Dimension: Inferences from Animal Toxicology and
                      Mechanistic Studies	  3-17
      3.3     Exposure Estimates in the Epidemiology Literature  	  3-18

-------
                               CONTENTS (Continued)
4.    Comments on Topic Area 2: The Proposed Exposure Index	 4-1

      4.1     Responses to Charge Question 4 	 4-1
      4.2     Responses to Charge Question 5 	 4-3
      4.3     Responses to Charge Question 6 	 4-4

5.    Comments on Topic Area 3: General Questions  	 5-1

      5.1     Responses to Charge Question 7	 5-1
      5.2     Responses to Charge Question 8 	 5-2
      5.3     Responses to Charge Question 9 	 5-3
      5.4     Responses to Charge Question 10  	 5-4
      5.5     Responses to Charge Question 12  	 5-6

6.    Comments on Topic Area 4: Conclusions and Recommendations	 6-1

      6.1     Responses to Charge Question 11  	 6-1
      6.2     Development of Final Conclusions and Recommendations	 6-6

7.    References      	 7-1
Appendices

Appendix A
Appendix B

Appendix C
Appendix D
Appendix E
Appendix F
List of Expert Panelists
Premeeting Comments, Alphabetized by Author (includes bios of panelists and the
charge to the reviewers)
List of Registered Observers of the Peer Consultation Workshop
Agenda for the Peer Consultation Workshop
Observer Comments Provided at the Peer Consultation Workshop
Observer Post-Meeting Comments

-------
                 LIST OF ABBREVIATIONS
ATSDR
EPA
ERG
IARC
IRIS
NIOSH
PCM
SEM
SVF
TEM
Agency for Toxic Substances and Disease Registry
U.S. Environmental Protection Agency
Eastern Research Group, Inc.
International Agency for Research on Cancer
Integrated Risk Information System
National Institute for Occupational Safely and Health
phase contrast microscopy
scanning electron microscopy
synlhetic vitreous fibers
transmission electron microscopy
micrometers
                             in

-------
                                EXECUTIVE SUMMARY

Eleven expert panelists participated in a peer consultation workshop to review a proposed protocol to
assess asbestos-related risks. The protocol is documented in the report, "Technical Support Document
for a Protocol to Assess Asbestos-Related Risk, Parts I and H" (Herman and Crump 1999,2001). At
the end of the 214-day workshop, which was open to the public, the expert panelists drafted the
following summary of their findings:

The peer consultation panel strongly endorsed the conceptual approach of developing an updated
cancer risk assessment methodology that takes into account fiber type and fiber dimension. The
opportunity is at hand to use substantial new information from epidemiology, experimental toxicology,
and exposure characterization on what continues to be an extremely important societal issue—assessing
the health risks associated with environmental and occupational exposures to asbestos. The panel
recommended that EPA proceed in an expeditious manner to consider the panelists' conclusions and
recommendations with a goal of having an updated asbestos risk assessment methodology. It is
important that EPA devote sufficient resources so that this important task can be accomplished in a
timely and scientifically sound manner. The panel urges that additional analyses underpinning the
document, preparation of documentation, and further review be carried out in an open and transparent
manner.
Prior to the workshop, the participants received draft copies of the "Methodology for Conducting Risk
Assessments at Asbestos Superfund Sites Part 1: Protocol" and "Part 2: Technical Background
Document" The panelists generally found that these documents did not provide a complete and
transparent description of how the data were analyzed to support the conclusions presented. The
incomplete documentation of methodology precluded the replication of the findings, in advance of the
meeting, by several panelists. The methodology used was clarified by the comprehensive presentations
that Drs. Herman and Crump made at the workshop. However, future drafts of these documents must

-------
clearly describe the methodologies and include sufficient data, peihaps in appendices, such that the
findings can be replicated.                                     •                     \
The panelists made the following conclusions and recommendations:
      Measurement methods. Continuing advances have been made in the application of exposure
      measurement technology for asbestos fibers during the past two decades. These advances
      include the use of transmission electron microscopy (TEM) and allied techniques (e.g., energy
      dispersive x-ray detection, or EDS) as an alternative to phase contrast microscopy (PCM),
      thereby allowing the bivariate (i.e., length and width) characterization of fibers and fiber type.
      The proposed risk assessment methodology incorporates these advances in the development of
      an exposure index. The panel was in agreement that this aspect of the new risk assessment
      methodology represents a substantial advance over the existing methodology.       :•

      Integration of exposure and risk assessment models. A key aspect of the proposed risk
      assessment methodology is a linking of specific exposure characterization methodology with
      exposure-response coefficients. It has been emphasized that any change in the exposure
      characterization metrics must be accompanied by changes in the exposure-response coefficients
      of the risk assessment models. This was emphasized in the report and the panelists endorsed this
      view.                                                             ,          <

      Access to additional raw data sets. The panelists strongly recommended that EPA make
      every attempt to acquire and analyze raw data sets from key human epidemiological studies.
      Where possible, it would also be desirable to obtain bivariate (i.e., length and diameter) fiber
      exposure information for these re-analyses. Several panelists believed that review of additional
      data sets offers substantial opportunity for improving the proposed risk assessment   j
      methodology. In the event that raw data cannot be obtained due to confidentiality reasons or
      other restrictions, the panelists suggested that the authors consider asking those who have
      access to the data to conduct the necessary statistical analyses and communicate their results
      directly to EPA for further consideration.
                                                                                 !'
      Fiber diameter. The proposed risk assessment methodology uses a diameter cut-off of 0.5
      micrometers (urn) for considering fibers. The report states that fibers 0.7 |j,m in diameter can
      reach the respiratory zone of the lung. A few panel members indicated that the fiber diameter
      cut-off could be as high as 1.5 jim during oral breathing. The 0.4 ^m cut-off came from rat data,
      but larger diameters would be expected to be respirable in humans. There was general
      agreement that the diameter cut-off should be between 0.5 and 1.5 jim This issue is deserving
      of further analysis.
                                             VI

-------
Fiber length. The Berman and Crump analyses made a significant contribution by obtaining and
analyzing membrane filters from the animal inhalation studies in Edinburgh and conducting
quality-assured bivariate length and distribution analyses by TEM—thereby greatly reducing the
uncertainly of the exposure side of the exposure-response relationship for chronic fiber exposure
in rats. Unfortunately, correspondingly detailed information on bivariate size distribution is not
available for humans. This leads to the need to use the animal data, although one must always
recognize the uncertainties associated with interspecies extrapolations such as anatomic
characteristics and respirability between species. Future analyses may benefit from using other
available laboratory animal data sets and human data sets.

The fiber length distributions for the human cohort exposures are much more uncertain. For the
Wittenoom, Quebec, and South Carolina cohorts, there are limited fiber length distribution data
based on TEM analysis from historic membrane filter samples, but only fiber categories longer
than 5 \nm and longer than 10 ^m were counted. For all other cohorts, the measurements were
limited to PCM fiber counts for all fibers greater than 5 nm in length in some, and particle counts
(lOx objective) on midget impinger samples in others. Both methods do not measure thin fibers,
do not discriminate between asbestos and other mineral particles, and provide no information on
the concentrations of fibers longer than 10,20, or 40 urn, or inter-laboratory variations in
optical resolution and counting rules. As one approach to addressing the varying uncertainty  in
assessing exposure in the different studies, Berman and Crump used the available information to
make adjustments to the uncertainty ranges in the exposure-response coefficients.  The
workshop panel welcomed this initiative but suggested alternative approaches (see "Methods,"
below).

Some panelists felt that an Exposure Assessment Workshop, with participants having a broad
range of expertise, could evaluate the uncertainties in historic occupational data sets' exposure
measurements. They felt such a workshop could result in a more confident assessment of
exposure-response relationships for populations exposed to a variety of amphiboles, chrysotile,
and mixtures. With incorporation of other available knowledge on fiber type, process, smoking
(if available), and the relative number of excess lung cancer and mesothelioma, it may well be
possible to gain a much clearer understanding of the roles of these variables as causal factors for
these asbestos-associated cancers. In addition, the workshop would  prove valuable in further
discussion of mineralogical, geological, and industrial hygiene issues  with regard to application of
the model to risk assessment in environmental sites of concern.

The Berman and Crump index assigns zero risk to fibers less than 5 \j.m in length  Fibers
between 5 and 10 (im are assigned a risk that is one three-hundredth of the risk assigned to
fibers longer than 10 \an. Panelists agreed that there is a considerably greater risk  for lung
cancer for fibers longer than 10 \ixn However, the panel was uncertain as to an exact cut size
for length and the magnitude of the relative potency. The panelists also agreed that the available
                                        vn

-------
 data suggest that the risk for fibers less than 5 \im in length is veiy low and could be zero. This
 specific issue was addressed by an expert panel convened by the Agency for Toxic Substances
 and Disease Registry (ATSDR) in October 2002. Some panelists suggested that, for
 mesothelioma, greater weight should perhaps be assigned to fibers in the 5 to 10 |im1ength
 range and to thinner fibers.
                                                       i
 Fiber type. For mesothelioma, the panelists supported the use of different relative carcinogenic
 potencies for different fiber types. The panelists unanimously agreed that the available
 epidemiology studies provide compelling evidence that the carcinogenic potency of ajnphibole
 fibers is two orders of magnitude greater than that for chrysotile fibers. There was some
 discussion about the precise ratio expressed due to questions about the availability of exposure
 data in existing studies (e.g., Wittenoom). There was recognition that time since first exposure is
 an important factor in determining risk for mesothelioma and some discussion is needed on the
 importance of duration and intensity of exposure.

 For lung cancer, ihe panelists had differing opinions on the inferences that can be made on the
 relative potency of chrysotile and amphibole fibers. Some panelists supported Ihe finding lhat
 amphibole fibers are 5 times or more potent for lung cancer than are chrysotile fibers. Other
 panelists did not think Ihe statistical analyses in the draft methodology document supports this
 relative potency and wondered if additional review of the epidemiological data might identity
 factors other than fiber type (e.g., industry considered) that provide further insights on the
 matter. These other factors can then be considered when the risk assessment is applied.

 Cleavage fragments. The panel knew of little data to directly address the question'as to
 whether cleavage fragments of equal durability and dimension as fibers would have similar or
 dissimilar potency for lung cancer. The general view is that data indicate that durability and
 dimension are critical to pulmonary pathogenesis. Therefore, it is prudent at this time to  assume
 equivalent potency for cancer in the absence of other information to (he contrary. Consideration
 of conducting a rat inhalation study using tremolite cleavage fragments was recommended to
 address this issue. For mesolhelioma, it was viewed that thin fibers greater than 5 urn in length
 are more important. Cleavage fragments that do not meet these criteria would not contribute to
 risk of mesolheiioma                                                         ;

 Other amphiboles. The panel agreed with the report's  conclusion that the potency of currently
regulated and unregulated amphibole fibers should be'considered equal based on the reasoning
that similar durability and dimension would be expected to result in similar pathogenicity.

Methods. The panelists extensively discussed the approach to conducting the meta-ahalysis of
the large number of epidemiological studies. A number of the panelists urged that consideration
be given to using more traditional approaches that would include development and application of
specific criteria for inclusion of studies into the exposure-response analysis, examination of
                                       via

-------

heterogeneity and sources of the heterogeneity, and the use of sensitivity analysis to identify
influential studies. .

The panelists also urged, in the study-specific analysis, exploration of alternative exposure-
response models other than the lung cancer and mesothelioma risk models EPA has been using
since 1986. This would possibly include non-linear response models (e.g., log-linear models),
examination of separate effects for concentration and duration, time since first exposure, time
since cessation of exposure, possibly dropping the "a factor," and different methods for
measurement error. The adequacy of different models should be examined using goodness of fit
statistics across all studies. The possibility of internal analyses should be re-examined (i.e., it may
be possible to obtain partial data, such as age-specific person years data, from authors).
Exploration of non-linearity should also include shape of the curve in the low exposure area

The panelists also urged alternative approaches to meta-analyses. In particular, panelists
recommended meta-regression using original (untransformed) exposure-response coefficients, in
which predictor variables include the estimated percentage of amphiboles, percentage of fiber
greater than 10 jim, and categorical grouping of studies according to quality. Original exposure-
response coefficient variances should be used in conjunction with random effects models in
which residual inter-study variation is estimated. Analyses restricted to long latency and a
predictor variable for industry type should be considered. A priori distribution for inter-study
residual variance might also be considered. Meta-regression will allow simple inspection of
likelihoods to consider Hie importance of different predictor variables. Sensitivity analyses should
be conducted in which tiie inclusion or exclusion of specific studies or groups of studies is
evaluated.

Cigarette smoking. Most panelists felt strongly that future analyses need to pay more attention
to the effects of smoking on the lung cancer exposure-response model and extrapolations to
risk. However, the current data sets have variable and limited information available on smoking.
The panelists noted that smoking is the primary cause for lung cancer, but the lung cancer dose-
response relationship for smoking is complex due to 1he effects of smoking duration, intensity,
and cessation.

The impact of smoking has effects on both the estimation and the application of the model for
projecting risk of lung cancer due to asbestos exposure. This may be an especially critical issue
for low-exposure extrapolation. With respect to estimation, accepting the form of the proposed
model, the effect of smoking may require different KL values for smokers and non-smokers. The
panelists recognized that there is limited epidemiologic data to address this issue, but
recommend that it be investigated. With respect to applying the model to make risk projections
for any  future cohort, the background rate of lung cancer employed in the model needs to be
carefully determined to capture the smoking behavior of the cohort.

-------
      Localized tremolite exposures. During the course of public comments, the panel received
      input from several individuals who expressed concerns about environmental exposures to
      tremolite asbestos from localized geologic formations in California Tlie individuals suggested
      that inadequate attention had been given to characterization of the exposures to residents of
      these communities. While the panel was not in a position or charged with the evaluation of ihis
      issue, the panel did feel that this was a potentially serious matter deserving of attention by the
      appropriate public health authorities. Evaluation of these kinds of situations would benefit from
      the use of the improved risk assessment methodology being considered.
The remainder of this report summarizes the discussions and observations that led to these findings,
                                                                                i>
reviews the panelists' comments on many topics not listed in this executive summary, and documents

the observer comments provided at the workshop.                                    >

-------
                                   1. INTRODUCTION

This report summarizes a peer consultation by 11 expert panelists of a proposed protocol to assess
asbestos-related risks. Contactors to Ihe U.S. Environmental Protection Agency (EPA) developed the
proposed protocol, which is documented in a report titled: 'Technical Support Document for a
Protocol to Assess Asbestos-Related Risk" (Herman and Crump 2001). The purpose of the peer
consultation workshop was to provide EPA feedback on the scientific merit of the proposed protocol.
The peer consultation workshop took place in a meeting open to the public on February 25-27,2003,
in San Francisco, California

This report summarizes the technical discussions among the expert panelists and documents comments
provided by observers. These discussions largely focused on three topic areas: interpretations of the
epidemiology and toxicology literature, the proposed exposure index, and general questions about key
assumptions and inferences in the protocol. The remainder of this introductory section presents
background information on Ihe protocol (Section 1.1),  describes the scope of the peer consultation
workshop (Section 1.2), and reviews the organization of this report (Section 1.3).

1.1   Background

EPA's current assessment of asbestos toxicity is based  primarily on an asbestos review completed in
1986  (EPA 1986) and has not changed substantially since that time. The 1986 assessment considers six
mineral forms of asbestos and all asbestos fiber sizes longer than 5 micrometers (urn) to be of equal
carcinogenic potency. However,  since  1986, asbestos measurement techniques and the understanding
of how asbestos exposure contributes to disease have improved substantially. To incorporate the
knowledge gained over the last 17 years into the agency's toxicity assessment for asbestos, EPA
                                                            *
contracted with Aeolus, Inc., to develop a proposed methodology for conducting asbestos risk
assessments. The proposed methodology distinguishes between fiber sizes and fiber types in estimating
                                            1-1

-------
potential health risks related to asbestos exposure. The methodology also proposes anew exposure
index for estimating carcinogenic risk

As a key step in determining the scientific merit of the proposed risk assessment methodology, EPA
                                                                               i.
decided to obtain expert input on the draft report through a peer consultation workshop. The purpose
of the workshop was to obtain feedback from subject-matter experts during the development stage of
the proposed risk assessment methodology; the workshop was not an official peer review. Eastern
Research Group, Inc. (ERG), organized and implemented the peer consultation workshop under a
contract to EPA.                                                                 '

1.2   Scope of the Peer Consultation Workshop

The peer consultation involved many activities before the workshop (see Section 1.2.1), at the
workshop (see Section 1.2.2), and after the workshop (see Section  1.2.3). The following subsections
describe these activities.

      1.2.1   Activities Prior to the Peer Consultation Workshop
This section describes the major activities ERG and the expert panelists conducted prior tolhe peer
consultation workshop:
      Select expert panelists. ERG selected the expert panelists for the peer consultation workshop.
      ERG sought to compile a panel of experts wife broad experience and expertise in the following
      disciplines: toxicology, epidemiology, biostatistics, asbestos sampling and analytical methods,
      EPA's human health risk assessment guidelines, and asbestos-related environmental 'and
      occupational health issues. Appendix A lists the expert panelists ERG selected, and Appendix B
      includes brief biographies that summarize the panelists' areas of expertise.
      Every panelist is either a senior scientist, physician, or researcher with extensive experience in
      the aforementioned fields, as demonstrated by peer-reviewed publications, awards, and service
                                            1-2

-------
      to relevant professional societies. To ensure the peer consultation offered a balanced
      perspective, ERG intentionally selected expert panelists with a broad range of affiliations (e.g.,
      academia, consulting, state and federal agencies). When searching for panelists, ERG asked all
      candidates to disclose real or perceived conflicts of interest

      Prepare a charge to the expert panelists. ERG worked with EPA to prepare written
      guidelines (commonly called a "charge") for the peer consultation workshop. The charge
      includes 12 specific questions, organized into 4 topic areas. Discussions at the workshop largely
      addressed Hie technical issues raised in the charge, but the expert panelists were encouraged to
      discuss other relevant matters that were not specifically addressed in the charge questions. A
      copy of the charge is included in Appendix B.

      Distribute review documents and other relevant information. Several weeks prior to the
      peer consultation workshop, ERG sent every panelist  copies of the charge and the proposed
      risk assessment methodology (Berman and Crump 2001). These items formed the basis of tie
      technical discussions at the workshop. In addition, ERG distributed several additional
      publications on related topics (see Table 1, at Ihe end of this section, for list of the publications).
      The supplemental publications were provided largely in response to panelists' requests for
      further background information  on selected issues. The panelists also circulated publications
      amongst themselves on specific topics. Finally, one of the meeting chairs noted for Hie record
      that, upon arriving in San Francisco, he also received a memo and copies of many abstracts and
      other information from Gate Jenkins of EPA. The meeting chair offered to share these materials
      with other panelists during 1he workshop.

      Obtain and compile the panelists' premeeting comments. After receiving the workshop
      materials, 1he panelists were asked to prepare their initial responses to the charge questions.
      Booklets containing the premeeting comments were distributed to the expert panelists before the
      workshop and  were made available to observers at the workshop. These initial comments are
      included in 1his report, without modification, as Appendix B. It should be noted that the
      premeeting comments are preliminary in nature. Some  panelists' technical findings may have
      'changed after the premeeting comments were submitted.
      1.2.2   Activities at the Peer Consultation Workshop


The 11 expert panelists and approximately 75 observers attended the peer consultation workshop,

which was held at 1he Westin St Francis Hotel in San Francisco, California, on February 25-27, 2003.

The workshop was open to the public, and the workshop dates and times were announced in the
                                             1-3

-------
 Federal Register. Appendix C lists the observers who confirmed their attendance at the workshop
 registration desk. The workshop schedule generally followed the agenda, presented here as Appendix
 D.

 The workshop began with introductory remarks from Ms. Jan Conriery (ERG), the facilitator of the
 peer consultation. Ms. Connery welcomed the expert panelists and observers, stated the purpose of the
 workshop, identified the document being reviewed, and explained the procedure for observers to make
 comments. Mr. Richard Troast (EPA) then provided background information on the review, document
                                                                                i
 and EPA's ongoing efforts to assess asbestos toxicity (see Section 1.1). Mr. Troast identified the main
 differences between EPA's existing asbestos risk assessment methodology (EPA 1986) and the
 proposed methodology (Berman and Crump 2001).  Mr. Troast noted that the expert panelists'
 feedback will ultimately help EPA complete its update of asbestos health risks for the Integrated Risk
 Information System (IRIS); he clarified that the final IRIS update will be subject to peer review or
 Science Advisory Board review before being implemented. Following these opening remarks, Dr.
 Wayne Berrnan and Dr. Kenny Crump—the authors of the proposed methodology—presented
 detailed information on the review document; Section 2 of this report summarizes their presentations.

 After the background presentation, Dr. Roger McClellan and Dr. Leslie Stayner chaired the .technical
 discussions that followed. For the remainder of the meeting, the panelists engaged in free-flowing
 discussions when answering the charge questions and addressing additional topics not specified in the
 charge. Observers were given fee opportunity to provide verbal comments three different times during
 the workshop; these observer comments are documented in Appendix E. Representatives from EPA
 and the document authors provided clarifications on the proposed methodology periodically (throughout
the 2'/2-day workshop.

      1.2.3    Activities Following the Peer Consultation Workshop
                                            1-4

-------
The primary activity following the peer consultation workshop was preparing this summaiy report A
technical writer from ERG who attended Hie meeting prepared a draft of this report, which ERG
distributed to the 11 expert panelists and asked them to verify that the draft accurately reflects the tone
and substance of Ihe panelists' discussions at the workshop. After incorporating the panelists'
suggested revisions to the draft report, ERG submitted the final report (i.e., this report) to EPA.

1.3   Report Organization

The structure of this report follows the order of the technical discussions during the meeting. Section 2
summarizes Dr. Berman and Crump's background presentations. Sections 3 through 6 are records of
the panelists' discussions on 1ne four main topic areas: interpretations of the epidemiology and
toxicology literature (Section 3), the proposed exposure index (Section 4), general questions (Section
5), and conclusions and recommendations (Section 6). Finally, Section 7 provides references for all
documents cited in the text.

The appendices to this report include background information on the peer consultation workshop. This
information includes items that were on display at the workshop and items generated since the
workshop (e.g., a final list of attendees). The appendices contain the following information:
       List of the expert panelists (Appendix A).
       The panelists' premeeting comments, the charge to the reviewers, and brief bios of the expert
       panelists (Appendix B),
       List of registered observers of the peer consultation workshop (Appendix C).
       Agenda for the peer consultation workshop (Appendix D).
       Observer comments provided at the peer consultation workshop (Appendix E).
       Observer post-meeting comments (Appendix F).
                                             1-5

-------
                                       Table 1
                    References ERG Provided to the Expert Panelists
Herman, DW and Crump K. 1999. Methodology for Conducting Risk Assessments at Asbestos
Superfund Sites; Part 1: Protocol Final Draft. Prepared for U.S. Environmental Protection Agency.
February 15, 1999.
Berman, DW and Crump K. 2001. Technical Support Document for a Protocol to Assess
Asbestos-Related Risk. Final Draft Prepared for U.S. Department of Transportation and1 U.S.
Environmental Protection Agency. September 4,2001.         •                    •,
Berman, DW, Crump, 1C, Chatfield, E, Davis, J. and A. Jones. 1995. The Sizes, Shapes;' and
Mineralogy of Asbestos Structures that Induce Lung Tumors or Mesothelioma in AF/HAN Rats
Following Inhalation. Risk Analysis. 15:2,181-195.                                 •
Berman, DW. 1995. Errata Risk Analysis. 15:4, 541.
Committee on Nonoccupational Health Risks of Asbestiform Fibers. Breslow, L., Chairman. 1984.
Asbestiform Fibers Nonoccupational Health Risks. Washington, DC: National Academy Press.
EPA 1986. Airborne Asbestos Health Assessment Update. U.S. Environmental Protection Agency.
EPA 600/8-84-003F. 1986.
NIOSH Interdivisional Fiber Subcommittee Report. Prepared by the NIOSH Interdivisiorial Fiber
Subcommittee. 1999.
                                         1-6

-------
                 2.  BACKGROUND ON THE PROPOSED PROTOCOL
                        TO ASSESS ASBESTOS-RELATED RISK

This section summarizes presentations given by the principal authors of 1he proposed risk assessment
methodology. These presentations were given because several panelists asked ERG, prior to the peer
consultation workshop, if the authors would provide detailed background information on how the
methodology was developed. This section reviews the major presentation topics, but does not present
the panelists' comments on the proposed protocol. Sections 3 through 6 document the expert panelists'
technical feedback on Ihe protocol.
      Motivation for developing the proposed protocol Dr. Herman identified several reasons for
      developing the updated protocol for assessing asbestos-related risks. These reasons include
      EPA's existing asbestos models being inconsistent with inferences from the scientific literature,
      the need for having uniformly-applied sampling and analytical procedures to measure asbestos
      characteristics most predictive of risk, and the belief that EPA's current asbestos risk
      assessment methodology may not be adequately protective in some circumstances. To improve
      upon the current methodology, the authors intended to develop a risk assessment model that
      adequately predicts cancer risk in all studied environments and can therefore be applied with
      much greater confidence to environments that have not been studied. Dr. Berman outlined the
      general  approach taken to develop the proposed protocol, as summarized in the following
      bulleted items.

      Dr. Berman provided background information on and definitions for asbestos, other fibrous
      structures, asbestos morphology, and cleavage fragments. He also described tie capabilities and
      limitations of the analytical techniques lhat have been used to characterize asbestos exposures,
      such as  midget impingers, phase contrast microscopy (PCM), scanning electron microscopy
      (SEM),  and transmission electron microscopy (TEM). Dr. Berman explained how differences in
      these analytical techniques must be critically evaluated when comparing results reported in all
      epidemiological and other types of studies that examine asbestos exposure. Dr. Berman also
      stressed that it is not just differences in analytical techniques, but choice of specific methods for
      each analytical technique that affects results. Further information on these topics is included in
      Chapter 4 of the proposed protocol (Berman and Crump 2001).

      Re-analysis of human epidemiological data. Dr. Crump described how the authors
      evaluated the human epidemiological data He displayed a list of the studies that were
      considered, noting that he had access to raw, individual-level data for three occupational
      cohorts: chrysotile textile workers in South Carolina, United States; crocidolite miners in

                                            2-1

-------
Wittenoom, Australia; and chiysotile miners and millers in Quebec, Canada All data1 sets with
exposure data were considered in the analysis, and criteria were not established for selecting
studies. Dr. Crump then presented findings for asbestos-related risks for lung cancer! and
mesothelioma                                                              '
                                                                           i
For lung cancer, Dr. Crump first reviewed EPA's existing lung cancer model for asbestos
exposure (see equation 6.1 in the proposed protocol), which relates the relative risk of lung
cancer mortality linearly to cumulative asbestos exposure, with a 10-year lag time. Dr. Crump
noted that the model predicts that relative risk for developing lung cancer remains constant after
asbestos exposure ceases—an assumption he showed was reasonably consistent with findings
from epidemiological studies. Dr. Crump also discussed how the model assesses interactions
between exposures to cigarette smoke and to asbestos—an issue die panelists revisited several
times later in the workshop (e.g., see Section 3.1.1 and the executive summary).  Dr. Crump
presented a series of tables and figures demonstrating the adequacy of multiple lung cancer
models: first using EPA's existing lung cancer model, next using  a modified version of the model
that accounts  for differences in the background rates of lung cancer, and finally using'the
proposed lung cancer model, which considers an exposure index lhat assigns greater
carcinogenic potency to amphibole fibers and to longer fibers.                    ';

Similarly,  Dr. Crump reviewed the performance of EPA's mesothelioma model for asbestos
exposures (see equation 6.11 in the proposed protocol), which predicts that mesothelioma risks
vary linearly wilh the average asbestos exposure and increase quadratically with time from onset
of exposure. Dr. Crump presented several tables and graphs indicating how well EPA's existing
model and the proposed protocol fit the human epidemiological data He made several
conclusions about the existing risk model, including that mesothelioma risk coefficients varied
considerably across the cohorts and the risk coefficients were generally higher for cohorts
exposed primarily to amphibole fibers, compared to those exposed primarily to chiysotile  fibers.
Dr. Crump also noted that the data did not support consideration of a sub-linear or threshold
dose-response relationship. This latter point generated considerable discussion later in the
workshop (e.g., see Section 4.3).
                                                      j
Dr. Crump then described the meta-analysis the authors conducted to evaluate the relative
potency of amphibole and chiysotile fibers. First, he explained how the authors weighted the
different studies in the meta-analysis, based on uncertainty factors assigned to the individual
studies. Dr. Crump identified the four uncertainty factors and described generally how each
factor was assigned. Sources of uncertainty included representativeness of air sampling data, the
availability of conversion factors to express exposures in terms of PCM concentrations, and
whether data on exposure duration were available. Dr. Crump then highlighted the main
conclusions from the meta-analysis. For lung cancer, the meta-analysis suggested that amphibole
fibers are approximately five times more potent than are chrysotile fibers, but the difference in
potency was not statistically significant (i.e., the authors could not reject the hypothesis .that
                                       2-2

-------
chiysotile fibers and amphibole fibers are equally potent). For mesothelioma, the meta-analysis
suggested 1hat chiysotile fibers are 0.002 times as potent as amphibole fibers, and the difference
in potency was statistically significant

Inferences drawn from the broader literature. Dr. Herman described how the authors
incorporated inferences from the broader scientific literature into the proposed protocol. He
reviewed key findings on how various mechanisms are biologically related to how asbestos
causes disease. These mechanisms included respiration, deposition, degradation, clearance,
translocation, and tissue-specific biological responses. Chapter 7 of the review document
provides detailed information on the relevance of Ihese mechanisms, with emphasis on 1he
influence of fiber type and fiber dimension.

Derivation of the exposure index.  Dr. Herman explained how the authors derived the
exposure index, which is largely based on an earlier re-analysis (Herman et al-1995) of six
animal inhalation studies conducted by a single laboratory. That re-analysis found that lung tumor
incidence is adequately predicted using an exposure index that assigns no carcinogenic potency
to fibers shorter than 5 ^m, relatively low carcinogenic potency to fibers with lengths between 5
and 40 um and diameters less than 0.4 nm, and the greatest carcinogenic potency to fibers
longer than 40 ^-m and thinner than 0.4 jim However, these findings could not be applied
directly  to the human epidemiological data, because the epidemiological studies dp not include
exposure measurements that quantify the relative amounts of asbestos fibers shorter and longer
than40|a,m

Dr. Herman noted that the proposed protocol includes an ad hoc assumption that the fiber size
weighting factors optimized from the  laboratory animal studies can be applied to humans, but
with a length cut-off of 10 urn in the exposure index, rather than a cut-off of 40 jim Dr. Herman
emphasized that this assumption was made to model the critical characteristics of asbestos in a
manner that reasonably captures cancer risks observed across multiple epidemiological studies.
He acknowledged that asbestos potency is likely a continuous function of fiber length, but the
exposure measurements from the available animal and epidemiological studies do not support
incorporating such a continuous function in the exposure-response model. The panelists
commented on the proposed exposure index when discussing topic area 3 (see Section 4).

Dr. Herman also noted lhat the authors selected a conservative set of dose-response coefficients
(see Table 6-30 of the review document), rather than using the optimized ones from the animal
studies (see Table 6-29). However, the conservative and optimized dose-response coefficients
were reasonably consistent none of the conservative coefficients differed by more than a factor
of 4 from the corresponding optimized ones.

Conclusions regarding proposed protocol Dr. Herman indicated that the proposed protocol
is substantially more consistent with inferences documented in the scientific literature (i.e., that

                                       2-3

-------
long, Ihin structures contribute most to risk) than EPA's existing risk assessment methodology.
Further, die proposed protocol provides a better fit to cancer risks observed in the human
epidemiological studies 1han does EPA's existing model, and the proposed protocol-appears to
underestimate risks of lung cancer and mesolhelioma less frequently and to a lesser degree than
the existing approach. Finally,  by recommending use of a standardized analytical method that
links directly to the exposure index, the proposed protocol will help ensure that future risk
assessments are conducted in a consistent fashion and their results can be readily compared
from one study to the next.                                                    i
                                       2-4

-------
          3. COMMENTS ON TOPIC AREA 1: INTERPRETATIONS OF THE
                  EPIDEMIOLOGY AND TOXICOLOGY. LITERATURE
                                     v
This section summarizes the panelists' discussions on the interpretations of the epidemiology and
toxicology literature. The meeting co-chairs—Dr. McClellan and Dr. Stayner—facilitated the
discussions on this topic area, which focused first on lung cancer (see Section 3,1) and then on
mesothelioma (see Section 3.2). This section presents a record of discussion of the topics mentioned
during the workshop. Several panelists referred to their premeeting comments (see Appendix B) for
additional suggestions for how the review of epidemiology and toxicology literature can be improved.

3.1   Lung Cancer

The panelists discussed at length whether the epidemiology and toxicology literature support the
proposed protocol's finding for how lung cancer potency varies with fiber type and fiber length. This
section summarizes these discussions, first on fiber type (Sections 3.1.1 and 3.1.2) and then on fiber
length (Sections 3.1.3 and 3.1.4). General issues regarding the lung cancer evaluation are presented in
Section 3.1.5.

      3.1.1   Lung Cancer and Fiber Type: Inferences from die Epidemiology Literature

According to the proposed risk assessment methodology,  amphibole fibers have a 5-fold greater lung
cancer potency than do chrysotile fibers. The panelists had differing opinions  on whether this finding is
consistent with the epidemiology literature. On the one hand, some panelists indicated that the
epidemiology  literature is consistent with amphibole fibers being more potent for lung cancer, though the
magnitude of this increase may not be known precisely. One panelist noted, for example, that multiple
analyses (e.g., Hodgson and Darnton 2000, Berman and Crump 2001, and the statistical analyses a
panelist presented during this discussion) all point to a consistent increased lung cancer potency for
amphibole fibers compared to chrysotile fibers, albeit a small increase. On the other hand, other
                                            3-1

-------
panelists did not believe the epidemiology literature supports this conclusion, for reasons stated below.
Finally, other panelists were not convinced that the epidemiology literature supports the higher lung
cancer potency for amphibole fibers, but they believed the difference in potency seems likely based on
evidence from the animal toxicology studies (see Section 3.1.3) and lung burden studies. A summary of
the panelists' discussion on this topic follows:                                          »
       Comments on specific publications. Several panelists cited specific studies to support their
       positions on the relative lung cancer potency of chrysotile and amphibole fibers, but the panelists
       often had differing opinions on the inferences Ihat should be drawn. The panelists mentioned the
       following specific studies:
                                                                                   j
       *•       Some panelists noted that a recent re-analysis of 17 cohorts (Hodgson and Damton
              2000) indicates that the lung cancer potency for amphibole fibers is 10 to 50 times
              greater than that for chrysotile fibers. One panelist did not agree with this finding, due to
              the crude approach the article uses to characterize relative potency. Specifically, this
              panelist noted that carcinogenic potency was calculated by dividing the overall relative
              risk for a given cohort by the average exposure for the entire cohort, even for cohorts
              where the data support more sophisticated exposure-response modeling. He was
              particularly concerned about the authors' decision to omit the cohort of South Carolina
              textile workers from the meta-analysis. This decision was apparently based on the
              South Carolina cohort being  an outlier, due to its much higher lung cancer potency
              when compared to other studies. The panelist noted, however, that the lung cancer risk
              for the South Carolina cohort is not unusually high when compared to other' cohorts of
              textile workers. The panelist  was concerned that omitting this study might have biased
              the article's finding regarding relative lung cancer potency. No other panelists discussed
              the review article.                                                     !.

       ••       One panelist cited a study  of Quebec chrysotile miners and millers (Liddell et al. 1997,
              1998) that reports that increased lung cancer risk was limited to the mining region with
              the highest level of tremolite asbestos, after correction for smoking and exposure. The
              article was distributed to the  panelists on the first day of the workshop, but no panelists
              commented further on the  study.

       »•       One panelist noted that his review of multiple textile cohorts (Stayner, Dankovic, and
              Lemen 1996) found relatively small differences in lung cancer potency, even though
              some of the cohorts were exposed to asbestos mixtures containing different proportions
              of amphibole fibers.
                                             3-2

-------
-       One panelist indicated that further evidence on how fiber types relates to lung cancer
        potency can be gleaned from epidemiological studies that were not included in the
        meta-analysis due to inadequate exposure data for exposure-response modeling.
        Examples include a study of non-occupationally exposed women from two chrysotile
        asbestos mining regions (Camus et al. 1998) and a study of railroad workers employed
        by shops that processed different proportions of amphibole fibers (Ohlson et al. 1984).
        Bo1h studies, she noted, provide evidence that amphibole fibers exhibit greater lung
        cancer potency. This panelist added that studies of auto mechanics have provided no
        convincing evidence of increased lung cancer due to chrysotile exposure, though she
        acknowledged that Ihe absence of an effect might reflect the short fiber length in the
        friction brake products. One panelist cautioned about inferring too much from these
        studies regarding fiber type because they were not controlled for other factors, such as
        fiber lengtii and level of exposure.

••       One panelist added that a recent study of a cohort of Chinese asbestos plant workers
        (Yano et al. 2001) should be considered in future  updates to the proposed protocol;
        the workers in the cohort had increased risks for lung cancer and were reportedly
        exposed to "amphibole-free" chrysotile asbestos. However, another panelist cited a
        publication (Tossavainen et al. 2001) that indicates that asbestos from many Chinese
        chrysotile mines actually does contain varying amounts of amphibole fibers.

        Several panelists noted that the proposed protocol's meta-analysis found a 5-fold
    j   difference in lung cancer potency between  amphibole and chrysotile fibers. However,
        other panelists indicated that the reported difference was not statistically significant
        Some panelists had additional reservations about the authors' meta-analysis, as
        summarized in the following bulleted items.

Comments on the meta-analysis approach. Several panelists commented on alternate
approaches the authors could have used to conduct their meta-analysis of the epidemiology
studies. One panelist noted that the lung cancer potencies reported by the various studies exhibit
considerable heterogeneity. In such cases, meta-regression is conventionally used to identify
which factors account for the variability in the results (i.e., in the lung  cancer potencies). This
panelist suggested that Ihe meta-analysis should have considered other factors in addition to
fiber type and dimension; such other factors could include industry, follow-up time for the
cohort, and estimated percentage of amphibole fibers in Ihe exposures, to the extent that data on
these other factors  are available.

To demonstrate how more detailed investigation might reveal further insights, one panelist
presented his  own initial statistical analysis of the epidemiological studies. This analysis used a
fixed effects model and a random effects model, both inverse weighted by Ihe variance of the
studies. His analysis examined how industry and fiber type contribute to the heterogeneity

                                       3-3

-------
observed among the cohorts and found that the industry of the cohort appears to be a stronger
predictor than fiber type. The panelist explained that the purpose of displaying his statistical
analysis was to highlight how other approaches to conducting meta-analysis can offer';different
insights on the epidemiological data. This panelist recommended that the authors conduct similar
meta-regression analyses to investigate the importance of various variables on the lung cancer
potency.                                                                    ;

This panelist also demonstrated how a sensitivity analysis might yield additional information on
influential studies. Using a fixed effects model, the panelist first showed how lung cancer potency
factors (KL) vary with exposure to chrysotile fibers, amphibole fibers, and mixed fiber types.
When all epidemiological studies were considered in his analysis, the amphibole fibers were
found to be three times more potent than the chrysotile fibers. When the cohort of chrysotile
miners and millers from Quebec was omitted from this analysis, however, the amphitiole fibers
were found to be nearly two times less potent than the chrysotile fibers. Conversely, 'when the
cohort of textile workers from South Carolina was omitted, Ihe amphibole fibers were found to
be more than ten times more potent than the chrysotile fibers. Given that the conclusions drawn
about the relative potency of chrysotile and amphibole fibers appear to be highly sensitive to
whether single studies are omitted from the analysis, this panelist was more skeptical jabout
whether the increased potency of amphibole fibers is a robust finding. He recommended that the
authors, when completing the proposed protocol, conduct similar sensitivity analyses^ help
reveal the factors or studies that appear to contribute most to' lung cancer.

Another panelist agreed with this feedback, and provided further comments on the meta-
analysis, noting that these analyses typically start with establishing criteria for study inclusion
After selecting studies to evaluate, she said, various statistical analyses can be used to test
hypotheses and to understand the concordance and  disparity among the individual studies.  The
panelist thought such an approach is needed to help understand the variability in potency factors
observed across the multiple studies and to identify for further analysis the studies found to be
most descriptive of exposure-response. To  clarify the authors' approach, Dr. Berman indicated
that the meta-analysis considered any published epidemiological study with sufficient :quantitative
exposure data that allowed for a reasonable estimate of the exposure-response relationship;
uncertainty factors were than assigned to give greatest weight to the most robust studies. In
response, additional  panelists concurred with the original comment that meta-analyses
conventionally begin with establishing explicit study inclusion criteria These panelists iclarified
that they are not advocating removing a majority of studies currently considered in the proposed
protocol, but rather being more judicious in selecting the studies to evaluate.

One panelist offered additional comments on the meta-analysis. He supported, for instance, the
use of sensitivity analyses, and encouraged the authors to conduct additional analyses to identify
influential studies, factors'that contribute to risk, and the impact of different weighting factors.
The panelist also rioted that more sophisticated statistical methodologies (e.g., Bayesian
                                       3-4

-------
•
	_„,	,	0	r	,	
discrete values, which might offer greater understanding of the inferences that can be drawn from
the epidemiological studies.

Disparate findings from the South Carolina and Quebec cohorts. Multiple panelists noted
that the issue of the relative lung cancer potency of chrysotile and amphibole fibers depends
largely on how one interprets the disparate findings from the cohort of textile workers in South
Carolina and 1he cohort of chrysotile miners and millers in Quebec. Two of these panelists
indicated that the relative potency issue likely will not be resolved until the underlying reasons for
the differences between these two studies are better understood The other panelist viewed the
difference in potency observed across industries (i.e., mining versus textile) as a more important
matter than the difference between the two specific cohorts. When discussing these studies, two
panelists indicated that the increased lung cancer risk for the South Carolina cohort might be
attributed to exposure to amphibole fibers, which are known to be found in trace levels in
commercial chrysotile.

Relevance of fiber durability. One panelist noted that the issue of fiber durability often enters
the debate on the relative lung cancer potency of chrysotile and amphibole fibers. Though he
agreed that the animal toxicology data indicate that amphibole fibers are more persistent than
chrysotile fibers, the panelist noted that trends among the human epidemiological
data—particularly the fact that lung cancer risk does not appear to decrease with time since last
exposure, even for chrysotile—suggest that the lower durability of the chrysotile fibers might not
be important.

Influence of smoking. The panelists had differing  opinions on how the proposed protocol
should address cigarette smoking. In terms of inferences drawn from the epidemiological
literature, two panelists noted that very limited data are available on smoking, making
quantitative analysis of its interactions with asbestos exposures difficult. Specifically, only one
study includes detailed information on smoking, but that study found no difference in lung cancer
potency between smokers and non-smokers. During this discussion, Dr. Berman explained that
the proposed protocol assumes a multiplicative interaction between smoking and asbestos
exposure, consistent with EPA's 1986 model. Dr. Berman noted that a multiplicative factor in
the model, a, represents the background risk in the studied cohort relative to the risk in the
comparison population, and both groups include smokers; he added that the influence of
smoking is addressed implicitly in the model because it is a relative risk model in which the effect
of asbestos is multiplied to the background risk that is present. A panelist clarified, however, that
neither the potency factors nor a were derived based on observations of smoking prevalence in
the epidemiological studies.

One panelist emphasized that the confounding effects of smoking greatly complicates the analysis
of lung cancer potency. He noted that the relative lung cancer risk from asbestos exposure is

                                       3-5

-------
       considerably lower than that for cigarette smoking. As a result, the panelist wondered how the
       meta-analysis can truly discern the relative potency of the asbestos fiber types from studies that
       present no information on cigarette smoking. This panelist provided an example to illustrate his
       concern: if a given cohort has between 5 and 10% more smokers than the typical population,
       this increased prevalence of smoking alone could totally confound relative risks attributed to
       asbestos. The panelist indicated that all future analyses of epidemiological data will suffer from
       similar limitations, so long as detailed information on smoking is not available.       ''
       General comments.  During ibis discussion, some panelists offered several general comments
       that apply to the entire proposed protocol. These comments included concerns about the
       transparency of the analyses, questions about data tables being inconsistent with text in the body
       of the report and some panelists' inability to reproduce certain findings from the available data.
       These general comments are reflected in the  executive summary of this report,      :
       3.1.2   Lung Cancer and Fiber Type: Inferences from Animal Toxicology and
              Mechanistic Studies
The panelists offered varying insights on the inferences that can, or should, be drawn from animal
toxicology studies and mechanistic studies regarding the relative lung cancer potency for chrysotile and
amphibole fibers.

Citing various publications (e.g., Lippmann 1994), multiple panelists noted that the animal toxicology
studies do not support the 5-fold difference in lung cancer potency between chrysotile and amphibole
fibers. Two panelists added that the absence of different potencies might result from Ihe animal studies
being of too short duration (typically no longer than 2 years) for the greater dissolution of chrysotile
fibers to be an important factor. Another panelist added feat exposure levels in some animal studies  are
not relevant to human exposures; as an example, he noted that a recent rat inhalation study .(Hesterberg
et al. 1998) involved exposure levels at 11,000 fibers per cubic centimeter. These panelists indicated
that the animal studies are generally more informative of how lung cancer potency varies with fiber
length (see Section 3.1.4), and are less informative on how potency varies with fiber type.
                                             3-6

-------
The panelists noted that in vitro studies exhibit various findings, depending on the study design and
endpoint assessed. One panelist, for instance, indicated that some in vitro studies suggest that
chrysotile fibers are actually more potent than amphibole fibers. Other panelists added that many in
vitro studies show crocidolite being considerably more toxic than chrysotile. These panelist cautioned
against drawing firm conclusions from the in vitro studies, however, given that Ihe study duration is far
too short for any impact of dissolution to be observed Finally, another panelist referred to the
International Agency for Research on Cancer (IARC) consensus statement on fiber carcinogenesis for
an overview of inferences that can be drawn from mechanistic studies: "Overall, the available evidence
in favor of or against any of these mechanisms leading to the development of lung cancer and
mesolhelioma in either animals or humans is evaluated as weak" (IARC 1996).

Based on the previous comments., the panelists cautioned about attempting to draw inferences from the
animal toxicology for several reasons. One panelist indicated that the animal studies have limited utility
because lung cancer in humans results from a complex set of exposures, including cigarette smoke, and
because rats, when compared to humans, develop different types of tumors at different sites. Another
panelist reiterated that the duration of most animal studies precludes one from observing dissolution
effects. Given these limitations, two panelists emphasized that conclusions should be based primarily on
die epidemiological data, especially considering the volume of human data that are available. Though
not disagreeing with this recommendation, one panelist noted that the exposure index—one of the
major outcomes of the proposed protocol—is, in fact, based on observations from animal studies.
       3.1.3   Lung Cancer and Fiber Dimension: Inferences from the Epidemiology
               Literature
 The panelists made several observations regarding what can be inferred from the epidemiology
 literature on how lung cancer potency varies with fiber dimension, though they first noted that most
 published epidemiology studies do not include detailed data on the distribution of fiber dimensions to
 which cohorts were exposed. Overall, the panelists generally agreed that indirect evidence from the
                                             3-7

-------
epidemiological studies supports the proposed protocol's finding that longer fibers have greater

carcinogenic potency for lung cancer. They added, however, that the epidemiology literature provides

no evidence to support or refute tiie magnitude of Hie relative potencies used in the proposed protocol

(i.e., fibers longer than 10 jim being 300 times more potent than those wife lengths between 5 and 10
                                                                                  «.'
urn). The panelists made no comments about fiber diameter when discussing this matter. Specific

discussion topics follow:
       Observations from the epidemiology literature. The panelists identified several studies that
       provide general insights on the role of fiber size in lung cancer. One panelist, for instance, noted
       that cohorts of textile workers, which were believed to be exposed to relatively longer asbestos
       fibers, exhibit higher lung cancer relative risks than do cohorts of miners or cement product
       workers. Another panelist indicated that studies of taconite miners from Minnesota (Cooper et
       al. 1988) and gold miners from South Dakota (McDonald et al. 1978) found no increased lung
       cancer risks among the cohorts,  which were known to be exposed primarily to fibers shorter
       than 5 jim (see Dr. Case's premeeting comments for further information on these studies). This
       panelist added that the Minnesota Department of Health is currently updating the study on
       taconite miners and a publication is pending. Another panelist added that epidemiology studies
       of workers exposed to asbestos  from friction brake products  show no clear evidence of
       increased lung cancer. This panelist acknowledged that these epidemiology studies do not
       include exposure measurements, but other studies of this work environment have indicated that
       the asbestos fibers in friction brake products are predominantly short chrysotile fibers.
      Relevance of fibrous structures shorter than 5 ^m. Some panelists noted thatnb
      epidemiology studies have examined the relative potency specifically of fibrous structures shorter
      than 5 jura, thus no conclusions could be drawn from the epidemiology studies alone; While not
      disagreeing with this observation, one panelist reminded panelists that airborne particles and
      fibers have a broad distribution of fiber lengths, with a clear majority (75-90%) of fibrous
      structures being shorter than 5 |im. This panelist added that indirect inferences can tie drawn
      from the epidemiology studies listed in the previous bulleted item. Another panelist noted that the
      fibrous structures shorter than 5 ^m behave more like particles rather than fibers, at least in
      terms of lung deposition and clearance patterns.  Finally, two panelists indicated that an ATSDR
      expert panel recently evaluated the issue of relative potency of fibers shorter than 5 urn;
      however, the final report from that expert panel meeting was not available until after the peer
      consultation workshop. The final report has since been released, and a conclusion from that
      panel was that "there is a strong weight of evidence that asbestos and synthetic vitreous fibers
      shorter than 5  ^m are unlikely to cause cancer in humans" (ERG 2003).
                                             3-8

-------
      Statistical analyses in the proposed protocol As indirect evidence that longer fibers have             ^^
      greater carcinogenic potency, one panelist indicated that the exposure-response modeling by
      Drs. Berman and Crump showed an improved fit to the observed relative risk from
      epidemiology studies when using an exposure index that assigns greater weight to longer fibers
      and no risk to fibers shorter than 5 jim Another panelist concurred, but added that the authors
      could have attempted to determine the specific weighting (i.e., between longer and shorter
      fibers) that would optimize the fit to the epidemiological studies.
      3.1.4   Lung Cancer and Fiber Dimension: Inferences from Animal Toxicology and
              Mechanistic Studies
The panelists generally agreed that the animal toxicology studies and mechanistic studies indicate that

fiber dimension—especially fiber length—plays an important role, both in terms of dosimetry and

palhogenesis. However, panelists had differing opinions on the specific cut-offs that should be used for

fiber diameters and lengths in the exposure-response modeling (though panelists generally concurred

that fibers shorter than 5 jim should be assigned zero potency).
      Fiber length. Multiple panelists noted that the animal toxicology studies provide compelling
      evidence that lung cancer potency increases with fiber length. Another panelist agreed, but had
      reservations about assigning no potency to fibrous structures shorter than 5 jim, based on a
      recent study of refractory ceramic fibers (Bellman et al. 2001) that found that the incidence of
      inflammation and fibrosis appears to be related to the presence of small fibers in the lung. This
      panelist indicated that exposure to small fibers likely has some bearing on the oxidative stress
      state and inflammation in the lung, and he suspected that the exposure-response relationship for
      long fibers might depend on co-exposures or past exposures to shorter fibers. Based on these
      observations, the panelist was hesitant to exclude fibrous structures shorter than 5 ym from the
      proposed risk assessment methodology. On the other hand, another panelist added that animal
      toxicology studies have shown that fibrosis endpoints are strongly related to fiber length, with
      exposures to shorter fibers showing less evidence of fibrosis or lung damage. The panelists
      revisited the significance of fibers shorter than 5 nm when discussing the proposed exposure
      index (see Section 4).

      Fiber diameter. The panelists offered several comments on the role of fiber diameter in the
      proposed protocol. Noting that fibers with diameters up to 1.5 ^m are capable of penetrating to
      sensitive portions of the lung during oral inhalation, one panelist indicated that this range of fiber
      diameters should not be excluded from future risk assessments. Other panelists shared the

                                             3-9

-------
       concern of assigning no lung cancer potency to respirable fibers with diameters greater than 0.5
       nm, especially considering that respirahility patterns in laboratory animals differ from'those in
       humans (i.e., thicker fibers are more likely to deposit in the human lung than they areiin the rat
       lung).                                                                       j

       The panelists also discussed a statement in the proposed protocol that "few fibers thicker than
       0.7 nm appear to reach the deep lung." First, one panelist indicated that the proposed protocol
       includes outdated information on fiber deposition patterns; he recommended that the authors
       obtain more current insights from specific publications (e.g., Lippmann 1994) and from the latest
       lung dosimetry model developed by the International Commission on Radiological Protection.
       Second, another panelist questioned the relevance of deposition in the deep lung, because
       humans tend to develop bronchogenic carcinomas, while rats develop  bronchoalveojar
       carcinomas. Anolher panelist cautioned against inferring that asbestos fibers must deposit on
       bronchial airways to cause lung cancer in humans, noting that significant accumulation of
       asbestos fibers does not occur in Ihe airways where carcinomas develop in humans, due
       primarily to mucociliary clearance; this panelist suspected that deposition of fibers in the deep
       lung is likely related to lung cancer formation in humans, though the mechanisms of  ':
       carcinogenesis are not fully understood.
      3.1.5   Other Issues Related to Lung Cancer
The panelists discussed several additional issues related to the proposed protocol's evaluation of lung
cancer potency. Most of the discussion focused on the utility of non-linear exposure-response
modeling, but other topics were also addressed:
      Consideration of non-linear exposure-response models. The panelists had differing
      opinions on the extent to which the proposed protocol should consider non-linear exposure-
      response modeling. On Ihe one hand, one panelist strongly recommended that EPA consider
      exploring the applicability of non-linear exposure-response models, given his concerns with
      linear low-exposure extrapolation. This panelist acknowledged that the revised linear; model in
      the proposed protocol clearly provides an improved statistical fit to the epidemiological data
      when compared to EPA's 1986 lung cancer model, but he advocated more detailed1 exploration
      of non-linear cancer risk models,  particularly to account for observations of cohorts with low
      exposures. This panelist was particularly concerned about the cancer risks that woulcl be
      predicted for low exposures: because the slope in any  linear lung cancer model will be
      determined largely by highly-exposed individuals, he questioned whether the slope derived from
                                            3-10

-------
high exposures truly applies to lowly-exposed individuals. To demonstrate his concern, this
panelist indicated that the epidemiological studies consistently show that cohorts (or subsets of
cohorts) with low exposure generally exhibit no increased lung cancer risk (standardized
mortality ratios not statistically different from 1.0). To account for the possibility of a threshold
or non-linearity in the exposure-response relationship, this panelist recommended that EPA
investigate alternate exposure-response models, such as linear-linear models (i.e., models with
two linear exposure-response regions having different slopes) or log-linear models.

Other panelists generally supported these comments. One panelist, for instance, noted that
EPA's Draft Revised Guidelines for Carcinogen Risk Assessment indicates that exposure-
response relationships should first be evaluated over the range of exposure observations, and
then various approaches to extrapolate to exposure levels outside (i.e., below) this range should
be investigated. Another panelist added that some studies finding no evidence of lung cancer
risks among large cohorts with low exposures should factor into the decision of whether the lung
cancer model should include thresholds; he cited a study of non-occupationally exposed women
from chrysotile mining regions in Canada (Camus et al. 1998) to illustrate his concern. Other
panelists noted that the utility of this study is limited, because exposures were not measured for
individuals; further, a panelist clarified that approximately 5% of the individuals considered in this
study were occupationally exposed. Finally, one panelist indicated that evidence from the
epidemiology literature strongly suggests there are asbestos exposure levels below which lung
cancer will not occur; ihis panelist added that he is unaware of any epidemiological study that
has found evidence of lung cancer risk at exposure levels below 25 fiber-years. He
recommended that the proposed protocol at least acknowledge the lowest exposure level at
which lung cancer effects have been demonstrated.

On the other hand, some panelists were not convinced of the utility of conducting detailed
analyses at low exposures and investigating possible thresholds.  One panelist, for instance,
indicated that a meaningful quantitative analysis of potential thresholds will not be possible, so
long as the authors do not have access to raw data from additional epidemiological studies.
Further, this panelist suspected that the protocol authors would find considerable heterogeneity
among exposure-response slopes for low exposures, and he questioned what conclusions could
be drawn by focusing exclusively on the low exposure region Anolher panelist agreed, adding
that the failure to find significantly increased cancer risks among lowly-exposed cohorts very
likely results from poor statistical power and other uncertainties, and not necessarily from the
presence of an actual exposure threshold for asbestos-related lung cancer. Finally, one panelist
indicated that the National Institute for Occupational Safety and Health (N1OSH) previously
examined a threshold model for the cohort of South Carolina textile workers, and that analysis
revealed that the best fit of the exposure-response  data was a threshold of zero (i.e., the best fit
indicated that there was no threshold).
                                       3-11

-------
       Consideration of cigarette smoking. Several times during the workshop, the panelists
       debated the ability of the proposed risk assessment model to address interactions between
       cigarette smoking and asbestos exposure One panelist recommended that the authors review a
       recent study that examined the role of cigarette smoking on lung cancer among chrysptile miners
       and millers in Quebec, Canada (Liddell and Armstrong 2002). Although the panelists generally
       agreed that smoking is an important consideration for developing and applying the model, some
       panelists were not convinced that the available data are sufficient to develop an exposure-
       response model that accurately portrays ihe interactive effects of asbestos exposure and
       smoking. The panelists further discussed this issue further later in the workshop.    !

       Transparency of the proposed protocol Several panelists indicated that the review of
       epidemiological data in Ihe proposed protocol is not presented in a transparent fashioa One
       panelist, for instance, sought more information on the uncertainty factors used in the meta-
       analysis, such as what ranges of factors were considered, what criteria were used to assign the
       factors, and a table of the factors that were eventually applied. This panelist also recommended
       that the proposed protocol  identify the a-values that were determined for each epidemiological
       study and provide explanations for any cases when these values are unexpectedly large. Another
       panelist indicated that the proposed protocol should more clearly differentiate conclusions that
       are based on  a meta-analysis of many epidemiological studies from conclusions that;are based
       on a detailed  review of just one or two studies.            •                     ,,

       The need to obtain additional raw data sets. The panelists unanimously agreed that EPA
       should make  every effort to try to obtain additional raw data sets for  the epidemiology studies,
       such that the  authors can further test how adequately the proposed risk assessment rnodel
       predicts risk.  The executive summary of this report presents the panelists' specific  I1
       recommendation on this issue.
3.2   Mesothelioma


The following paragraphs document the panelists' responses to charge questions regarding inferences
from tiie epidemiology and toxicology literature on how mesothelioma potency varies with fiber type

(Sections 3.2.1 and 3.2.2) and fiber length (3.2.3 and 3.2.4).


      3.2.1   Mesothelioma and Fiber Type: Inferences from the Epidemiology Literature
                                            3-12

-------
The expert panelists unanimously agreed that Ihe epidemiology literature provides compelling evidence
Ihat amphibole fibers have far greater mesothelioma potency than do chrysotile fibers—a finding
reported both in the review document (Berman and Crump 2001) and a recent re-analysis of 17 cohort
studies (Hodgson and Darnton 2000) that reported at least a 500-fold difference in potency. Two
panelists commented further that the epidemiology literature provides no scientific support for chrysotile
exposures having a role in causation of mesothelioma—an observation that is generally consistent with
the meta-analysis in Ihe proposed protocol, which failed to reject the hypothesis that chrysotile fibers
have zero potency for mesothelioma


The most notable response to this charge question was the agreement among most panelists that
amphibole fibers are at least 500 times more potent than chrysotile fibers for mesothelioma, as
supported by two separate reviews of epidemioldgical studies.-The panelists made additional comments
on specific matters when responding to this question, as summarized below, but the key point in this
discussion was the agreement that chrysotile is a far less important cause of mesothelioma than are
amphiboles.
      Relative roles of chrysotile and amphibole. One panelist indicated that cohort studies with
      individual-level exposure-response data and the broader epidemiology literature both provide no
      evidence of increased mesothelioma risk due to chrysotile exposure. Further, this panelist noted
      that 33 of 41 mesothelioma cases previously identified as occurring among workers primarily
      exposed to chrysotile fibers (Stayner et al. 1996) were later reported as likely resulting from
      exposures to tremolite fibers found in the chrysotile mines (McDonald et al.  1997). This panelist
      noted that a recent finding of a small mesothelioma risk from chrysotile (Hodgson and Damton
      2000) results entirely on the assumption that the 33 mesothelioma cases mentioned above result
      entirely from chrysotile exposures. Based on these observations, this panelist indicated that the
      literature suggests Ihat chrysotile exposures have limited, if any, role in causing mesothelioma He
      nonetheless supported the relative potency attributed to chrysotile in the proposed protocol as a
      conservative measure in the overall risk assessment process.

      Specific comments on the Connecticut friction products workers. Another panelist
      commented on an epidemiological study of a cohort of workers employed at a friction products
      plant in Connecticut The panelist noted that the original study (McDonald et al. 1984) did not
      identify any deaths from mesothelioma, but review of the state cancer registry (Teta et al. 1983)

                                            3-13

-------

       revealed that three Connecticut residents who died of mesothelioma were employed'by 1he
       same friction products company. One of these employees had amphibole exposures during the
       time he worked for a textile plant that was under the same parent company that owned and
       operated the friction products plant The other two cases, the panelist noted, were females who
       indeed worked at the friction products plant. A pathology review found that one of these cases
       was a woman with probable pleura! mesothelioma and 5 years of exposure; the other case was
       a peritoneal mesothelioma in a woman who also had asbestosis, and worked as a clerk for 30
       years. This panelist noted that it was questionable to attribute the latter two mesothelioma
       diagnoses to the chrysotile exposures at the friction products plant, though she added that this
       possibility cannot be definitively ruled out This panelist encouraged that future review of this
       epidemiological study should be revised given this new information.                j,

       Comments on the proposed 500-fold difference in relative potency. The panelists had
       several comments on the finding in the proposed risk assessment methodology that amphibole
       fibers are 500 times more potent for mesoHielioma than are  chrysotile fibers. Several, panelists
       noted that this finding is consistent with that of a recent re-analyses of 17 epidemiological studies
       (Hodgson and Darnton 2000). Though not disagreeing that amphibole fibers are clearly more
       potent, one panelist was concerned that the risk coefficients (KM) were largely derived from
       data sets with inadequate exposure-response information for mesothelioma,  and assumptions
       had to be made to determine critical inputs to the mesoflielioma model (e.g., average exposure,
       duration of exposure).                                                       i;

       Other panelists commented on specific sections in the proposed protocol. One panelist, for
       example, recommended that the authors check the accuracy of data presented in Table 6-16
       and Table 6-29 of the report, which are not  reported consistently. Another panelist suggested
       that the authors better explain why separate risk coefficients for amphiboles  and chrysotile were
       calculated for some cohorts (e.g., Hughes et al. 1987) but not for others (e.g., Berry'and
       Newhouse 1983), even though the exposure information available for the studies appears to be
       comparable. Finally, one panelist recommended that the authors of the proposed protocol
       consider questions recently raised (Rogers and Major 2002) about the quality of the exposure
       data originally reported for the Wittenoom cohort (De Klerk et al. 1989) when evaluating
       exposure-response relationships for mesothelioma.         ,                     ;
      3.2.2   Mesothelioma and Fiber Type: Inferences from Animal Toxicology and
              Mechanistic Studies

The panelists discussed die inferences provided by animal toxicology data and mechanistic data

regarding relative mesolhelioma potency of different asbestos fiber types. Overall, two panelists
                                            3-14

-------
commented lhat 1he human epidemiological data clearly establish that exposures to amphibole asbestos
fibers pose a greater mesothelioma risk than do exposures to chrysotile fibers. They added that the
animal toxicology data are generally supportive of this finding, but the animal data suffer from some
limitations. Two panelists, for instance, noted that the utility of animal toxicology studies is United by the
fact that rodents are rather insensitive to mesothelioma. These panelists added that the animal
toxicology studies involving intra-tracheal instillation or peritoneal injection are not directly relevant to
tiie inhalation exposures that occur in humans. These limitations notwithstanding, Hie panelists raised the
following points when discussing the animal toxicology and mechanistic studies:

One panelist referred to one of his earlier publications (Lippmann 1994) for further insights on the
occurrence of mesothelioma in animal studies. At that time, this panelist noted, Ihe animal inhalation
studies found fewer than 10 cases of mesothelioma, and the number of cases appeared to be greatest
among animals that were exposed to mixtures containing higher proportions of amphibole fibers. He
found Ihis consistent with the influence of fiber type observed in the human epidemiological data (see
Section 3.2.1).

During this discussion, one panelist reviewed a publication (Suzuki and Yuen 2001) that was mentioned
earlier in the workshop. The publication documents the amounts and types of asbestos fibers measured
in samples of pleura! plaques and tumor tissue collected for legal cases. These analyses reportedly
found relatively large amounts of short, thin chrysotile fibers in the pleura, suggesting that these fibers
should not be excluded from the group of fibers believed to induce mesothelioma The panelist had
several criticisms of the study. First, he indicated that the samples  were analyzed vising anon-standard
technique, without any controls. Second, he questioned the major finding of fibers being detected in the
pleura, because most of the samples analyzed were actually tumor tissue, in which he would not expect
to find fibers. The panelist suspected that the chrysotile fibers reportedly found in the study likely result
from specimen contamination—a bias that would have been more apparent had rigorous quality control
procedures been followed. Finally, the panelist noted that a more  rigorous study (Boutin et al. 1996) of
                                             3-15

-------
asbestos fibers in the parietal pleura found a mixture of fibers, including long amphibole fibers, among
living patients with asbestos-related conditions. Based on ihese concerns, Ihe panelist concluded that
the publication of concern (Suzuki and Yuen 2001) is seriously flawed and its recommended should be
excluded from EPA's analyses.                                                     !

A specific issue raised regarding the analytical technique in Ihe study (Suzuki and Yuen 2001) was that
                                                                                 {.
water was used during die digestion process. Noting that water may contain large amounts (>30,000
fibers/L) of small asbestos fibers, another panelist suspected that the fibers detected in the study might
have resulted from contamination introduced during the digestion process. Because control samples
                                                                                 i'
were not analyzed, Ihe panelist said the study offers no evidence lhat the fibers detected truly' were in
the original pleura! plaques or tumor tissues. He added lhat studies of lung-retained asbestos fibers
                                                                                 i.
routinely detect primarily short, chrysotile fibers, and that the presence of the short fibers in the pleura!
tissue—even if the measurements from the study are valid—would not necessarily prove lhat short
fibers cause mesothelioma
                                                                                 .1
                                                                                 I
       3.2.3   Mesothelioma and Fiber Dimension: Inferences from the Epidemiology
              Literature

The panelists commented briefly on how the human epidemic-logical data characterize the role of fiber
size on mesothelioma risk Noting that exposure measurements in most every epidemiological study do
                                                                                 i
not characterize fiber length distribution, one panelist indicated that these studies provide no: direct
evidence of how fiber length is related to mesothelioma He added that fee studies ofTer'conflicting
indirect evidence of the role of fiber length. Specifically, the higher mesothelioma risk coefficient among
textile workers in South Carolina, when compared to that for the chrysotile miners and millers in
Quebec, could be supportive of longer fibers beirig more potent, since exposures in South Carolina had
a larger percentage of long fibers. However, a cohort of cement plant workers in New  Orleans was
found to have a higher mesotheliorna risk coefficient than that of the South Carolina cohort, even though
the South Carolina workers were exposed to higher percentages of long fibers. Finally,  as indirect
                                            3-16

-------
evidence that carcinogenic potency increases with fiber length, this panelist noted that the mesolhelioma
risk model using the proposed exposure index, which is heavily weighted by long fibers, provided a
considerably improved fit to the epidemiological data        .

The panelists briefly revisited the inferences that can be drawn from studies of lung-retained fibers. One
panelist again commented that results from a recent study (Suzuki and Yuen 2001) should be viewed
wilh caution. He added that several other lung pathology studies (e.g., McDonald et al. 1989, Rogers et
al. 1991, Rodelsperger et al. 1999) have been conducted using more rigorous methods, such as using
appropriate controls for age, sex, and hospital. These studies all showed that risk of mesolhelioma was
considerably higher for individuals wilh larger amounts of long fibers retained in their lungs.

One panelist indicated that results from a study of lung-retained fibers (Timbrel! et al. 1988) suggest
fiber diameter plays a rule in mesothelioma risk: the study observed no mesothelioma cases among a
population highly exposed to anthophyllite fibers, which tend to be thicker fibers. Citing his earlier
review of mesolhelioma cases (Lippmann 1988), the panelist also noted that crocidolite fibers are both
thinner than and more potent than amosite fibers, which further supports the hypothesis that
carcinogenic potency for asbestos decreases with increasing fiber diameter.
       3.2.4   Mesothelioma and Fiber Dimension: Inferences from Animal Toxicology and
              Mechanistic Studies
The panelists made few observations on findings from animal toxicology studies regarding mesothelioma
and fiber length One panelist indicated that findings from the animal toxicology studies generally
support the overall finding that mesothelioma risks are greatest for long, thin fibers. However, another
panelist noted that his earlier review of mesothelioma risks (Lippmann 1988) hypothesized that the
critical fibers for mesolhelioma induction are those with lengths between 5 and 10 ^m. This panelist
added that fibers of this dimension are more likely to translocate to the pleura than are longer fibers, but
                                             3-17

-------
he acknowledged that it is unclear whether fibers must first translocate to the pleura in order to cause
mesothelioma                                                                    ].

Some panelists indicated that fiber durability likely plays a role in inducing mesothelioma, based on the
fact that mesothelioma is more easily induced in animals using administration methods (e.g.,iperitoneal
injection) that remove the importance of dissolution.

3.3    Exposure Estimates in the Epidemiology Literature
                                     <                       <
The panelists raised numerous issues when responding to the third charge question: 'To what extent are
the exposure estimates documented in the asbestos epidemiology literature reliable?" Recognizing that
the exposure estimates from the epidemiology studies are critical inputs to the exposure-response
assessment, the panelists expressed concern about the exposure data: few studies provide detailed
information on fiber size distribution; many studies report exposures using outdated sampling and
analytical methodologies (e.g., midget impinger); individual-level data are not available for most studies;
and many studies do not report detailed information on parameters (e.g., exposure levels, exposure
duration) needed to evaluate exposure-response relationships, particularly for mesothelioma Their
                                                                                  {
specific concerns on these and other matters follow:
      Concerns regarding exposure estimates in specific studies. Some panelists expressed
      concern about the assumptions made to interpret the exposure data originally reported in the
      epidemiology studies. One panelist reviewed specific examples of these concerns:   'j
      »       The original study of workers at a Connecticut friction products plant (McDonald et al.
              1984) reports exposures measured by midget impingers (in units of mmpcf),1 with no
              information on how to convert this to PCM measurements, and the original publication
              includes limited data on exposure duration.                             '•
      »       The original study of workers at a New Jersey insulation factory (Seidman et al. 1986)
              did not report any exposure measurements from the factory studied, and data collected
                                            3-18

-------
       from another plant with similar operations were used to characterize exposure-response
       for this cohort

«•      The original study of workers at a Texas insulation factory (Levin et al. 1998) reported
       a range of exposure levels (15-91 fibers/mL), and the authors of the proposed protocol
       assigned an average exposure level (45 fibers/mL) to the entire cohort.

*      The original study of U.S. insulation applicators (Selikoff and Seidman 1991) has no
       information on exposure. The proposed protocol assumes that all workers were
       exposed to 15 fibers/mL for 25 years, based on a separate review of exposures among
       insulation workers (Nicholson 1976).

»>      The original study of retirees from the U.S. Asbestos Products Company (Enterline et
       al. 1986) reported exposures based on midget impinger sampling, with no information
       on how to convert these exposures to PCM measurements.

••      According to a recent letter to the editor (Rogers and Major 2002), the original study
       of the Wittenoom cohort (De Klerk et al. 1989) might have overestimated exposures,
       possibly by as much as a factor of 10.

The previous comments led to a discussion on whether certain studies should be excluded from
the meta-analysis used in the proposed protocol (see next bulleted item). Prior to this discussion,
one panelist expressed concern about being overly critical of the exposure estimates used for
many of the studies listed above; he emphasized that all exposure estimates appear to be based
on a critical review of the literature, and no estimates are completely arbitrary, as some of the
panelists' comments implied.

Comments on using study inclusion criteria for the meta analysis. Given die concerns
about Ihe quality of exposure data reported in some epidemiology studies, the panelists  debated
whether future revisions of 1he proposed protocol should exclude certain studies from the
exposure-response analysis. The panelists were divided on this matter.

On Ihe one hand, several panelists recommended that the authors develop and apply study
inclusion criteria in the exposure-response evaluation, as is commonly done when conducting a
meta-analysis. One panelist, for instance, recommended assessing exposure-response
relationships for only those studies found to have adequate exposure data, and then using a
sensitivity analysis to examine the effect of excluding studies with inadequate exposure data
These panelists clarified that they are not advocating disregarding the majority of studies; rather,
they are suggesting simply that the authors of the proposed protocol use study inclusion criteria
and sensitivity analyses to ensure that the conclusions are based on the best available exposure
data

                                      3-19

-------
On the other hand, several panelists supported 1he current approach of using as many studies as
possible and accounting for the quality of the exposure measurements in the uncertainty factors.
One panelist, for example, commended the authors for being as inclusive as possible'when
reviewing the studies; he supported the approach of recognizing the limitations of the available
exposure data and accounting for these limitations in the uncertainly factors that were ultimately
used to weight the studies in the meta-analysis. This panelist acknowledged that the exposure
estimates in some of die epidemiological studies might be rough estimates, but he emphasized
that the estimates are not worthless and should not be discarded. Other panelists concurred with
these comments, and did not support applying overly restrictive study inclusion criteria.
                                                                            i
Comments on the uncertainty factors assigned to each study. The panelists made several
comments on the uncertainty factors that the authors assigned to each study. Dr. Herman first
explained the four uncertainly factors: the first factor (Fl) characterizes the confidence in
exposure estimates; the second factor (F2) represents the confidence in the conversion to PCM
measurements from other exposure metrics (typically midget impinger analyses); the tod factor
(F3) characterizes the confidence the authors had on worker history data; and the fourth factor
(F4) was a non-exposure related factor to account for other uncertainties (e.g., lack^of
information on confounders, incomplete or inaccurate mortality ascertainment). Dr. Berman
described generally how the individual uncertainty factors were assigned and noted that each
factor could range from 1 to 5.                                                !

The panelists' comments primarily focused on Ihe transparency of how uncertainty factors were
presented and incorporated into the meta-analysis. Multiple panelists, for instance,  '
recommended that future revisions to the proposed protocol include a table that lists jthe
uncertainty factors assigned to each study. Further, one panelist suggested that the revised
protocol describe the assumptions inherent in the uncertainty factor weighting approach, such as
explaining why some factors are assigned values over a broader range than others (e.g., why Fl
values span a broader range than F4 values) and describing why the individual uncertainty
factors have equal weights in generating the composite uncertainty factor. Another panelist
agreed, and added that the revised protocol should more explicitly describe how the] uncertainty
factors were combined  into the composite factor and how this composite factors affects the
weighting of studies in the meta-analysis.  Expanding on this point, another panelist suggested that
Ihe final document more clearly explain that the final estimates of cancer risk coefficients (K.L*
and KM*) are actually weighted averages of the epidemiological studies, with the weights
assigned to each study being a function of that study's uncertainty. This panelist also..
recommended that the revised document clearly state how, if at all, Ihe fraction of amphibole
fibers and Ihe fraction of fibers longer than 10 jim are reflected in the uncertainty factors.

Some panelists debated the utility of alternate approaches that could be used to assign
uncertainty factors. Two panelists noted that the approach used to assigning uncertainty factors
is somewhat subjective, because different groups of analysts would likely assign different
                                      3-20

-------
      uncertainly factors. To avoid the appearance of arbitrariness, these panelists suggested using
      alternate meta-analysis approaches that do not require using uncertainty factors. They noted, for
      example, that the authors could use a random effects model in which residual inter-study
      variation is estimated. Another suggestion was to conduct sensitivity analyses examining the
      effects of including or excluding studies, depending on the uncertainty factors assigned to them.

      Another panelist disagreed with these comments and supported the analyses in the proposed
      protocol; this panelist indicated that the authors had no choice but to make judgments based on
      the information documented in the epidemiology literature. He suggested that EPA consider
      convening a separate expert panel to assign uncertainty factors, if panelists do not support those
      selected by Drs. Berman and Crump.

      Assumptions made to convert exposure estimates from midget impinger sampling.
      Several panelists noted that the original publications for many epidemiology studies document
      exposure estimates based only on midget impinger sampling and do not include any information
      on how to convert these exposures to levels that would be measured by more modem methods
      (e.g., PCM, TEM). The panelists noted that the conversion factor (from mmpcf to fibers/mL)
      can vary considerably from one occupational setting to the next

      Interpretations of the study of South Carolina textile workers. The panelists had different
      opinions on interpretations of the study of South Carolina textile workers (Dement et al. 1994).
      One panelist, for instance, found this particular study to be an outlier among the other
      epidemiplogical studies, and he recommended that the authors exclude this study from the
      exposure-response analysis until the causes for the increased relative risks observed for this
      cohort are better understood. Another panelist suggested that the proposed protocol should
      classify the South-Carolina cohort as being exposed to mixed asbestos fibers, rather than being
      exposed to chrysotile fibers. He indicated that some workers in the cohort were exposed to
      amosite and crocidolite, in addition to being exposed to chrysotile.1

      Other panelists, however, did not think the South Carolina study should be excluded from
      EPA's analysis. One panelist was troubled about criticisms of the exposure estimates for this
      cohort, given that tills is one of few studies in which co-located samples were collected and
      analyzed using different methods, thus providing site-specific data for converting midget impinger
        1 After reviewing a draft of this report, one panelist indicated that it is important to note that exposure data
for the South Carolina cohort are available from more than just one reference (Dement et al. 1994). He suggested that
EPA use data from studies conducted by McDonald in the 1980s of a parallel cohort in the same plant. However, he
cautioned EPA against treating multiple studies of the same relatively small group of workers as separate studies,
considering the large overlap of workers studied by the two groups of investigators. This panelist encouraged EPA
to consider other data sources for this cohort, given that a recent re-analysis of epidemiological studies (Hodgson
and Damton 2000) severely criticized the data source EPA uses (Dement et al. 1994), to the point of those data being
dropped from the recent re-analysis altogether.

                                              3-21

-------
sampling results to PCM measurements. Another panelist challenged suggestions that the South
Carolina study is an outlier; he indicated that the South Carolina study is one of the more
rigorous epidemiology studies available for asbestos exposures, and he found no valid scientific
reasons for discarding it During this discussion, one panelist point out in response that the South
Carolina study is indeed an outlier among Ihe textile cohorts, with a slope which is higher than
eilher of the two textile cohorts; this panelist did acknowledge that the lung cancer risk among
the textile cohorts is greater than that among the mining cohorts. This panelist added that
scientists need a better explanation for why the lung cancer risk among the Soufli Carolina
cohort is greater than that of other cohorts before the South Carolina study can achieve
credibility, especially considering that exposures in South Carolina were supposedly to "pure"
chrysotile.
                                       3-22

-------
       4. COMMENTS ON TOPIC AREA 2: THE PROPOSED EXPOSURE INDEX

This section summarizes ihe panelists' responses to the charge questions pertaining to the proposed
exposure index. Section 4.1,4.2, and 4.3 document the panelists' responses to charge questions 4,5,
and 6, respectively.

4.1   Responses to Charge Question 4

Charge question 4 asks: "The proposed exposure index does not include contributions from fibers
shorter than 5 ^m. Please comment on whether the epidemiology and toxicology literature support the
conclusion that asbestos fibers  shorter than 5 \im present little or no carcinogenic risk." The panelists
discussed this matter earlier in the workshop (see Sections 3.1.3 and 3.1.4 for these comments), and
provided additional insights on the matter. Overall, the panelists agreed that carcinogenic potency
increases with fiber length, particularly for lung cancer. Most panelists supported assigning no potency
to fibrous structures smaller than 5 ^m. Some panelists agreed that the short fibrous structures are
clearly less potent ttian long fibers, but they had reservations about assigning zero potency to the
structures smaller than 5 pin; these panelists acknowledged that the toxicity of the short fibrous
structures might be adequately addressed by EPA's air quality standards for paniculate matter.  Specific
comments on this charge question follow:
      Reference to ATSDR's expert panel workshop on the role of fiber length. Two panelists
      noted that ATSDR convened an expert panel in October 2002 to discuss the role of fiber length
      on toxicity, and much of that discussion specifically addressed fibrous structures smaller than 5
      |im A main conclusion of that panel was that there is "a strong weight of evidence that asbestos
      and synthetic vitreous fibers shorter than 5 urn are unlikely to cause cancer in humans" (ERG
      2003). The panelists encouraged EPA to review the summary report prepared for that
      workshop, which was officially released on March 17,2003, and is available on-line at:
      www. atsdr. cdc.gov/HAC/asbestospanel.
      Evidence from, epidemiological studies. One panelist indicated that the epidemiological
      studies do not provide direct evidence of the role of fibrous structures shorter than 5
                                            4-1

-------
However, the panelist indicated that a growing body of evidence suggests that the cohorts -
predominantly exposed to shorter fibers (e.g., friction brake workers, gold miners., taeonite
miners) do not have statistically significant increased cancer risks. This panelist added that the
mechanistic studies provide the strongest evidence for assigning no potency to fibrous structures
(see next bulleted item). Another panelist agreed with these statements, and added that his
interpretation of data compiled by the National Cancer Institute provide additional indirect
evidence of short fibrous structures presenting little or no carcinogenic risk (see page 102 of the
premeeting comments in Appendix B).                                         I

The panelists briefly revisited the findings from a recent publication (Suzuki and Yuen 2001) that
reported finding relatively large amounts of short, thin chrysotile fibers in malignant mesothelioma
tissue. Several panelists encouraged that these findings not be considered in the risk assessment
methodology for reasons cited earlier in the workshop (see Section 3.2.2).

Evidence from mechanistic studies. The panelists offered different interpretations of
mechanistic studies. One panelist indicated that mechanistic studies have shown that shorter
fibers are cleared more readily than long fibers from the alveolar region of the lung by
phagocytosis, and therefore provide supporting evidence that short fibers play little or no role in
carcinogenic risk. This panelist acknowledged that extremely high doses of particular matter and
other non-fibrous structures can generate  biological responses (e.g., inflammation), but he
doubted that such "overload" conditions would be relevant to the environmental exposures that
the proposed protocol will be used to evaluate.             (

Another panelist agreed that long fibers are clearly more potent than short fibrous structures, but
he questioned the conclusion that short fibrous structures have no impact on carcinogenic risk.
This panelist noted that mechanistic studies have demonstrated that short fibrous structures and
spherical particles, like silica, can elicit the same toxic responses (e.g., generate reactive species,
stimulate proliferative factors) identified for asbestos fibers. This panelist added, referring to his
premeeting comments, that exposure to short fibers could cause inflammation and generation of
oxidative species that might increase the response to long fibers (see Bellman et al. 2001).
Overall, this panelist acknowledged that long fibers are more persistent than short fibers in the
lung and should be weighted more heavily in the exposure index, but he was hesitant to assign
the short fibrous structures zero potency.                   •          .           it

Implications on sampling and analytical methods. One panelist commented on the
practical implications, from a sampling perspective, of any changes to the exposure index. This
panelist indicated that measuring all fibers (including structures shorter than 5 nm) in i
environmental samples would not only be expensive, but also would compromise the sensitivity
of measuring the longer fibers that are most predictive of cancer risk. This panelist  >
acknowledged that human exposure is predominantly to fibrous structures less than 5 ^m, but he
noted that the amounts of short fibrous structures retained by the lung tend to be  very; strongly

                                       4-2

-------
      correlated with the amounts of long fibers retained by the lung. Due to this correlation, this
      panelist noted that measuring long fibers with sufficient accuracy would allow one to estimate
      amounts of short fibrous structures in a sample. This panelist added, however, that he sees no
      benefit of characterizing exposures to fibrous structures smaller than 5 \an, given the conclusion
      that such fibers do not cause cancer (ERG 2003).
4.2   Responses to Charge Question 5  ,


Charge question 5 asks: "The proposed exposure index is weighed heavily by fibers longer than 10
Specifically, Equation 7.13 suggests that the carcinogenic potency of fibers longer than 10 \an is more
than 300 times greater than that of fibers with lengths between 5 and 10 jira How consistent is this
difference in carcinogenic potency with the epidemiology and toxicology literature?" The panelists'
responses to this question follow:
       Consistency with epidemiological literature. The panelists noted that the original
       epidemiology studies did not collect exposure information that provides direct evidence of the
       relative potency assigned to the two different fiber length categories: fibers longer than 10 urn,
       and fibers with lengths between 5 and 10 jim. During this discussion, one panelist recommended
       that EPA consider the results of a case-control study (Rogers et al. 1991) that suggests that
       mesothelioma risks are greater for individuals with larger amounts of the shorter fibers (i.e.,
       between 5 and 10 |im) retained in their lungs. Another panelist was not convinced of the findings
       from this study, due to possible biases from selection of controls not matched for hospital of
       origin. This panelist encouraged EPA to refer to more rigorous lung-retained fiber studies (e.g.3
       McDonald et al. 1989, R6delsperger et al. 1999) that have found that the maj ority of cancer
       risk for mesothelioma is attributed to exposures to longer fibers, even when measurements of
       short fibers are taken into account.

       Questions about the fiber length-dependence used for mesothelioma. Some panelists
       were not convinced that the relative potencies assigned to different fiber lengths were
       appropriate for mesolhelioma. One panelist, for instance, noted that his previous review of the
       literature (Lippmann 1994) suggests that cancer risk for mesothelioma is most closely associated
       with exposure to fibers between 5 and 10 ^m long. He indicated that this assessment is
       consistent with other human lung evaluations (e.g., Timbrell et al. 1988), which have reported
       that fibers retained by the lung tend to be longer than fibers that translocate to the pleura This
       panelist added that the epidemiology literature clearly suggests that lung cancer and
                                             4-3

-------
       mesolhelioinahave different risk factors, as the relative amounts of lung cancer and '
       mesothelioina cases vaiy considerably from one cohort to the next. Based on these concerns,
       this panelist suggested that EPA consider developing separate fiber length weighting schemes for
       lung cancer and mesothelioma
                                                                                 I;
       Another panelist indicated that the epidemiology studies provide indirect evidence that
       carcinogenic potency appears to increase with fiber length. Specifically, he noted that the studies
       consistently show that mesothelioma has a very long latency period—a trend that suggests that
       the most durable fibers (i.e., the longer fibers) are the most potent. The panelist added that the
       analyses in the proposed protocol provide further indirect evidence of mesothelioma risks
       increasing with fiber length: when the exposure index was used in the mesothelioma model, the
       proposed risk assessment methodology generated an improved fit to the epidemiological data
                                                                                 i
       During this discussion, a panelist cautioned about inferring that only those fibers that reach the
       pleura are capable of causing mesothelioma, because researchers have not determined the exact
       mechanisms by which mesothelioma is induced. Further, he cautioned about inferring too much
       from a single study (Timbrell et al. 1988), given that many additional studies are available on
       lung-retained fibers.                                                         !j
                                                                                 |:
       Questions about the relevance of animal toxicology data. Some panelists expressed
       concern about basing the proposed weighting factors for different fiber lengths on observations
       from animal data First, one panelist noted that the weighting factors were derived strictly based
       on lung cancers observed in laboratory animals, and he questioned whether one can assume that
       the weighting factors can be defensibly applied to mesothelioma Second, other panelists noted
       that extrapolating the weighting factors from rodents to humans also involves uncertainty, due to
       inter-species differences in respiratory anatomy, macrophage sizes, and sites of lungtcancers.

       Suggested follow-up analyses. Given the concerns about basing the proposed exposure index
       entirely on data from animal toxicology studies, two panelists recommended that EPA attempt to
       optimize the weighting factois applied to different fiber length categories using the available
       human epidemiological data One panelist suggested that this optimization could be performed
       using the data compiled in Table 6-15 in the proposed protocol, which presents estimates of the
       fiber length distribution for different occupational cohorts. A panelist also suggested that EPA
       consider deriving separate weighting factors for lung cancer and mesothelioma,  rather than
       assuming the same fiber length dependence for both outcomes.
4.3    Responses to Charge Question 6
                                             4-4

-------
I
                   Charge question 6 asks: "Please explain whether the proposed exposure index will allow meaningful
                   comparisons between current environmental exposures to asbestos and historical exposures to asbestos
                   that occurred in 1he work place," The panelists discussed several topics when addressing the question,
                   because some panelists had different impressions of what the question was asking. Some panelists
                   viewed the question as asking about the validity of low-dose linear extrapolations (see Section 3.1.5 for
                   more information on this topic), and others viewed the question as asking about whether the proposed
                   methodology is an improvement over EPA's current risk assessment model. A summary of the
                   panelists'specific responses follows:
                          Is the proposed exposure index an improvement to asbestos risk assessment? When
                          answering this charge question, multiple panelists focused on whether the proposed exposure
                          index is an improvement over EPA's 1986 asbestos risk models. These panelists agreed that the
                          proposed approach is more consistent with the overall literature on health risks from asbestos,
                          which show that cancer risks vary with fiber type and fiber dimension. Two panelists were
                          hesitant to call Ihe proposed approach an improvement for evaluating mesofhelioma risks,
                          because the fiber length weighting factors are based entirely on lung cancer data in animals.
                          These panelists were particularly concerned that the proposed methodology might assign lower
                          risks for mesothelioma in certain circumstances, because the fiber-length dependence in the
                          methodology is not based on any lexicological or epidemiological studies of mesothelioma

                          Does the proposed risk assessment model support extrapolation from occupational
                          exposures to environmental exposures? Some panelists commented on the applicability of
                          the proposed risk assessment model to exposure doses below the ranges considered in the
                          occupational studies. Referring to observer comments provided earlier in the workshop, two
                          panelists indicated lhat some environmental exposures in areas with naturally-occurring asbestos
                          do not appear to be considerably lower than those experienced by occupational cohorts.
                          Another panelist agreed, and cautioned about distinguishing environmental exposures from
                          occupational exposures; he instead encouraged EPA and the panelists to focus on the exposure
                          magnitude, regardless of whether it was experienced in an occupational or environmental setting.

                          One panelist recommended that EPA investigate how cancer risks for lung cancer and
                          mesothelioma vary between EPA's 1986 model and the proposed risk assessment
                          methodology: for different distributions of fiber types and dimensions, does the proposed
                          methodology predict higher or lower risks than the 1986 model? Dr. Berman indicated that the
                          proposed methodology, when compared to EPA's 1986 model, generally predicts substantially
                          higher risks for environments with longer, thinner fibers and environments with larger amounts of
                                                                4-5

-------
amphibole fibers and predicts somewhat lower risks for environments with shorter, thicker fibers
and environments that contain only chrysotile fibers. One panelist recommended that future
revisions to die proposed protocol include sample calculations, perhaps in an appendix, for
several hypothetical environments to demonstrate how estimated cancer risks compare between
the new methodology and the 1986 model. ,                                   ;
                                     4-6

-------
              5.  COMMENTS ON TOPIC AREA 3: GENERAL QUESTIONS


This section summarizes the panelists' responses to charge questions 7-10 and 12. Responses to

charge question 11 are included in Section 6, because this charge question sought the panelists' overall

impressions of Ihe proposed risk assessment methodology, rather than focusing on any one specific

issue.


5.1   Responses to Charge Question 7


This charge question asks: "The proposed risk assessment approach assigns carcinogenic potency to

individual fibers and to cleavage fragments (or 'bundles that are components of more complex
structures'). Please comment on whether cleavage fragments of asbestos are as toxicologically

significant as fibers of the same size range." The panelists raised Ihe following points when responding:
       Terminology used in the charge question. One panelist took strong exception to the
       wording in this question (see pages 30-33 in Appendix B) and strongly recommended that the
       panelists use correct terminology during their discussions. This panelist noted, for instance, that
       cleavage fragments are not equivalent to bundles, nor do cleavage fragments meet the regulatory
       definition of asbestos, as the charge question implies. He clarified that he defines cleavage
       fragments as non-asbestiform amphiboles lhat are derived from massive amphibole structures.
       This panelist was concerned that none of the panelists at the workshop has the mineralogical
       expertise needed to address issues pertaining to cleavage fragments. Another panelist echoed
       these concerns and agreed that this charge question raises complex issues.

       Significance of cleavage fragments with respect to human health effects. The previous
       concerns notwithstanding, several panelists commented on the role of cleavage fragments in the
       proposed risk assessment methodology. One panelist, for example, indicated that there is no
       reason to believe lhat cleavage fragments would behave any differently in the human lung than
       asbestiform fibers of the same dimensions and durability; he added that this conclusion was also
       reached by the American Thoracic Society Committee in 1990 (Weill et al. 1990). This panelist
       acknowledged, however, that expert mineralogists have differing opinions on the role of
       cleavage fragments. Several other panelists agreed that it is reasonable to assume that cleavage
       fragments and asbestos fibers of the same dimension and durability would elicit similar toxic
       responses.
                                             5-1

-------
      Review of selected epidemiological and toxlcological studies. The panelists briefly
      discussed what information has been published on the toxiciry of cleavage fragments.' One
      panelist indicated that Appendix B in Ihe proposed protocol (see pages B-3 through B-10)
      interprets results from an animal study (Davis et al. 1991) that evaluated exposures to six
      tremolite samples, including some that were primarily cleavage fragments. This panelist noted
      that the study provides evidence that cleavage fragments can cause mesothelioma in animals.

      Another panelist, however, cautioned against inferring too much from this animal study for
      several reasons: the study was not peer reviewed; the fiber measurements in the study reportedly
      suffered from poor reproducibility; and the mesotheliomas observed in the study might have
      reflected use of infra-peritoneal injection model as the dose administration method. This panelist
      recommended that EPA conduct a more detailed review on the few studies that have examined
      the toxicity of cleavage fragments, possibly considering epidemiological studies of taconite
      miners from Minnesota (Higgjns et al.  1983) and ciimmingtonite-grunerite miners from South
      Dakota (McDonald et al. 1978); he noted that a pending publication presents updated risks
      among the taconite miners.

      Practical implications of measuring cleavage fragments in environmental samples. One
      panelist added, and another agreed, that measuring cleavage fragments in environmental samples
      presents some challenges, because microscopists cannot consistently distinguish cleavage
      fragments from asbestiform fibers, even when using TEM.                        '
5.2    Responses to Charge Question 8


Charge question 8 asks: "Please comment on whether the proposed cancer assessment approach is
relevant to all amphibole fibers or only to the five types of amphibole fibers (actinolite, amosite,
anthophyllite, crocidolite, tremolite) designated in federal regulations." The panelists made the following
                                                                                 j,
general comments in response:
      Review of evidence from toxicological and epidemiological studies. The panelists
      identified few studies that address the toxicity of amphibole fibers other than actinolite, amosite,
      anthophyllite, crocidolite, and tremolite. One panelist indicated that animal toxicology,(studies
      have demonstrated that synthetic vitreous fibers with differing^ernistry, but having similar
      durability and dimensions, generally exhibit similar potency for fibrosis, lung cancer, and
      mesothelioma Another panelist added that lung cancer and mesothelioma exposure-response

                                            5-2

-------
      relationships for a cohort of vermiculite miners from Libby, Montana, have been published for
      both asbestiform richterite and winchite.

      Appropriateness of applying the model to non-asbestiform amphiboles. Several panelists
      agreed that the proposed risk assessment methodology is relevant to amphibole fibers other than
      ihose listed in the federal regulations. The panelists noted that, in the absence of more detailed
      information on the matter, it is prudent to assume that fibers of similar dimension and durability
      will exhibit similar toxic effects. Two panelists expressed some hesitation on applying the
      proposed model to the non-asbestiform amphiboles: one panelist asked how confidently one can
      apply the cancer risk coefficients to amphibole fibers that have not been studied, and another
      panelist indicated he was not convinced that the model should be applied to the other
      amphiboles,  let alone for the amphiboles that are listed in the federal regulations.

      Given the amount of naturally occurring amphiboles in the Earth's crust, one panelist suggested
      that the proposed protocol clearly state that the non-asbestiform amphiboles being evaluated are
      only Ihose wilh Ihe same dimensional characteristics and biodurability as the corresponding
      asbestiform amphiboles.
5.3   Responses to Charge Question 9                           ••      ..


Charge question 9 asks: "The review document recommends that asbestos samples be analyzed by

transmission electron microscopy (TEM) and count only those fibers (or bundles) longer than 5 \im

Such counting practices will provide no information on the amount of asbestos fibers shorter lhan 5 urn.

To what extent would data on shorter fibers in samples be useful for future evaluations (e.g., validation

of the cancer risk assessment methodology, assessment of non-cancer endpoints)?"


The panelists expressed varying opinions on this matter: some panelists saw no benefit of measuring

fibrous structures shorter than 5 |im, based on responses to earlier charge questions (see Sections

3.1.3,3.1.4, and 4.1); other panelists indicated that there is some utility to collecting information on

shorter fibrous structures, particularly if the incremental analytical costs are not prohibitively expensive

and if counting short fibers does not compromise accurate counts of longer fibers. The panelists raised

the following specific issues when discussing measurement methods:
                                             5-3

-------
      Support for using TEM in future sampling efforts. The panelists unanimously supported the
      recommendation in the proposed protocol of using TEM, rather than PCM or somelother
      method, to characterize exposures in future risk assessments. The panelists also emphasized that
      future measurement methodologies must focus on generating accurate counts of the most
      biologically active fibers, or fibers longer than 5
      Practical implications of counting fibers shorter than 5 \int One panelist indicated that
      analyzing samples for fibrous structures shorter than 5 p.m would compromise analysts' ability to
      accurately count the amounts of longer fibers that are of greater biological concern. Some
      panelists and an observer further discussed the costs associated with counting fibers in multiple
      length categories, including shorter than 5 firn. The panelists did not cite firm cost figures for
      these analyses. However, noting that environmental samples typically contain more than 90%
      short fibrous structures, one panelist suspected that counting the shorter structures would
      considerably increase the time a microscopist needs to analyze samples, and therefore also
      would considerably increase the cost of the analysis. A panelist indicated tirat the costs and
      benefits of counting fibers shorter than 5 fim might be more appropriately debated between
      microscopists and risk assessors, with inputs from industrial hygienists and mineralogists.

      Relevance of fibers shorter than 5 \j.mfor non-cancer endpoints. One panelist; noted that
      exposures to fibrous structures shorter than 5 jim can contribute to asbestosis in occupationally
      exposed individuals (Lippmann 1988), but he doubted that the exposure levels found to be
      associated with asbestosis would be experienced in non-occupational settings. Another panelist
      added that the role of shorter fibrous structures for other non-cancer endpoints is not known,
      such as the pleura! abnormalities and active pleura! fibrosis observed in Libby, Montana, No
      panelists were aware of any authoritative statements made on the role that short fibers play, if
      any, on these other non-cancer endpoints. During this discussion, one panelist indicated that the
      toxicity of fibrous structures shorter than 5 |im might be adequately addressed by EPA's
      particulate matter standards.
5.4   Responses to Charge Question 10


Charge question 10 asks: "The proposed risk assessment methodology suggests that exposure
estimates should be based only on fibers longer than 5 p,m and thinner than 0.5 jim. Is this cut-off for

fiber diameter appropriate?" Before the panelists responded to the question, Dr. Berman first clarified
that the exposure index optimized from the animal studies (see Equation 7.12 in the proposed protocol)
                                             5-4

-------
assigns a far greater carcinogenic potency to fibers longer than .40 \ao, wilh diameters less than 0.4 urn;
he noted that the proposed diameter cut-off (0.5 jim) was based on an ad hoc adjustment.

The panelists agreed that Hie proposed cut-off for fiber diameter (0.5 urn) would likely include most
fibers of health concern; however, they also unanimously agreed that the exposure index should not
exclude thicker fibers that are known to be respirable in humans. The main argument given for
increasing the cut-off is that fibers with diameters as large as 1.5 (im (or with aerodynamic diameters as
large as 4.5 \s.m) can penetrate to small lurig airways in humans. Other panelists provided additional
specific comments, generally supporting inclusion of thicker fibers in the proposed exposure index. One
panelist, for example, advised against basing the fiber diameter cut-off strictly on observations from rat
inhalation studies, due to inter-species differences in respirability. Further, noting that the proposed cut-
off for fiber diameter would likely exclude some amosite fibers and a considerable portion of tremolite
fibers with known carcinogenic potency, another panelist encouraged that the proposed exposure index
include contributions from thicker fibers.

The panelists noted that consideration of fibers thicker than 0.5 \im was viewed as being most
important for the lung cancer risk assessment model, as risks for mesothelioma appear to be more
closely linked  to exposures to long, thin fibers (see Section 3.2.3). Further, some panelists suspected
mat increasing the fiber diameter cut-off in the exposure index should be accompanied by changes to
the exposure-response coefficients in the risk assessment models, but the panelists did not unanimously
agree on this issue.
                                              5-5

-------
5.5.   Responses to Charge Question 12

Charge question 12 asks: "Section 8.2 of the review document presents three options for assessing
cancer risks from asbestos exposure. Please comment on the technical merit of the proposed risk
assessment options." The panelists briefly reviewed the strengths and weaknesses of the three options
presented in the proposed protocol for assessing asbestos-related cancer risks. The panelists agreed
that the first option—direct use of EPA's lung cancer and mesothelioma risk assessment   !
models—allows for the greatest flexibility in evaluating site-specific exposure scenarios, particularly
those with time-varying exposures. Dr. Crump indicated that he envisioned this option being coded into
                                                                                  i
a computer program, into which users enter their site-specific exposure information. Most panelists
endorsed developing such a program. The panelists did not reject use of the second and third options,
provided that EPA ensures that all three options generate equivalent risk estimates for the same
exposure scenario.

The one issue discussed in greater detail was how sensitive predictions using the first option are to the
mortality rates used in the evaluation Noting that mortality rates as functions of age and sexjdifFer from
one location to the next, this panelist encouraged EPA to consider carefully whether nationwide
mortality estimates would be programmed into the risk assessment model or whether risk assessors
would have the option of entering site-specific mortality rates. The panelist also suggested that the
authors of the risk assessment conduct sensitivity analyses to quantify how strongly the mortality data
affect cancer risk estimates. These comments also raised questions about the fact that two populations
with different underlying mortality rates could have different cancer risks, even though their asbestos
        t
exposure levels are equivalent
                                             5-6

-------
  6. COMMENTS ON TOPIC AREA 4: CONCLUSIONS AND RECOMMENDATIONS


This section reviews the panelists' individual conclusions and recommendations regarding the proposed
protocol (Section 6.1), as well as how Hie panelists developed their overall conclusions and
recommendations that appear in the executive summary of this report (Section 6.2).


6.1   Responses to Charge Question 11


Charge question 11 asks: "Discuss whether the proposed cancer assessment approach, as a whole, is a
reasonable evaluation of the available health effects data What aspects of the proposed cancer

assessment approach, if any, are inconsistent with 1he epidemiology or toxicology literature for
asbestos?" The panelists offered individual summary statements, which were not discussed or debated
among the panel. Following is a summary of the panelists' individual summary statements in the order
they were given:
      Dr. Lippmann's summary statement Dr. Lippmann commended Drs. Herman and Crump
      on developing the proposed risk assessment protocol and supported use of a model that
      accounts for the factors (e.g., fiber type and dimension) that are most predictive of cancer risk.
      Dr. Lippmann supported the authors' attempt to make full use of the existing data and to
      interpret the results from Ihe epidemiological studies. He strongly recommended that EPA make
      every effort to obtain individual-level data from additional epidemiological studies. Dr. Lippmann
      suggested that a follow-up workshop with experts in exposure assessment could help EPA
      evaluate the uncertainties in exposure measurements from historic occupational data sets. Dr.
      Lippmann supported an observer's suggestion to conduct an animal inhalation study  using
      tremolite cleavage fragments to help resolve the issue of 1hese fragments' carcinogenic potency.
      Overall, he encouraged lhat future work on the proposed protocol continue, through use of
      additional expert panels, to make more informed usage of the human exposure data.

      Dr. Teta's summary statement. Dr. Teta indicated that the proposed protocol is an impressive
      integration of Ihe animal toxicology data and the human epidemiology data. She commended the
      authors for developing a scientific methodology that successfully reduces the variability in results
      across the epidemiological studies, suggesting that the studies might be more consistent than
      were previously thought. Dr. Teta recommended improvements to the meta-analysis of
                                            6-1

-------
epidemiological studies, such as establishing and applying criteria for use of human data in
characterizing exposure-response relationships. Overall, Dr. Teta found no inconsistencies
between the proposed protocol and the larger body of epidemiology literature, including studies
of cohorts (e.g., gas mask workers, railroad workers, friction brake workers) that do not have
well-defined exposure information. Though not disagreeing with the utility of other panelists'
recommendations, such as re-analyzing data from additional epidemiological studies and
convening additional expert panels, Dr. Teta encouraged EPA to move forward expedrtiously
with completing the proposed protocol and discouraged implementing additional steps that might
delay the overall project                                                      ;,

Dr. Hoel's summary statement. Dr. Hoel encouraged the use of more sophisticated modeling
that incorporates data on exposure-response (including non-linear models), duration.of
exposure, cessation of exposure, and uncertainty in exposure. Dr. Hoel also'strongly]1
recommended that EPA attempt to obtain individual-level data from additional epidemiology
studies, or at least obtain partial data sets. He encouraged Drs. Berman and Crump to use more
sophisticated uncertainty analysis techniques, such as generating prior and posterior distributions
of uncertainty. To ensure that the lung cancer model is not confounded by cigarette smoking, Dr.
Hoel recommended that Drs. Berman and Crump more closely evaluate all available; data on the
interactions between asbestos exposure and cigarette smoking.                    ;

Dr. Steenland's summary statement Dr. Steenland indicated that the proposed protocol is a
step forward in asbestos risk assessment; however, he had several recommendations' for
improving the analysis of epidemiological studies. For instance, Dr. Steenland suggested that the
authors conduct meta-regression analyses using the original exposure-response coefficients, in
which predictor variables  include fiber size, fiber type, the estimated percentage of amphiboles,
percentage of fiber greater than 10 (im, and categorical grouping of studies according to quality.
He indicated that these factors can be examined using both fixed effects and randomjeffects
models. Dr. Steenland recommended that the proposed protocol explicitly state and defend the
basis for choosing the 10  ^m cut-off for fiber length in the exposure index. He suggested that
EPA should consider using Bayesian techniques or olher methods to determine which' relative
potencies assigned to different fiber length categories optimize the model's fit to the
epidemiological data

Focusing on specific topics, Dr. Steenland indicated that he disagrees with the approach of
assigning amphibole fibers five times greater lung cancer potency than chrysotile fibers,
especially considering that the statistical analysis in the proposed protocol could not reject the
hypothesis that amphibole fibers and chrysotile fibers  are equally potent Further, he advocated
suggestions of exploring the adequacy of other exposure-response models (e.g., nonjlinear
models). Finally, Dr. Steenland suspected that cigarette smoking likely will not be a confounding
factor in exposure-response analyses for two reasons.  First, he noted that differences in smoking
practices between working populations and general populations typically do not cause

                                       6-2

-------
substantial differences in standardized mortality ratios. Second, he indicated that it is highly
unlikely that prevalence of smoking varies with workers' exposure levels. Dr. Steenland
encouraged that EPA refer to a recent publication (Liddell and Armstrong 2002) for similar
insights on interactions between asbestos exposure and cigarette smoking.

Dr, Crapo's summary statement. Dr. Crapo complimented Drs. Herman and Crump on
preparing the cancer risk assessment methodology, and he supported the general approach of
expressing cancer risk as a function of asbestos fiber type and fiber dimension. Dr. Crapo
indicated that the proposed protocol reaches several defensible conclusions, such as assigning
greater mesothelioma potency to amphibole fibers and to longer fibers while assigning no risk to
fibers less than 5 jim in length. However, he was concerned about some specific issues that are
not yet adequately resolved. For instance, Dr. Crapo felt additional data are needed to
rigorously define how mesothelioma potency varies with fiber length (i.e., fibers longer than 10
(j,m being 300 times more potent 1han fibers with lengths between 5 and 10 nm). Dr. Crapo
recommended that EPA, when revising 1he proposed protocol, explore more sophisticated
modeling techniques, including non-linear exposure-response models and consideration
threshold effects. He supported more detailed analyses of interactions  between asbestos
exposure and cigarette smoking, again through the use of non-linear models.

Dr. Sherman's summary statement Dr. Sherman first indicated that she concurred with
several recommendations made by Drs. Hoel and Steenland. She focused her summary
statements on the proposed exposure index, recommending that Drs. Herman and Crump use
the epidemiology data to  further investigate other formulations of an exposure index. Dr.
Sherman recommended, for example, examining the goodness of fit of other formulations of the
exposure index (e.g., assigning zero potency to all fibers shorter than 10 \an).  Further, she
recommended that the authors attempt to optimize the potency weighting factors in the exposure
index to the epidemiological data Finally, given that panelists expressed concern regarding how
potency varies wilh fiber length for mesothelioma, Dr. Sherman suggested that Drs. Herman and
Crump consider developing two different exposure indexes—one optimized for lung cancer, and
the other for mesothelioma Dr. Sherman added that she generally supported the lung cancer
and mesothelioma exposure-response models, and questioned whether using more complicated
models would necessarily lead to a better understanding of the data,

Dr. Castranova's summary statement. Dr. Castranova concluded that the proposed protocol
is a significant advance in asbestos risk assessment methodology. He strongly supported the
recommendation that future measurements be performed using TEM, rather than PCM. Dr.
Castranova also supported the approach of assigning equal carcinogenic potency to cleavage
fragments and asbestos fibers of similar dimension—a finding, he noted, that could be tested in
an animal inhalation study, Further, Dr. Castranova agreed that non-asbestiform amphiboles and
asbestos amphiboles of the same dimension should be assigned equal carcinogenic potency. Dr.
Castranova indicated that the epidemiology and toxicology literature clearly indicate that
                                      6-3

-------
       mesothelioma potency varies with fiber type, but he was not convinced that Ibis literature
       supports a difference in lung cancer potency between amphibole and chrysotile fibers.

       Dr. Price's summary statement Dr. Price found the proposed protocol to be an impressive
       compilation of the epidemiology and toxicology literature into a cancer risk assessment model
       that addresses most, but not all, risk factors debated since EPA's 1986 model. Dr. Price urged
       EPA to explore exposure-response models other than the models that involve linear.; low-dose
       extrapolations, which he viewed as being inconsistent with the epidemiology literature. Dr. Price
       indicated that future revisions to the protocol should definitely consider non-linear models and
       threshold effects.

       As an additional comment, Dr. Price emphasized that the two main elements of the  (
       protocol—the proposed exposure index and the exposure-response analysis—are closely inter-
       related and subsequent changes to the proposed exposure index could affect the robustness of
       the overall modeling effort. As an example of his concern, Dr. Price noted that increasing the
       fiber diameter cut-off in the exposure index from 0.5 nm to 1.5  (am could (according to an
       observer comment) lead to dramatic differences in the number of cleavage fragments' counted in
       environment samples; however, he indicated that 1he animal studies used to derive the original
       exposure index did not include cleavage fragments. Such scenarios raise questions about using
       an exposure index derived from very specific exposure conditions in animal studies to evaluate
       human health risks associated with exposures of an entirely different character. Dr. Price
       encouraged further study of cleavage fragments, perhaps in an animal inhalation study, to resolve
       the role of cleavage fragments.                                                  :

       Dr. Case's summary statement Dr. Case congratulated Drs.  Herman and Crump for
       compiling what he viewed as a reasonable evaluation of the available toxicology and.'
       epidemiology literature, and he strongly supported the general approach of factoring'fiber type
       and fiber dimension into cancer risk assessment. Dr. Case indicated that he agreed with the
       finding that amphibole fibers have slightly  greater lung cancer potency than do chrysotile fibers,
       although he believed that fiber dose, fiber length, and especially smoking history and;type of
       industry have greater importance in this regard. Dr. Case recognized that how one views the
       differences between the Quebec and South Carolina cohorts affects the conclusions ;drawn on
       this issue, and he encouraged EPA to classify the cohort of South Carolina textile workers as
       being exposed to mixed asbestos fibers, rather than being exposed to only chrysotile'fibers.2
         When presenting the summary statements, one panelist (LS) indicated that NIOSH is re-analyzing filters
that were collected in the 1960s from the South Carolina textile plant, and these re-analyses should indicate the
distribution of fiber types in this cohort's eKposures. Another panelist (BC) noted that these re-analyses will not
characterize earlier exposures to amosite fibers, which are believed to have occurred primarily before 1950 (based on
findings from studies of lung-retained fibers).


                                              6-4

-------
Dr. Case made several recommendations for further evaluating the existing epidemiological data
and for collecting additional data First, Dr. Case indicated that it is critically important for any
lung cancer risk model to consider confounding effects of cigarette smoking, and he encouraged
EPA to incorporate interactions with cigarette smoking into the lung cancer model to 1he greatest
extent possible. Second, Dr. Case supported Dr. Lippmann's recommendation of convening an
additional expert panel workshop to critically review inferences that should be drawn from the
exposure measurements made in the epidemiological studies; such a panel, Dr. Case noted,
would require inputs from experts in mineralogy, industrial hygiene, and measurement
methodologies. Third, he supported comments recommending that EPA examine non-linear and
threshold exposure-response models. Finally, Dr. Case agreed that conducting an animal
inhalation study is probably the best way to examine whether tremolite cleavage fragments
produce lung cancer, but did not advocate using rat inhalation studies to examine whether these
fragments induce mesothelioma, because results from rat inhalation studies have been shown to
be a poor model for mesolhelioma in humans. He added, however, that it would quite probably
be impossible to design an experiment in which rats were exposed only to "cleavage fragments"
or "non-asbestiform fibers" with no asbestiform fibers present at all.

Dr. Stayner's summary statement Dr. Stayner supported the general concept of
incorporating fiber type and fiber dimension into cancer risk assessment, but he recommended
that additional work be conducted before EPA accepts the proposed protocol as a new risk
assessment paradigm Dr. Stayner indicated 1hat his  confidence in the proposed protocol varies
between the lung cancer and mesothelioma models.

For lung cancer, Dr. Stayner indicated that the available epidemiological data should be able to
support anew risk assessment model, but he recommended that EPA consider the panelists'
many recommendations for how the meta-analysis can be improved (e.g., using different
statistical models, developing and applying minimal study inclusion criteria, conducting additional
sensitivity analyses). Concurring with Dr. Steenland's summary statement, Dr. Stayner added
that cigarette smoking is very unlikely to be a confounding factor in the lung cancer model and he
questioned whether the available data would support a quantitative assessment of the interaction
effects. While Dr. Stayner supported the recommendation for evaluating non-linear.exposure-
response models, he noted that the individual-level data needed to construct these models are
not available for most epidemiological studies. Dr. Stayner added that obtaining raw data from
additional occupational cohorts would provide the best opportunity for more detailed
exploration of non-linear exposure-response relationships.

Dr. Stayner expressed greater concern about the foundation of the mesothelioma risk model. He
indicated, for instance, that the relative potencies included in fee proposed exposure index are
based entirely on toxicology studies for lung cancer,  and not on any epidemiology or toxicology
studies specific to mesothelioma Despite these concerns about the biological basis for the
proposed mesolhelioma model, Dr.  Stayner noted that the proposed model does provide an

                                       6-5

-------
       improved fit to ihe findings from the epidemiological studies. He recommended that EPA
       consider optimizing the relative potencies in the exposure index to the human data, especially if
       EPA can access raw data from additional occupational cohorts to evaluate how exposure-
       response varies with fiber size and fiber type.                                   •

       Dr. McClettan's summary statement Dr. McClellan congratulated Drs Berman and Crump
       for integrating the lexicological and epidemiological data into a reasonable evaluation of asbestos
       cancer risks. Overall, Dr. McClellan found the proposed protocol to be a substantial;
       improvement over EPA's 1986 models and urged EPA to continue to move forward1 .with
       completing the protocol based on the panelists' feedback. Though he found the presentation of
       information in the draft document to lack transparency on many important matters, Dr.
       McClellan indicated that the authors' presentations at the workshop addressed many ;of his
       concerns regarding the transparency of how the proposed model was developed. One
       suggested improvement to the protocol's transparency was to clearly describe what literature
       were reviewed and to specify what studies actually factored into the quantitative analyses.

       Addressing specific topics, Dr. McClellan indicated that the analyses in the proposed protocol
       adequately characterize Ihe general roles that fiber type and fiber dimension play in cancer risk.
       He supported suggestions for involving additional experts, perhaps in another expert panel
       review, to further review interpretations of the epidemiological studies. Further, Dr. McClellan
       agreed with other panelists' recommendation that EPA explore the utility of non-linear
       exposure-response models, consistent with the agency's proposed revised Cancer Risk
       Assessment Guidelines. If linear, low-dose extrapolation models are ultimately used, he
       suggested that EPA explicitly acknowledge the uncertainties associated with such anjapproach.
       Dr. McClellan indicated that obtaining raw data from additional epidemiological studies might be
       particularly helpful in the exposure-response modeling. Finally, Dr. McClellan emphasized that
       the exposure characterization in the proposed protocol is closely linked to the exposure-
       response assessment; thus, the authors must carefully consider how revisions to the exposure
       characterization affect Ihe assumptions in the exposure-response assessment, and vice versa.
6.2   Development of Final Conclusions and Recommendations


After presenting their individual conclusions and recommendations, the panelists worked together to
draft summary statements for the peer consultation workshop. Every panelist was asked to write a brief
synopsis of a particular topic debated during the workshop. These draft statements were then displayed
to the entire panel and observers, edited by the panelists, and then compiled into this document's
                                            6-6

-------
executive summary, which should be viewed as the expert panel's final conclusions and
recommendations regarding the proposed protocol.
                                            6-7

-------
                                    7. REFERENCES

B Bellmann, H Muhle, O Creutzenberg, et al. 2001. Effects of nonfibrous particles on ceramic fiber
(RCF1) toxicity in rats. Inhalation Toxicology 13(10):877-901.

DW Herman, KS Crump, EJ Chatfield, JMG Davis, AD Jones. 1995. The Sizes, Shapes, and
Mineralogy of Asbestos Structures that Induce Lung Tumors or Mesothelioma in AF/HAN Rats
Following Inhalation Risk Analysis 15(2).

DW Berman and KS Crump 1999. Methodology for Conducting Risk Assessments at Asbestos
Superfund Sites; Part 1: Protocol. Final Draft, Prepared for U.S. Environmental Protection Agency.
February 15,1999.                                                             |

DW Berman and KS Crump 2001. Technical Support Document for a Protocol to Assess Asbestos-
Related Risk. Final Draft. Prepared for U.S. Environmental Protection Agency and U.S. Department of
Transportation. September 4, 2001.                                               !
                                                                              I:
G Berry and ML Newhouse. 1983. Mortality of Workers Manufacturing Friction Materials iUsing
Asbestos. British Journal of Industrial Medicine 40:1-7.

C Boutin, P Dumortier, F Rey, et al. 1996. Black spots concentrate oncogenic asbestos fibers in the
parietal pleura American Journal of Respiratory Critical Care and Medicine 153:444-449.

M Camus, J Siematycki, B Meek. 1998. Nonoccupational exposure to chrysotile asbestos and the risk
of lung cancer. New England Journal of Medicine 338:1565-1571.                     '<

WC Cooper, O Wong, and R Graebner. 1988. Mortality of workers in two Minnesota taconite mining
and milling operations. Journal of Occupational Medicine 30(6):506-511.               ;'

JMG Davis, J Addison, C Mclntosh, BG Miller, and K Niven. 1991. Variations in the Carcinogenicity
of Tremolite Dust Samples of Differing Morphology. Annals of the New York Academy of Sciences,
473^190.                                                  '

PE Enterline, J Harley, and V Henderson. 1986. Asbestos and Cancer—A Cohort Followed to Death.
Graduate School of Public Health, University of Pittsburgh.                            j

N De Klerk, B Armstrong, A Musk, M Hobbs. 1989. Cancer mortality in relation to measures of
exposure to crocidolite at Wittenoom Gorge in Western Australia British Journal of Industrial Medicine
46:529-536.                                                                   "
                                           7-1

-------
JM Dement, DP Brown, A Okun. 1994. Follow-up Study of Chiysotile Asbestos Textile Workers:
Cohort Mortality and Case-Control Analysis. American Journal of Industrial Medicine 26:431-447.

EPA 1986. Airborne Asbestos Health Assessment Update. U.S. Environmental Protection Agency.   •
EPA 600/8-84-003F. 1986.

ERG. 2003. Report on the Expert Panel on Health Effects of Asbestos and Synthetic Vitreous Fibers:
The Influence of Fiber Length. Prepared by Eastern Research Group, Inc., for the Agency for Toxic
Substances and Disease Registry. March 17,2003.

TW Hesterberg, GA Hart, J Chevalier, et al. 1998. The importance of fiber biopersistence and lung
dose in determining the  chronic inhalation effects of X607, RCF1, and chrysotile asbestos in rate.
Toxicology and Applied Pharmacology 153:68-82.

ITT Higgins, JH Glassman, MS Oh, and RG Cornell. 1983. Mortality of reserve mining company
employees in relation to taconite dust exposure. American Journal of Epidemiology 118(5):710-719.

J Hodgson and A Damton. 2000. The Quantitative Risk of Mesothelioma and Lung Cancer in Relation
to Asbestos Exposure. Annals of Occupational Hygiene 44(8):5 65-201.

JM Hughes, H Weill, YY Harnmad. 1987. Mortality of Workers Employed at Two Asbestos Cement
Plants. British Journal of Industrial Medicine 44:161-174.

IARC. 1996. AB Kane, P Boffetta, R Saracci, and JD Wilboum. Mechanisms of Fibre
Carcinogenesis. International Agency for Research on Cancer. Oxford University Press 140:1-9.

JL Levin, JW McLarty, GA Hurst, AN Smith, and AL Frank. 1998. Tyler Asbestos Workers:
Mortality Experience in a Cohort Exposed to Amosite. Occupational and Environmental Medicine
55:155-160.

FD Liddell and BG Armstrong. 2002. The combination of effects on lung cancer of cigarette smoking
and exposure in Quebec chrysotile miners and millers. Annals of Occupational Hygiene 46(1):5-13.

FD Liddell, AD McDonald, and JC McDonald. 1997. The 1891-1920 Birth Cohort of Quebec
Chiysotile Miners and Millers: Development from 1904 and Mortality to 1992. Annals of Occupational
Hygiene 41:13-36.

FD Liddell, AD McDonald,  and JC McDonald 1998. Dust exposure and lung cancer in Quebec
chrysotile miners and millers. Annals of Occupational Hygiene 42(1):7-20.

M Lippmann. 1988. Review: Asbestos exposure indices.  Environmental Res 46:86-106.
                                           7-2

-------
M Lippmann. 1994. Deposition and retention of fibers: Effects on incidence of lung cancer and
mesothelioma Occupational and Environmental Medicine 51:793-798.                  :
                                                                              •f
                                                                              •+
JC McDonald, GW Gibbs, FD Liddell, and AD McDonald. 1978. Morality after long exposure to
cummingtonite-grunerite. American Review of Respiratory Disease 118(2):271-277.      '

AD McDonald, JS Fry, AJ Woolley, and JC McDonald. 1984. Dust Exposure and Mortality in an
American Chrysotile Asbestos Friction Products Plant British Journal of Industrial Medicine
41:151-157.

JC McDonald, B Armstrong, B Case, et al. 1989. Mesothelioma and asbestos fiber type. Evidence
from lung tissue analysis. Cancer 63:1544-1547.

AD McDonald, BW Case, A Churg, A Dufresne, GW Gibbs, P Sebastien, and JC McDonald. 1997.
Mesothelioma in Quebec chrysotile miners and millers: epidemiology and aetiology. Annals of
Occupational Hygiene 41(6):707-719.

JC McDonald, J Harris, and B Armstrong. 2002. Cohort mortality study of vermiculite miners exposed
to fibrous tremolite: an update. Annals of Occupational Hygiene 46(S 1): 93-94.

WJ Nicholson. 1976. Part III: Recent Approaches to the Control of Carcinogenic Exposures. Case
Study 1: Asbestos—The TLV Approach. Annals of New York Academy of Science 271:152-169.
                                                                              i
C-G Ohlson, T Rydman, L Sundell, et al. 1984. Decreased lung function in long-term asbestos cement
workers: A cross-sectional study. American Journal of Industrial Medicine 5:359-366.    i!

K Ro'delsperger, H-J Woitowitz, B Briickel, et al. 1999. Dose-response relationship between
amphibole fiber lung burden and mesotiielioma Cancer Detection and Prevention 23(3): 183-193.

AJ Rogers, J Leigh, G Berry, et al. 1991. Relationship between lung asbestos fiber type and
concentration and relative risk of mesothelioma. Cancer 67:1912-1920.

A Rogers and G Major. 2002. Letter to the Editor: The Quantitative Risks of Mesothelioma and Lung
Cancer in Relation to Asbestos Exposure: The Wittenoom Data Annals of Occupational Hygiene
46(1):127-129.

H Seidman, IJ Selikoff, and SK Gelb. 1986. Mortality Experience of Amosite Asbestos Factory
Workers: Dose-Response Relationships 5 to 40 Years after onset of Short-Term Work Exposure.
American Journal of Industrial Medicine 10:479-514.
                                           7-3

-------
IJ Selikoff and H Seidman. 1991. Asbestos-Associated Deaths among Insulation Workers in the
United States and Canada, 1967-1987. Annals of the New York Academy of Sciences 643:1-14.

LT Stayner, DA Dankovic, RA Lemen.  1996. Occupational Exposures to Chrysotile. Asbestos and
Cancer Risk: A Review of the Amphibole Hypothesis. American Journal of Public Heallh
86(2): 176-186.

V Timbrell, T Ashcroft, B Goldstein, et al. 1988. Relationships between retained amphibole fibers and
fibrosis inhuman lung tissue specimens. In: Inhaled Particles VI. Annals of Occupational Hygiene
32(S1):323-340.

Y Suzuki and S Yuen. 2001. Asbestos Tissue Burden Study on Human Malignant Mesothelioma
Industrial Health 39:150-160.

MJ Teta, HC Lewinsohn, JW Meigs, et al. 1983. Mesothelioma in Connecticut: 1955-1977:
Occupational and geographic associations. Journal of Occupational Medicine 25(10):749-756.

A Tossavainen, M Kotilainen, K Takahashi, G Pan,  and E Vanhala, 2001. Amphibole Fibers in
Chinese Chrysotile Asbestos. Annals of Occupational Hygiene 45(2):145-152.

H Weill, JL Abraham, JR Balmes, B Case, AM Churg, J Hughes, M Schenker, and P Sebastien.
1990. Health Effects of Tremolite. Official statement of the American Thoracic Society. American
Review of Respiratory Disease 142(6): 1453-1458.

E Yano, ZM Wang, XR Wang, MZ Wang, and YJ  Lan. 2001. Cancer Mortality among Workers
Exposed to Amphibole-free Chrysotile Asbestos. American Journal of Epidemiology 154(6):538-543.
                                           7-4

-------
The following Appendices:

Appendix A - List of Expert Panelists
Appendix B.- Consultants' Premeeting Comments
Appendix C - List of Registered Observers
Appendix D - Agenda
Appendix E - Observer Comments Provided at the Peer Consultation Workshop
Appendix F - Observer Post-Meeting Comments

Are available at: http://www.epa.gov/superfund/programs/risk/asbestos/index.htm

-------
     Appendix A




List of Expert Panelists

-------
                                  Appendix B

              Premeeting Comments, Alphabetized by Author
         (includes bios of panelists and the charge to the panelists)
Note: This appendix is a copy of Ihe booklet of the premeeting comments that ERG distributed at the
     peer consultation workshop. One panelist (Dr. Bruce Case) submitted an edited form of his
     premeeting comments to ERG at the workshop. That edited version appears in this appendix.

-------
                       Appendix C




List of Registered Observers of the Peer Consultation Workshop

-------
              Appendix D
Agenda for the Peer Consultation Workshop

-------
                                   Appendix £

     Observer Comments Provided at the Peer Consultation Workshop
Note: The peer consultation workshop included three observer comment periods, one on the first day
      of the workshop and two on the second day of the workshop. This appendix includes verbatim
      transcripts (to the extent that specific remarks were audible from recordings) of the observer
      comments, in the order the comments were given.

-------
          Appendix F




Observer Post-Meeting Comments

-------
                       APPENDIX C:
     COMPENDIUM OF MODEL FITS TO ANIMAL
  INHALATION DATA IN SUPPORT OF THE BERMAN
    ET AL. (1995) STUDY AND POST-STUDY WORK

The attached tables are a compendium of raw outputs for the fits of various (exposure index)
models to the Davis et al. animal inhalation studies.  Each entry lists the date of the run, the size
categories included in the run, the maximum likelihood estimate for the run, the degrees of
freedom, the P-value for the fit, and the coefficients representing the relative potency assigned to
each size category for the model.
                              c.1

-------
                                                            o in o  *r  o ro  I-H
                                                                 I      W
                                                                 1  CM  CSJ
                                                                 I  O  PO
                                                            CM T  UT<
                                                                                                                                                         (N
                                                                                                                                                         o
                                       .. O      -91
                                     O O   h  CD
                                     «3 O  O  ^r
                                                              n eft  i   o in o  i
                                                              ' ^ O  O  I  O
                                                              o r- o
                                                            I   I  VO  I  O
                                                            i  en "a1 m • o  i
G
1
O
O
>. o
o o
o
rH
o
rH
V
o
o
O
0 0
LO fi
o ••
o o o
1
CM 1 O
• CM
o o
i i
in
O rHJ
1
in o o
rH 0* •
m
o
o
fO
'«
in
0
o in
*. o
en H
O CM
CO
G
* CM
O
                                                                                                                                             CM  W


                                                                                                                                             A  -*

0
CM



o
CM
I
O
rH
^
O


LT)
fc
in
v

5?
QU tij


ftj 0
J
rH
O
G
- G
rH G
. G 0
O - **
V A

0

0 1
m o
A **)
0

GOO

0 0
CM CM


5 8

o
L) 0 U
CO O t/J
o rt
$ ?
< m
V CO

rH CM
V •



s. (*>
O O
A W
0
-0
O 


O

1
0



o
1
o
CM

«•
CM
V

O
S i


U f>
CO CD
o
i CM
o m
CM in


0
CM
I
O
r-t
„
O 0
rH G
1 O
LO O
„

v

g
a*
o
CJ 0
W 0
O CM
*a- O
AJj
» CM
O **

1 O
o -



o

1
0
CM m
ID
»• CM
O CN
,
G


U

O
u o
CO O

o
CM
A


O
CM
I -
0

fc
O
rH

in
„,
m
v

S ?

CD
U 0
to ro
f)O *. rH
CM CM •
- 1 ( O
G O O V
CO rH rH
I CD O CD *.
O ^" w. CM » P*> O


w * 1 . 1 • 0
O LO 1/1 ^
CM
1 » -• *


1 1 II
«• W S td ^ U O



in vo S f*5 S co o
* *a» IH -^ n r- 1
in o O o
V CM
M U CM U
£ O O S
U 1 O I 0 O

m — eo — o

s* £ ** t^ " 0 t°
O
v
o *.
0 0


> G



O

t
o

0

o o
1
o
CM

G £

CM

CD f*
o"

O O 0



o

1
0

s: o

O Ou O

o 2
IH
O

O
O



0 ^ rH
G

o o

o

0

1
o
CM
S •

<*4 O

S G
W
0

O
o
a
— o

s? §
• CM
0 0
V I
[4
•. LO
O en






O

I

O

o o
1
0
CM

g

O

B* G
O
V

G G

f O




O

1
0



O rH
|
O
CM
CM
5 ?

CM

t3 ^
O 0
V tH
1
•- Cd

CM CM



G CM
CM
1
O   in CM i
             i O rH I
I "S, W  ^--iJ^.iJ'^vl_J  ^vt-l^,t^ ^>». tJ  ^%  ^ "N. C*  ^v O '^

lOOGOOlOOOOOOOCpOOCMGOJOOO
• t   OK  OB  r- e   o  *  oe   o e   o  e   coc    •  c  o e
1      '.>'...       f     rH       •
                                                                                                                  r'OOOO'

-------
PJ
(j

-------
o
o
o
0
o
o
o
o
o
o
o
o
0
o
o
o
o
0
o
o
o
o
0
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
in
o
rH
o
en
en
to

^
CD
m
o
rH
'!»
o
r-*
o
o
o
o
m

m
CO
CM
CO
o
o
o
0
U3
rt

r-
                                             iH      O     r-t
                                                                                                                                                     'o
                                                    3
                                                    O

                                                    4-»
                              3 CM     CM
                          CM  " O O  O
                          O      1  IO   1
       CM  ~ M 0  CM     CM
       O      OVOOOO
        1  O  1  A   1  tO   1
WW^DW     W A  W
m  A <*»  A  -"a1  *  o     m
        "
                                                                                                                               A
                                                                                                                         •. O  -.  O   ••      --
                                                                                                                        «-c o CM  o co o m
                                                                                                                         . o  •  o   •  o   *
0
\o
1
o
in
o
luation?
30-40, 4C
Ed -
fl
0
r- o
. V£>
CM 1
O
m
CM
O «.
1 0
O •&
f*> O
o
ro
^
•* ro
ID o
0 1 •
*• CM
CM O O
o m )
m o *
. ^ (SJ
CM 1
0
ro eg
\S> 0 en
U>
O iH
1 
O fO
1
o
ft
-
r*> rH
o in
o .
1
o
in CM
0
_ i
o
1
o
ro
* CM
tO to
O
1 . kO .
m o
-.CM CM
oo * a
if) I O 1
*r *r O O
•.in CM
1 •*•
O 1
rO 10 0 CO
0 CM 1 f"J
• CM -
£ r
0
in
CM
*. O
O 1
i?
1
o
o m
1 0
CM •
O •
O CM
in o
t CM
0
f*>
O 1
1 CO
CM «3
^ .

in
»• CM
O O
in i
- in
o -
^*- CM
1
O
(0 CO
o en

H
(M
O
E
CO
Eu
O
Q
fi
H
e

o
w
bu
INSTEAD C
—
-v.
<
a
o
A
O
m
o
m
1 CM
0 0
tt
0 -H
1 ID
a
o
A
O
o
m
o
in
t
0

o
1
%
0
A
O
to
o
in
0*
in
CM 1
o o
Cd'
^••0
CO 1


CM
O
w
o
CM
^
0
u>
A
O
in
cT
m CM
1 o
O 1
1/1
0 <0
1
V • V . V O V
o o o o
(O ID kp IP
A A A A
O O O O O O
O O O O O O O
in * in * in o if)
o o o c?
in in in m
i i j t
o o o o
o o
O O O O O f> O
( • 1 • 1 CD t
    o     ooo

  r o  o o  i   o o*
                                                                                           f> CO <

                                                                                                           *      VCM-.CM*.      *      *.
                                                                                           OCMOCMOCMOOOOO     O     O

                                                                                           1   I    I   I   I   I   I  Ei]  |  Ei]   t       I       I
                                                                                           QUOCdOtiJOCMOr-OOO     O
                                                                                           CMCOCM-rcM-^rCMOCMincMOCMOCM
                          OOO   I  O  rH  ^ r-l  •
                                                                                            M fo
                                                                                                    . CO '
                                                                    CM
                                                                   . CM
                                                                           CD
                                                                                  CM
                                                                                                                  lomomomotn
                                                                                                                  fOiHOrHQrHQT~l
                                                                                                                    O t   OS  O  t  O C

o     o     OCMO     OP^oinOf-iomoooin
t   O t   O e   -toe
                                                                                   CM
                                                                                   en
                                                                                          CM
                                                         CM
                                                         en
92
o-j
CFi
92
CM
cn
                                        -  01  r* o^
                                        )  rH  O rH                        .   .   _
                                          X.   I  X.    ^»     X,     X,   |  V.  |  X,  |   '
                                        |  00  tO O>    CO     CD     CD  (d CD td CD [J  i
                                        lOrOO    O     O     OE~OOOni
                                          XCMX.OX.OX.Ct^xp.'v^cnx.eM'
                                          o(po\0i9iu>oou>aD'
                                        'Omooooooocnotnocfii
                                                 -oeotot
                                                                                      (O
                                                                                                                          u>
                                                                                                                                 CD
                                                                                                                                       <  o ua
                                                                                                                                         o o
                                                                                                                                         O C

-------
                                                                                o
                                                                                o
                                                                                o
o
o
o
o
o
o
                                                                                                                                               o     -*
                             ^*     O     <-l
                                                                                                       O CM  T     O
                                         u
                                         c»l     O
                                         •fl-     O
                                         r-i     o      a     ai
                                         ^r     o      o     •«•
                                                                                                   O   I  O      rH
                                                                                                                                 co id »-t       »     i

                                                                                                                                 rH *D CM  CD O» CO  i

                                                                                                                                    *O     *H     Q
CN
o ^*
I o
£jj CD
Pi CM
• 1
.

i • i m i r-
CN rH
i-t m CM
CM »n o
. &}
fl i*1 CM CM
(  CM CM CM •*
O O CM O ID
f) i r- i CM
§ ° SS S M
CN CM
O
o
o
o

r-l
O
O ••.
O rH
o
ID
A
*•
O to
0 1
o o
. m
o
i
o
m
0 *.
0 0
• i
o
0
O CM
O
0 -
• o
t
o
o
CO
in o
o
o
o
o


o
o
o
o


o
o
o

o
o
o

CM
-*
r-H
rS
s


o
CM
o
A
-
to
I
O
in
o
o
o
)
o
o
1
CM
O
7
O
U
to








cn
CM

r-
CM
eg
o
l
o
CD
CN

^
O
A
o
V
o
A
-
1
o
m
o
o
o
t
o
m
o
t
CM
O
CM
I
0
I
9774E-0

0 O
W 1
in eg
i
o
w u
CN PH
en








eg
o
l
CM
0\
iH
CN
1
CD
eo
CM
CN
O
U
CO

CN
O
A
CM
V
A
*
1
o
i
-3*
A
r*>
O
CM
t
I
o
o
o
o


o
o


o
o

o
o


o
0
CO
CO

CN
A
O
o
A
o
o
m
o
o
o
o
1
o
o
I
0
0
o
o


o
o


o

o
1
ro
t>
H

o
0
LO
0
o
o
o
o
M
o
A
c*) o
• o
V o
A
-
1 O
in o
o
i
o
1 0
o *
CO
Q
eg o
CN '
1
r-
A *•
CM
O * O |
O 
1 O 1
o * o

CN ^ CM
. o •
O | O
1 U 1
. f ,
CO
0 O
V V
OOO
o
0 O
o
o
10
cn


o
o
o


o

o
0

o
o
o
1
bJ
o
1
o
o
o
eg
o
•
0
"
o
A
o
V
(D
i-t
O
lifferenl
u
o
o


o
o
o
o


o
o
o
o

o
1
m
CD

0
0
o
o
0
o
o

o
I
o
o
v
o
o
o
TH
&
o
A
o
r>

i
o
o
o
2:
o
V)
4978E-0
without
o
I
CD -
rH O
c- o
1
o
o
o ^.
o o
o o

o
0 A
o
o o
o
o
O 1
o o
0 CN
o
o
CM
fO O
0 TH
is
o>
10 u
HN W
7535E-0


vo
r-


O
o
0
o
o
CN
O
O
o •-
o •
• o
v
o
1
0 «r
O
o *
o o
1
o
IO
cn U
CM CO
                                                                                                                                                               10

                                                                                                                                                               6
O t  O B   O
                          os  p-  E  o t
                                                       me   OB  ^<
                                                                             OB   ot  o  1=   ot  oc  O
                                                                                                                                 O\ O  O O O  O O

                                                                                                                                 O O  rH O r-t  O iH

                                                                                                                                 C   OB  OB   OB
                                                    OCMOtHO
                                                           '"
  -  t  O B   ^3-

-------
    O      V     *9
                                                                                                                              o      «s-     r-
                                                                       O
                                                                        I
                                                                       td
                                  ^  I  0\      Ij"
                                                                                                                              •*      •*     o
                                                                                                                       trt     i-f      rH
                                    ' o •*      kO     •*     **
                                         ^f  tN O CM  CM  CM "I      ^*     rHOiHrMCM
                                  V     CM  [il CV I
                                  O O LO  •»" U"l i
                                 OWOtdObdOPdO
CTiO^OOOOOCMOr~OuOOCriO«>OOJOCOO   -Or-OiTJO   -OO
                                     .H     ^-     m     VO       •      '      I      tH       •      I

-------
                                                           o
                                                           :
C-     lA
o     r-
  .  CM
I-H  o m
OCMOCMO
                              OtOOCMO
CO CM
O [-
1 CD
> CM •
) rH
> CM
CM O
W
3 CO -^
incOOOCMrHrOCM rH
O O Cn O -"J" O T rO
o • -tj* • en . o o co
o f*) CM m CM o
O t- CO T O
O in CO rH CM O CM
O TO O i-H O rH 1 1
1 1 1 W W
cnor-com^oa-^coO
0
o
o
o
n
0
o
o
r-t
rH
CM m
O ro
I en
Id r-
r- • r-
en en H
. r- o
ID CM ^T
rH O rH
CO 1 CO
r- rH Ut CO
en r~
1 CO t
en f-
rH CD LO CO
CM O CM
rH 1 f-
CO W O
o m o
i w
O CO V
CO
CM
O
O (Tv
0
(Tv
t-
rH 0
1
W
CD
cn o  *°
^
U1
r-
o
CO CO
rH
en 
rH

rH
rH
\
tD
O
e

ro
CM
ro
MP
O
O
o
•a- o
O
A

O
V *O
r-
*. ro
G CO

A


o

1
O
m o
o
- o
o o
m '
i
o
••*
CM
- O
O 1
T U
, ^
o cn
(O O
o
O CM

1
o
CM
hi
o o
CM 0
1 O
0 0
"


o
rH


0
U 0
cn o
*
CM O

w o
cn ••
in ID o
O r-t O
O B O
• o
CM
o\
CM cn
O rH

W rH
00 rH
ro ^x O
CM to O
•V O G
* s a
m
5-8, >8 (without) "
.4395 1.5005E-02

in
1 O
CM O
O
.. o
CM •
1


O



•m* o
. o
O O
1 O
o



CO

o o
1 O
CM G
• O
"



CM
o
v
0
* o
o o
CM O
A

G
CM



U 0
W O

CM
CM
vo
o


r-t O
B O
O
CM
c\
en
rH

t*H
rH
•"-^ O
W G
O O
f O

O G
G O
O G
O O
+j * ZJ
g 8
SCM £3
O 4J
-H 1 -H
5 W 5
— -V — ro
T CM
A Ch A ^
*. cn •.
1 1
in in

tn in
i o i cn
CM O Csj ^O
O rH
-* o - m
CM • CM •
1 1


o o



•** O -a* O
. cn » G
O T O O
1 CO | 0
o o



to ro

o o o o
1 O 1 O
CM O CM O
• O ' O
"



CM CM
0 0


*. o *. o
o o o o
ro o •* O
A • A

O G
ro ^i


O O
O G U O
W O CO O

m G
VO QO
O O


rH G rH O
t o e o

CM CM
cn cn
01 cn
rH rH
"""«. "V^,
rH rH
rH rH
*v. O "*x O
VD O VD O
O O O O
t G t G

CM
O
A

O
e

o
O
o
o
o
o
o
o

o
o
o
o




CM
o



CO
CO
to
,_;





en
crs

en






0
O
o
G






O
0
o
o




o
o
o
o



CM
O
•)
w
vp
o
CM
T!
rH
O
O
o
G
O
o
o
G
fO -
o
«. o
rO 1
. [d
O co
V ^
in
«. r-
o •

A

*• CM
o o

1 W
O *£>
LT) CM
t—
.. en
0 •
1
o



o

1 U)
o cn


o"
CO

O
CM
». 0
O O

1 G
0 •


O
rH

S o
0
L) 0
W G

O
cn
O
" 0
IjD O
r-t O
e o

CM
en
cn CM
T-H O
•v^ 1
^W
rH -V
— «, ro
UP cn
O u
O

O
5
e-«
o>
52;
rH

CD

rH


Oj
S


o


w

03*
S
W


J?
o


U


es
in


CM
rH
B

CM
cn
cn
rH
•>»»,
CM

'v.
1C
O
e




























U
W


s

w
o
M
s
2


^
D


U
w

m
.
o> «>
o o
. e
rH
CM
O
A
CM
G
"O
a
rB
o
A


o

1
o
in


o
tO
1
o


o

1
o



CO






G

O
0



CM
i
to
cn

en


o
l



•<*

CO








0

o o o o
o o o o o o
O G G G O O
O O O O C" O
0 G
CO CO CO 00
to Tf A A A A
• O - O O CM
OOOCP ** CO «,1D isO fcCF"!
AOAGCor-comcoocom
O O 1 rH t CM 1 a 1 r-
«. • «, « in oj LO m in 'in •
o o in 10 in in
i i t i
CM ro CM CM CM CM
•a -o o o
c a H i * t *
rOrH cUMPCMtdCMUCMOeMQ
ScoScn'^otro^o^o
A -A •ocoomo -G -


G G s .. *. «.

II ....
O O O O O O
m in i i i t
O OrO CO rOOfOG
^o ««o »o «o »o «o
ooooor-oooooo
tnomo rH o o o
1 • I . ro o


GO 1 1 1 1 QO-WQ-^Tj-rO
*3« ^ CM CNJ CM CM ^ 1 ^ 1 * 1 • O •
1 1 • • • • roW-0-WOWOIO

cOOtOO O O O OOCMOCM r- CD


O «O •CMOcMOCM "CM -"O -13 -CtJ '(TJCMfO
1 1 GOOO (C fOOOroG
OOVVVV •* ^ xp
CMCM- OCMOCMACMAA

•• •• rt JnOG C A'A'**IV^W
o o rt nj i tti nj EX] uJ o u o i o
CMOCMO Cil O O».CM«.CJl^'O^1W^1
1 O 1 ooooenepoooOrHOro f r- t rH I

rHOT^OAOAtOAGA0 1 m 1 rOrH^CMfOCM

o o o o co o o * *. o otno

V V V V V V V VCMV-^V V
CO fO G rH
£ £ SGSG£ S E £i£i£ £


oorHcooomeomr*

COOCQOCOmCnoWOOTOtOrHCflrHCnrOW^DW
w ro
^ "ro° S ^CMO^O ^o^^oj" S
kD f-ICO O rHICMICO rOliOIcO -!p
CM CsJUOCM to COCdcOtdcO cOW-cOWfO <*>
.. O •• •«• •• •• •« CO " G " O " ID " CO ••


E OK t-H t O i o: r~ s CMC OB co- U> s cn c
O 0 • - ' - - 0
CMCMCMCMCM.s 
-------
3 O
0 0
3 O
u o

n
o
i
cd
^> *o

N ^
> CO
i)


N
3

^
H O
M O
:n o
«r o
H








»
D O

O
o
o
o





0
o
o
o








00
m
i-H



rH




in
m
rH

O
o
o
o





r*
to
rH
•V








O
O
rH



r-l





m
in
i-H

O
o
o
o





o
o
o
o







o
o
o
o









^
*o
in

o
o
0
o





t"-
n

v








o
o
rH



H





cn
in
rH

O
O
o
o





cn
i-H
i-H
•fl*




O
I


CO
m

cn


^




m
r-l
ID
rH

Ul
in
CD
CM
CD
rH
tn
o
1
6J
o
(n
o
vo
rH


CN
O


o
CD

ID
^r







cn
o
r*
rH

M
ys
rH
m
rH
CD




O
o
o
o







CM


CM
cn
r-t CN
j

r-

m
CM

CN
VD O
rH 1
• U
VD
rH
O
O
O
o





o
o
o
o









o
rH









CD















O
1

t-
o>
CO





o
1


o
.
CM



CD
CTv
O
o
O
o





o
o
o
o
o
1
fa]
CM fl
o o
1 0

CD •

CM

cn o
i


tn
.
CN

rH
r-J
rH
r-
o
o
O
O





O
o
o
o
o

bl
CM rH
o cn
1 CD
W r-
rH 4


m
CO O
i

VD
*r<
.
CM

CO
m
VD

o
o
o
o





o
o
o
o







CN
rH

rH









ffi














CM
1
W

rH
rH
cn

cn


CM
1
Cd
f-

m
CM




in
r-l
fa)
ID
rH
O
CM
in




o
o
o
o


cn
CN CN
O V
1 rH

CO
rH

ua eg
CD |

to

m
CM

CN

O 1
10

(*]
m
cn
O
r~
CD
CM
O
|

(T)
cn
o
F"
CD

rH


rH

O
O
o
O CN
1

m

m
m


in
*"".
0
t-t
tol
rH
CD
CO
cn
•«ji
CN
O
|
U]
T-t
CO
CD
cn
,3.


CN
o


m

CO
o
m







CM
o














CM
i
tt
CO

CN


en


CN
1

O
CM
m
CN

CM
O
PI
OT
CM
tJ
cn
o
rH
f>
rH




o
o
o
o
o
r
w.
CN \£>
o m
1 0}
W T
** '
r- "a-
ID
in
m o
i

to
VO
.
CN


CN

CO
                             en CM  y?
                             CM CD  c^-
                             O  I   CN
cNonocMOCM
 1   •   I    -I    •   I
                                            CNOCMOCMOCNCNCMCN
                                                          4J W  ^J      *

                                                       rH  (T) tO  El) O JJ
                                                       PO  U O  O O Of  O
                                                       •*     O     000
flrHPOOmOOlW
 I   T  I   O   I  O   I   *r
omooooor-c
CNnCNOCNOCN'tf
                       ^ o   » o   » o   -.
                      O  I  O  • I  O  I   O
       rO  td
        I   in
       O  CD
       CM  CD
       n Dd  i
        I' V
       o o  <
       CN CO  1
i   % O  I
i  O O
t  f> O  '
                      o in o  f> o co  o
 I   O  I  O   I  O
o  o

 I   O
in  o
o  o
rH  O
 I  O
in  o
                                    o o  o o
                                                    «. o
                                                   CD  O
                                                    1   O
          •. o   •» o   -
         o  o o o  o o
         rH  O ,-1 O  rH O
          I   O  I  O   I  O
              O     rH     rH

-------
i  O Ch O
i   I  W  I
I  W CO Id
                         r-  CM  CD
                         CD  O  CO
                         CO   I   CM
                         CO  Id  rH
                                                              «H     «H     tO
                                                                                                                      td      I
                                                                                                                      •  i  ^n
 t4  O  H    On  Q .H  V-l  CTi  ^ cn
    O     •«*     O  O  O  O CM  i
Q  CM O  CM  O CM     iH     rH
r*)   -CM    *  r-1   • in   -c*l   .1
                                                                                                                                             nj  o  rt) O
                                                                                                                                                                at
                                                                                                                                                                6
-M  M  i3  M  tj M J-t
•H  OJ -H  -<*•  -H O "H  O
 5  co  5  «>  3 cn  Jo
    O     O     O>     O
o  -v o  o  o u
                      51
                     Id

                 5  c-
                     tn
               i  O  O

                 A  rH
CD  CM
A  O
     I
TJ  UJ
£  CO
nj  in
    «-i
o  o%

A  *»•
                                                                                                           A     O >
                                                                                                           is     •*,
                                                                                                                                            O
                                                                                                                                            IT)
                                                                                                                                            A  •
                                                                                                                                            ^c  O  ^f o
                                                                                                                                                O     O
                                                                                                                                              ^ o   * o
                                                                                                                                            o  o  o o
m
CM
0 1
1 
CM
i
CM
CO
CM
tn
CN
CM
1
M
CO
<£>
in
o
CO
A
T3
9
o
o
T
in
CM
1
Ed
CO
CD
r-
m
10
CM
1
td
r-
r-
0
CO
A
T3
a
nj
o
^
o
CO
CM
0
Id
r-
rH
iH
•V
in
3
Ed
tn
LO

CO
o
V
Id
£
nj
O
V
A
O
•^

CM
0
PJ
»n
T-H
r-

o
V
T3
a
CD
O
•*
A
o
->s«
1

g
[d
cn
rn
r~
tr-
co
CM
O
Ed
U3
O

V
CO
A
•o
a
flT
o
f*>
A
O
en
i

CM
1
Cd
CT»
CO
in
CO
1-i
CM
1
u
ro
V
CO
A
TJ
K
tti
O
in
A
O
tn
i
o

CM
O
td
CPi
f-H
O
CM
CM
esj
0
in

V
CO
A
-o
c

1
O
CM
O
CM
O
0
in

0
o
o
o
0
o

o
o
m
i
o
CM
0
CM
1
O
O
in

o
o
o
o
m
t
CM
V
CO
A
*d
£
nJ
0
T
A
O
1

CM
I
Ed
CM
r-
lO
(O
CM
1
O
to
n m
o o
(*j r*)
1 O 1 0
o o o o
CM O CM O
O O
Q " 0
CM CM
o c*i o n
-.Ed *. Ed
O Ch O CTi
TH r- t-i r-
m CM in CM
                                                                   CM   *• CM   *
 coowr-
                                                 r- o o  o r»  i   o   i  c?      cooco
                                                 ir>   i  LO   i  m cj  csi  cd CM      1/1  i   m
                                                 " td  "  W  •• rH   "  •»•  "  rH  ..  fj •«
                                                                                                                          coocno     v
                                                                                                                CM t   •* t  «* e   CD

-------
> o
> o
> o
> 0
o
o
o
o
o
0
o
o
o
o
0
o
o
o
o
o
o
o
o
o
o
o
*o
o
o
o
o
o
o
Q
o
o
o
o
o
o
o
o
o
o
0
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
.8155E-0
o
1
til
T
o
ro
to
o
1
b)
to
rH
rH
(>     CO     CD

C% rH  UJ rH  CM
fM rH  lO'iH  m
t?  I   ID  |  "I-
COinrHOCOCDrHrH
    CO     CO     CM      rH
    co     in     o      co
    f-H     o     m      m
                                                   CD O   O  >•!
       t
           CM
         li O  CO     CD

        f5«A     A
      1 — cn   «. o   * o


      I   • in  in o  m o
TJ     TJ     CM     CM
 CS  cn  « o   1  O   I  co

    CD     o   • o   • r-
      ' O O  O G  O CD .—-
in
       in
                                             i*--O*--O   wO  *O
                                                  O     O  CO O CO O
                                                 i O  tO O   I  O  I  O
                                                 • o   ' o  in o m o
                                       "O O  T)     *d     CM     CM
                                        COCOdO   I   *r   I   «r
                                        rto  nlo  oJo^H^rH
                                           O     O     O   «  cn   *  cn
                                             •OOOOOCOOCD
                                                                                                            O     (
                                                                                                           i O  CD I
                                                                                                           , O  A '
                                                                                                                       CTJ
                                                                                                                  !  O
                                                                                                                    o o
                                                   in
                                                          tn
                                                                 in
                                                                     O O  O O O  CM
                                                                                                          . o   -
                                                                                                          i  O  O
                                                                                                          '  O  *r
                                                   ^r     ^r   •  T
                                                                                                        O   *•
                                                                                                        O O C
                                                                                                        O *r (
                                                                                                        O  I  C
 I   O  I   O

CM  O CM  CM   --CM
                      O fd  V O V  W
                                V     fO  CO
                                    CM     O
                                 *  O   *  I
                                co  I  O  Cd

                              t      co  I  r-

                              i  a  ^ CM  *H
                              I  flj  ^r
                                                                         I  CM  I  CM  ro
                                                                        CM  O CM O   •
                                                                 O       •   I    '  I   O
                                                                  i
                         •er     •sr     i
                       i  A     A
                            CM     C-J '
                          - O   «. O i
                       I  O  I  O  I
OOOOOOOO   (   rH
rHOrHO^O^OOCO
 IO  lOAOAQrHVQ
in  o m  o     o     o
     •      ««.'"k>fcCO
 *     v     O     O     O
OrHOrHOOOO   I

 I   CO  I   fO  A  O  A  O  rH
                                    O CM O  CM   ^
                                                           * ^-   » "T  O
                                                                                                                CO  I   I  O  •-
                                                                                                                          O  I   O  I
                                                                                                          •>• O   •• ID  O  I
                                                                                                                  101        w  O   *
                                                                                                                   o  m T o  o o
                                                                                                                  i O     CPi T-I  o H
                                                                                                                   O   -. rH  I  O  t
                                                                                                                             o  «.
                                                                                                                             O O  CO O
                                                                                                                             O rH  CO V
                                                                                                                             O  |   V> A
                                                                                                        S  o  2 o  .
                                                                                                                            *  u  <^ o  •
                                                                                                                            i  co  T-H cn  i
                                                          in o  tn o
                                                         CN  r-f CM rH  CM
                                                                                      CN  r*> CM  CM
                                                                                                                             cn ^*  CPV •*  c
                                                                                                        CMIdCMWCM

-------
                         u
                  O     fj-
                  o     t-
                                                                     •s^      ^J"     O
           tH  CO CO  1
         >)<£     '
           pi]  «3     I
                                            H CM  in  5T «
O  «3 O     en CM  O m
ocMOt  me—  o (^
OiHO  — ^CMQCM
                                              .
                                            T3Of-DI>OCOOCOOCOOeDOO
.0000 .0000 .0000
50 and <0.3, >8 (chrysotile) (avg) "
A

o"

l
O
o •*
o
o ••
0 0
o
f*)
„
o
CO

\£) O

CO
o
CM
1
0
iH


O O
O rH
O 1
o m

tn
V
o 35

fa) ft*
CD
\0 U
*r w
.0000 .0000 .0000
0.4-2, 2-5, 5-8, >8 (chrysotile) (

"*

o
i
O CO
o •
o o
o


O *H
CM cn
m
* •
CM
1
0
rH


O O
rH CO

m p-

*•
V
a
O*
o
O 0
0) G
[*4 G
E
fO
rH
O
X.

CD

rt


n)

>)
p
O

ffl
i-H

Jj
P

^
O


o
CM
A

•>•
CM
V



u
to
h AR>-100 or W>=8 (14 CAT.) (no disc
.0000 .2507 .0000
50 with <0.3, >B (no discharged) "
.0000 .1023 .3292
50 with <0.4, >8 (no discharged) *
.0000 .2528 .5300
50 with 8 (no discharged) "
.6633E-02 .4443 .0000
h <0.5 and AF>=3,>8(no discharged)"
t6633E-02 ,4443 .0000
h >8,<0.5and AK>=10{no discharged) H
. 4356 6 . 4002E-03 . 0000
h <0.5and AR>=20,>8(no discharged)"
.0000 .4509 .0000

"s - - s "2 *S
o^ o o
o in in m o o o
m o i i i in in in
AOOGOOO A A A

OO 0 O OOQO.HO'O
m • «.o «.o woinoinomo
1 O*O-OOIOtiH|O
o -^ -v •* • o • o » o •
•v 1 I 1 •* TT xr

O cOCMCOCMf^l O O O
1 .. I *. | s, o 1 O 1 1
OOOWOUOIOIO O

•- G 1 CD | QO 1 O ^O*-O «cn
OOO\OO^'OcOOry>«^w^>wgwo


^
Q>


ffl

u


II
A
S

3
-H



.-H

—*
O
CM
A

O
CM

O


»•
tH
V



to

(no WDC chrysotile or tremolite) *

**>

A

pi


•3

•-H
CO


o

A

p


5i
52

M
O


-
8
o
o
a

O
M
 oinoy5Oin^*u")cn
 O«HOHOrHCMr-)Ln
 OE   Ot  O*   rHB   CM
                                                  -E   O t   O
                                                                                                CM CM CM  CM CM CM  CM
    cMt^llM     CMUCMM     *^
    CMCDCM     CMrHCMCO     CM
 O"x.o^O^CMjX,cn     V
 ovoiHyso^oinvor*     ^o
 oocnooo«>oo     o
        rO
                       CO
CDCMOCMWCMP-CM     CM     CM      CM     CM     CM     CM     CM


oo   .or-ocMOoooooooocnocnonio
                                                                                           •*-*  I  •*•*  I  'x.  I
                                                                                          I fo U  m td co  ttl
                                                                                          i CJ CD  CM CO CM  "JU
                                                                                                                          ^r- o  m  o «

-------
o
m
o
o
o
o
CM
O
1
r-
in
Ol
cn
00
r-
CO
0
CM
I
i
u
iJ
0
o
d
CO
A
•H
.-H
O
CM
A
O
oo
1
o
o
rH
V
Eli
s
o
S
8
O
O
d
8
c
r-l
m w
• > CD
l£> TH
& CM
CM
m CM
ro fj
CM «*>
"*** rO

CM
m
rH
m
CM
in
cn
o
\
O
in
(O
cn
CN
to*
CO
cn
CM
I
AR>=3 and W>=0.2) (no tremolite or WDC chrysotile}
,c
3
r-i
n)
o
A
O
CM
Oi
f
CM
"-. 1
-r td
•-, m
o o
rH
*3 and W>=0.2) (no tremolite or WDC chrysotile) *
o
A»
CM
t
0 •
rH
O
rH
V
8
<*
" O
01 |
B r
fV
C-J rO
O lA
CM
CO
CN
cn
0
1
Ed
O
o
c-
CM
O
CD
to
=3 and W>=0.2) (no tremolite or WDC chrysotile)
5

n)
rH
A
CQ
\D
10

cn
•a- Id
**-, O
O O
rH
OD
LD
rH
cn
rH
f)
o
1
bl
IT)
O
01
rH
O
1
td
i-H
o
m
CO
o
V£>
r-
o
0
OJ
(all with AR>=3 and W>-0.2} (no tremolite or WDC chrysotile)
A
0
o
rH
rH
V
£
CO
m td
»' r-f
•*»• rH
s CD
CM
CM «>
0 «>
CM
O If)
O r-
rH
0 0
r*J
O
l> rH
t-t
O O

r-' CM'
cn ui
cn CM
i i
m
. CM 0
0 0 V
V 1
w ^
3 o m
ri £ A
Tf C"
£d O
O oo A
"7" CO
I cn *.
m • O
V |
V
8£^
^ ?-
cn o
.. o ••
cot
S n S
?ll
p «a o
•*
o
o
o
1
o
m
o
i
rH
O

r~
cn
EJl
1
A
*.
O
V
>
0
A
0
1 U
^ OH
3K
1 »
O 
O
M
^H
n)
TJ
S
A
O
V
s
O
A
O
t
td CM
?£
"co
0 0
1 0
Cxi ••
CO f
cn
CN
cn o

o
o
m
0
i
CO
rH
O

in
«,
• ^r
1
ir
u
u
Jj
n
-H
>H
ffi
•a
n3
A
O
V
S
o
A
O
I
S£
•"?.
CM ••
0 0
1 0
n? t
cn
cn
CN
in o

ID
r-
rH
0
s
o

w
LO
CM
1
m
11
Tf
O
A
*
0
A
0
l
i U i
[4 Cu W
sy s

i
A
in
c
m
m
d
V
s
O
A
O
1
y%
^ j,
0 0
1 0
Cd ..
to fc
m
CM
10 0

o
o
o
1
u
0
(*>
o
1
u
(H
0

r-
w
cj>
1
in
A
o
rH
1
a
nj
d
V

0
A
0
u) a.
3^
rt^,
O O
1 O
w ••
00 B
cn
cn
CM
cn o

o
0

eg ••
O 0
t o

cn
CM
01 0

0 0 O 0
o o o o
o o o o
o o o o
CM CO ro
o o o
1 1 1
CD o r- *f
m o CM in
•v o m CM
c\i o r* ^r
I-H rH r-
CM m CM
O O O
1 1 1
6J td M
*i» CM rH O>
1 V i-H rH kD
10 r* c*7 o
o ^ in CM
rH O% m
O CD cn o
rH r-* E- <*)
CD O CM rH
o
CM O -CD
1
CM CM cn tr
r- ^" at> rH
r* in « **)
CN £N CM CM
1 1 1 1
th W <0.5 and AR>=30, >5
0, 20-30, 30-40, 40-50, >50 (no discharged chrysotile)
2.4594E-02 7.7840E-03
th W <0,3, >5 (and all structures AK>=3) (no discharged)
th W <0.5 and AR>-20, >5 {no discharged chrysotile)
o
O 0 O
A 0 «J A A
r-* CO
'*. J CO - -
o in • o o
t *. \ i
m in in tn.i
V *j^cjT-


-------
              OJ O  CM


              CM O  CM
                                                                                                                                                  (j
—.     ,-.     CO
•r4 -H V
O O T3
M JQ CO
>< >» ftf O
MM O
•S •§ o°
in
•0 "C A
M M O
(Q tfl u")
A X! i o in m
O o Q o rH A ""> A
n n «* o ii A
•*H 'H O A *• ••
T3 "O •• O QL Q - O
O • lOj CM lO CN
O O "* M 11 It
— — ' ^ £ £ £
co m in rfl sd i< ><
m in in A A
A A A * co *o 73 TJ
o .. c a c:
•> %. •. fO o ro ro o rt nJ nj
tO n> CO ' i O •« • V
OOOCJOI 1 -
O C fl -H --t -H -H -^
AAA ^.oooaoooo
ooo V S A A A A A A A
til O O 0 0 O 0 O
LO 1 t 1 1 1 1 1
— -—,-^vcoin m m m in in m




UcMUCJuSucouSoSucMUCMUt^U^U
vicMWOt/jcntQr- Wf^ cor^ c/)^rmaacacMCft<0<£
rH 	

oojr-i TH CM CM co o •* •>!• m t-(
CD 1 CO O CD O GO OOOQrOOOOOQOOrH


S^SmSSrHoSc^Se^HHSSrHUIHOrH
•C1*SBrHF°e^BSetna^tStSt
A
S
O
A
O
t





U
U

CO
rH



O

m
A
rO
0
V
TJ
CD
0
A
0
T
in




w
AJ

CD
O
CO


o

A
01
• H
O
flj
Q)
09
5
ro
O
V
(fl
O


•*
in


o o

U
fO pQ
in [JH


•v
O CN

to .-
0 0
t-
and <0.3, >5 (no discharged chrysotile)
o
A
o
I
m
CM
S O
U I
pit f^
U o


CTi
^*
CM 0


O H
^
and <0.3, >5 (no discharged chrysotile}
0 and <0,3 (there are no fibers >5) (no
o *r
A
O ^r
i m
tn
CM ^



CO U
£»





n



0



o

>40 with >5, <0.3 (no discharged chrysol
*.
i

^D


o E o


CO O CO
in co CD



o to o


c^ S S
en . r-
A
CO
O
V
x:
•r"
in

in
V




r>




in


rH








r>
I
U

CO






o
m
tn
in
A
o
CM
11
•o
c
r0
ro
O
V
••H
«n


\f




t>
03


^*
m


in
rH

                                                                                            CM CN  CM CM CM
                                                                                                               I O  CO O rH
                                                                                                               I t   rH t   03
tfl  CT* VD
o  t- o
S   CTi t

-------
o
1
O O O 
O O O cn
O O O IT)
m
o o o
t i i
W Ed Id
5
but exclude all M and *
2 studies)
A " CM
O CJ
CM «• ^P •• *
H CO CM *"> fO
K O G O 0
< V • V V
r^
,C ij ^ ,3
S 3 S 3 3
O O UJ 0 O
A A *H A A
O O « O O
1 1 1 1
in m in m
ri CM CM
o £ o E E o K
1 D i CJ O 1 O

5 S 3 o
o in o in to o vo
S.n'S'^o^S^
**• r- co m
i*J CM CM CM CM CM CM
X. X. -v. V^ |
\p VD tO 'f M
CM CM CM CM U>
, O
M
CM
rH

m
cn

CN
rH
fO
O
O
I



2 studies)

A
o
V
5
s
o
A
O
I
in
U
M

CO
\0
in

CM
cn
x. i
o
1
Ed
(M
ro
rH
O
1
S
o
0
CO
"I*
CD
o
in
cn
I



chrysostile only)

A
n
o
V
0
A
0
1
0 «
CO »H

O

" t-f
in r-

 f)
m CM
CM rH

CO tD
O ^"
**• rH
rt
CM rHi
m to
CM CO
r- CM
o o
o o
rH f»1
1 J



amphiboles only)
tR>=20, >5 (2 studies)

1 s
0 0
V V
5 5
0 O
A A
0 O
t 1
in in
CM
U i£ O
W H Cfl

o eg

m CM m
**•
CM CM CM
cn cn c\i
-•» x. i
CM CM r-
o
1
CJ
CTI
O
CM
rH

CO
c,

vo
(D
CM
O
O
1



n"
CD
-H
TJ
CM
A
O
CM
<
rd
0
V
S
o
A
O
1
0

CM
to
m

en CM
*-. 1
CM tO
O O
1 1
W W
CM. O O
^j. o n
0 0
1 1
W fi]
R S S
co CM
o
i
•4* i/) 
en CM H
in CM co
rH r- CM
cn o o
cn o O
1 t I
2 studies) *
E* >i
a a >i
iR>=20, >5 [chrysotile c
iR>=20, >5 (amphiboles c
CS only), <0.3 (FBC on]
"* "• -a
| | S
u
t*i f> ••-*
o O m
V V A
5 5 £
o o o
A A A
O O O
1 1 1
in in in
CM
in •"*•
U CM U ^ U

CM 0 \£>
ID 1 ID O GO
.* 0 " U ••
in P- in co r*-
8 CM B tj t

CN
CO

CM
CO
CO

rH
en
>
c
o
1
o
V
1
TJ
m
u
in
A
5
3
o
A
0
U
w

s
CO
£

CM
-«. f
CM rO
o
W
^,
m
r-
CM
O
1
in
rH
in
r-l
CM
O
1
0
CO
<»•
CO
o
in
r-l
en
CD
rH
1
chrysotile only) *

>
c
0
u
B
o
V
>
c
o
w
u
•a
re
o
m
A
3
o
A
o
1
in
O

CM
O>
r-

OJ
cn
cn
\
CM
O ,
J
U]
eg •
CO
CM
O
0
o
'
o
rH
n
CM
CM
0
O
1
amphiboles only) '

>i
ci
0
o
V
s
i-H
§
u
Tf
U
m
A
4-»
S
0
A
0
I
in
e>g
CM Ed
** ID
o u cn

CO
«3«
O OX O
W - Ed
CM •*
d CM CM
cn
X.
CM
O
Ed
CO
CM
CD
CO

f»>
O
CO
CM
<")
CM
CM
1

,™.
03
no discharged) (2 studa

o
V
in
A
j_j
o
A
O
in
^

rH
m
tn
i>

CM
-20, >5 (no dischargs

O n)
V
tn o
A V
J3 J3
4J 4J
3 S
0 0
A A
0 0
1 1
in in
CM
O CM U
W cn w

r- CM
CM ro
o o r*
" W ••
r- in in
o
<\J ft CM
en ch CM
\ ~x |
CM CM CM
-s,  CM
X. 1
"X O
o
1
rH
CO
o
I
Cd
en
p-
O
CM
CJ
CM
<*>
n
tn
I
- no discharged)
>«
G
Q>
_4
•H
4J
O
£TJ
&
X!
U
m
A

CJ
rt(
o
V
4J
S
o
A
o
1
5
cu
U

r-
p-
in
rH

CM
en
cn
x,
^v
o

-------
U
m
ID
w
rH
m
rH
cn
en
r-t
c-
in

"O
3
CM.
I
ns
o
•H
o
"w
•H
g,

1
-2
"O
O
a
o
a>
•H
•H
O
"m
Cll
o
tit
03
Jj
u
A
0
CM
n
g
o
A
0
m
5
pu
U
CO

n
CD
lO

.9317E-02
"06/26/1992*
o
o
o
o
CM
O
W
cn
U>
CO
TH
O%
u>
in


(0
1
s
_,
o
•H



V>
1
•H
ED
O
•H
(0
•H
0
A
in
3
o
A
o
l
W ft*

«
O CO
1 T

O
CM CN
rH
O
o
o
o

o
o
o
p^
vC>
0


TO
1
-harged chrysotile) T
-H
•o
o
A
V
4J
•H
*
in
m

CN
O CO
l -*•

o
9403 2
"06/26/1992"
o
o
o
o
ro
o
i
u
CO
iH
01
TH
t—
<£
O


CO
?
4J
O
£1
M
U
Cn
M
U
•H
O
A
CM i
t
Id O
o -
• o
VO V
rH O
C>J 0
1

<.-,£
I CO
W -*
IP.O

.9186E-03 -1
"06/26/1992*
DJ
tH
in
m
r*l
CN
CM
in
cn
CO
o
LJ
t-
0
*o

CM
1
"oJ1
•H
0)
u
-H
4J
n
&
u
0
i
)2 S.5809E-02
.ructurea AR>-3
id
T-l >-<
oj O
A
O 0*

0
o
o
W CD

.6766E-G3
"06/26/1992*
3351E
rO
cn
r-
TH
CM
O
J
W
CM
^
rH
»H

CM
rH
CM
M
t
*F THE -SUM OF THE NUMBER OF STRUCTURES'
INSTEAD C
)F SURFACE AREA

M
5
cu

ID
iS
c- o

.3164E-02 4
"06/30/1992*
w
^
CM
o
r-
^
o
o
m
0
0
0


o
r^
rH
Ol
1
a
1
UJ
O
PM
O
O
O
>F VOLUME INSTE;

IH
p
CLI

r-
CN
^
O
«H

.0475E-06
"06/30/1992*
.0362E-06
bd
CN
CD
VD
t-
O
O
o
CN
G
I
PJ
Ch
O
in
o
CM
0
cn
rH
CD
CO
CM
1
en
nj
u
•D
O
O
7
Q]
M
a
2
tn
-H
O
V
in
A
Tf
0
A
o
1
5
cu
ID

I—
O
cn O
v 1
in *
r-t O
r-
"07/01/1992*
3941 2
w
O O tH
O O Ch
o o cn
o o o
CO
O O «H
O O CO
o o cn
r- CM O
in rH O
r- -* o
iH (O O

1-1
CM CM
• EJ t- bl •
VD ^D
CM CM
^D O CD O CM
m U CM pa o
CN <-a- CM CM CN
l • i • l
CM CN
and W <0.3, >=S
.0000 6.1Q57E-02 .1258
and W <0.3, >=5 {no discharged) *
.0000 8.2033E-02 .1382
hrysotile)
o o
11 O II O Q)
A O A O D*
O O 14
O • O • J3
•* •** o
O O -^t
-. *. o
OOP!
1 O t O
o o O O m
-CN o c\j o n
• A
o o *.
CM CM co
I t CN •
rH TH 1 V
•* o * cn 3
rH 0 *H rH -O
t 0 | CA J3
m -m • nj
rH
m 10 ii
V g Vg A
U Id U W O
in -*r
O 0 U TH u
in w
in m ^«
O CM CM
to in TH
r-4 O TH O CO
TH O TH Q rH

"07/02/1992'
.1629E-03
"07/02/1992
.5393E-04
"01/02/1992'
[4
r-
i*
«n
eo
CM
0
cn
cn
O
O
o
0


0
CM
1

;hrysotile
no discharged c.
O
A
O
i
in
U

CN
T^
TH
c
.0646E-03
"07/02/1992'
ID
T^

O

-------
                                                                              o
                                                                              o
                                                                              o
                                                                                                        r-l O

                                                                                                           O

                                                                                                           O
                                                                 CD O     cn O
                                                                 -* o     r* o
                                                                 CM o     CM o
                                                                  10      i  o
                                                                                        cn o    CM  o
                                                                                        r-o    r-o
                                                                                        CM O    eg  O
                                                                                        I  O     I   O
                                                              CM O
                                                              r- o
                                                              CM O


J5
•H
4J
O

5
»
S1

J3
o
01
•d




O
r-
»H
e
eg
o
i

CM
o
m
e
CM
O
fe]

o •«*• «o
o CM •«•
o CM cn
CM
o
1

CM
Cft
r-
E



m





o
o
o
o



^r"
*»•
CM
CM
t


O
O

tn o
c
CM
O
I
U
o
CM

f-



O
o
o
o
                                                                                           o  * o o   ^    <
                                                                                            I  o  i   i  ro

                                                                                        o  cn in co co  o o <
                                                                       ioiou>»oin
                          -P     o
                          O     F3
                          m     —'
                                                                           o co
                                                                           I
                                                                           w o
TJ


 CP
TJ


 O
               • CM  O
             o o  m
             V  i ~j
                W T3
                                •s     i
                                TJ     >i
                                                                                      «Oc\)  IOr-
                                                     »        m

                                                       O O  r-l O
                                                     III      t
                                                     • (£| Cd   "" E»J

                                                     ' r- o  t  *r
TJ


 0     §
V

•o
fi
a
                                                                      lOVOin   • O  O 49
                                                                      ' O     O IO O O«O
                                                                      lOTJOrHVOOO
                                                                      I O  C  O lrt    O  O ^*
                                                                                           o ^
                                                                                           O t-t
                                                                                           O 10
                                                       O O  Q  O
                                                       O O  CB  O
                                                       o o     o
                                                                                                                           V

                                                                                                                           •a
V

3

•o
•tf

o
             O iH O
             •v in •*
              I  Cs]  I
              » 
2
r-l

n m n
1 *-s, t
[j CM u3
r- o m
m *>, r-
if O O
' R
U
rH
"
t-H

og
cn
cn m
x» i
CN M
O O
"•-. CD
O •*
B -
U
m
•*
m

eg
*••. 1
CM W
O cn
•>. m
O M>
P
O
o
*
r,

CM
crt
cn
O
o
e
r- (J
r-l W
eg
0-0
i m i
i
1 "•-. 1
W eg fcd
VD O t*
O3 S^ ^.
O\ O CO
. c
u «r
CO ID
m
" CM
CO rH
O
C\J CO
cn
CM
O
O CD
e cn
U
W
cn
^
c*>

CJ
O fJ
o o
U U
rH CO
' S
n r,

eg oj
O •*• O O
O CO O U>
B • t
U
W
to
*
™

CM
O
0
e
U
CO
to
*
n

ej
CM O
to o o
O3 O rM CO O
en i-
O C4

O O O
°°r-=>-«-
ot oos o o e OOB o^ce
O OO OO OO OtH
CM CM CM eg CM
CM CM CAI CM CM
o o o o o
oooooooooooooo
OB ooe oot oos oo*

-------
in    •*    t-i
                               CM 
CM *v
rH CM

O ^H
O rH
0 0
o o
o o
A
O
V
tJ
ra
o
N
(J
CO
r-
fO
CM
ch
0

/06/1992
0
S
(TJ
O
-P
T3
0
5
§
S
S
CM
0
m
c—
m
o
rH

62E-02
/06/1992t
t- p

in
n
A
S
03
§
U
CM
r-
m
o
rH

91E-05
/06/1992'
CO O
* E:
nf
o
-fl
4J
•H
Sg
O
4J
03
i-H
rt
5
CLi
U
IT)
CO
rH
rH

•74E-Q3
/06/1992
m o
• c

If)
A
S
0)
u
0
o
rH
H
f-l

91E-05
/06/1992
CO O
• t
o
» V - CO
O 1 O O
T 0 •+ |
i CM i ta
OOO OOOCOfO
CMOO «.OOCMCMin
O O O O O VD CM
»OOCMOO «.r-r~
CM O CM rH
1 rH 1
0 0
H  CJX
fO O\   V)
•^ O O O
r — ^oox.oox,o o
inooooooooo
'S O O S OOft OO
A
o
CM
O
CM
t
0
rH
1
in
i
V
E
O
CO
CM
rH

/07/1992
o
A
0
O O (
OOO
O O CM
0 O
0
CM
1
CM O
O rH
O CO rH
O en i
o •* in
-* *
V
g
cu
0 CO
O rH B
e
in
V£>
O O rH
O O
0
0
/07/1992
OOO
O O E
O
1
OOO
0 O CM
O O
a o -
CM
1
0
rH
O
O 0 1
o o in
o o
in
S
On
0 0
O O e
E
in
 O •! CV] •. G*l •.
o o or- o I-H o
•*& o TO «*• in -«r
1 O 1 CM 1 U> i
CM CM CM *n CM
cT CMOcOCMOCO 0
CM OCMOOCMO CM
1 1 1 t 1 1 1 1
O rJ O W EJ O t^ O
vSO-sf fcCO^r •.CMQ *
OCMtOOCO^DOrHrHO
1 CO 1 CM -V 1 CO |
CM CM CM
S o S OS OK
O cO o i£> O fO
t OU3E O^-B O"*E
* CM t CM T CM E
•W O CM rH
10 in o o
••* O O O O t-H T-*
CM 1 
-------
) O CM
0
O
O rH
0
0
o n
0
0
o
o
0
o
o
0
o
o
o
o
o
o
O CM O CD O CD O
O' O O o '
o o o o
0
o
0
0
0
0
o
o
o
o
0
o
       Iri o
       r)   •
       O «H
                                                                      I  rH  I  O  t Cn  I  rH
                                                                      u^Edr-idioucn
                                                            O O
                                                            »H o
                                                               O
               m
                                 n
  • cd
•«r o    f*> o    CM in
r- ^    r- O    f< O
CM n    CM o    eg Q
 I CO     I  O     I  CO

   CO
                                          CD o
                                          cn o
                                          MO
                                           10
                                    iH O
                                    CD  I
                                      • fd
                                    n ID
                                    r- O
                                    CM 01
                                     I O
                                    CM O
                                    <3\ O
                                    CM O
                                     4  O
                                                                      I  O
                                                                                                         IO
                                                                                                         o
•  U">  P-  Ch     rH O     fl\  O


  CO  I   ^     10     10
                                                                                 eg O
                                                                                 r- O
                                                                                 CM Q
                                                           r-
                                                           CM
                                                                                  01 CM  Cl
                                                                                  a) o ,c
                                                                                  U  I  4->
                                                                                  3 tu  cn
                                                                                  4-> r-  « o
                                                                                  U CM  o o
                                                                                    Sen •-* O
                                                                                    m    o
                                                                                  fl o
                                                                                  0) o
                                                                                 -HO
to

X
0>
p-H
ft
> O £ Q
3 TO O O
-j , o (jo



T O
* t
O [d
V \0
o






o
V O
o
* o

X
0)


g
o o
u o
o


CM
o
rH rH |
II 1 Ed

o • o t~
% O O O H rH
jj
flit
X
(V


g
o
y

cn
X



g
i O
u

ni





O
o
o
ID .
S

s

X
0) O
.-i O
fX O
\£> .
5



X
0 O
-H O
EX o
T .
o


II
A
o
•d o
f? O

3
=»
D
M
3
I
d
•n
n
i
n

v



0
1
N
D in

^ cn
n

•T
o
rH
A
1
O
U
c
IIS
O
II
A

O


0
CM


O
CM
i
o
rH

O
g
m
o
o
o
o
o
o
o
0
o
o
o
o








u>
*r
*r


rH
o
o
V
•o
£3
ft
O
II
A
o"
1
. O
CM


O
CM

O
rH


O
TH
in
^
m
o
o
o
o
o
o
o
o
o
o
o
o








o
o
0











CM
0
1
w
o

CD
en
n
o
o
o
o
o
o
o








o
o
0











CM
0
1

CO

rH
o>

o
V
X
0)
i-H
L>
£3
4V
0
II
A
o"
l
o
CM


O
CM

O
TH


o
TH
in
w
in
o
o
o
o
o
o
o
o
o
o
o








o
o
0












0
1
Ed

r-
r>
CM

o
o
V
S
o
•o
s
o
A


O

1
0
CM


0
CM
1
O

<-
^
1
O
o
o
 m
o o
i
v [±1
rH r-
II CM
A 0
^r
G m
o
A

0 0


o *
CM

0
CM
1
o
rH

0 0
t 0
^
in
O
1 0
o o
OJ O
o +
CM
O
O
rH
1 O
in o
m o
V -
•o
o o
I n 4




• ^-4 4 4



CM .a CM
O J-J O
t -H 1

n o CM
in * o CM
„, ^H o rH
TH S •*
O
1
II C-
m
% CD
1 IT)
O
• O
0 0
V 0
o
•a •
O fl
H t




O 4
CM tO

0
CM
1
O
rH

O O
1 0
^
in
CM
o
CM
1
o
rH
O
rH
4
m
1
i
o
CM
O O
[d A

CO O
CM n
«> 1
• o
n CM

0
CM CM
O 1
1 o
[d TH

cn o
•3* |
CM ••

o ••*• uj o
O rH (X O
O T 6  O
u     i
04    DJ
   O CM
03 o n
CM O r-
t:  O in
U     I
CU    fd
   O r-
Ol O M3
en o o
                                         ;5    ?5
w cn o '
Oi cn CM '
f  \O O I
i O  M  i
' CM  &4  i
i O  8   i
                                                                                   i    ?S
                                                                                   I    U 04
                                                                                    a ^
n o
O rH
1 1
in CM
cn o
r- 0
in in
rH rH






O O
o o
o o
o o
n
u>
m




CM
u\
cn
rH
•"-«.
r-
o

c-
o
e


o
o
o
o







o
o
o
o
T
£*
CD
o
E

CM
cn
CTt
rH
*v.
m
O


o
t


o
o
o
o







o
o
o
o

o
1
td


o
(*>
t-
1





o
o
o
o
f*>
CM
T
CD
O
s

CM
cn
cn

"-»,
CO
o

r-
o
E


o
O

o







o
o
o
rH
o

CN O

cn
CM
cn
Ol
rH
--•.
CO
o
O 'v.
O r-
o o
O B



o
o
o







o
o
o
o

o
1
u


n
in

f





o
o
o
o
iH O rH CMOntM -V 0 r-t rH
O IrH 1 CM ICMinrHn ^ "*• ini rH
cnoncnocMcnocMO^ocnooonoo^'QQ^'OO ^tDOtn
OQOQOCMOOTOOOOrHOrHOOrHOOrHOOrHtDOrH
E OCMC OCOB OTS Ot Ot Ot OOt OOS QOE VQOB
om ou3 OCM o o o o* oo oo ino
CMCMCMCMCMrHCMCMtMCM CM CM CMt£>CM
en i cri cnicncncTiOT o\ cn cn o\
cn cn o> encncnenfiai cn a\ cn
rH rH rH .-tr-lrHiHlOTH rH rH rH
"^ \ ">*. 'v ^v. -v. x^ | -v. -v. 'x. **%.
CO CO CO 03 CD CO CD M CD CD CO CO
O O O OOOOrHO O O O
*NxOO'^s.OO^'1siOO'1NiOs-sO^-«.O>^.OO'^OO'vx.OO'^OO'X.
E"-oor~ ocMr~ oor- ot~-or~- on- Or-if- ocpr* oor"Oor~
ooooor^ooooooooooocnoooooooooo
t oos ocnB oot ot OB OB o *E o o s OOB o o t

-------

                                                                              o o
                                                                              CD  I
                                                                              ^ W
                                                                                                 o      o
                                                                                                 V      O
                                                                                                 CM      O
                                                                           O   - O     i*     *f
                                                                           cn  CD r-     r-     r-
                                                                           csi     r*     o     o
                                                                                                                    Eil  O
                                                                                                                    *n   '
                                                                                                                    O
                                                                                                         rH O     O  O
                                                                                                         rH O     rH  O
                                                                                                            00
                                                                                                            o         o
CM  ^r
o  ^o
 I   cn
W   •
m  CM  o
co  t-  o
U)  CM  O
CM   I   O
OJ  CO
O  X!
 I   4J
                   CM O

                   CM O
                                                                                                         CO  O     t*>  O
                                                                            m      —     —
leng

0000

10
                                                                               °£
                                                                                                  rH     m  o  tn ui  m  o
                                                                                                                                       4-JO
                                                                                                                                        t?io
                                                                                                                                        CO
X
CO OJ O
* 8
•Q
C CM
n) o
1
tH W
ID H CO
CM A O
fl VD
CM - CD
. rH
TO

o



o • o
o o o
o v o
00

e
rd


-^ O
II 1
A W
o ; o
O o r*
o ^ r-
O 1 rH
• O *
CM •*

O
7
O
rH
0 * O
o o o
O rH O
O 1 O
- m •

m
V

5

rH O
«i> U 0
CM W O
^ E O
t~
CO
O 1/1 O

pt3 «« pj
co in •*
LO i-H f-

-a1 rO
r*- CM i^>

cn
rH
03
O
O -»x. O
O c— O
o o o
O B O














CM
o
1

CO
in

r>


CM


CD
rH


CM


CM
0
Ed
TH
CM
V
r-



o
o
o
o





o
o
o
0
c/i i
u w
13 CM
C 0
rt ^
U CM
•a
EH CM
1
II c-
AE
•• CD
rH

O



• O
O O
V 0
0
•o •
a
nt

O CO CM CO
•^•00 •
H 1 1 O
A W Sd V
.. I- TH
O CO CO 3!
-^ 0 OJ
1 O \O
O • • "O
CM 10 m c
(U
O
CM OJ
1 O O
O 1 **
rH fd II
-. o cn A
O o cn
rH O rH «,
) O ^ O
m • • -*
-. CM 1
m in
v
CM


O V
U o cn D
w o -^ co
. .
03 ^*
o CM
TH O O
*H 1 fO
en to o rH
O VQ O rH
t ID o e
tn o
CM us CM
cn en

-*, -x.
O O
•x^ o o -x.
o o o o
B O O B

















A




fl

o

ii
A

O

1
in


g


U
to

CD
CM
O
CO
rH
rH
S

CM

cn
01
O
o -v.
rH O
H E
















in
A


0
v



TJ
a
ni

o


in


g


CJ

01
CM
O
4-O
rH
TH
e

CM
cn
CM en
t ~x
U en
•^ o

m o
a saw
-^ \D 01 Ql
tn CD x o 0) M
5 en a? o u ^
CO i
(-3 II • OH TO
A 0 -H 
r»l TJ CM r<> A
d o * • i
o o ni i o o co
V TO W T V
1 O ^-10 0
£ LOOOCMOO^ V
CM O •* cn CM O [5
• l| Ol O
T3 «. A • *• " "O
dm in o a "d
nj CM - CM oJ rt
i CM o i n)
fo in o ^* CM o fo
O O rH 1 1 O tH O O
T 1 W O 1 1 ^- O
ii w * in CM w •• w n -^
Ammo CMOCMA n
CDrHLO^OTH-V A
»C~ICOOCri|CM - •.
OO1CD >CMrOincMO O

1 CM *. O CM - CO 1 |
m eo i-H in in in
V V
CM CM CM

CLittlCLiOP^ DjldOuWCLi
^« o ^r cn G
UcnuOU^UcMUOU
WWJCOOWlDWr- WfOCO

rH *£> f CO CO CM
fO CM •*»• ^ 1^r co
OC3U1 uOOH UIOCO
f^> 1 O O 1 r*> lO 1 rH
tHiricMOcM'>'i^



-4-J


CM


in
A



O
i


Tf

nj

o

n
A
o

1
in


u


u
cn

CM

CO
rH
in
rH


CM
Ch
CM cn
foj 01
fO O
ro "*-.
rH O
• £
rH W
1 o r-
- o m
O O CM
CM
0 CM
1 O
in i
rH td
O O
If)
rH in
O
V CM
o
"O 1
C M

O O






O

1
o
o r~
% o o
O O cn
CM • •

0

CM
•. O
O 1
^ u
1 CO O
in r- o


V TH



rH

o o
woo


CM
ijj
T

rH O O
too
o o

-------
                                                                                     O
                                                                                     O
                                                                                     o
                                  
                                                                          CO td  r-t
  10  r- o
| ^  OJ O
    •   I  O
0.15, 0.15-0, 3, 0.3-1, 1-5, >=5 (Mesotheliomas) "
00 .0000 .0000 .0000
v o
0



o
l\
A
1o
o o
V 0
1 0
• o •


o
1
o
rH
» O
o o
rH O
1 O
00 -5.0341E-OB 4.6448E-03 5.0542E-04
=5 ,1-5,0, 3-1, 0,15-0. 3, <0. 15 (Mesotheliomas) "

o



e>
0
A



O (
* o
CM

O

o
rH
O -
0 0

O )
o
o
o
o
0
o
o
o
0
0

o









o



ej
0
1
U]
(JO
O

r-
t
Cvl •
O O
0 — O
CD O
i-l 4J
0 rH 0
1 |
M 'P til
0 "N. 0
fo S oi
co in
•V X CO
. 0> '
iH -H VO
of
1-* U
1
eg co
CM in ID
(M ' CO
lO
0 -» O

QUO




**• o
II 1




0 1 V



o
CM
1
o

o * o
o o o

O 1 O
B
0
03

G
i-H
VD
"is °
0
X O
QJ .
O CM
U 0
1
m
f> CM
H' rH
A •
^ r
o m
1 rH
U] •
• CD •• O
in rH O
fO -• O
r-



O «r
t II




fO 1 OD
• O •


O
CO CM
0 )
1 0
J-d rH
CO * O
in o o

en | o
E
0
n
0>
£§
o
II
A
)
m
rH O
, o
» o
10 0
rH '
V
O
•*
II O
A.g
0 0
o
o o
I eg

IO O O

tn


0 V O
1 1




CM 4-> CM



—(
CO
O VO
1 -v.
co 2
•HO

VD ffl O

o
01
•To
^-- o
o
m o
V
r
in
H O
• O
- O
1*
A
t-8
0
R O
o" o
o
o o
J CM
W J
CM * o
O O 0
rH |



o v
1




r- -P o



V£> ""*
O VO
1 *••.
W S
rH 0

rH Q) O

O
O
VD




O
1




kD



CO
O
1
tJ
CM

CD

e
O
01
If
to o
1
II O
* o
in o
r-f •
V
-*
o
^»
II 0
A 0
O 0
1
o
CM
o
eg

o o




V 0
1

o


jJ CO



^
VD


O
X 0
d> o
-1 O
o
J

iH




o
1
w
p-


10 ^1

•^ o

V

a
m
c—
r- 0
CO CsJ O
co eg
CM
O
t
bl
ro
m r-
c
0)
kD
J-.! O
4J 0
-rH O
S o
d
0
X
0)
In
0 rH
U •
13
C
ft
o o
v-a o
*
-qt
IF
A

0
W O
1 O
0 O
CM O


O
eg
1
o


* 0
o o
rH O
f O
in *
01
+J
.H 0
O
to o
^
a
o
X in
Q) o>
r-l O
p, co
c]
• o
0 0
1 0
o
eg
o
v§
m o
O
*T
H
A

o

1 0
0 O
CM 0
O

O
eg
iN
rH O
1

O ID
rH O
i eg
un «*
•>=1, 0.4-1, <0.4 (no complex structures)
00 1.1999E-02 2.79B7E-02 ,0000
u» o
o
*p
u
A
eg
*. O
0 I
*r W
t a\
O ro

O

o in
CM
O



O O
rH O
1 O
in o



.
o
V

S
•0




T*
a
a
o
eg
II
A


4-1
C



.
O O
V 'V

5 5
•O T3

is S


•0 T3

n) n.
000

I I II
A A A

X! J3 rC
4J JJ -tJ
Ui O* O
C C C
o
U 0
w o
6 O
B
CM
rH
CO O
CM 1
in
"07/10/1392*
0000 1.
CD
m B
e
eg
in
O
o c
o
eg
&\
CT»
*•>.
O
** o
•V B
VD O U
VD o w
•V O B
r- t
CO
m
o
CM
O O B
o o
oooo
0000
"07/10/1992"
o o
O O O
o o e
e
m
to
TH
o o *
o o
OOOO
oooo
"07/10/1992'
o o
0 0 U
O O E
t
^1*
O
1 CM
O rH t
O O
OOOO
OOOO 1
"07/10/1992*
o o
o r- u
B
eg
CM
CM
fO
O O B
o o
OOOO
OOOO
"07/10/1992*
o o
o o u
o o t
B
rH
CM
VD
O 

m S' to o rH O e o o "07/16/1992* oooo - u u u u « to w » «n ^ in u> CM eg CM ^ eg 04 VD 1 kD 1 ^ O VD rH pr] rH W rH f rH rHC>T-HODc-Hin»— teg T-ICMrHOr-'ChfHCg CfOCCDeCDCW) . VD rH CM eg rg rg ev o\ o\ CT\ eg <-^ ert rH O rH O rH rH ^v 1 -^ 1 ^s. "N. O bJ O Ert O O CM eg eg eo CM CM "*vU>"viCO>>NCM-^s.'«' t~-cMr~cor-nr-^' OcDOlDOrHOin t . £ , t rH e T-I


-------

0
o
o
o




rH
rH





*O
rH

V
to






O
rj\
(/)

lO
CD
CM
1










































•"*

O
H
A

.C
jj
TJ
,^t
S


•o
a
m

o

CM
1

























































0
iH

A
,C
-P
O«
fl
CD
J



^
U
CO

CM

r»
CM




0
<-H





0
CM

CM
rH






ro
to
Ol

CM
t~
CM
1




in
M
A

o
iT
It
^
S




rH
C
0

X
Q)

ft
Ei
O
O

TJ
£
nj


ro

o
v

o
•^
11
A
,c
4-1
-H
3

03
m
b*

m

o
V
•.
o
•«*
1
m

5
•3

m
•Jt
fa



oi
U
to

CD
tH
r-
rH




CD






in
in

*H
rH






r-
ro
in

CM
r-
CM
1










(O
V
o*

1
LO
II
cq



CO

V

o
«?
1
LT>

tn



in
A

O
'jj'
A
U
X
0)
•H



U



fO
• m
V CM
«. in
0 rH
^< .
A
CO
CM
O
- t
(0 U3
V ^o
o o1*

A CM
II
f«
•^
CM o
o S I
1 U U
ul CM ro
in c~
r- U ID
tH U3 (O


rH
P-
CN




0
rH






rH

CM
«Hi






in
fO
o\

CM
p-

O
fO
1










































in
CM

O

V

5
T3
•H
s

t3
§

CO

II
A
A
-iJ
Oi
£
 • >
Q, .-H
y> -H
U *J
X O
cu tn
^ &
-1 O O X3 O O
moo u o o
•*•- o o *- o o
e o o s o o
r- O
• o
en .
0
rH





0
o
CM o
CM O
.
•?•
CM
CM -^
1 0
1

f o m
o tfi

O iH

rH

CM
0
i

O rH

o m
O Q"i

in in
""?
o c-
V 0
fc 1
CO til

O O Ch
1 O rH
LO O CM
* rH
O
r-<
1
fO

O O O
% o o
in o o
1 O O
rH * .

m
n
A

t3

fl] O O
0 0
O O O
** o o
11 • •
A

O*
T
1
o
CM
•. O O
o o o
CM O O
1 O O
O ' .
iH

o
rH
1
in
in o o
V o o
o o
S o o

0*
o
CO fO
o
— 1
>i W
r-t r- o
a e^ o
o m o
m o
OJ + .
-H CT»
•H
4J
O
n
&
X! CM O
U 0 0
— ^ O
C tH O
O O
•ro O
Xf

ad





0
O
VD O
T O

cn
en
rH 
TO o tn
o
ro

O

in
rH O O
• O O
o o o
-. o o
m - -
TH

o
V CM
O
•a i
c w
n o «-t
0 0
O O IT)
•*»• o o
1 • •
O CM
CM

o
CM
I
o
TH 0 O
* o o
o o o
rH O O
t • -
in

m
V

0
•«#
BOO

A o S
£ 0 0
U ' -
Cu
D
CO


^ 0 0
d o o
o o o
O 0
Q) - .
-H
-H
+J
o
VI
&
€S§
— 0 0
too

CM
CD
O
in




(O






rH
CM
fO

CM



{-
O
+
M
^"
o\
in
^
t*
r-

CM
1























0)
rH
••H
4-f
O
n
>1
M
^3
U

4J
a
(U
y
X
0)
i-H
rH
«c


*>>

O

V

in
II
A

£
•.H CM
3 0
1
O bl
-3« 0\
II Oi
A ^
U>

o to
•*
t
IT) CM
0

U N
PU CO
*£>
U ID
« \O
e .
i
W
to
\o
e?i
o

CO


ro






U?
CM
CO

CD



P-
O
£

O
Ul
•*
r-
r-

CM
i









o
*
T3

CD
O


CM
•a

O
o
o
tH
*
rH
T3



>*
i-H
G
O

<1>

•H
4J
O
&
x:
u


in

A



0
V

5

S •*
0
0 1
if W
I m
in r*
•w
*. (O
O '
•* ro
U
A
CM

CU W
CO
U «
CO W
S U3

O
CD
CM
V>

O


O
TH





in
o
o

CO






t£>
in
T-t

rH
t-
CM
1









































in
A


to

V


j;

•o
-r-l
3

o

i
in

m
V

rQ
4J
Cn
fl
0)
r-3

O
t
Csf
o
r»-
CM CM
\o t

rH CM
s O

CM
O>
O% CM
rH O
*X. 1
O W
cM-cn
"x. O
r- O
O CTl
S
CO
CM CM
V& 1
rH W
.' CO
S O

CM
O>
. r-
P" CO
o 
en
rH
\
T-\
CM
"v. fO
t- in
o •*
t rH
0
CM
CTi O
rH 1
.. U
f ^fl
in
CM CM
cr«
&<
rH
^NV
rH
CM
^v O
r- fO
O m
* CD
VD
CO CO
CO 1
co W


•
CM
CA
CT\ CM
rH O
"v. |
VD M
O rH
^V P-
CO rH
O V
C
^
•"»"
o


t ul
ao
CM
en
en
TH
^
^r
rH
"^ O
CM O
H 0
t 0
e
to
10


E O
0
CM
0L
cn

"V,
in
rH
"v. O
CM O
tH O
* O
e
r-

P-
rH

o e
o
CM
en
en
rH
X,
tn
rH
o •*>.
O CM
O rH
o e


o

M
o 
rH
"v.
r-
r-t
•">.
CM
rH
C





0 0
0 0
0 0






o o
o o
o o
o o
e
CD

••**
rO

o in o
c ^ o
CM O
CM
<3\
en
rH
\
CD
rH
"x. O O
CM O O
rH O O
S O O
r-
tO
CM
O
^ r^
rH O
•
CO
 CM
rH O
"*^ I
CD W
O O
^. o
o o
rH CD
C •
t CM
to
rO
O

e CM
CM
to
eh
o\
T-t
"X.
CM
rH
•x co
O CD
rH CO
e CM

-------
                                                               o o
                                                               o o
                                                               4 O
                                                               bi   -
                                                               O O
                                                               O
                                                               O
                                                               o o
                                                                . o
                                                               O 4-
                                                                  (d
                                                                  O
                                                               O O
                                                               o o
                                                               4- O
                                                               W   -
                                                               o o
                                                           td
                                                           O
                                                       o  o
                                                       o  o
                                                       +  o
                                                       W   •
                                                       o  o
                                                       o  o
                                                       o  o
                                                       4  O
                                                       Ed   •
                                                       O  O
O     rH     «-»
O  O
  •  O
O  -I-
    w
o  o
o  o
4  O
o  o
o  o
4  O
W   •
o  o
o
o
o  o
                                                O
                                            o  o
                                            o  o
                                            4-  O
                                            [d   '
                                            O  O
    Ed
    O
O  O
o  o
4  O
Ed   •
O  O
                                                                                                              o o
                                                                                                                . o
                                                                                                              O 4
o  o
o  o
4-  O
Id   •
O  O
                                                                                                                  O
                                                                                                              CM O
                                                                                                              O O
                          CO O
                             o
                             o

                             o


                             CM
                                                                • W
                                                               o o
                                                               iH O
                                            in  o
                                            CD  4
                                            O  [d
                                              •  O
                                            in  o
                                            CD  O
                                            CJ  O
                                             1    •



























A


m
V A



jci rC





*«. •»
0 0
CM ••*



W O W 0
 «H bJ
t- ». nr *• m
f*> ^ o u"> m
en rH CM rH rH
rH * VO S 10
03 C O S rH
-J1 .q.


rH rH O
^v ^s. t
en CO td
rH 


IP
a

TJ
(4
nf



V V




4J 4-3
•o -a
-M -iH



O 0

A A *N



CM CM M
1 1 0
in in CM
LO
Jj ^5 rH
-P 4J

Q) flJ O


t 1 t-

u u m

E S.

t o e r-
e rH e o


en CM en o
rH O rH O

PI Id m cd
CM to CM o
•••», ij> -x^ Qj
O CM O O
E • t
CM O
1 0
rH O
4
* td .
rH O
1 O
m o
• o

- o
 td
• o
o
•. o
en o

1 O
in CM
rH O
• O 1
0 bl

in bl rH
rH O fO
. O en
1 0
o o to

" O
& CM
o
"> O 1
00 bl
•* 4- W
AW to
o o
.. o in
0 0
1
0 O
CM

-. o
0 0
OJ 4 en
1 Cd to
o o m
•Hi 0 CM
O
* O 0
o •
rH O
i a
in o
0 4
•* O td
U1 4 O
V W 0
o o
^ o
o o
— o
50
0
§004-
O O M
S 4- 4- 0
n td Cd O
— o o o
000
U 0 0 -
W O O CP
R
o o
E O
to O
in o o 4-
0440
^* Ed W O
• • o o o
CD O O O
o o o -
E O O O
too


en o o
rH O O

O Cd b] m
en o o •*
x. o o r-
O O 0 O
E
o o
-. o
H 0
1 4
r*l W
• 0
0
H 0
m o

1 O
m CM
rH O
• 0 1
0 U
* H- CM
m td vo
rH O (M
• O to
1 0
00 CM

.. 0
3 CM
0
-* o t

** H- O
A Ed rH
o to
*. O rH
O 0
^ O CM

O 0
CM

* o
O 0
CM 4 CM
1 Ed CM
O O f>
rH o in
o
<- O O
0 •
rH O
1 O
in o
0 4

lO 4- O
V td O
o o
•• O 0
H? O
o o
-~ o
Jtt 0
-W O
O O O 4
O O O Cd
S 4 4- O
«1 Id W O
•— o o o
o o o
U o o *
to o o o
E •
o o
s

CM O O •*
rH 0 0 *
e o O O
e o O
<»« o

Ol O C>J 4
rH O O [d
V. 4 1 O
m u M o
o o en o
*•%, o en o
O O 00 O
E
O rH

















in m
A A


m f>

V V



| S

•H -H



o o

A A


o o


in in


5 CM ff CM
O -P O
O» f Cn 1
d td a td
0) r- 0) in
»-3 O h4 •*

,1 rH | 0
m frj 	 (*)
,0 .a
-P -U
O CM O CM
G O O O

01 Id W fd

in r*>
U <3- U r~
W rH W CM
C .ft •
CM CJ
E t
in o

CM to

..to --en
fn !-* en o
rH CM rH CO
B PO t m
'E o' * 0



rH rH
>»» X.
in in
o en o to
•>* 04 "X to
o to o to
e . t
o o
o
o
H-
Q) fd
n o
0 0
T3 0
o
a> •
J3 0
B  Cn  O o in
rt o o *
Q O O O
t .
0 0
B O
- rH O
CM O +
O 4 O
n td o
.. o r- o
01 O O O
O O rH
t o n o
TOO
-rf 0

cr> O o 4
t-t O O 6d
~x 4 4 O
r-l W W O
rH O O O
'v. O O O
o o o o

o d
^


« CM
CM

m o


m o
rH O
' 4-
b]
* o
in o
rH O
. o
V •
o


o
• *. o
0 4

A 0
0
*. 0
o o

1 o
o
CM
0
*. 0
0 H-
CM td
1 O
0 0
rH 0
o

o o

1
to o
o
*. 4

vg
• • o
h-3 O
—. o
5
G 0


01 Ml

o
w o


o

to
CM O
CD H-

" O

0 0
E O
e o

en

rH O
"^ 4
rH b]
H O
"v, O
o o
C *
o

-------
o
o
o
o
o
o
o
o
bl
O
O
o
o
o
CM
0
3160E
\
CM
o
Ed
co
*£>
c~
 CD IT) **•
V CM V v>
cv *r
.. in .- -j-

4-J CM 4J rH
1=1 a
CD CM ID CM
1*3 O i-3 O
1 )
i Ed i b]
cn CM
O <0 W CM
01 p- Ou en


m 10


r-t O -* O




•* 0 ^ 0

t o t o



cn O cn O
rH O iH O
**-» H- "^ 4-

0 0 O O

cn o en o

CM rH CM rH CO •*

O O CM

CM CM CM
O O O
t 1 1
Ed W b]
fO CD i-t

r- r-- (-•
CM CM W

O
O
£
m r- o
o en o

Ch en o

1-s o *« O ** o
U U U

*. o * o - o

i0 H- i# 4- «) -f.

000
*«• o ** o •* o
o o o o o o
«9" O T O ^ O
A • A • A •
o o o

O 0 0
••sr O i* O -* O
1 O 1 O | O
O -r O + O +
CM Ed CM W CMJ bl
O O O
o o o o o o
CM O CM O CM O
I O 1 O | O
o • o • o •
rH O rH O 1-H O


o o o o o o
r-« 0 rH O HO
1 + 1 + 1 +
m bl in Ed m b]
0 O O

m o m o m o
v o v o v'o

.. O "0 " O
555

C 0 C O P
0) + 0) H> tiJ
o o *r
1 O 1 O I r-
0 O ^
w o to o o co

o o o
4^ 4^ 4J
O o u o u o
0) O OJ 0 V •+• \ -T ''S. |
CD b] CO Cd CO CO
O O O O O <£>
•^ o -^ o -^. o

O O O O O rH
^ O CO
H 1 C\
o foi vo
tO •* CM rH

CM o

CN CM
o o
t 1
M bl
rH CO

r- r-
10 CM

O
£ o a

o r»
«• O *. CJ>



yoyo

o o
CQ o pa o
Ml + »* 4-

o^o
.» o +*. o
o o o o
t*< o •*!* O
A • A *
0 0

o o
W O i* O
1 O 1 O
O H- O 4-
CM b] CM bl
o o
o o o o
CM O CM O
I O 1 O
o • o •
rH O rH O


O O O O
rH O rH O
14-14-

o o

in o m O
V o v o


5 5o

d co
m o> 4-
J rf W
TT O
1 1> 1 O
^- o
O O M O
CQ • CM -
O 0
4J~ 4J
U O U O
 O O
"v. O -v. O

O rH O O
rH rH
CM rH CM rH

O O

CM CM
O O
1 1
td W
tn  o O
-tr o -*1 O
111 +
O W O W
CM co CM o
o o
O rH O O
CM ^o CM o

O *P O O
rH i-l

>, O *. O
o o o o
rH 4- iH 4-
1 bl 1 Ed
m o m o

K O - 0
in o in o


.d ff
Jj T 4-> O
CT O cJi O
fS 1 Pf •+•



m o rH o
-• O " O
CO 4- en -+-
CM b) CM Ed



B O B O



CM



II
A
rH



V '
o
-^
4J 0
tf o


o

o o

A *
o

o

1 O


o
o o

4 O
o
rH 0


0 0
TH O
1 4-

o
* 0



• • o
5o

13 0
35
o
1 0
o

a* -
o
4^
u o
Q] O

•H Cd

C 0
M 0
— ' O

O


•
co r*
cn tf)
10 ^o

o

CM
o
i

CO

. J-
A CM
— n

X
Ty U
U A
r~



- , v
i°-

o 5
CD O Tf


O

O O O

A • A
o

o o

1 O 1


o
o o o

1 O 1
o • o
rH O iH


O O O
rH O rH
1 + 1

o





lot


Q) H- (P
O
I 0 |
o

b • CO
0
4J i7
U 0 O
fl> 0 (B

«H M -i-H

c o rt
MOM
— o —

O


O CM rH
.. o ••

CM W CM



tot


0\ Ol

rH O rH
•s. + x*

r-f O rH


° ° f
O rH O
i m i
b] co td

rH CM

CM CM
O O
t 1
b] W
O (*!

m •• rH
CD A CD

0 0
o X o
wow
0 0
0*0



o y o

0 0
0 03 0


O O

o o o

• A •
o o

o

O 1 O


o o
o o o

o r o
• O '
0 TH O


o o o
0 rH O
4- 1 4-

O O





J=
4J

C
a>

co i CM
cn fo


o o
4~J
o u o
0 CD 0

W -H bl

O fi O
OHO
O — O

0 0

'*•
O fO

+ •cn
w CM



O S V£>


CTi

O rH O
1 "^ 1

rH rH CM



r- i
CM rH



CM
O
I
Ed


CO
CM



U
A




V •
0

4J O



o

o o

A -
o

o

t O
o +

o
o o
CM O
1 O
o •
rH O


O O
rH O
t 4-

O





£1
Oi
C



CN

cn •
o
^J
u o
Q) O

•H Ed
•o o
PI 0
M O
— 0

o

0
CM


rH



£ rH


cn

rH O


CM O


O O
t- t CM 1
CM rH CM CM

rH rH

CM CM
O O
1 E

"91 ^«

CD m
CN CM

rH O\
«H O
II 1
II A td
A ^ CO
•^ «. o


V - V •
o 
-------
0
1
w
CO

O
rH
CM
o
i
Cd
m
in
CO
vo
ro




CM
m
CO
O
en
in
Ul
CM
«i>
CM
in
m
r-
CM
1
o
+
W
O
o
o
o
ro
O

W
rO
CO
CO
c-
y>





O
CO
0
«H
CO
r-
Ch
CM

CM
r*
r-
CM
i
Ed
CD
O
CO




D
O
0


o
o
+
M
0
0
o
o
o
TH
r-l
C-0

CO
CM
m
fO
rH
CO
ro
1

CM
m
CM




CD
U>
•^
r-
0
o
o
+
M
0
0
o
o
o
H
rH
m
m
0
ID
r-
ro
O
O
cO
1
O
u
CO
v>
o
CM
O
1
t3
tfl
m
CO
«)
ro




CM
CO
CD
rH
O
0%
m
m
CM
\D
CM
in
CO
r-
CM
1
O
4-
w
o
o
o
CO
o
1
Ed
to
CO
to
r~
VD




CM
cn
r~
T-^
O
0
rH
m
CD
CO
cn
•v
LO
^«
r-
CM
1
O
1
Fd
en
^.
en




v
o
•«r
fO
O
m
O
i
U
en
en
0
CO
CTi
cn
m
(C
rH
cn
<0
r-
CD
r-
CM
\
o
M

o
CD
CM
o
\
Ed
CM
en

VJ
cn
cn
H
u>
rJ
0
f-
f~
CD
r-
CM
1
o
1
Ed

•V
rO
CM
O
t
Ed
m
CM


10
r-l
0
*
cn
CO
CM
to
O*
cn
rO
r-
CM
i
o
+
Ed
O
o
o
o
ro
o
1
Cd
fO
in
CO
CD
CO
CM
O
I
U]
CM
fO
<£>
O
CO
O
,-H
OD
r»
cn
CM
^-
CM
r-
r~
CM
I
rH
OD
o
OD
f>



0
0
o


0
o
•r
Ed
o
o
o
o
0
1-t
1-1
ro

m
CM
If)
ro
rH
CO
CO
t

CM
tn
CM
O



CD
VP
*»"
r-
o
o
o
-+•
Ed
o
o
o
O
O
r-f
r-t
in
in
o
vo
r-
co
O
O
rO
1
                                                                                                                                                    o
                                                                                                                                                    ti
u ii
A A
V V
a A
\ \

A A

V •.
A
V
!

A

m
A
V
4J
m

A

m
A
V
•H
O

A
V
O
II*
A
V
••H
O

A

0
A
V
. TJ
O
rO
A

'-.
II
A
V •'
A
C **
•d
-H V
O 4->
tO TI
A •£
3

A
V

T)
•H


II II
A A
V V
£ 5
•H -^
S 3
in 10
f) f>
A A

«. *
A
V
•H
in
n
A

in
II
A
V
"s
in

A

in
A
V
o

A

0
A II II
A A
V
V V
/
-H T3 £
. * '^ 1
0 - - j
m 0 0 4
A A •;
* ;
o •- *
\ A
J V
3 -p
^ "S
£ 3

                                                                                                                                                           •3
                                                                                                                                                            (JP
5       ,£.£.£.£.£:,££:,£
    Ed  0)  td  0> td  (D bJ   CD



 lOICNienirHlCOirOlr-ttOI


W   -U   •«   «U   -W   -u   -en  «U   •«
Oi^ftOCMCHrHCOCMCuCMCOiHCurotO^&j
  S  CM  4-1  CM
 S'
 t^  Ed
4JCM4-JCM4-lCM4-JCM4-lllQ>
^-lEdlHU  X-l  fcj  S-J Id  S-l Cd   MEd  tJ


                         «j  Q in  Q  m D
                                                  5     £,
                                                   J

                                                    I

                                                   U
                                                          CJ*O  C7»O  OiO  C3tO  &tO  Oio  tn O  Cn o  Cn

                                                          ffl bJ  (DEd  4>td  Q)  CtJ  fl)EdEd  0)Cd  dl  PJ  QJ
   I  CM

'  [J   .  ,
                                                                             •O   -co   «O   •«   -U   -to
                                                                         oo  uo  oo  uo  uo  uo  uo  uo  u
CM  ro co  ro ro
                  CM     CM      CN     CM     CM

                      in     C3     rH     ro     '
         t  cn to cn  c> CM  '
                                                                                                            r~'  I   r-   I   r-  I  r-
  ..-,%%s
                                                                                                                            ii

-------
i-t      «*      O
td
-er
CD
rO
r~
u?



m
CTI
r*
r-t
o
o
T-H
m
CO
CT*
in
•*r
CM
1













•*
o
•V
fO
o
to
O
M
CM
O
CM
m
en
01
m
u>
CO
r-
00
CM
1













u
iH
CTS

V
en
cn
rH
\o
rH
r-
CO
CM
t













IJJ
O
CM
«3
CM
rO
CM
o
fej
•*r
O
m
o
".0

4-»
01
a
•1— 1
•3
PL)
tn
CM
CO
ro
rH
CM
o
U
rH
US
r-
co
w.
o
CM
CM
o
in
m
CM
i





^^
n

CO
m
CM
i
TJ
(U
4J
VI
?
•o
rt

jj
0
C

-^1
o
•H
M-4
4H

0
0


0)
3
f-H
n.
>
Q.
U<
Q
i
g
0)
•H
^
t
0
tJ













tJ
o
o
o
o
o
o
o
fcd
O
0
o
o
o
CM

O
u>
o
(O
u>
CM
I













O
o
o


o
o
tx]
o
o
o
'o
o
CM
i-H
O
<^>
0
m
\o
CM
1













O
o
o


o
o
w
o
o
o
o
*o
C-J
y?
(Ti
CM
t-«
m
m
in
CO
1













O
O
o


o
o
M
o
o
o
o
o
CM
CM
O
i-l
m
CM
CO
1













O
o
o


o
o
T4
O
o
o
o
o
CM
ro
en
«n
m
01
(C
m
i













o
o
o


o
o
U)
o
o
o
o
o
CM
CMJ-)CM4JCN.-4JCM-UCM4JCM4JCM4J

icic.tiic.cjidittia
W dS W £D W 0) td (D W Q][dQ>td(])bJ flj
CMJo.-^t^jfi-qvD^inji-ij-^'i^
i-iCMCDfOinvf>rOo>
OICMlOTIrHlrOI-OlrHIOI
• U -W 'U 'to 'U -U) «U • W
•"tf'COCMt^rHWCMAHCMOlrHdWfOCOrHCU


o uooouo oooooo uou



CMCM^r-ICMHOfO
CMCMCMCMCMCMCMCM

-HfOlDOBrH'a'P-OX


' (O|O OlO OIOIO1O
bJOMO OWO OWotdOtaJO
eyiQ3cocDr* cof~-coi>cor~cop~oo
OOCMOCMOCnOrHOrHOOJOOOO
rH S CDE CMB O t r4B OE r~B ^B






"X, '"V. "x. Xi "^» "X. "^ "x.
oooooooo
CM'^> **
 • O
^ w as «

CM -4-* CM -P CM
o y o y o
t a> i a) <
til LH pd M [i]

en Q 03 D r*~-
m — m •— in
r- m P-
CM !
T3 Ow

S W

^_i

'M m
m

U CM

S? "
tu o>
*J r-i
n) B
u



*O cn
-H .-i
S "x
t CM
& CM
cn o
§^













o

II
A
£>
•v tn

1 CN

(V) ..
W .-1
in E


VD
a>
CM cn
O eH
1 "X.
a> CM
rH CM
rH O


fO
(O
O
V

tfl
5
T3
•H
f

•o
a
ra

o

II
A
n
•3
tyi
B
(D
1-*

g


W
fe


u>
tn



en
rH
t


UP

^r1  CM
•V OJ
00 *v.

§
r-1

5
cu

w
cu


r-


CM


-------
                      a
                      o
                      o
W
o
o
                                                                                                               o

                                                                                                               o
                                                                           o
                                                                           CO
                                                                            I
                                                                                  01      cH     rH
                                                                                                                                                           8
                                                                                                                                           o


                                                                                                                                           A

in
o
II
A
n
JS
jj
T3
3
TJ
It)
o
A
5
&
d
0)
w
ft*

f-
ID
CO
f>
CF»

u>
en
eft ir
"v. 1
CM  m 0) m (p
J3 m .C •* .d W>
V V V *d I-H "Q ^" "d f)
-H Ol -H tO -H OD
m m w 5 • 5 • £ •
tf J3 JS CO CO f
S S T3 C! d (3
•H -H -H Cti v v v
M ^< V3 eTi M O
dcocico c r- GO a o do
0) fl) 0> Q)  r- I-H o
U}OCO<^D3^tCOr-tWCOWCD
Ai-^Ojroo^-^friyseLiUJO^in
f*> ro f*> O O O
if] to to i/>ou")oino
(N « IN CJ + W + CSI +
fO <*) f*l fOWfOWf^td
^W^CD^OtJ»O^OChO

CTv O\ CTi O> <7i <7»
•V. -** 'iN. •x.4-'Vl4-x^4-
(M CM 
(ft
•~v. |



O
rH
II
A
*2
4->
HI
£
a>
t_>
w

r-f
O
M
M
cn

u>
CTl
*^ 1

CO
o
V


L>
w

CM
O
fO
ffl

u>
en
i "^ i
P 
1 "v, 1

•V
0
II
A

f>
CTx

IP
an.
•**,
o
A
<^
md widths «
o
TH
A
, O
. r-J
V
4J
0)
u
to

f>
o
M
fO

O
1 ~>.
o o M Tf in
A A O ^ O •* O
v o v o v
1 1
*y tO 0] Q) PJ 0) 01
,C rH X! 0 £
•-H f^ -H M >H
ro w 3 . S . 3
5 5 •s1"*^
TJ -a c d a
••H --H fB CJ tti Cs] nJ
33 oo
O 1 O ( O
T* T3 H OJ rH 0) H
c p n to n r- s
o o - o *. 01 >.
rH rH O • O • O
A A I > "
ifj m m
^» >. ^ * ^<
OOOOO-. *. *
i I-H i T-H r m m m
ovtuvtev v v
r-* O 0 O "*
iner-do>c dOd
d) fl) {y a) fl)
CM CJ CNt O
O O O O
Gfl>DwO D
r-t tn i*> o o
OUCO(JOJUOUO\U
OW
O O O O O O O
ro co <*1 + M + fO
co M M w r*> ca PO
o^o^o^i^g*
r-f . . . .
10 tO U> laO «J3
O\ crx o> ox o\
O tTi OCSItTiOO^Otft
•r^v. "X 1 X. -t- ^v, •*• "s.
                .

O10O»-HO(NOCOOOOOOO

-------
i CM
) 0
1
J 0>
J CM
) CM
J \O
i n
CM
O
r
OJ
r-
w
rH
,3.

m
y»
^
*D

I—
n
CO
t-
(M
O
|
H)
o
at
<£>
a?
o
o
4.
w
o
o
o
o
o
o
•f
w
o
o
o
o
o
o
4-
Id
o
o
o
o
o
o
-f-
w
o
o
o
o
o
o
-f
u
o
o
o
o
o
o
+
w
o
o
o
o



o
o
o

CM
o
1
a>
n
iH
n-)
<*1

w>
en
  t— I O r-l O  rH  O

-------
        id
o      o
o      o
       o
        fl
       A
       5
       •o

       '3

       TJ
        G
        (It
       v   J,
        W  CM
                a  i> a
                4J [M U
                T3  (

         U •a'  Dl O  Ol r-t
                         i  O  O  O

                         1   iV  i

                         •  *  AS
                         . Tl*   s.  rH


                           CM   I   CM
      V


.    i  01  O
:  ^  xl  m
I  rH  4-1  C^

fit--  OHO


I  O  0)  O
            1   U   I   I
            0>  PJ  0>  I







o
1
0)
r-
O
S.

in
o
in
CM
rH

£
T-» O
-X J
Co H>
CM U>
O CO
§

in
o
in
CM
rH

£
f-4 O
"•••. 1
CD <|3
CM CM
*^. P-
O ro
S

in
O
L/l

0
o
4-
O
o


o
H-
b)
O
O
o
1

p-
o
in
rH

S
r-t
CO
CM
o
CD
CM
0
O
+
CO
o
o


o
u
o
o
o
Q

o
tn
CM

Cf>
0/28/l(
o
0
o
4-
O
.

O
•+•
u
o
o
o

-------
                              APPENDIX D:

     THE VARIATION IN KL VALUES DERIVED FOR
  CHRYSOTILE MINERS AND CHRYSOTILE TEXTILE

                                WORKERS

The difference between the observed risk of lung cancer for comparable levels of chrysotile
exposure among Quebec miners (most recent followup: Liddell et al. 1997) and South Carolina
textile workers (Dement et al. 1994; McDonald et al. 1983a) has been the focus of much
attention.  Reasonably good agreement between results from the Quebec studies and another
study of chrysotile miners in Italy (Piolatto et al. 1990) coupled with reasonably good agreement
between results from the South Carolina Plant and results from textile plants in Mannheim,
Pennsylvania (McDonald et al.  1983b) and in Roachdale, England (Peto 1980a,b; Peto et al.
1985) suggest that the difference between Quebec and South Carolina may reflect a general
difference between the two industries (see Table 7-6 and Section 7.2.2). This appears true
despite the fact, for example, that cohorts at two of the textile plants were apparently exposed to
significant amounts of amphibole in addition to chrysotile (see Appendix A and Section 7.2.2).

Three main hypotheses have been advanced to explain the difference in the risk per unit
exposure observed among miners and textile workers (see, for example, Sebastien et al. 1989).
These are:

       (1)    the low reliability of exposure estimates in the various studies;

       (2)    differences in fiber size distributions in the two industries (with textile-related
             exposures presumably involving greater fractions of longer fibers); or

       (3)    simultaneous exposure to a co-carcinogen (i.e., oil that may have been sprayed on
             the asbestos fibers) in the textile industry.

It has also been proposed that differences in the concentration of long tremolite (amphibole)
fibers in dusts from each of the two industries might represent an explanatory factor (see, for
example, McDonald 1998b). However, this would also require a large relative difference
between the potencies of tremolite (amphiboles) and chrysotile toward the induction of lung
cancer. This latter issue is addressed further in Sections 7.4-7.6. McDonald (1998b) also
presents an overview of the current status of each of the hypotheses described above.

In an attempt to distinguish among the above-listed hypotheses, Sebastien et al. (1989)
conducted a study to determine lung fiber concentrations in tissue samples from deceased
members of the cohorts studied from both the Quebec mines (specifically, from the Thetford
mine) and the South Carolina textile plant. These researchers ultimately analyzed tissue samples
from 72 members of the South Carolina cohort and 89 members of the Thetford (Quebec) cohort.
Because the tissue samples came from cohort members, they could be matched with estimates of
the exposure experienced by each of the individuals as well as details concerning the age at first
employment, the age at death, the years of employment, and the number of years following
employment until death.

                                        D-1

-------
In the Sebastien et al. (1989) study, tissue samples were obtained in formalin-fixed or paraffin
blocks, which were then digested in bleach, filtered, and analyzed by TEM.  Tissue samples
were apparently "opportunistic."  Only fibers longer than 5 |im with an aspect ratio >3:1 were
included in the count  For consideration of the limitations associated with such preparations, see
Section 5.2.

Results from matching of tissue samples with the histories of corresponding cohort members
indicate that tissue samples obtained from each cohort covered a broad range of exposure levels,
duration of exposure, and years since the end of exposure. They also indicate that South
Carolina cohort members included in the Sebastien et al. (1989) study experienced, oh average,
13.5 years of exposure with 18.1 years between the end of exposure and death.  In contrast,
Thetford workers included in this study experienced an average of 32.6 years of exposure with
only 11.6 years between the end of exposure and  death. Corresponding to differences 'in
exposure levels observed across the two cohorts in the original epidemiology studies, mean
exposure levels experienced by Thetford cohort members included in this study were about 10
times mean exposure levels experienced by South Carolina workers (19.5 mpcf vs. 1.9 mpcf).
                                                                             i'
Because Sebastien and coworkers recognized the general lack of a good model describing the
retention and clearance of asbestos fibers in the lungs at the time their study was conducted, they
performed most of their analyses either on pairs of members (one from each cohort) matched for
duration of exposure and time since end of exposure or on groups  of members from each cohort
similarly stratified by duration of exposure and time since end of exposure.

Results from their study  indicate that, overall, lung  burdens observed among Thetford'cohort
members are substantially higher than those observed among South Carolina cohort members.
Geometric mean lung chrysotile concentrations are  reported to be  5.3 and 0.63 fibers/jig dry lung
tissue in Thetford workers and South Carolina workers, respectively. Furthermore, despite
tremolite representing only a minor contaminant in  the chrysotile from Quebec and the dusts to
which the miners were exposed (Sebastien et al. 1986), the majority of fibers observed in the
lungs of Thetford miners were in fact tremolite (mean concentration 18.4 f/jig dry lung).  Since
the raw material used in the South Carolina plant came largely from Quebec, tremolite was also
expected to be a minor contaminant in the dusts to which textile workers were exposed. Yet
among these workers also, tremolite represented a substantial fraction of the lung fibers observed
(mean concentration 0.36 f/jig). Thus, the ratio of tremolite concentrations observed among
Thetford miners and that observed among South Carolina workers (18.4:0.36, or 51) is even
more extreme than the ratio observed for chrysotile (8.4).

To evaluate the first of the above-listed hypotheses, it is instructive to compare the ratios of
chrysotile or tremolite fibers  observed in the lungs of deceased workers from Thetford and South
Carolina, respectively, with the overall exposures that each received. A rough estimate of
cumulative exposure for each set of workers  in the Sebastien et al. (1989) study representing
each cohort can be derived as the product of the mean duration of exposure and the mean
intensity of exposure.  Thus,  for example, mean cumulative exposure in Thetford wasl32.6
yearsx!9.5 mpcf or 635.7 mpcf-yrs. Similarly, for South Carolina, mean cumulative exposure
was 25.65 mpcf-yrs, which gives a Thetford/South Carolina ratio of 24.8. This presumably
represents the relative cumulative exposure to chrysotile. For tremolite, Sebastien and
coworkers report that, based on a regression analysis, the fraction  of tremolite fibers among total

                                           D-2            ;

-------
asbestos fibers were likely only 0.4 times as much in South Carolina as in Thetford (where they
likely averaged 1% of total fibers). Therefore, the ratio of cumulative exposures to tremolite for
the sets of cohort members studied by Sebastian and coworkers is likely 62.

Comparing the ratio of Thetford:South Carolina lung burden estimates with the ratios of the
corresponding cumulative exposures, it appears that the chrysotile lung burden ratio (8.4) is only
a third of the ratio predicted based on cumulative exposure (24.8). However, the ratio of lung
tremolite concentrations (51) is much closer to the corresponding cumulative exposure ratio (62).
It thus appears that, although, airborne concentrations may not closely track the exposures that
led to the observed lung burdens for individuals (see below), the overall trend in exposures
predicted by airborne measurements is approximately correct.  It is therefore likely that overall
exposure concentrations in Thetford were in fact substantially higher than in South Carolina (in
agreement with airborne measurements).  Thus,  we concur with Sebastien et al. that the
unreliability of exposure estimates in these two cohorts is unlikely to explain the observed
difference in the risk per unit of exposure observed for each cohort.

Importantly, although the general trend in relative overall exposure levels predicted by airborne
measurements between Thetford and South Carolina appear to have been confirmed by mean
lung fiber concentrations in the Sebastien et al. (1989) study, the estimated exposures correlate
poorly with lung burdens for any particular individual. To demonstrate this, we analyzed the
Thetford:South Carolina ratios of lung chrysotile concentrations and, separately, lung tremolite
concentrations reported by Sebastien et al. for their set of 32 matched pairs of cohort workers to
determine whether trends in these ratios adequately matched trends in the corresponding
estimated airborne exposure level ratios for the same matched pairs.  To do this, we subjected the
ratios presented in Table 7 of the Sebastien et al. (1989) study to a Rank Von Neuman test
(Gilbert 1987). Results indicate that trends in neither lung chrysotile concentration ratios nor
lung tremolite concentration ratios can be predicted by the observed trend in the estimated
airborne concentration ratios among these 32 matched pairs.

There are numerous sources of potential uncertainty that may mask the relationship between
airborne exposure estimates and resulting lung burdens (Section 5.2). Potentially the largest of
these is the variation expected among lung burden estimates derived from use of "opportunistic"
tissue samples, which are not controlled for the portion of the respiratory tree represented by the
sample. Even for samples collected from adjacent locations in lung parenchyma, observed fiber
concentrations may vary substantially and such variation is magnified between samples taken
from different individuals at locations in the lung that may not in any way correspond to their
relative position in the respiratory tree.

Other potentially important sources of variation that may mask the relationship between airborne
exposure concentrations and resulting lung burden estimates may primarily involve limitations in
the degree to which the airborne estimates from an epidemiology study represent actual
exposures to the individual members of a study cohort (Section 5.1). The following factors may
all contribute to the uncertainty of exposure estimates:potential differences between individual
exposures versus area concentrations (which are what is typically measured), the adequacy of
extrapolation to the earliest exposures in a cohort (when measurements were generally not
available), or the adequacy of estimating job x time matrices for individual workers that can then
be integrated  with work area exposure estimates to  derive individual exposure estimates.

                                          D-3

-------
 The second of the above-listed hypotheses, involves potential differences in the size of structures
 that may have been present in the airborne concentrations in Thetford and South Carolina, which
 may not have been adequately represented by the exposure measurements.  More generally, this
 is a question of the degree to which measured exposures in the two environments adequately
 reflect potential differences in the character of exposure that relate to biological activity.
                                                                              I.
 Sebastien et al. (1989) considered this second hypothesis by generating and comparing size
 distributions for the fibers observed in the lungs of workers from Thetford and, separately, South
 Carolina.  Importantly, the size distributions for each cohort were generated by including the first
 five fibers observed from every member of that cohort, without regard to the duration of
 exposure, level of exposure, or time since exposure experienced by each cohort member.
 Therefore, the size distributions obtained are "averaged" over very different time frames during
 which differing degrees of fiber retention and clearance will have taken place, each of which
 potentially alters the distributions of fiber sizes (Section 6.2).  Thus,  the two distributions
 generated are each actually collections of samples from multiple, varied size distributions (rather
 than single distributions) and this likely masks distinctions between the two work environments.
 It is therefore not surprising that the authors found relatively little differences in the two size
 distributions.                                              ,

 The portion of the generated size distributions that are least likely to have been affected by the .
 limitations due to the manner in which they are generated (as Sebastien et al. suggest)]is the
 fraction of tremolite (amphibole) fibers longer than 20 urn  This is because (1) tremolite fibers
 (unlike chrysotile) are biodurable and (2) biodurable fibers longer than approximately;20 (im
 have been shown to clear from the lung only very slowly, if at all (Section 6.2). Thus,- the
 Thetford: South Carolina ratio of long tremolite fibers may provide the best indication pf the
 relative exposures to long fibers in the two environments.
                                                                              '!
 Table D-l presents the estimated, relative concentrations of specific lengths of fibers observed in
 lung tissue among Thetford miners  and South Carolina workers, respectively.  The length
 category for various fibers is presented in the last column of the table. The estimated ,.
 concentrations, presented in Columns 2 (for Thetford) and 3 (for South Carolina) of this table
 were derived as follows. For the first length category (L>5 um), concentrations are taken
 directly from Table 5 of the Sebastien et al. (1989) paper (the geometric means are presented).
 Concentrations for the remaining length categories were estimated by multiplying the ,
 concentrations for this first length category by the fraction of the size distribution represented by
 each succeeding length category (as provided in Table 4 of the Sebastien  et al. paper).' So that
the relative precision of these concentration estimates can be evaluated, an estimate of the
numbers of fibers included in each length category (from the total used to derive the size
distribution in Table 4 of Sebastien et al.) are provided in Columns 6 (for Thetford) and 7 (for
 South Carolina), respectively.  The Thetford: South Carolina ratios  of the concentrations of fibers
in each length category (for each fiber type) are provided in Column  5 of the table.    '
                                          D-4

-------
 Table D-l. Estimated Concentrations of Sized Fibers Observed in the Lungs of Thetford
                      Miners and South Carolina Textile Workers"
                   MEAN LUNG
                 CONCENTRATION
NUMBER OF FIBERS
Fiber Type
Chrys
Trem
Chrys
Trem
Chrys
Trem
Chrys
Trem
Thetford
5.3
18.4
1.73
3.90
0.59
0.72
0.16
0.037
South
Carolina
0.63
0.38
0.17
0.091
0.070
0.024
0.031
0.008
Units
f/ug.lung
f/\ig lung
%glung.
f/ug lung
f/[ig lung
f/H-g lung
f/H-g lung
f/tig lung
Ratio:
Th/SC
8.41
48.42
10.00
42.95
8.41
30.46
5.15
4.40
Thetford
371
405
121
86
41
16
11
1
South
Carolina
226
175
62
42
25
11
11
4
Size Range of
Fibersb
Length>5 |im

Length>8 ^m

Length>13jim

Length>20|im

"Derived from data presented in Tables 4 and 5 of Sebastien et al. (1989)
bGeometric mean

It is instructive to compare the ratios presented in Table D-l to the Thetford:South Carolina
ratios of mean cumulative exposures estimated above for chrysotile and tremolite among the
cohort members included in the Sebastien et al. (1989) study (24.8 and 62, respectively). As
indicated in Table D-l, for chrysotile, the ratio remains approximately constant at about 9
(varying only between 8.4 and 10) for all of the size ranges reported except the longest. For the
longest category (L>20), however, the ratio drops to 5.  Because fibers longer than 20 um are
expected to be the most persistent in the body (Section 6.2), it may be that the ratio of 5 best
represents the relative concentration of long chrysotile structures among the two sets of cohort
members.

Because this ratio (for the long fibers found in the lung) is  only approximately 1/5 of the
estimated ratio for the cumulative exposure to chrysotile (24.8), this suggests that the South
Carolina cohort may indeed have been exposed to dusts enriched in long fibers relative to dusts
experienced at Thetford. Because the estimate of this ratio is based on counts of at least 11
fibers from Thetford and South Carolina, respectively, it is unlikely that this ratio will vary by
more than a factor of 2 or 3 (the 95% CI  around 11 fibers, based on a Poisson distribution is
6-19).

The trend with tremolite is even more striking.  Moreover, as previously indicated, because
tremolite fibers are biodurable, it is the tremolite fibers longer than 20 um that may best
represent the ratio of long fibers to which these two groups of cohort members were exposed.
The ratios observed among tremolite fibers steadily decrease from approximately 50 for fibers
longer than 5 um to 4.4 for fibers longer than 20 um, although this last value is uncertain (due to
it being based on only 1 fiber observed among Thetford-derived lungs and only 4 fibers among
South Carolina-derived lungs). In fact these  data are statistically consistent even with a ratio
considerably less than I, (i.e., with a considerably higher concentration of long tremolite fibers
in South Carolina than in Quebec).  Given that the ratio of the original cumulative exposures for
tremolite was estimated to be 62, mat the ratio of long tremolite fibers is only 4.4 suggests that
dusts in South Carolina may have been highly enriched in long fibers.
                                          D-5

-------
Observations that the fibers to which textile workers were exposed were longer and tijinner than
those found in mining are further supported by various published size distributions of fibers
determined in air samples collected in these environments (see, for example, Gibbs and Hwang
1975,1980). Also, as noted in Crump (1986), the raw fiber purchased by textile plants was
commonly described as the longest grade of product (see Table 22 of Crump). Size issues are
addressed further in Section 7.4.

The data in a more recent study by Case et al. (2000) demonstrates even more strongly that
South Carolina textile workers were exposed to fibers that were substantially longer tljan those
inhaled by Quebec miners and millers. In this study, lung fiber contents were determined for 64
deceased textile workers and 43 deceased chrysotile miners and millers, respectively, which
represent randomly selected subsets of the workers, miners, and millers for whom lung burdens
were previously described by Sebastien et al. (1989), as discussed above.           i:

In the Case et al. (2000) study, analyses were conducted on sets of TEM specimen grills that had
originally been prepared in the Sebastien et al. (1989) study, thus selection of subjects, and the
preparation of samples in this study is the same as described above for the Sebastien et al. study.
However, Case et al. focused specifically on the counting of fibers longer than 18  \im.\

Results from the Case et al. (2000) study are summarized in Table D-2. As indicated in the
second column of Table D-2, the mean cumulative exposure to which the selected cohort
members from Quebec and South Carolina were exposed in this study was 186 and 3.63 mpcf-y
(millions of particles per cubic ft-years), respectively. This gives a Quebec/South Carolina ratio
of approximately 51. In contrast the Quebec/South Carolina ratios of the concentrations of
asbestos fibers observed in lungs among these selected cohort members are substantially smaller
(4.28 for long chrysotile, 12.04 for long tremolite, and 5.45 for long amphibole).  This'implies
mat the lungs of South Carolina workers are substantially enriched in these long fibers  relative to
the lungs of Quebec miners and millers. Moreover, because substantial numbers of long fibers
were counted in these analyses, the uncertainty of these ratios is relatively small.    :

   TABLE D-2. ESTIMATED MEAN AIRBORNE EXPOSURE CONCENTRATIONS
 AND ASSOCIATED LUNG FIBER BURDENS FOR A SELECTED SET OF TEXTILE
                       WORKERS, MINERS, AND MILLERS"             ,j
Location
Quebec Mining
SC Textiles
Ratio
Mean Airborne
Exposure
Concentration
(mpcfy)
186
3.63
51.24
Lung
Chrysotile
Content
(long fibers)
(f/ug)
0.231
0.054
4.28
Lung
Tremolite
Content
(long fibers)
!(f/Jig)
. 0.325
0.027
12.04
Lung Total
Amphibole
Content
(long 'fibers)
(f/|ig)
0;349
0.'064
5^45
 aDerived from data presented in Table 2 of Case et al. (2000)
                                        D-6

-------
If the estimated KL's derived for Quebec miners (0.00029) and South Carolina textile workers
(0.021), as reported in Table 7-6, are adjusted to account for the relative concentrations of long
fibers reported by Case et al. the disparity in these KL estimates effectively disappears.  If
adjusted as described in Section 7.4.2, the new KL«'s for Quebec (0.234) and for South Carolina
(1.21) now differ by only a factor of 5 (rather than the original factor of 72). Thus, accounting
for long structures appears to reconcile these potency estimates.

The data presented by Case et al. (2000) also indicates that the lungs of textile workers  in South
Carolina (but not Uiose of Quebec miners) contain substantial concentrations of commercial
amphibole asbestos fibers (amosite and crocidolite) in addition to tremolite. In fact, the majority
of the amphibole fibers observed in lungs from South Carolina workers were composed of the
commercial amphibole types. This suggests, among other things, that the exposure environment
in South 'Carolina should actually be characterized as a mixed exposure environment rather than
a chrysotile exposure environment.  As indicated in the following two paragraphs, however,
conclusions concerning the nature of the general exposure environments in Quebec mines or the
South Carolina textile mill that are based only on observations among the small subsets of these
cohorts examined by Case et al. may not be robust.

Importantly, Case et al. indicate in their paper that, because they observed substantially greater
absolute numbers of long fibers in the lungs of Quebec miners than in the lungs of South
Carolina workers, they conclude that (regardless of the above analysis), Quebec miners were still
exposed to a greater absolute number of long fibers than South Carolina workers.  However, this
does not appear to be a valid conclusion that can be derived from the data provided in the paper.

We compared the mean exposure concentrations reported for the subset of Quebec miners and
South Carolina textile workers that Case et al. (2000) examined (Table D-2) to the distribution of
exposures reported among the entire cohorts in Quebec (Table A-2) and South Carolina (Table
A-8), respectively.  Results  suggest that, exposures for the subset of Quebec cohort members
included in the Case et al. study are higher than approximately 75% of the exposures
experienced by the overall cohort. In  contrast, exposures for the subset of the textile worker
cohort examined by Case et al.  are lower man approximately 50% of exposures experienced in
the overall cohort.  Thus, given that the mean exposures experienced by the subsets of each
cohort examined by Case et al. do not reflect mean exposures for the respective cohorts as a
whole, it is not reasonable to compare absolute numbers of structures observed in the lungs of
these workers and draw general conclusions about the relative, absolute exposures among the
entire, respective cohorts.

At this point it is worth mentioning some of the potential differences in the characteristics of
mining dusts and textile mill dusts that may affect biological activity, but that may not be
adequately delineated when measuring exposures by PCM (in f/ml) and almost certainly not
delineated when exposures are measured by midget impinger (in mpcf), see Section 4.3. During
the mining of asbestos, only a small fraction of the rock (generally no more than 10%) that is
mined is typically composed of the fibers of interest

While the host rock in a mine may be  of similar chemical composition, it generally represents an
entirely different crystalline habit. Nevertheless, a large fraction of the dust that is created
during mining is likely composed of fragments from the host rock and many of these fragments

                                           D-7

-------
will be of a size that would be included in the particles counted by midget impinger.   '
Furthermore, at least some fraction of the fragments created by the crushing and cutting of the
host rock will be elongated "cleavage" fragments (Section 4.0) so that at least some fraction of
these may be included even in PCM counts, despite many of them being either too thick to be
respirable,  or too short or thick to be biologically active (see Section 6.2). Note, although
Sebastien et al. (1989) employed TEM to characterize fibers in the study, they apparently
employed a fiber definition that was sufficiently broad that they too  would have counted large
numbers of structures that may be too short or too thick to contribute to biological activity.
                                                                              i
In comparison, the dusts created in a textile factory are likely composed almost exclusively of
true asbestos fibers. The raw material received by the factory will already have been milled and
beneficiated to remove the vast majority of non-fibrous material.  It  is therefore, much less likely
that extraneous fragments (even cleavage fragments) exist that might be counted either by
midget impinger or PCM.  We make this point because, if this represents the true situation, it
would be expected that risk per unit exposure estimates (i.e., exposure-response factors) derived
from any mining site, may be smaller than estimates derived for the  same fiber type in |
occupational environments where only finished fiber is used.  Thus,  another interpretation of the
variation observed among estimated KL values for amphiboles (reported in Section 7.2J2) is that
mining values are somewhat low.  As later described (Section 7.3.2), the same may be true for
amphibole  KM values. The implications of this possibility are discussed further in each
respective section.
                                                                              i
Note, although the Sebastien et al. (1989) paper suggests that (mpcf) exposure estimates from
Thetford and South Carolina grossly suggest the relative range of lung burdens observed, there is
too much scatter in the data to determine how closely the air ratios track the lung burden ratios.
For example, ratios derived from arithmetic means (rather than the geometric means) for the
Sebastien et al. data are substantially different Moreover, as indicated above, there may be
substantially different size distributions in the two environments, which might at least in part be
explained by the inclusion of large numbers of cleavage fragments (with dimensions  ]
inappropriate for biological activity) in the mining environment.

Although the third of the above-listed hypotheses was not addressed by Sebastien and ;
coworkers, the question of whether a co-carcinogen contributes to the overall observedjung
cancer rate  among textile workers has been considered by several other researchers. To test the
hypothesis  of whether oils potentially contributed to disease in South Carolina, Dement and
Brown (1994) performed a nested case-control study among a subset of the cohort members
previously  studied by Dement et al. (most recent update, 1994). In this analysis, Dement and
Brown qualitatively assessed the probability of mineral oil exposure for cases and controls based
on knowledge of historic descriptions of mineral oil use. The extent of such exposure was then
further categorized into three strata: none or little, moderate, or heavy, based on where'each
worker was longest employed.  Cases and controls were then further categorized based'on years
at risk and level of asbestos exposure.  Results from this nested analysis indicated no significant
change in the estimated exposure-response slope for asbestos after adjusting for mineral oil
exposure.

Additional, albeit qualitative, evidence that oils may not represent an adequate explanation for
the relative lung cancer risks observed in mining and textiles is provided by McDonald (1998b).

                                          D-8             ;                    ':

-------
McDonald suggests that oils were not used in the Roachdale plant until 1974.  Therefore, due to
latency, it is unlikely that the use of such oils would have had a substantial impact on the
observed lung cancer cases at the point in time that the study was conducted (Peto 1980a,b; Peto
etal. 1985).

Taken as a whole, the evidence presented in this section suggests that the relative distribution of
fiber sizes found in dusts in the textile industry and the mining industry, respectively, may be the
leading hypothesis for explaining the observed differences in lung cancer risk per unit of
exposure between these two industries.

REFERENCES

Case BW; Dufresne A; McDonald AD; McDonald JC; Sebastien P. Asbestos Fiber Type and
Length in Lungs of Chrysotile Textile and Production Workers:  Fibers Longer than 18 \an.
Inhalation Toxicology. l(Suppl 1):411^H8. 2000,

Crump KS. Asbestos Potency Assessment for EPA Hearing. Prepared for Asbestos Information
Association/North America 116pp. 1986.

Dement JM; Brown DP. Lung Cancer Mortality Among Asbestos Textile Workers: A Review
and Update. Annals of Occupational Hygiene. 38(4):525-532.  1994.

Dement JM; Brown DP; Okun A. Follow-up Study of Chrysotile Asbestos Textile Workers:
Cohort Mortality and Case-Control Analysis. American Journal of Industrial Medicine.
26:431-447. 1994.

Gibbs GW; Hwang CY. Physical Parameters of Airborne Asbestos Fibres in Various Work
Environments - Preliminary Findings. American Industrial Hygiene Association Journal.
36(6):459-466. 1975.

Gibbs GW; Hwang CY. Dimensions of Airborne Asbestos Fibers. In Biological Effects of
Mineral Fibers. Wagner JC (ed.). I ARC Scientific Publication, pp. 69-78.  1980.

Gilbert O. Statistical Method for Environmental Pollution Monitoring.  VanNostrand Reinhold,
New York. 1987.

Liddell FDK; McDonald AD; McDonald JC. The 1891-1920 Birth Cohort of Quebec Chrysotile
Miners and Millers: Development From 1904 and Mortality to 1992. Annals of Occupational
Hygiene. 41:13-36.  1997.

McDonald AD; Fry JS; Wooley AJ; McDonald JC. Dust Exposure and Mortality in an
American Chrysotile Textile Plant.  British Journal of Indus trial Medicine. 39:361-367. 1983a.

McDonald AD; Fry JS; Woolley AJ; McDonald JC. Dust Exposure and Mortality in an
American Factory Using Chrysotile, Amosite, and Crocidolite in Mainly Textile Manufacture.
British Journal of Industrial Medicine. 40:368-374.  1983b.
                                         D-9

-------
McDonald JC. Invited Editorial: Unfinished Business - The Asbestos Textiles Mystery. Annals
of Occupational Hygiene. 42(l):3-5. 1998b.

Peto J. Lung Cancer Mortality in Relation to Measured Dust Levels in an Asbestos Textile
Factory.  In Biological Effects of Mineral Fibres, Wagner JC (ed.).  I ARC Scientific !
Publications, pp.  829-836.  1980a.

Peto J. The Incidence of Pleural Mesothelioma in Chrysotile Asbestos Textile Workers. In
Biological Effects of Mineral Fibres.  Wagner JC (ed.). IARC Scientific Publications;  pp.
703-711.  1980b.                                                           '

Peto J; Doll R; Hermon C; Binns W; Clayton R; Goffe T.  Relationship of Mortality to Measures
of Environmental Asbestos Pollution in an Asbestos Textile Factory. Annals of Occupational
Hygiene. 29(3):305-355. 1985.                                              '

Piolatto G; Negri E; LaVecchia C; Pira E; Decarli A; Peto J. An -Update of Cancer Mortality
Among Chrysotile Asbestos Miners in Balangero, Northern Italy. British Journal of industrial
Medicine. 47:810-814. 1990.

Sebastien P; Plourde M; Robb R; Ross M; Nadon B; Wypruk T.' Ambient Air Asbestos Survey
in Quebec Mining Towns. Part II: Main Study. Environmental Protection Service, Environment
Canada. EPS 5/AP/RQ/2E.  July.  1986.

Sebastien P; McDonald JC; McDonald AD; Case B; Harley R. Respiratory Cancer iniChrysotile
Textile and Mining Industries: Exposure Inferences from Lung Analysis. British Journal of
Industrial Medicine. 46:180-187. 1989.
                                        D-10

-------
                             APPENDIX E:
  CALCULATION OF LIFETIME RISKS OF DYING OF
       LUNG CANCER OR MESOTHELIOMA FROM
                      ASBESTOS EXPOSURE

This appendix describes how additional lifetime risk of lung cancer and mesothelioma are
calculated from the estimated KL, the potency for lung cancer, and KM, the potency for
mesothelioma.  Let SE(t | x) be the probability of surviving to age t, given survival to age x 
-------
i* observational interval, given survival to the beginning of the interval, is calculated as
(risk per person-year times years of observation), and the probability of surviving this age
interval, given survival to the beginning of the interval, is calculated as S^l-b^Aj. The
probability of surviving to the beginning of the 1th interval given survival to age x, is calculated
recursively as                                                               't
                                      M                                    •  (Eq.E-4)

                                     m
                                      ;=o

where, by definition, S0=l. The probability PO(X!, x2) of dying of lung cancer between'x, andx2,
given survival to x1? is the sum over each observational interval of the probability of surviving to
the beginning of the age-interval times the probability of dying of lung cancer in the interval
given survival to the beginning of the interval, or
                                                                            '  (Eq.E-5)
This expression represents a discrete approximation to the integral (Eq. E-2).        :;

We now indicate how this expression is modified to account for exposure. First suppose the
exposure pattern E is a step function defined by constant exposure to f (in units of the optimal
exposure index) between ages el and e2, with no exposure at other ages.  According to" the lung
cancer model (Eq. A-l), in the presence of exposure the mortality rate ai for the i* observational
interval is increased to ai*(l+KL*dj), where di is the cumulative exposure lagged 10 years for
this interval.  In the implementation of this algorithm, d; is calculated as
      0,

      /

      f*(e2-e},)
                                          ifel+lQ
-------
account for the dose-related effects of both lung cancer and mesothelioma upon survival,
Si=l-bi*A i is replaced by
               St(E}= \-(b, -a, *KL*d, -KM *G)* A,
(Eq. E-8)
Similarly, the probability of dying of mesothelioma from exposure pattern E between the ages of
Xj and x2, given survival to Xj, is calculated as
                                 M
                                                                              (Eq. E-9)
                                 ;=o
The oldest (n"1) age-interval is unbounded above. In the implementation of the algorithm, a
width of 1/bn is assigned to this interval, which is an estimate of the average survival time in this
age-interval. When, as is typical, the oldest interval is for ages * 85 years, this assignment only
affects the calculation when the fbllowup period extends past 85 years (x2>85), and then only
minimally.

When used to estimate risk from continuous exposure (24 hours/day, 7 days/week), KL and KM
were adjusted upward by multiplying by 365/240 (to adjust from an assumed occupational
exposure of 240 days/year to 365 days/year) and by 2.0 (to adjust from an assumed exposure
during work hours to 24 hours/day, assuming that the amount of air breathed during 24 hours is
roughly double the amount breathed during a single work shift.

This algorithm is expanded to handle dose patterns composed of any linear combination of step
functions simply by replacing dj and Q4 by the sum of the corresponding terms resulting from
each step function that composes the linear combination.  Since any exposure pattern of interest
can be approximated to any degree or accuracy by a linear combination of step functions, the
algorithm can consequently estimate risk from any exposure pattern of interest.

Age-specific mortality rates for both lung cancer (aj and all-causes (bj) are needed to calculate
asbestos-related risk using the above approach.  In order to account for differences in asbestos-
related risk between males and females and—particularly for lung cancer—between smokers and
non-smokers, it is necessary to apply sex- and smoking-specific mortality rates. Lung cancer
and all-cause mortality rates for U.S. males and females for the year 2000 (CDC 2003) are
provided in Table E-l. Also provided in this  table are corresponding rates for never-smokers
and current smokers, which were calculated from the U.S. 2000 rates, data on the effect of
smoking obtained from the Cancer Prevention Study II (CPS-II) of the American Cancer Society
(Thun et al. 1997a), and information on the prevalence of smoking obtained from the National
Health Interview Survey (NHIS) (Trosclair et al. 2002). The following paragraphs describe how
these smoking-specific rates were calculated.
                                          E-3

-------
   Table £-1. Mortality Rates for All Causes and Lung Cancer per 100,000 Population per Year


Age


U.S. 2000
All Causes
Non-
smokers


Smokers


U.S. 2000
Lung Cancer
Non-
smokers
!

Smokers
Males
1
5
130
288
15-20
20-25
25-30
30-35
35-40
40-45
45-50
50-55
55-60
60-65
65-70
70-75
75-80
80-85
85+
799.9
36.5
18.3
25.0
94.9
142.0
141.9
157.3
209.8
303.3
461.7
658.3
1007.5
1565.5
2399.3
3705.4
5591.2
8956.9
16605.4
799.9
36.5
18.3
25.0
94.9 >
93.8
93.7
103.9
138.4
192.5
314.9
430.1
671.4
1087.1
1690.8
2661.6
4334.9
7257.3
15651.1
799.9
36.5
18.3
25.0
94.9
281.4
281.2
311.7
416.3
623.7
886.0
1318.1
1979.1
2948.5
4447.7
6723.1
9223.2
13870.5
19364.3
. 0
0
0
' 0
0
0
• 0.3
0.8
3.3
10.4
26.0
54.6
118
206.2
327.5
444.0
507.3
549.6
499.0
0
0
0
0
0
0
0.1
0.2
0.9
2.7
5.8
8.1
15.7
23.6
42.6
59.7
80.9
128.4
144.1
; o
0
. . o
0
': 0
0
0.9
; 2.5
10.4
!, 32.6
'• 84.4
i, 189.1
413.8
734.2
1151.3
1555
:' 1740
; 1767.3
•.' 1525.2
Females
1
5
130
288
15-20
20-25
25-30
30-35
35-40
40-45
45-50
50-55
55-60
60-65
65-70
70-75
75-80
80-85
85+
654.3
29.1
14.5
16.6
40.0
48.2
56.5
76.0
115.1
172.2
254.0
386.3
611.8
982.0
1527.5
2381.8
3812.6
6444.8
14768.6
654.3
29.1
14.5
16.6
40.0
48.2
56.5
76.0
112.7
171.7
207.6
324.1
486.0
774.5
1199.7
1943.6
3230.0
5753.4
13829.2
654.3
29.1
14.5
16.6
40.0
48.2
56.5
76.0
124.2
174.2
428.5
620.3
1085.1
1762.6
2760.7
4030.4
6004.2
9045.8
18302.3
0
0
0
0
0
0
0
0.9
2.6
8.1
16.6
35.1
70.9
122.3
181.6
238.7
268.6
272.8
213.5
0
0
0
0
0
0
0
0.3
0.8
2.6
3.2
11.2
16.5
32.1
41.4
81.6
79.6
118.0
84.7
0
0
o
0
o
0
0
3.2
9.3
29.0
67.2
125.0
275.4
461.5
709.2
829.6
979.6
855.3
698.2
CPS-II CThun et al. 1997a) prospectively followed more than one million persons in the U.S.
                                         E-4

-------
beginning in 1982.  Subjects were recruited by volunteers and were s30 years of age at the time
of enrollment.  Smoking status was determined by a questionnaire administered at the time of
enrollment. Never smokers were defined as persons who had never smoked any tobacco
product, and current smokers as persons who were cigarette smokers at the time of enrollment.
Although follow-up has now been extended, following a recommendation of Dr. Thun (Thun
2003), the present calculations are based upon follow-up through 1988 (Thun et al. 1997a).
There are two reasons for this: (1) follow-up past 1988 mostly involves older ages for which
sufficient numbers of deaths had already occurred prior to 1988 to insure adequate statistical
stability of mortality rates; and (2) more importantly, since smoking histories were not updated,
with longer follow-up there is greater misclassification of persons who were classified as current
smokers at time of enrollment, but who may have quit smoking during the follow-up.

Based on follow-up of the CPS-II cohort through 1988, Thun et al. (1997a) present age-specific
mortality rates for a number of causes of death, including lung cancer and all-cause mortality, by
5-year age-intervals beginning at age 30. Separate tabulations are provided for never smokers
and for current smokers in both males and females (reproduced in Table E-2).  These rates are
not necessarily representative of the general U.S. population. For example, a member of the
CPS-II cohort is more likely to be college-educated, married, middle-class, and white (Thun et
al. 1997b).  Note also, mat, despite the fact that smoking is a well-documented health risk,
female smokers in CPS-II (Table E-2) had lower all-cause mortality than U.S. women in general
(Table E-l). Consequently, rather than applying the CPS-II  rates directly to the U.S. population,
they are used only to estimate age- and sex-specific relative risks resulting from smoking. These
relative risks are used in conjunction with estimates of the current fraction of smokers to
partition the U.S. 2000 mortality rates between non-smokers and smokers.  For a given age, sex
and mortality cause (lung cancer or all-cause mortality), we write

                                                                            (Eq. E-10)
             2000     NS                                  -
                                   ,„,    ""       ""     ""* where r2000 is the U.S.
 2000 mortality rate for the age, sex and cause category, pSM is the proportion of smokers in the
 U.S. population, and RRSM is the relative risk for smoking obtained from the CPS-II data
 (mortality rate in current smokers divided by mortality rate in non-smokers). The U.S. mortality
 rate for non-smokers, rNS, is estimated by solving this equation.  The corresponding U.S. rate for
 smokers is men estimated as the product of the rate in non-smokers and the relative risk for
 smoking, rNS*RRSM.
                                          E-5

-------
          Table E-2.  CPS-II Mortality Rates for Ail Causes and Lung Cancer
                            per 100,000 Population per Year*
All Causes
Age
Non-smokers
Smokers
Lung Cancer11
1
Non-smokers (adj) Smokers (adj)
Males
25-30
30-35
35-40
40-45
45-50
50-55
55-60
60-65
65-70
70-75
75-80
80-85
85+
—
72.9
93.7
151.8
221.4
367.7
672.6
-1096.7
1846.6
3441.2
5466.5
11141.6
—
219.3
303.6
427.1
678.5
1083.8
1824.2
2884.9
4664.5
7321.7
10447.8
13784.9
. (0.02)
(0.2)
: 4.6(0.6)
0.0(1.5)
6.0(2.8)
5.5(4.9)
5.3(7.8)
11.6
21.5
34.9 -
52
89.2
86.8,
• (0-3)
, (2.2)
? 5.9(7.4)
18.7(17.5)
41.4
j, 115.3
' 206.1
j. 361.1
581.6
909
1118.3
1227.7 '
,. 919
Females
30-35
35-40
40-45
45-50
50-55
55-60
60-65
65-70
70-75
75-80
80-85
85+
80.6
109.3
122.4
182.1
268.2
411.4
666.5
1073.9
1838.7
3154.2
8069.2
88.8
110.9
252.6
348.5
598.8
936.3
1533.7
2227
3417.9
4959.2
10679.2
i (0.03)
2.0(0.2)
0.0(0.8)
' 1.9(2.0)
'. 5.8
> '7.2
12.3
16.7
30.5
32.5
57.6
60.6
(0.4)
1 4.0(2.8)
!' 8.9(9.5)
42.4
' 64.7
;' 119.9
: 176.6
286.3
310
r 400
417.6
499.6
Thun et al. (1997a).
bAdjusted rates (in parentheses) were used to calculate the rates in Table E-l.
See text for adjustment method.
                                          E-6

-------
To implement this approach an estimate is needed for pSM, the proportion of smokers in the U.S.
population. Based on the NHIS administered in 2000 to a nationally representative sample of the
U.S. non-institutionalized population over 18 years of age, the proportion of current smokers was
0.257 among men and 0.210 among women.  Smoking prevalence was fairly age-independent,
except among persons greater than 65 years of age.  Among men the proportions of current
smokers was 0.285, 0.297, 0.264, and 0.102 among men aged 18-24, 25-44,45-64 and ^65,
respectively. The corresponding proportions for women were 0.251, 0.245,0.216, and 6.093
(Trosclair et al. 2002).  The oldest category likely includes a sizable percentage of former
smokers whose mortality rates are influenced by their former smoking habits.  Because of this
and related problems, it was decided not to age-adjust smoking rates, but simply to apply the
overall rates from the NHIS survey. Consequently, the proportion of smokers was assume to be
pSM=0.257 in men and pSM=0.210 in women.

Smoking-specific mortality rates were not available from CPS-II below the age of 35.
Additionally, in both males and females the CPS-II lung cancer rates in the lowest age categories
were based on fewer than 10 deaths, and consequently quite uncertain.  In these age categories,
the lung cancer rates were adjusted using a cubic function of (age less the oldest age at which the
2000 U.S. rate was zero), keeping the total expected number of lung cancer deaths in these
categories equal to the observed number.  The resulting adjusted rates are shown in parentheses
in Table E-2,  Equation E-l 0 was applied to these adjusted rates.

Turning now to all-cause mortality, for males between the ages of 35 and 60, the CPS-II all-
cause mortality rates in smokers were approximately three times the rates in non-smokers
(RR~3).  Consequently, to estimate smoking-specific rates below the age of 35, equation E-10
was applied using RR=3 between the ages of 20 and 35 and RR=1 for earlier ages. For women,
since the all-cause mortality rates in smokers and non-smokers differed by less than 10%
between the ages of 35 and 45, the rates in smokers and non-smokers were assumed to be equal
below the age of 35. The resulting smoking-specific rates are shown in Table E-l. The
difference in estimated rates between smokers and non-smokers is not necessarily solely due to
smoking; other differences in lifestyle between smoker and non-smokers likely contributed,
particularly among males.

REFERENCES

Center for Disease Control National Center for Health Statistics (CDC).  GMWK210.  Death
Rates for 113 Selected Causes, by 5-year Age Groups, Race, and Sex: United States, 1999-2000.
http://wvtw.cdc.gov/nchs/datawh/statab/unpuba/mortabs/gmwk210  10.htm. 2003.

Thun MJ. Personal Communication to Dr. Kenny Crump. 2003.

Thun MJ; Day-Lally C; Myers  DG; Calle EE; Flanders WD; Zhu BP; Namboodiri MM; Heath
CW. Trends in Tobacco Smoking and Mortality from Cigarette Use in Cancer Prevention
Studies, I (1959 through 1965)  and II (1982 through 1988).  In Changes  in Cigarette-Related
Disease Risks and Their Implication for Prevention and Control.  Burns D; Garfinkel L; Samet
J; (eds.).  Smoking and Tobacco Control Monograph, No. 8, National Institutes of Health,
National Cancer Institute. Government Printing Office. Bethesda, MD. 1997a.
                                         E-7

-------
Thun MJ; Peto R; Lopez AD; Monaco JH; Henley SJ; Heath CW; Doll R. Alcohol Consumption
and Mortality among Middle-aged and Elderly U.D. Adults. The New England Journal of
Medicine.  337(24): 1705-1714.  1997b.                                         •

Trosclair A; Husten C; Pederson L; Dhillon I.  Cigarette Smoking among Adults - United States,
2000. Morbidity and Mortality Weekly Report.  Surveillance Summaries. 51(29):642-j645.
2002.

                                         E-8

-------

-------