Health Risk Assessment of 1, 3-Butadiene, Review Draft


              United States
              Environmental Protection
              Agency
                Office of Research and
                Development
                Washington DC 2046O
v>EPA
Health Risk
Assessment of
1,3-Butadiene
EPA/600/P-98/001A
January 1998
External Review Draft
                                            Draft
                                            (Do  Not
                                            Cite or
                                            Quote)
                              Notice
              This document is a preliminary draft. It has not been formally
             released by EPA and should not at this stage be construed to
             represent Agency policy. It is being circulated for comment on its
             technical accuracy and policy implications.

-------

-------
DRAFT                                                          NCEA-W-0267
DO NOT CITE OR QUOTE                                            January 1998
                                                           External Review Draft
                        Health Risk Assessment
                            of 1,3-Butadiene
                                   NOTICE
THIS DOCUMENT IS A PRELIMINARY DRAFT. It has not been formally released by the
U.S. Environmental Protection Agency and should not at this stage be construed to represent
Agency policy. It is being circulated for comment on its technical accuracy and policy
implications.
                    National Center for Environmental Assessment
                        Office of Research and Development
                       U.S. Environmental Protection Agency
                               Washington, DC

-------
                                  DISCLAIMER
      This document is an external draft for review purposes only and does not constitute U.S.
Environmental Protection Agency policy. Mention of trade names or commercial products does
not constitute endorsement or recommendation for use.
1/28/98                                  ii        DRAFT-DO NOT CITE OR QUOTE

-------
•
                                   CONTENTS

 LIST OF TABLES . . ...... . ........................ ........................ _   viii

 LIST OF FIGURES .......... ........... ..................................... xii
AUTHORS, CONTRIBUTORS, AND REVIEWERS
•
 1. INTRODUCTION ........... .................. . ....... . .                   M
   1.1. BACKGROUND .......................................... '.'.'.'.'.'.'.'.'.'.'.'.  l-l
   1.2. SUMMARY OF PAST CARCINOGEN ASSESSMENTS OF 1,3-BUTADIENE '.'.'.  1-1
       1.2.1. Summary of EPA's Carcinogen Assessment (U.S. EPA, 1985) ...... . ......  1-2
       1.2.2. Summary of lARC's Evaluation of 1,3-Butadiene (IARC, 1986, 1992)  ......  1-5
       1 .2.3. Summary of the National Institute for Occupational Safety and Health
            (NIOSH) Evaluation of 1,3-Butadiene (NIOSH, 1991a)  ........... . ......  1-7
       1.2.4. California Air Resources Board (GARB, 1991) ........... ..............  1-8
       1 .2.5. Summary of Findings by U.S. Occupational Safety and Health
            Administration (OSHA)  ...................... ......               1_10
   1.3. DISCUSSION ........................ ............ ..........'......... l-U

2. OVERVIEW OF EXPOSURE TO 1,3-BUTADIENE ................ ....          2-1
   2.1. PHYSICAL/CHEMICAL PROPERTIES  ...... ..................       '  "  2-1
   2.2. PRODUCTION AND USE ................. ..........  . ............ '...'... 2-1
       2.2. 1 . Styrene-Butadiene Latex and Rubber Production ........................ 2-2
       2.2.2. Polybutadiene Production . ....... ............. . .................... 2-3
       2.2.3. Neoprene Rubber Production ....................................... 2-3
       2.2.4. Acrylonitrile-Butadiene (ABS) Resin Production ....................... 2-3
       2.2.5. Nitrile Elastomer Production ........... ...... .... ................ . . 2-3
       2.2.6. Adiponitrile Production  ..... .................... ........           2-4
   2.3. SOURCES AND EMISSION  ....... ............... ..................... 2-4
       2.3.1. Mobile Sources .................................................. 2-4
            2.3.1.1. On-Road Mobile Sources ................................... 2-5
            2.3.1.2. Nonroad Mobile Sources  ......... .......................... 2-5
            2.3.1.3. Aircraft  ....................... ____ . ..................... 2-6
       2.3.2.  Miscellaneous Sources ............. ............. ....      ........ 2-6
            2.3.2.1. Miscellaneous Chemical Production ........................... 2-6
            2.3.2.2. Secondary Lead Smelters ..................... .............. 2-6
            2.3.2.3. Petroleum Refining ........... ............................. 2-7
       2.3.3.  Combustion Sources ............................... ............... 2-7
            2.3.3.1. Tire Burning .............................................. 2-7
            2.3.3.2. Biomass Burning .......................................... 2-7
1/28/98                                 iii        DRAFT-DO NOT CITE OR QUOTE

-------
                           CONTENTS (continued)
  2.4. AMBIENT CONCENTRATION OF 1,3-BUTADIENE ....................... 2-8
      2.4.1. Air [[[ 2-8
           2.4.1.1. Ambient Monitoring Data ................................... 2-8
           2.4.1.2. Ambient Source Apportionment ............................. 2-23
      2.4.2. Indoor Exposure to 1,3-Butadiene  .................................. 2-23
      2.4.3. Water [[[ 2-24
      2.4.4. Food  [[[ 2-24
  2.5. PATHWAYS OF EXPOSURE .......................................... 2-24

3. METABOLISM AND PHARMACOKINETICS ................................. 3-1
  3.1. OVERVIEW OF PHARMACOKINETIC STUDIES .......................... 3-1
      3.1.1. Pathways Elucidation ............................................. 3-2
      3.1.2. Species Differences ............................................... 3-2
           3.1.2.1. In Vitro Metabolism ....................................... 3-2
           3.1.2.2. In Vivo Pharmacokinetics .................................. 3-18
  3.2. MOLECULAR DOSIMETRY .......................................... 3-41
  3.3. STRUCTURE-ACTIVITY RELATIONSHIPS  ............................. 3-42
  3.4. DISCUSSION AND CONCLUSIONS .................................... 3-45

4. MUTAGENICITY  [[[ 4-1
  4.1. INTRODUCTION  [[[ 4-1
  4.2. GENE MUTATIONS .................................................. 4-1
  4.3. CYTOGENETIC EFFECTS— HUMAN .................................... 4-4
  4.4. CYTOGENETIC EFFECTS— RODENT ................................... 4-5
  4.5. SUMMARY [[[ 4-6

5. REPRODUCTIVE AND DEVELOPMENTAL EFFECTS  ......................... 5-1
  5.1. REPRODUCTIVE EFFECTS ............................................ 5-1
      5.1.1. Carpenter et al., 1944 ............................................. 5-1
      5.1.2. Owen et al.,  1987; Owen and Glaister, 1990 ........................... 5-2
      5.1.3. NTP, 1984 [[[ 5-2
      5.1.4. NTP, 1993 [[[ 5-3
      5.1.5. Hackett et al., 1988a .............................................. 5-5
      5.1.6. Hackett et al., 1988b ........................................ . ..... 5-7
      5.1.7. Anderson et al., 1993 ............................................. 5-8
      5.1.8. Adler et al.,  1994 ................................................. 5-8

-------
                              CONTENTS (continued)


   5.3. STRUCTURE-ACTIVITY RELATIONSHIPS	       5-24
       5.3.1. NTP, 1986	'"' 5i24
       5.3.2. Melnick et al, 1994	" 5.26
       5.3.3. Doerr et al., 1996	'' 5.25
   5.4. SUMMARY AND CONCLUSIONS	...'.'.'.'.'.'.'.'.'.'.'.'.'. 5-27

6. TOXICITY IN ANIMALS 	                6-l
   6.1. SUBCHRONIC TOXICITY	  	6-1
   6.2. CHRONIC TOXICITY	     " "	6_j
   6.3. CARCINOGENICITY	  6-8
       6.3.1. 2-Year Study (NTP, 1993)	...6-8
       6.3.2. 2-Year Stop-Exposure Study (NTP, 1993)	6-19
       6.3.3. Summary of NTP (1993) Study	6-28
       6.3.4. 1-Year Study (Irons et al., 1989; Irons, 1990)	  6-28
   6.4. RELATED COMPOUNDS	      6-29
   6.5. DISCUSSION AND CONCLUSIONS	6-30

7. EPIDEMIOLOGIC STUDIES OF CARCINOGENICITY	7-1
   7.1. MONOMER PRODUCTION	7_1
       7.1.1. Texaco Cohort	 j.\
            7.1.1.1. Downs etal., 1987: Mortality Among Workers at a
                   Butadiene Facility	7_1
            7.1.1.2.  Divine, 1990: An Update on Mortality Among Workers at a
                   1,3-Butadiene Facility—Preliminary Results	7-4
            7.1.1.3.  Divine et al., 1993:  Cancer Mortality  Among Workers at a
                   Butadiene Production Facility	7-6
            7.1.1.4.  Divine and Hartman, 1996: Mortality Update of Butadiene
                   Production Workers	7.7
      7.1.2. Shell Oil Refinery Cohort	7-8
            7.1.2.1.  Cowles et al., 1994: Mortality, Morbidity, and Hematological
                   Results From a Cohort of Long-Term  Workers Involved in
                   1,3-Butadiene Monomer Production	7-8
      7.1.3. Union Carbide Cohort	7.9
            7.1.3.1.  Ward et al., 1995: Mortality Study of Workers in 1,3-Butadiene
                   Production Units Identified From a Chemical Workers Cohort
                   Ward et al., 1996c: Mortality Study of Workers Employed in
                   1,3-Butadiene Production Units Identified From a Large
                   Chemical Workers Cohort	7.9
  7.2. POLYMER PRODUCTION	\\\ 7.10
      7.2.1. Cohort Identified by Johns Hopkins University (JHU)  Investigators	7-10
            7.2.1.1. Matanoski and Schwartz, 1987:  Mortality of Workers in
                   Styrene-Butadiene Polymer Production	7-10


1/28/98                                 v        DRAFT-DO NOT CITE OR QUOTE

-------
                             CONTENTS (continued)


           7.2.1.2. Matanoski et al, 1989: Epidemiologic Data Related to Health
                   Effects of 1,3-Butadiene	7-14
                   Matanoski et al., 1990: Mortality of a Cohort of Workers in the
                   Styrene-Butadiene Polymer Manufacturing Industry (1943-1982) . . 7-14
           7.2.1.3. Matanoski etal., 1989: Epidemiologic Data Related to
                   Health Effects of 1,3-Butadiene  	7-16
                   Santos-Burgoa et al., 1992: Lymphohematopoietic Cancer in
                   Styrene-Butadiene Polymerization Workers	7-16
           7.2.1.4. Matanoski etal., 1993: Cancer Epidemiology Among
                   Styrene-Butadiene Rubber Workers	7-19
      7.2.2. Cohort Identified by University of Alabama (UAB) Investigators	7-20
           7.2.2.1. Delzell et al.,  1996: A Follow-Up Study of Synthetic
                   Rubber Workers	7-20
           7.2.2.2. Macaluso etal., 1996: Leukemia and Cumulative Exposure to
                   Butadiene, Styrene, and Benzene Among Workers in the
                   Synthetic Rubber Industry	7-23
  7.3. SUMMARY AND DISCUSSION	7-25
      7.3.1. Monomer Production	7-31
      7.3.2. Polymer Production	7-32
      7.3.3. Relevant Methodologic Issues and Discussion	7-34
      7.3.4. Criteria of Causal Inference	7-37

8. PHARMACOKINETIC MODELING 	8-1
  8.1. INTRODUCTION	8-1
  8.2. PBPK MODELS FOR 1,3-BUTADIENE	8-2
      8.2.1. Hattis and Wasson (1987)	8-2
      8.2.2. Hallenbeck (1992)	8-4
      8.2.3. Kohn and Melnick (1993)	8-4
      8.2.4. Johanson andFilser (1993)	8-8
      8.2.5. Evelo et al. (1993)	8-12
      8.2.6. Medinsky et al. (1994)	8-15
  8.3. SUMMARY	8-19
  8.4. CONCLUSIONS	8-21

9. QUANTITATIVE RISK ASSESSMENT FOR 1,3-BUTADIENE	9-1
  9.1. EPIDEMIOLOGICALLY BASED CANCER RISK ASSESSMENT	9-1
      9.1.1. Exposure-Response Modeling	9-1
      9.1.2. Prediction of Lifetime Excess Risk of Leukemia	9-3
      9.1.3. Sources of Uncertainty 	9-10
      9.1.4. Summary and Conclusions  	9-12
  9.2. CANCER RISK ESTIMATES BASED ON RODENT BIOASSAYS	9-15
      9.2.1. Rat-Based Estimates	9-15


1/28/98                                  vi        DRAFT-DO NOT CITE OR QUOTE

-------
                            CONTENTS (continued)


      9.2.2. Mouse-Based Estimates	9_16
            9.2.2.1.  Quantal	 9.15
            9.2.2.2.  Time-to-Tumor	9-19
    -  9.2.3. Discussion	9-24
  9.3. REPRODUCTIVE AND DEVELOPMENTAL TOXICITY		'.'.'.'. 9-27
      9.3.1. Introduction	9_27
      9.3.2. Fetal Weight Modeling	9.31
      9.3.3. Male-Mediated Developmental Toxicity Modeling	9-35
      9.3.4. Ovarian, Uterine, and Testicular Atrophy Modeling  	9-40
      9.3.5. Summary and Conclusions	9.45
  9.4. REFERENCE CONCENTRATIONS FOR REPRODUCTIVE AND
      DEVELOPMENTAL EFFECTS	9-46
      9.4.1. Introduction	9.45
      9.4.2. Calculation of RfCs	9-46
      9.4.3. Discussion	9.49
      9.4.4. Conclusions	9.51
  9.5. CONCLUSIONS ON QUANTITATIVE RISK ESTIMATES	9-52

10.  WEIGHT OF EVIDENCE	10-1
    10.1.  EVALUATION	10-1
    10.2.  CONCLUSION  	......	10-1

11.  RISK CHARACTERIZATION	H-l
    11.1.  INTRODUCTION	11-1
    11.2.  EXPOSURE OVERVIEW ..	11-2
    11.3.  CANCERHAZARD ASSESSMENT	11-2
         11.3.1. Human Evidence	11-2
         11.3.2. Animal Data	.. 11-8
         11.3.3. Other Supportive Data	11-9
         11.3.4. Classification	11-9
    11.4.  QUANTITATIVE RISK ESTIMATION FOR CANCER	11-10
    11.5.  SUMMARY OF REPRODUCTIVE/DEVELOPMENTAL EFFECTS	11-13
    11.6.  QUANTITATIVE ESTIMATION (RfC) FOR REPRODUCTIVE/
         DEVELOPMENTAL EFFECTS	11-14
    11.7.  SPECIAL SUBPOPULATIONS	11-15
         11.7.1. Sensitive Subpopulations	 11-15
         11.7.2. Highly Exposed Subpopulations	11-16
    11.8.  FUTURE RESEARCH NEEDS	11-16
    11.9.  SUMMARY AND CONCLUSIONS	11-17

12.  REFERENCES . .  . . ,	12-1
1/28/98                              vii       DRAFT-DO NOT CITE OR QUOTE

-------
                                  LIST OF TABLES
1-1    Carcinogenicity assessments of 1,3-butadiene	1-12

2-1A  Summary of 1,3-butadiene ambient data from the EPA Aerometric Information
       Retrieval System (AIRS) for 1988 to 1991  ....-	2-10
2-1B  Summary of 1,3-butadiene ambient data from the EPA Aerometric Information
       Retrieval System (AIRS) for 1992 to 1994	2-12
2-2    Summary of 1,3-butadiene ambient data from the Urban Air Toxics Monitoring
       Program (UATMP)	2-16
2-3    Summary of outdoor urban data from the National Ambient Volatile Organic
       Compounds (NAVOC) Database	2-18
2-4    Summary of air monitoring program results for 1,3-butadiene	2-19
2-5    Summary of 1,3-butadiene ambient data from the EPA Aerometric Information
       Retrieval System (AIRS) based on sampling locations	2-21
2-6    Summary of 1,3-butadiene data from Table 2-2 based on sampling locations	2-22
2-7    Summary of 1,3-butadiene data from Table 2-3 based on sampling locations	2-22
2-8    Summary of the relative  contributions to ambient 1,3-butadiene emissions given
       as percent of total mg/yr  	2-24

3-1    Metabolic pathways of 1,3-butadiene metabolism	3-4
3-2    Species comparison of reaction rates for epoxidation, GSH conjugation, and
       hydrolysis reactions involved in the metabolism of 1,3-butadiene	3-13
3-3    Rate constants for in vivo hepatic clearance of 1,3-butadiene and EB
       (extrapolated from in vitro)	3-19
3-4    Summary of closed-chamber inhalation studies  	3-21
3-5    Toxicokinetic parameters for uptake and elimination of 1,3-butadiene in
       mice and rats  	3-23
3-6    Toxicokinetic parameters for the uptake and elimination of epoxybutene in
       rats and mice  	3-23
3-7    Summary of nose-only inhalation studies	3-24
3-8    Comparison of 1,3-butadiene, epoxybutene, and diepoxybutane blood
       concentration data from different species of laboratory animals exposed to
       1,3-butadiene by inhalation	3-30
3-9    Tissue levels of epoxybutene and diepoxybutane (pmol/g tissue) in male rats
       and male mice exposed by inhalation to 62.5 ppm 1,3-butadiene for 4 h	3-32
3-10   Tissue levels of epoxybutene and diepoxybutane (pmol/g tissue) in male
       and female rats exposed by inhalation to 62.5 ppm 1,3-butadiene for 6 h  	3-33
3-11   Excretion of I4C by monkeys exposed to l,3-[14C]-butadiene	3-36

5-1    Reproductive tract lesions hi female  B6C3F, mice exposed to  1,3-butadiene
       by inhalation	5-4
5-2    Reproductive tract lesions in male B6C3Fr mice exposed to 1,3-butadiene
       by inhalation	5-6


1/28/98                                  viii        DRAFT-DO NOT CITE OR QUOTE

-------
                            LIST OF TABLES (continued)
5-3    Maternal toxicity in Sprague-Dawley CD rats exposed to 1,3-butadiene
       by inhalation	.	5-10
5-4    Developmental toxicity in Sprague-Dawley CD rats exposed to
       1,3-butadiene by inhalation	5-12
5-5    Malformations and variations in Sprague-Dawley CD rats exposed to
       1,3-butadiene by inhalation	5-13
5-6    Design of the developmental toxicity studies on 1,3-butadiene 	5-14
5-7    Maternal toxicity in Sprague-Dawley CD rats exposed to 1,3-butadiene
       by inhalation	5-15
5-8    Developmental toxicity in Sprague-Dawley CD rats exposed to
       1,3-butadiene by inhalation	5-17
5-9    Malformations and variations in Sprague-Dawley CD rats exposed to
       1,3-butadiene by inhalation	5-18
5-10   Maternal toxicity in pregnant CD-I mice exposed to 1,3-butadiene by inhalation ... 5-19
5-11   Developmental toxicity in CD-I mice exposed to 1,3-butadiene by inhalation  	5-20
5-12   Malformations and variations in CD-I mice exposed to 1,3-butadiene
       by inhalation	5-21
5-13   Reproductive and developmental toxicity of chemicals structurally
       similar to 1,3-butadiene	5-25

6-1    Survival of male and female B6C3FJ mice exposed to 1,3-butadiene by
       inhalation for 103 weeks	6-3
6-2    Incidence of hyperplasia in male and female B6C3Fj mice exposed to
       1,3-butadiene by inhalation for 103 weeks	6-7
6-3    Incidence of hyperplasia in male B6C3F, mice exposed to 1,3-butadiene by
       inhalation in the stop-exposure study	 6-8
6-4    Incidence of primary neoplasms in male B6C3FJ mice exposed to
       1,3-butadiene by inhalation for 103 weeks  	\ ,	 6-10
6-5    Incidence of primary neoplasms in female B6C3F! mice exposed to
       1,3-butadiene by inhalation for 103 weeks	6-12
6-6    Incidence of primary neoplasms in male B6C3F; mice exposed to
       1,3-butadiene by inhalation for 9 months and 15 months	6-14
6-7    Incidence of primary neoplasms in female B6C3Fj mice exposed to
       1,3-butadiene by inhalation for 9 months and 15 months	6-15
6-8    Incidence of primary neoplasms in male B6C3Fj mice exposed to
       1,3-butadiene by inhalation in the stop-exposure study	 6-21

7-1    Epidemiologic studies of the  health effects of exposure to 1,3-butadiene—
       monomer production	 7-11
7-2    Epidemiologic studies of the  health effects of exposure to 1,3-butadiene—
       polymer production	7-26
1/28/98                                    ix        DRAFT-DO NOT CITE OR QUOTE

-------
                             LIST OF TABLES (continued)
 8-1    Parameter values used in the Hattis and Wasson (1987) PBPK model	8-3
 8-2    Parameter values used in the Kohn and Melnick (1993) PBPK model	8-5
 8-3    Parameter values used in the Johanson and Filser (1993) PBPK model	8-10
 8-4    Parameter values used in the Evelo et al. (1993) PBPK model	8-14
 8-5    Parameter values used hi the Medinsky et al. (1993) PBPK model 	8-17

 9-1    Results from exposure-response models of continuous cumulative
       exposure to 1,3-butadiene and styrene using alternative structural forms
       reported by Delzell et al	9-3
 9-2    Results from "final" square root exposure-response model of continuous
       cumulative exposure to 1,3-butadiene reported by Delzell et al	9-4
 9-3    Maximum likelihood estimates (MLEs) of excess  risk with one-sided 95% upper
       confidence limits (95% UCL) from several models reported by Delzell et al. (1995)
       for continuous lifetime exposures to varying concentrations of 1,3-butadiene	9-9
 9-4    MLEs of parts per million continuous exposure concentrations associated with
       varying excess risk levels with one-sided 95% lower confidence limits (95% LCL)
       based on relative rate results of several models reported by Delzell et al. (1995)
       and U.S. population rates	9-10
 9-5    Maximum likelihood (ECp) and 95% lower-bound (LECp) estimates of
       the continuous exposure concentrations associated with varying levels of
       excess risk (p)  	9-11
 9-6    Cancer potency (unit risk) estimates based on linear extrapolation from
       the LECoi or EC01 calculated from the models presented by Delzell et al	9-15
 9-7    Dose-response data for linearized multistage model	9-17
 9-8    Parameter estimates for multistage Weibull time-to-tumor model based on
       female mouse tumor incidence, w/o 625 ppm group	9-21
 9-9    Parameter estimates for multistage Weibull time-to-tumor model based on
       male mouse tumor incidence, w/o 625 ppm group  	9-21
 9-10   Human unit cancer risk estimate (extra risk, computed for risks of 10"6)
       based on female mouse tumor incidences, w/o 625 ppm group using
       multistage Weibull time-to-tumor model	9-22
 9-11   Human unit cancer risk estimates (extra risk, computed for risks of 10'6)
       based on male mouse tumor incidences, w/o 625 ppm group using
       multistage Weibull time-to-tumor model	9-22
 9-12   Unit potency estimates (extra risk) summed across tumor sites  	9-25
 9-13   Prenatal (developmental) toxicity study (Hackett et al., 1987b)	9-28
 9-14   Male-mediated developmental toxicity (Anderson et al., 1993, 1995)	9-28
9-15   NTP chronic study (1993)	9-29
9-16   Fetal weight modeling (LOAEL = 40 ppm)	9-33
9-17   ECs and LECs for male-mediated developmental toxicity	 9-36
9-18   ECs and LECs for ovarian, uterine, and testicular atrophy using the quantal
       Weibull and log-logistic models 	9-41


 1/28/98                                   x        DRAFT-DO NOT CITE OR QUOTE

-------
                             LIST OF TABLES (continued)
9-19   Parameters for Weibull time-to-response model used to model reproductive
       effects observed in mice based on ppm butadiene exposure	9-44
9-20   Human benchmark 1,3-butadiene exposure concentrations calculated for
       reproductive effects observed in mice using a Weibull time-to-response
       model (extra risk)	9.45
9-21   Points of departure and RfC calculations for reproductive and
       developmental effects of 1,3-butadiene	 9-48

11-1   Summary of epidemiologic studies	11-4
11-2   Epidemiologic causality criteria	11-8
11-3   Estimates of upper bounds on human extra unit cancer risk (potency)
       from continuous lifetime exposure to 1,3-butadiene based on animal
       inhalation bioassays	11-11
1/28/98                                    xi        DRAFT-DO NOT CITE OR QUOTE

-------
                                  LIST OF FIGURES
3-1    Some pathways in the metabolism of butadiene	3-3
3-2    Two-compartment pharmacokinetic model for inhalation chamber	3-34

6-1    Kaplan-Meier survival curves for male B6C3F, mice exposed to
       1,3-butadiene by inhalation for 103 weeks  	6-4
6-2    Kaplan-Meier survival curves for female B6C3F] mice exposed to
       1,3-butadiene by inhalation for 103 weeks  	6-5
6-3    Kaplan-Meier survival curves for male mice in the stop-exposure
       inhalation study of 1,3-butadiene	6-20

9-1    Excess risk and 95% upper confidence limit excess risk estimates based
       on the multiplicative model reported by Delzell et al., 1995	9-6
9-2    Excess risk and 95% upper confidence limit excess risk estimates based
       on the power model reported by Delzell et al., 1995	9-6
9-3    Excess risk and 95% upper confidence limit excess risk estimates based
       on the linear excess relative rate model reported by Delzell et al., 1995	9-7
9-4    Excess risk and 95% upper confidence limit excess risk estimates based
       on the final square root model reported by Delzell et al., 1995	9-7
9-5    Excess risk and 95% upper confidence limit excess risk estimates based
       on the square root model reported by Delzell et al., 1995	 9-8
9-6    Observed versus predicted dose (exposure) probability P(d) of fetal weight
       reduction below the lOthpercentile of controls using log-logistic model	9-34
9-7    Observed versus predicted mean fetal weight per litter using continuous model .... 9-34
9-8    Observed versus predicted percent of mean fetal weights per  litter less
       than the 5th percentile of controls (P0 = 0.05) using hybrid model	9-35
9-9    Observed versus predicted mean number of implants (prenatal)
       using log-linear model	9-37
9-10   Observed versus predicted proportion of early and late deaths
       per implantation (prenatal) using log-linear model	9-37
9-11   Observed versus predicted proportion of live implants (prenatal)
       using log-linear model	9-38
9-12   Observed versus predicted mean number of implants (postnatal)
       using log-linear model	9-38
9-13   Observed versus predicted proportion of post-implantation losses (postnatal)
       using log-linear model	9-39
9-14   Observed versus predicted mean litter size at birth using log-linear model  	9-39
9-15   Observed versus predicted mean litter size at weaning using log-linear model 	9-40
9-16   Ovarian atrophy (groups 1-5) using log-logistic model	9-42
9-17   Uterine atrophy (groups 1-6) using quantal Weibull model 	9-43
9-18   Testicular atrophy (groups 1-6) using quantal Weibull model  	9-43
1/28/98                                    xii        DRAFT-DO NOT CITE OR QUOTE

-------
                                      PREFACE

       This Health Risk Assessment of 1,3 -Butadiene has been prepared to serve as a source
document for Agencywide use. This document was developed primarily for use by the U.S.
Environmental Protection Agency's (EPA) Office of Mobile Sources (QMS) to support decision
making regarding the Air Toxic Rule's Section 202L2 of the Clean Air Act Amendment.  Since
OMS requested that this assessment focus on mutagenicity, carcinogenicity, and
reproductive/developmental effects, an evaluation of other health hazards has not been included.
This document, therefore, is not a comprehensive health assessment. The exposure information
included here is an overview of the ambient exposures and exposure to populations adjacent to
emission sources, without any actual exposure assessment as such.
       In the development of this assessment, relevant scientific literature has been incorporated
from the period July 1, 1985, through January 31,1997. Key studies have been evaluated to
qualitatively describe the mutagenicity, reproductive/developmental effects, and carcinogenicity
of 1,3-butadiene. The assessment also includes a summary, conclusions, and risk
characterization. Measures of dose-risk relationships relevant to ambient air exposures are
discussed so that the adverse health effects can be placed in perspective with possible exposure
levels.
1/28/98                                   xiii       DRAFT-DO NOT CITE OR QUOTE

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

       This document was prepared by the National Center for Environmental Assessment-
Washington Office (NCEA-W) of EPA's Office of Research and Development. Sections of this
report were prepared by Oak Ridge National Laboratory (ORNL) under Interagency Agreement
No. DW89937638-01-0. Aparna M. Koppikar1 served as the Project Manager, providing overall
direction and technical assistance.

AUTHORS
       Chapter 1:  Aparna M. Koppikar, Kowetha Davidson2
       Chapter 2:  ChiehWu1
       Chapter 3:  Kim Hoang1, Carol Forsyth2, and Robert Young2
       Chapter 4:  Lawrence Valcovic1
       Chapter 5:  Kowetha Davidson
       Chapter 6:  Rosmarie Faust2
       Chapter 7:  Aparna M. Koppikar
       Chapter 8:  Jennifer Jinot1, Carol Kirnmel1
       Chapter 9:  Jennifer Jinot
       Chapter 10: Aparna M. Koppikar
       Chapter 11: Jennifer Jinot and Aparna M. Koppikar

CONTRIBUTORS
       Thomas M. Crisp1
       Dharm Singh1
       Steven Bayard (now at OSHA)
       Milton Siegal1
       John Schaum1
       Leslie Stayner3
       Stephen Gilbert3
       Randall Smith3
      'National Center for Environmental Assessment-Washington Office.
      2Oak Ridge National Laboratory.
      3National Institute for Occupational Safety and Health.
1/28/98                                 xiv       DRAFT-DO NOT CITE OR QUOTE

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

REVIEWERS
       David Bayliss (NCEA-W)
       Robert Bellies (NCEA-W)
       Pam Brodowicz (QMS)
       Robert Bruce (NCEA-Cin)
       James Cogliano (NCEA-W)
       Richard Cook (QMS)
       Michael Dellarco (NCEA-W)
       Gary Kimmel (NCEA-W)
       William Pepelko (NCEA-W)

       The authors would like to acknowledge the contributions of several people who have
made this report possible:
    •   Theresa Konoza of NCEA-W, who was responsible for coordinating and managing the
       production effort.
    •   The CDM Group, Inc., under the direction of Kay Marshall, who was responsible for
       editing, word processing, and literature searches.
1/28/98                                xv        DRAFT-DO NOT CITE OR QUOTE

-------

-------
1. INTRODUCTION

1.1. BACKGROUND
2 1,3-Butadiene (CH2=CH-CH=CH2, CAS No. 106-99-0) is a colorless gas produced by
3 three different processes: (1) oxidative dehydrogenation of n-butene (the Oxo-D or O-X-D
4 process), (2) catalytic dehydrogenation of n-butane and n-butene (the Houdry process), and (3)
5 recovery from the C4 coproduct (by-product) stream from the steam cracking process used to
6 manufacture ethylene (the ethylene coproduct process). This noncorrosive gas has a boiling
7 point of -4.4°C and a vapor pressure of 1,900 mm/Hg at 20°C (Kirshenbaum, 1978). 1,3-
8 Butadiene is highly volatile and has a low solubility in water; thus environmental release results
9 primarily in atmospheric contamination. Atmospheric destruction of 1,3-butadiene occurs
10 primarily by photoinitiated reactions. A significant amount of destruction also occurs by the gas
11 phase reaction with ozone and reaction with nitrate radicals at nighttime, particularly in urban
12 areas (U.S. DHHS, 1992). The major photooxidation products of 1,3-butadiene are acrolein and
13 formaldehyde (Maldotti et al, 1980).
14 Approximately 12 billion pounds of 1,3-butadiene are produced annually worldwide and
15 3 billion pounds in the United States (Morrow, 1990; USITC, 1990). It is used as an
16 intermediate in the production of polymers, elastomers, and other chemicals. The major uses-of
1,3-butadiene are in the manufacture of styrene-butadiene rubber (SBR) (synthetic rubber) and of
\8 thermoplastic resins. Elastomers of butadiene are used in the manufacture of tires, footwear,
19 sponges, hoses and piping, luggage, packaging, and a variety of other molded products. In
20 addition, 1,3-butadiene is used as an intermediate to produce a variety of industrial chemicals,
21 including the fungicides captan and captfol. The primary way the 1,3-butadiene is released in the
22 environment is via emissions from gasoline- and diesel-powered vehicles and equipment. Minor
23 releases occur in production processes, tobacco smoke, gasoline vapors, and vapors from the
24 burning of plastics as well as rubber (Miller, 1978). ,

25 1.2. SUMMARY OF PAST CARCINOGEN RISK ASSESSMENTS OF 1,3-BUTADEENE
26 The purpose of this section is to review past carcinogen risk assessments of 1,3-
27 butadiene. It should be noted that the Toxicological Profile for 1,3-butadiene (ATSDR, 1992),
28 profile of 1,3-butadiene to set the threshhold limit value (TLV) (ACGIH, 1994), and 1,3-
29 Butadiene OEL Criteria Document by the European Center for Ecotoxicology and Toxicology of
30 Chemicals (1997) are not reviewed here as they are not risk assessments.
1/28/98 1-1 DRAFT-DO NOT CITE OR QUOTE

-------
1 1.2.1. Summary of EPA's Carcinogen Assessment (U.S. EPA, 1985)
2 Pertinent studies reported before 1986 were reviewed in Mutagenicity and
3 Carcinogenicity Assessment of 1,3-Butadiene (U.S. EPA, 1985). This document was peer
4 reviewed by experts in the field, as well as in public sessions of the Environmental Health
5 Committee of EPA's Science Advisory Board. The studies presented in the 1985 document will
6 not be reviewed in the present document but are briefly summarized below.
7 EPA reviewed six epidemiological studies, which included four retrospective cohort
8 mortality studies, one nested case-control study, and an industrial hygiene and hematologic cross-
9 sectional survey. The first cohort study involved 6,678 hourly workers in a rubber tire
10 manufacturing plant in Akron, Ohio (McMichael et al., 1974). The standard mortality ratios
11 (SMRs) were calculated using the 1968 U.S. male population as the reference. Cause-specific
12 mortality was evaluated for 16 different occupational title groups (work areas) within the plant.
13 This study was followed up by a nested case-control study involving 455 of the 1,983 deaths
14 recorded between 1968 and 1973 (McMichael et al., 1976). The second cohort study was
15 conducted in 8,938 male workers in a rubber plant also located in Akron, Ohio (Andjelkovich et
16 al., 1976, 1977). The 1976 study used the U.S. male population as the reference for calculating
17 the SMRs, whereas the entire cohort was used to calculate the SMRs for 28 different work areas
18 for the 1977 study. The third cohort study included 2,756 workers at two styrene-butadiene
19 rubber facilities in eastern Texas (Meinhardt et al., 1982). The sex, age, race, and calendar time
20 cause-specific rates of the U.S. population were used to calculate the SMRs. The last and most
21 comprehensive study was conducted in 13,920 workers at one Canadian and seven U.S. styrene-
22 butadiene rubber plants (Matanoski et al., 1982). The SMRs for black and white workers were
23 calculated separately. The cross-sectional survey was conducted on workers in the same styrene-
24 butadiene rubber plant studied by McMichael et al. (Checkoway and Williams, 1982). Blood
25 samples were obtained to evaluate hematology parameters. The survey was not designed to
26 evaluate mortality experience and did not contribute to cancer risk evaluation of 1,3-butadiene.
27 Of the five epidemiologic studies that evaluated cause-specific mortality, three cohort
28 studies demonstrated statistically significant excess mortality due to cancers of the lymphatic and
29 hematopoietic tissues (Andjelkovich et al., 1976; McMichael et al., 1976; Meinhardt et al.,
30 1982). The fourth cohort study by Matanoski et al. (1982) also showed increased leukemia, but
31 failed to achieve statistical significance. Lastly, the nested case-control study by McMichael et
32 al. (1976) showed statistically significant increased standardized risk ratios for cancers of the
33 lymphatic and hematopoietic tissues among workers with exposures of 5 years or more in one
34 area of the plant (synthetic rubber plant area), as compared with either all the other workers or
35 the matched controls. Statistically significant excess cancer mortality was also observed for
36 gastrointestinal tract, respiratory tract, central nervous system, prostate, testicles, and urinary
1/28/98 1-2 DRAFT-DO NOT CITE OR QUOTE

-------
         bladder in one or more studies.  However, these excesses were not observed consistently across
         all the studies.
                Although excess mortality due to cancers of the lymphatic and hematopoietic tissues was
  4      observed consistently in all the evaluated studies, the methodologic limitations, such as too few
  5      deaths from specific cancers to evaluate the causal association; exclusion of large portions of the
  6      population due to lack of records; lack of adjustment for smoking; confounding by other
  7      exposures such as benzene or styrene; and excess cancer mortality at other sites prompted EPA to
  8      conclude that the evidence was inadequate for determining a causal association between
  9      exposure to 1,3-butadiene and cancer in humans.
 10             Two long-term animal studies presented strong evidence for the induction of cancers at
 11      multiple anatomical sites in both rats (HLE, 1981) and mice (NTP, 1984).  Sprague-Dawley rats
 12      were exposed by inhalation to 1,3-butadiene at concentrations of 1,000 or 8,000 ppm 6 h/day, 5
 13      d/week for 111 weeks and 105 weeks for males and females, respectively.  Statistically
 14      significant increased incidences in the following neoplasms were observed at one or both
 15      concentrations:  mammary gland tumors, thyroid follicular adenomas/carcinomas, and Zymbal
 16      gland carcinomas in female rats  and Leydig cell adenomas/carcinomas, pancreatic exocrine
 17      adenomas, and Zymbal gland tumors in male rats. In addition, gliomas occurred in four high-
 18      dose male rats. Nonneoplastic effects due to long-term exposure of rats to 1,3-butadiene
         included clinical signs of toxicity, an increase in liver weight in both sexes, marked to severe
         nephropathy in 27% of the high-dose  male rats compared with 9% or 10% of the controls, and
 21      alveolar metaplasia in male rats.
 22             Among E6C3F1 mice exposed by inhalation to  1,3-butadiene at 625 or 1,250 ppm for 6
 23      h/day, 5 d/week, neoplasms also developed at multiple anatomical sites; this study was
 24      terminated at week 60 to 61  because of high mortality in the treated groups, primarily due to
 25      neoplasms. There was an overall increase in the number of animals with primary neoplasms and
 26      animals with multiple neoplasms. Neoplasms showing statistically significant increased
 27      incidences among both male  and female mice were as follows:  malignant lymphomas,
 28      alveolar/bronchiolar adenomas/carcinomas, hemangiosarcomas of the heart, and forestomach
 29      papillomas/carcinomas.  In addition, mammary gland acinar cell carcinomas, ovarian granulosa
 30      cell carcinomas,  and hepatocellular adenomas/carcinomas occurred in female mice.
 31      Nonneoplastic effects observed were testicular atrophy, chronic inflammation,  fibrosis,
 32      cartilaginous metaplasia, osseous metaplasia, and atrophy of the sensory epithelium of the nasal
 33      cavity in male mice. Ovarian atrophy was observed in female mice.  Some discrepancies were
 34      noted for this study, but they were not considered to pose a significant impact  on the overall
J35      interpretation of the study.


         1/28/98                                    1-3       DRAFT-DO NOT CITE OR QUOTE

-------
1 EPA also reviewed data from metabolism and mutagenicity studies, concluding that
2 inhaled 1,3-butadiene is metabolized to mutagenic epoxide intermediates.
3 In addition, EPA reviewed the carcinogenicity of related compounds (4-vinyl-l-
4 cyclohexene, epoxybutene, <37-l,2:3,4-diepoxybutane, and f«eso-l,2:3,4-diepoxybutane). 4-
5 Vinyl-1-cyclohexene is carcinogenic in female mice (oral/gavage), based on increased incidences
6 of ovarian and adrenal gland neoplasms. Equivocal evidence was noted for malignant
7 lymphomas and alveolar/bronchiolar adenomas in male mice and clitoral gland neoplasms in
8 female rats (NTP, 1986). Skin painting of mice with meso-l,2:3,4-diepoxybutane induced
9 papillomas and squamous cell carcinomas (Van Duuren et al, 1963), and subcutaneous injection
10 with d/-l,2:3,4-diepoxybutane caused fibrosarcomas in mice and rats (Van Duuren et al., 1966).
11 Based on the studies in mice and rats, EPA concluded that there was sufficient evidence
12 for carcinogenicity of 1,3-butadiene in animals. EPA also concluded that evidence from
13 metabolism, mutagenicity, and carcinogenicity studies suggests that 1,3-butadiene presents a
14 genetic risk to humans.
15 Two developmental toxicity studies were reviewed. One study (HLE, 1981) was
16 conducted using pregnant female Sprague-Dawley rats exposed to 200, 1,000, and 8,000 ppm 6
17 h/day on gestation days 6-15. Developmental effects included slightly decreased fetal weight and
18 mean crown-rump length and increased skeletal variations and malformations. The other study
19 (Carpenter et al., 1944) was inadequately reported.
20 EPA presented the following conclusion regarding the qualitative evaluation of the data
21 for 1,3-butadiene: "On the basis of sufficient evidence from studies in two species of rodents,
22 and inadequate epidemiologic data, 1,3-butadiene can be classified as a probable human
23 carcinogen, Group B2." Using the classification scheme of the International Agency for
24 Research on Cancer (IARC), 1,3-butadiene would also be classified as a "probable" human
25 carcinogen, Group 2B.
26 The linearized multistage model was used to calculate the maximum likelihood estimate
27 for the incremental risk for 1,3-butadiene based upon the National Toxicology Program mouse
28 data (NTP, 1984), the HLE (1981) rat data, and internal dosimetry derived from data on mice and
29 rats exposed to varying concentrations of 1,3-butadiene for 6 h. The upper-limit unit risk of 6.4
30 x lO'^ppm)"1 was a geometric mean of the values calculated for male and female mice
31 separately. The unit risk extrapolated to humans was 2.5 * lO'^ppm)"1. This value was used to
32 predict human responses in the epidemiologic studies, which were then compared with the actual
33 responses. According to EPA, ". . . The comparisons were hampered by a scarcity of information
34 in the epidemiologic data concerning actual exposures, age distribution, and work histories. In
35 addition, because there was no consistent cancer response across all of the studies, the most
36 predominant response, cancer of the lymphatic and hematopoietic tissues, was chosen as being
1/28/98 1-4 DRAFT-DO NOT CITE OR QUOTE

-------
the target for 1,3 -butadiene. Based on the comparisons between the predicted and observed
human response, the extrapolated value from the mouse data was consistent with human
response, but in view of all the uncertainties and apparent inconsistencies in the epidemiologic
data, a fairly wide range of potency estimates and exposure scenarios would also be satisfactory. .
5 . ." (U.S. EPA, 1985).

6 1.2.2. Summary of lARC's Evaluation of 1,3-Butadiene (IARC, 1986,1992)
7 IARC reported the first evaluation of 1,3-butadiene as a separate chemical in 1986
8 (IARC, 1986). In an earlier report (IARC, 1982), 1,3-butadiene was evaluated as a chemical
9 used in the rubber industry. lARC's 1986 evaluation of the animal data consisted of the NTP
10 (1984) study using male and female B6C3F! mice exposed to 625 or 1,250 ppm 1,3-butadiene for
11 60 or 61 weeks and an abstract description of the HLE (1981) study in rats exposed to 1,000 or
12 8,000 ppm (Owen et al., 1985). The human data consisted only of a cohort study described by
13 Meinhardt et al. (1982) and a brief mention of the following studies of workers in the rubber
14 industry that were included in IARC's evaluation of the rubber industry: Andjelkovich et al.,
15 1976, 1977; McMichael et al., 1976; and Monson and Nakano, 1976. The supporting evidence
16 considered by IARC consisted of absorption, distribution, metabolism, and excretion (ADME)
17 data. The genotoxicity data showed that 1,3-butadiene was mutagenic in S. typhimurium with
metabolic activation, and the metabolites (1,2-epoxybutene and l,2:3,4-diepoxybutane) were
mutagenic in S. typhimurium without metabolic activation. IARC also evaluated data on acute,
20 reproductive, and developmental toxicity of 1,3-butadiene. IARC (1986) concluded that the
21 supporting evidence for genetic activity was "inadequate," the evidence for carcinogenicity in
22 experimental animals was "sufficient," and the evidence for carcinogenicity in humans was
23 "inadequate" (Group 2B).
24 IARC reevaluated the data on 1,3-butadiene and reported the results in 1992. Additional
25 animal and human studies were available for evaluation. In addition to the first NTP (1984)
26 study in mice, IARC (1992) evaluated a more recent NTP study reported by Melnick et al.
27 (1990a). In this study, male and female B6C3F! mice exposed by inhalation to 1,3-butadiene at
28 concentrations of 6.25 to 625 ppm for 2 years developed neoplasms at multiple sites and at all
29 concentrations. IARC also evaluated the published HLE (1981) long-term study showing tumors
30 developing at multiple sites in male and female Sprague-Dawley rats exposed to 1,000 or 8,000
31 ppm 1,3-butadiene (Owen et al., 1987) and a comparative study in B6C3FJ and NTH Swiss mice
32 examining the role of endogenous retroviruses on the induction of lymphomas by 1,3-butadiene
33 (Irons et al., 1987). IARC also presented some evidence showing that the metabolites 1,3-
epoxybutene and 1,2:3,4-diepoxybutane possessed carcinogenic activity.

1/28/98 1-5 DRAFT-DO NOT CITE OR QUOTE

-------
 1            Epidemiologic studies evaluated by IARC (1992) consisted primarily of the studies
 2     published since 1982.  The following studies were evaluated: (1) the mortality study conducted
 3     by Meinhardt et al. (1982) of workers in two styrene-butadiene rubber plants, but not the most
 4     recent update of this study by Lemen et al. (1990); (2) the mortality study by Downs et al. (1987)
 5     of workers who manufactured 1,3-butadiene monomer and the most recent update of this study
 6     by Divine (1990); (3) a mortality study by Matanoski et al. (1989) of workers in eight U.S. and
 7     Canadian styrene-butadiene rubber plants (update of the study by Matanoski and Schwartz,
 8     1987); (4) a nested case-control study of the 59 workers from the eight U.S. and Canadian plants
 9     who died from lymphopoietic cancer (Santos-Burgoa, 1988; Matanoski et al., 1990); (5) a nested
10     case-control study of rubber workers dying from various types of cancer including
11     lymphohematopoietic cancer (McMichael et al., 1976); and (6)  a population-based case-control
12     study of various types of cancers (excluding leukemia) conducted in Montreal, Canada
13     (Siemiatycki, 1991).
14            Supporting evidence evaluated by IARC (1992) included in vitro studies on the
15     metabolism of 1,3-butadiene using human liver and lung homogenates and comparative in vivo
16     and in vitro metabolism studies in mice, rats, and monkeys.  A detailed discussion on in vivo and
17     in vitro genetic toxicity of 1,3-butadiene and metabolites (1,2-epoxybutene and 1,2:3,4-
18     diepoxybutane) was presented along with other available information on short-term toxicity and
19     nonneoplastic effects of 1,3-butadiene in humans and experimental animals.
20            IARC  (1992) concluded that the evidence for the carcinogenicity of 1,3-butadiene in
21     humans is "limited" based on (1) a study showing an increased risk for lymphosarcoma and
22     reticulosarcoma among workers who manufacture 1,3-butadiene monomers; (2) a suggested
23     increased risk for leukemia among workers at one of two styrene-butadiene rubber plants studied;
24     (3) no increase of leukemia among the  entire cohort of workers at  eight U.S. and Canadian
25     styrene-butadiene rubber plants, but a significant risk of leukemia among a subgroup of
26     production workers; and (4) a large excess of lymphohematopoietic cancer nested among
27     workers exposed to 1,3-butadiene in styrene-butadiene rubber plants.  IARC also concluded that
28     the evidence for the carcinogenicity of 1,3-butadiene in experimental animals was "sufficient"
29     based on tumor induction at multiple sites in mice and rats, the  induction of neoplasms in mice at
30     all concentrations tested (6.25 to 1,250 ppm), the carcinogenicity of two metabolites of 1,3-
31     butadiene, and the detection of activated K-ras oncogenes in lymphomas, liver tumors, and lung
32     tumors induced by 1,3-butadiene.  Evidence from metabolism and genetic toxicity studies
33      supported the conclusions of the carcinogenicity studies.  IARC (1992) concluded that 1,3-
34     butadiene IB probably carcinogenic to humans (Group 2A).
        1/28/98                                   1-6       DRAFT-DO NOT CITE OR QUOTE

-------
  1      1.2.3. Summary of the National Institute for Occupational Safety and Health (NIOSH)
 .2            Evaluation of 1,3-Butadiene (NIOSH, 1991a)
 *3            NIO SH (1991 a) conducted a qualitative and quantitative assessment of the
  4      carcinogenicity of 1,3-butadiene.  The evaluation of animal data focused on the studies that could
  5      be used for quantitative assessment, namely the studies using Sprague-Dawley rats (Owen et al.,
  6      1987) and E6C3F1 mice (NTP, 1984; Melnick et al., 1990a). The qualitative evaluation of the
  7      evidence from human studies focused on the studies by Downs et al. (1987) and updated by
  8      Divine (1990); Meinhardt et al. (1982) and updated by Lemen et al. (1990); Matanoski and
  9      Schwartz (1987) and updated by Matanoski et al. (1990); and a case-control study of the
 10      lymphopoietic cancers (Santos-Burgoa, 1988) from the Matanoski cohort. According to NIOSH,
 11      the results of this nested case-control study "provide the strongest human evidence to date for an
 12      association between 1,3-butadiene and the risk of lymphopoietic neoplasms, particularly
 13      leukemia." NIOSH concluded that overall the epidemiologic studies showed an increase in
 14      lymphopoietic neoplasms, which is consistent with the induction of lymphomas in mice exposed
 15      to 1,3-butadiene.  However, NIOSH reported that the epidemiologic studies had certain
 16      limitations, such as the lack of historical exposure levels, the inclusion of both support and
 17      production personnel whose exposure would probably be minimal, and the inconsistent diagnosis
 18      of the different types of lymphohematopoietic neoplasms.
 19            NIOSH reported on metabolism, pharmacokinetics, and disposition studies; their
"20      evaluation focused primarily on studies that provided data for estimating metabolic rates at low
 21      concentrations and comparison of metabolic pathways and rates in different species (mice,  rats,
 22      monkeys, and humans). With respect to genetic toxicity, NIOSH did not focus on details of any
 23      studies, but noted that 1,3-butadiene is mutagenic in Salmonella with metabolic activation,
 24      whereas the metabolites are mutagenic without metabolic activation.
 25            NIOSH (1991a) concluded that the present evaluation supports the conclusion of a
 26      previous evaluation (NIOSH, 1984), which stated that "1,3-butadiene should be considered to
 27      represent a potential human health hazard with respect to carcinogenicity."  The basis for the
 28      conclusion was positive evidence of carcinogenicity in three long-term animal bioassays in two
 29      species, positive evidence of mutagenicity and genotoxicity, and less conclusive epidemiologic
 30      evidence of excess deaths from lymphopoietic neoplasms.
 31            NIOSH used data from the study in B6C3FJ mice (Melnick et al., 1990a) for its
 32      quantitative assessment because the lowest concentration (6.25 ppm) was similar to the proposed
 33      OSHA standard of 2 ppm.  WeibulPs one-, two-, and three-stage time-to-tumor models were
 34      used to derive the maximum likelihood and 95% confidence limit estimates on excess risk.  The
 35      models were fit for the individual tumors for which the incidences were significantly higher in
 |6      exposed groups than in control groups of male and female mice.  Hemangiosarcomas of the heart
         1/28/98                                   1-7       DRAFT-DO NOT CITE OR QUOTE

-------
1 and lymphomas were modeled as fatal lesions and all others as incidental lesions. The equivalent
2 human doses were calculated based on body weight to the three-fourths power (BW3/4) and
3 converted back to ppm exposures in the workplace for 45 years of exposure. The excess risk for
4 lifetime occupational exposure at 1 ppm was 305/10,000 based on lung neoplasms in females
5 (highest) and 0.03/10,000 based on heart hemangiosarcomas in females.
6 NIOSH (1991b) discussed the uncertainties associated with its assessment. The dose-
7 scaling method chosen and species differences in the metabolism of 1,3-butadiene were major
8 sources of uncertainty. Another source of uncertainty was the most relevant tumor site used to
9 predict human risk. The female lung was the most sensitive site, but based on the epidemiologic
10 evidence, lymphomas may be the most relevant neoplasms. Other sources of uncertainty were
11 the model selection: (1) whether the Weibull time-to-tumor model was the most appropriate and
12 which stage model to use, (2) the assumption regarding lethality of tumors and omission of the
13 high-dose group, and (3) estimation of the internal dose and the application of kinetic data.

14 1.2.4. California Air Resources Board (CARS, 1991)
15 The GARB (1991) evaluated the data on 1,3-butadiene and presented quantitative
16 estimates of the cancer risk from inhalation exposure to 1,3-butadiene in ambient air. The
17 literature review consisted of toxicokinetic data that focused on information presented by Bond et
18 al. (1986, 1987) for absorption and tissue distribution data and reports published between 1985
19 and 1991 for metabolism and excretion data. Acute, subchronic, and noncancer chronic toxicity
20 information was obtained from excerpts from EPA's 1985 carcinogen assessment document, and
21 reproductive/developmental toxicity data and genetic toxicity data were reported from the
22 primary literature. Genetic toxicity data focused on mutation tests in S. typhimurium, DNA
23 alkylation studies, SCE and chromosome aberration tests, and various in vivo studies.
24 Animal carcinogenicity studies evaluated by CARB included the two NTP studies in mice
25 (NTP, 1984; Huff et al., 1985; Melnick et al., 1990a), the inhalation study in rats (Owen et al.,
26 1987), the role of retroviruses in 1,3-butadiene-induced carcinogenesis (Irons et al., 1987), and
27 the expression of oncogenes in tumors induced by 1,3-butadiene (Goodrow et al., 1990). Human
28 studies evaluated by CARB started with the 1976 study by McMichael et al. and continued
29 through the 1990 reports by Lemen et al., Divine, and Matanoski et al. CARB discussed several
30 factors that must be considered when interpreting the epidemiologic studies: (1)
31 misclassification of exposure—unexposed individuals classified as exposed would bias the
32 results toward the null; (2) exclusion of most highly exposed workers—studies in which the
33 workers with the highest potential exposure (World War II workers) are excluded would be less
34 likely to see a significant effect; (3) no dose-response effect—the lack of a positive association
35 with duration of exposure should not discredit the study because the most recent NTP animal
1/28/98 1-8 DRAFT-DO NOT CITE OR QUOTE

-------
         study (Melnick et al., 1990a) demonstrated that short-term exposure to a high concentration of
         1,3 -butadiene could be more effective than long-term exposure to low concentrations; and (4)
         varying health endpoints—there were inconsistencies in the subtypes of lymphopoietic and
  4      hematopoietic cancers observed in the various studies, but nomenclature changed over time and
  5      there are probably close relationships between the different subtypes. CARB presented four
  6      points of evidence for an association between exposure to 1,3-butadiene and lymphopoietic and
  7      hematopoietic cancers in humans. The first point is that the strongest effect was observed in the
  8      cohort involved in the production of 1,3-butadiene monomer, and this cohort had the greatest
  9      potential for exposure to 1,3-butadiene in the absence of styrene.  The second is that the
 10      observations of cancers in cohorts having potential exposure to styrene and 1,3-butadiene are
 11      consistent with the findings from the cohort from the 1,3-butadiene monomer production facility.
 12      The third point is presented in the case-control study by Matanoski and Schwartz (1987)  and the
 13      cellular  study by Checkoway and Williams (19 82) in which both attributed the observed effects
 14      to 1,3-butadiene exposure and not to styrene exposure. The fourth point is that the cancers
 15      observed in humans are consistent with those observed in the mouse experiments.  CARB
 16      concluded that "the epidemiological studies reported to date give evidence for increased
 17     incidences of leukemia and/or lymphohematopoietic neoplasms resulting from exposure to
 18     vapors in styrene-butadiene rubber plants or butadiene production plants." They further stated
        that the  evidence for elevated rates of stomach and lung cancers is inconclusive.
              CARB conducted an extensive quantitative assessment of the risk from exposure to 1,3-
21     butadiene. The two mouse studies and the rat study were considered suitable for quantitative
22     evaluation. Dose estimations were based on experimental (applied) dose, continuous internal
23     dose, metabolized dose, target tissue dose, and molecular tissue dose. CARB used the retention
24     data from Bond et al. (1986) to  estimate  the daily dose adjusted for 7-day week exposures
25     (internal dose). The pharmacokinetics model of Hattis and Wasson (1987) was used to estimate
26     internal  exposure to metabolites, namely  butadiene monoepoxide (metabolized dose). The tissue
27     distribution data of Bond et al. (1986, 1987) were used to estimate the target tissue doses, which
28     were not used for risk estimation because the data were not reliable.  Sufficient data on DNA
29     adducts  were not available for deriving molecular tissue doses.
30           CARB fitted the experimental (applied dose), internal, and metabolized doses estimated
31      from the first mouse  study (NTP, 1984) to the linearized multistage (Global 86) and the Weibull
32     time-to-tumor models; the cancer potency estimates derived using the linearized multistage
33     model and Weibull's model gave similar results. The multistage model was used to derive
34     cancer potency values using the  second mouse study (Melnick et al., 1990a) and the rat study
35     (HUB, 1981).  Cancer potency estimates  were derived for each anatomical site separately  and for
        the total number of tumor-bearing animals with significantly increased tumors in both males and

        1/28/98                                    1-9        DRAFT-DO NOT CITE OR  QUOTE

-------
,-3
1 females. The human cancer potency estimates, based on 70 years of continuous exposures,
2 derived from the first mouse study using the total significant tumors, the internal dose, and the
3 multistage model were 0.32 (ppm)"1 or 0.59 (mg/kg/day)"1 for male mice and 0.18 (ppm)"1 or
4 0.33 (mg/kg/day)"1 for female mice. Cancer potencies derived from applied doses were about
5 10-fold lower, and those derived from metabolized doses were about 50% lower. The human
6 cancer potency estimates using the rat data (total significant tumors), internal dose, and the
7 multistage model were 1.8 x 10"3 (ppm)"1 or 8.4 x 10"3 (mg/kg/day)"1 for male rats and 3.5 x 10
8 (ppm)"1 or 1.6 x 10"2 (mg/kg/day)"1 for female rats. The estimates based on applied or
9 metabolized doses were much lower. The data from the second mouse study were analyzed
10 extensively; CARB concluded that the best human cancer potency estimates based on internal
11 doses and estimated using the multistage model (Global 86) were 0.37 (ppm)"1 or 3.4
12 (mg/kg/day)"1 derived for alveolar/bronchiolar adenoma/carcinoma in female mice. The
13 corresponding unit risk derived from the second mouse study was 1.6 x 10"4 (ug/m3)"1 and the
14 exposure for the risk at 10"6 was 6.0 x 10"3 ^ig/rn3. From their cancer potency values, CARB
15 estimated the lifetime extra risk associated with exposure to 1 ppb 1,3-butadiene to range from
16 9.8 x 10"6 to 3.7 x 10"4, which corresponds to 10 to 370 additional cases per 1 million
17 individuals.

18 1.2.5. Summary of Findings by U.S. Occupational Safety and Health Administration
19 (OSHA)
20 The most recent analysis of health effects of 1,3-butadiene by a government entity is by
21 OSHA (1996). While the analysis in general is similar to that of NIOSH, OSHA incorporated a
22 recent update of the large SBR polymer retrospective follow-up study that had been started by
23 Matanowski et al. This update, Delzell et al (1995), included not only an additional period of
24 follow-up, but also a detailed exposure history for 1,3-butadiene, styrene, and benzene for more
25 than 15,000 employees who had worked in SBR and related activities at the eight study plants.
26 Delzell et al. concluded that "This study found a positive association between employment in the
27 SBR industry and leukemia. The internal consistency and precision of the results indicate that the
28 association is due to occupational exposure. The most likely causal agent is BD or a combination
29 of BD and [styrene]. Exposure to [benzene] did not explain the leukemia excess." OSHA in its
30 analysis of the Delzell et al. and previous studies recognized these consistencies and similarly
31 concluded that "there is strong evidence that workplace exposure to BD poses an increased risk
32 of death from cancers of the lymphohematopoietic system. The epidemiologic findings
33 supplement the findings from the animal studies that demonstrate a dose response for multiple
34 tumors and particularly for lymphomas in mice exposed to BD" (OSHA, 1996, p. 56764).

1/28/98 1-10 DRAFT-DO NOT CITE OR QUOTE

-------
1 OSHA also examined the evidence for reproductive and developmental effects.
.2 Analyzing data from both the NTP I and the NTPII studies, OSHA noted the consistency and
*3 dose response and concluded "that exposure to relatively low levels of BD resulted in the
4 induction of ovarian atrophy in mice..." (OSHA, 1996, p. 56765). For the total database on these
5 and mutagenic effects, OSHA concluded that "these animal studies taken as a whole, offer
6 persuasive qualitative evidence that BD exposure can adversely affect reproduction in both male
7 and female rodents. The Agency also notes that BD is "mutagenic in both somatic and germ
8 cells" (p. 56767).
9 For quantitative risk assessment, OSHA's analysis was very similar to that of NIOSH
10 (1991 a) in its choice of data set (NTP II mouse study), model (multistage Weibull), treatment of
11 tumors (dose-response analysis on an individual basis), treatment of fatal vs. nonfatal in the time-
12 to-tumor analysis, choice of parsimonious model algorithm (fewest parameters of the multistage
13 model that provide an adequate fit to the data) and reporting of the ML estimates. The major
14 difference between the NIOSH and OSHA analyses was that OSHA used (mg/kg bw-day)
15 equivalence for species conversion instead of the BW3/4 conversion used by NIOSH. This change
1Q of species conversion factors and some minor modifications relating to animal weights and
17 breathing rates decreased OSHA's potency estimates by a factor of approximately 4 from the
18 NIOSH estimates. Based on the female mouse lung tumor response, the OSHA ML estimate of
potency was 8.1 x 10'3 (ppm)'1 for an occupational lifetime of exposure to 1 ppm, 5 days/week,
>0 50 weeks/year for 45 years. If this potency estimate is extrapolated to be based on a 70-year
21 continuous lifetime exposure, the OSHA estimate would be approximately 36.7 x 10"3 (ppm)"1.
22 Based on the OSHA risk assessment, their permissible exposure limit was lowered from 1,000
23 ppm to 1 ppm with a 15-min short-term exposure limit.

24 1.3. DISCUSSION
25 Six different carcinogenicity assessments of 1,3-butadiene, done by five different
26 agencies in different time periods, are summarized in this chapter. The major conclusions of
27 these evaluations are presented in Table 1-1.
28 Although no apparent agreement is evident from the table among the five agencies'
29 assessments, in fact they are very similar. Both EPA (1985) and IARC (1986) conclude that the
30 carcinogenicity evidence in humans is inadequate and in animals is sufficient. But due to
31 different classification systems, they get different alphabetical assignments, i.e., B2 and 2B,
32 which correspond to "probable" and "possible" descriptors, thus appearing to be in disagreement
33 with each other. NIOSH and OSHA both use the dichotomous descriptors with "potential
^34 occupational carcinogen" being the highest ranking.

1/28/98 1-11 DRAFT-DO NOT CITE OR QUOTE

-------
      Table 1-1.  Carcinogenicity assessments of 1,3-butadiene
Agency
(year)
EPA (1985)
IARC (1986)
IARC (1992)
NIOSH (1991a)
GARB (1991)
OSHA(1996)
Cancer
classification
"B2-probable human
carcinogen" — based
on inadequate human
and sufficient animal
evidence.
"IB-possible human
carcinogen" — based
on inadequate human
and sufficient animal
evidence.
"2A-probable human
carcinogen" — based
on limited human and
sufficient animal
evidence.
"Potential human
health hazard with
respect to
carcinogenicity."
No formal
classification given
Potential
occupational
carcinogen
Quantitative risk
Unit risk to humans — 2.5 *
10"1 (ppm)"1 based onNTP
(1984) mouse data.
No quantitative risk
presented.
No quantitative risk
presented.
Range of excess risk at 1
ppm is MLE of 305/10,000
based on female mouse
lung neoplasms to MLE of
0.03/10,000 based on heart
hemangiosarcomas in
females. Data from
Melnick et al. (1990a) used
for this quantitation.
Human cancer potency
based on mouse data from
Melnick etal. (1990a)
range for 1 ppb
exposure— 9.8 x 10-6to 3.7
x lO'4.
Human cancer potency
estimate based on female
mouse lung neoplasms.
MLE is 8.1 x 1Q-3 (ppm)-1.
Remarks
Cancer classification
using EPA
carcinogen
assessment
guidelines.
Cancer classification
using IARC system.
OSHA cancer policy
classification system
used.
Quantitative risk is .
for occupationally
exposed populations.
No formal cancer
classification system
used.
Quantitative risk is
for general
population.
"Convincing
evidence that BD is a
probable human
carcinogen."
Quantitative risk is
for occupationally
exposed populations.
1/28/98
1-12
DRAFT-DO NOT CITE OR QUOTE

-------
              OSHA, NIOSH, and CARB assessments all state that the human evidence is strongest for
        an association between butadiene exposure and the occurrence of lymphohematopoietic cancers.
        The same evidence is described as "limited" human evidence by IARC, which elevates the
 4      classification of this compound to "2A-probable human carcinogen." Furthermore, it should be
 5      noted that the  quantitative risk estimates appear to be different for OSHA/NIOSH and
 6      EPA/CARB. NIOSH/OSHA quantitative risk estimates are for occupationally exposed
 7      populations, while quantitative estimates of CARB are for general population (lifetime risk),
 8      even though they are derived from the same animal data.
 9            The apparent differences in these assessments thus can be explained by availability of the
10      studies at the time of evaluations, different cancer classification systems, and quantitative
11      assessments done for different purposes.
       1/28/98                                  1-13       DRAFT-DO NOT CITE OR QUOTE

-------
                         2. OVERVIEW OF EXPOSURE TO 1,3-BUTADIENE

 1            The purpose of this chapter is to present an overview of how human exposure to 1,3-
 2      butadiene occurs. The chapter summarizes physical/chemical properties, production/use,
 3      sources/emissions, and ambient air data. Pathways of exposure are briefly described, but no
 4      quantitative estimates of exposure levels and numbers of people exposed are presented.

 5      2.1. PHYSICAL/CHEMICAL PROPERTIES
 6            1,3-Butadiene (CH2=CH-CH=CH2, CAS No. 106-99-0) is a colorless gas with mildly
 7      aromatic odor (Sax and Lewis, 1987). It is a noncorrosive gas and has a molecular weight of
 8      54.09. Its boiling point is -4.4 °C (Weast, 1989) and its vapor pressure is 1,790 mm Hg (239
 9      kPa) at 20°C (Santodonato, 1985).  It is easily liquefied with a density of 0.6211 g/ml at
10      20°C/liquefied (Kirshenbaum, 1978; Verschueren, 1983). It is soluble in ethanol, diethyl ether,
11      and organic solvents (Verschueren, 1983; Sax and Lewis, 1987; Budavari,  1989) and is also very
12      slightly soluble in water with a solubility of 735 mg/1 at 20°C.  1,3-Butadiene has a flash point of
13      -76°C (Sax and Lewis,  1987) and is slowly dimerized to 4-vinyl-l-cyclohexene (U.S.
14      Occupational Safety and Health Administration, 1990) and may form peroxide upon exposure to
15      air (Kirshenbaum, 1978). Since 1,3-butadiene is a highly volatile gas, it is expected to partition
16      predominantly to the atmosphere and then undergo rapid destruction by photomitiated reactions.
17      The reaction with photochemically produced hydroxyl radicals has a calculated half-life of
18      approximately 6 h and is expected to be the dominant pathway for atmospheric removal (U.S.
19      Department of Health and Human Services [DHHS], 1992). Destruction of atmospheric 1,3-
20      butadiene by the gas-phase reaction with ozone and by the nighttime reaction with nitrate
21      radicals in urban areas is also expected to be significant (U.S. DHHS, 1992). The major
22      photooxidation products of 1,3-butadiene are acrolein and formaldehyde (Maldotti et al. 1980).
23            There are limited data on the fate of 1,3-butadiene in soil or water.  Based on its physical
24      properties, rapid volatilization of 1,3-butadiene from either soil or water to atmosphere is
25      expected to dominate over all other potential environmental processes.  Studies performed with
26      pure cultures indicate that 1,3-butadiene may be susceptible to microbial attack. Based on
27      estimated values, 1,3-butadiene is not expected to adsorb significantly to soil or sediment, nor is
28      it expected to bioconcentrate in fish or aquatic organisms (U.S. DHHS, 1992).

29      2.2. PRODUCTION AND USE
30            1,3-Butadiene was first produced in 1886 by the pyrolysis of petroleum hydrocarbons
31      (Kirshenbaum, 1978). Commercial production of 1,3-butadiene started in  the 1930s (Kosaric et

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

-------
al, 1987) and has been produced by three processes: catalytic dehydrogenation of ra-butane and
«-butene, oxidative dehydrogenation of w-butene, and recovery from the C4 coproduct (by-
product) stream from the steam cracking process used to manufacture ethylene. The ethylene
4 coproduct process accounts for approximately 95% of U.S. and 85% of worldwide production
5 (Morrow, 1990). Approximately 12 billion pounds of this gas are produced annually worldwide
6 and 3 billion pounds in the United States (Morrow, 1990; USITC, 1990).
7 1,3-Butadiene is used as an intermediate in the production of polymers, elastomers, and
8 other chemicals. The major uses of this chemical are in the manufacture of styrene-butadiene
9 rubber (synthetic rubber) and of thermoplastic resins. In 1990, 1,3-butadiene was used in the
10 United States for styrene-butadiene rubber (30%), polybutadiene rubber (20%), adiponitrile/
11 hexamethylenediamine (15%), styrene-butadiene latex (10%), neoprene rubber (5%),
12 acrylonitrile-butadiene-styrene resins (5%), exports (4%), nitrile rubber (3%), and other
1 3 (including specialty polymers) (8%) (Anon., 1991).

1 4 2.2.1. Styrene-Butadiene Latex and Rubber Production
1 5 Styrene-butadiene (SB) latex and rubber production is the major use for butadiene,
1 6 accounting for 40% of butadiene consumption. SB latex and rubber are used for a variety of
V7 products, including automobile tires, textiles, paper, and adhesives.
The 1994 EPA report Locating and Estimating Air Emissions From Sources of 1,3-
1 9 Butadiene lists SB latex and rubber production as the major contributor to industrial butadiene
20 emissions (EPA, 1994a). About 74% of the industrial emissions are from SB latex and rubber
21 production. There are at least 26 facilities in the United States that produce SB latex and rubber
22 (SRI International, 1993).
23 As stated previously, butadiene has a very low water solubility and high vapor pressure;
24 thus, if it were released to an aqueous waste stream, it would immediately evaporate. It is then
25 .logical to assume, and the data confirms that, the amount of butadiene found in secondary
26 sources such as waste water and solid waste is minimal or nonexistent. The majority of the
27 butadiene releases during industrial production occurs via process vents, so only emission factors
28 for process vents will be presented. The emission factors, as presented in the 1994 EPA report
29 for process vent butadiene released during SB latex and rubber production, range from 0.00024
30 to 94.34 Ib butadiene emitted/ton produced (mean of 7.10) measured at 18 facilities (EPA,
31 1994a).
1/28/98 2-2 DRAFT-DO NOT CITE OR QUOTE

-------
1 2.2.2. Polybutadiene Production
2 The second largest use for butadiene is in the production of polybutadiene, accounting for
3 over 20% of butadiene consumption. Polybutadiene is used in tire manufacturing and in the
4 high-impact resin industry. Four companies at five locations in the United States currently
5 produce polybutadiene. The estimate for process vent butadiene emissions from polybutadiene
6 production, as stated in the 1994 EPA report, ranges from 0.00008 to 36.06 Ib butadiene
7 emitted/ton produced (mean of 6.14) measured at six facilities (EPA, 1994a).

8 2.2.3. Neoprene Rubber Production
9 Neoprene, or polychloroprene, rubber production accounts for 5% of butadiene
10 consumption. Neoprene rubber is primarily used in the automotive industry as belts, cables,
11 hoses, and wires. Three facilities currently produce neoprene, though only two use butadiene as
12 a raw material and the other starts with chloroprene. The two facilities identified in the 1994
13 EPA report that used butadiene as a raw material yield estimated process vent butadiene
14 emissions from neoprene production ranging from 0.32 to 6.78 Ib butadiene emitted/ton
15 produced (mean of 4.04) (EPA, 1994a).

1 6 2.2.4. Acrylonitrile-Butadiene (ABS) Resin Production
17 ABS resins are used to make plastic components such as automotive parts, pipes and
18 fittings, appliances, telephones, and business machines, among many other uses. ABS
19 production accounts for 5% of butadiene consumption. Currently, there are 10 facilities that
20 produce ABS resin, only 6 of which use butadiene as a raw material. The estimate for process
21 vent butadiene emissions from ABS resin production ranges from 0.16 to 10.66 Ib butadiene
22 emitted/ton produced (mean of 4.22) measured at three facilities (EPA, 1994a).

23 2.2.5. Nitrile Elastomer Production
24 Nitrile elastomer or nitrile-butyl rubber is a specialty elastomer known for its oil-,
25 solvent-, and chemical-resistant properties. Some uses include hoses, belting, and cable
26 manufacturing and seals and gaskets. Nitrile elastomer is produced at nine facilities in the
27 United States and accounts for about 5% of total butadiene consumption. The estimate for
28 process vent butadiene emissions from nitrile elastomer production ranges from 0.0004 to 17.80
29 Ib butadiene emitted/ton produced measured at six facilities identified in the 1994 EPA report
30 (EPA, 1994a).
1/28/98 2-3 DRAFT-DO NOT CITE OR QUOTE

-------
2.2.6. Adiponitrile Production
Adiponitrile (hexanedinitrile) is primarily an intermediate used in the production of nylon
6,6. Three facilities produce adiponitrile, but only two of these facilities use butadiene in
4 production. This accounts for 12% of butadiene consumption. Despite the large usage of
5 butadiene in adiponitrile production, emissions appear to be fairly small. The estimate for
6 process vent butadiene emissions from adiponitrile production, based on actual emissions
7 reported at two facilities, is 0.12 Ib butadiene emitted/ton produced (EPA, 1994a).

8 2.3. SOURCES AND EMISSION
9 1,3-Butadiene may be released to the environment as an intentional or fugitive emission
10 during its production, use, storage, transport, or disposal. Its sources and emission to the
11 environment can be classified as industrial production and use (1.6%), mobile sources (78.8%),
12 and other miscellaneous combustion sources (19.6%) (EPA, 1994a).
13 Industrial butadiene emissions arise from process vents, equipment leaks, and secondary
14 sources such as waste water treatment. Since butadiene released to aqueous systems or entering
1 5 treatment plants is likely to evaporate completely, all emissions of butadiene can be considered
16 air emissions. Actual reported emissions of butadiene are available through the Toxic Release
17 Inventory, and the relative contribution of butadiene production to the national butadiene
emissions is 0.2% (EPA, 1994a).

19 2.3.1. Mobile Sources
20 Butadiene is formed as a product of incomplete combustion of fossil fuels and has been
21 reported in the emissions from gasoline and diesel vehicles, as well as aircraft. Emissions of
22 butadiene from combustion sources are commonly represented as a weight percent of total
23 organic gas emissions. The relative contribution of mobile sources to the national butadiene
24 emissions is 78.8%, which includes both on-road and nonroad engines. Levels of butadiene in
25 gasoline and diesel fuel are expected to be insignificant since butadiene tends to form a varnish
26 that can be harmful to engines; therefore, refiners try to minimize the butadiene content. Since
27 butadiene is not a component of gasoline, it is not present in mobile source evaporative or
28 refueling emissions and will be found only in exhaust emissions (EPA, 1992).
29 It should be noted that a recent reevaluation by Nordlinder et al. (1996) of the Concawe
30 report (1987) found that the concentrations of 1,3-butadiene in gasoline vapors were much lower
31 than had been reported. Two analyses by Lofgren et al. (1991) and Ramnas et al. (1994) also
32 found negligible amounts of 1,3-butadiene in gasoline vapors. When they compared the
concentrations of benzene and butadiene in gasoline, they found concentrations to be 3%-5% and

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

-------
1 <0.0005%, respectively. Based on these three reports, Nordlinder et al. (1996) concluded that
2 there is no significant amount of 1,3-butadiene present in gasoline vapors.

3 2.3.1.1. On-Road Mobile Sources
4 On-road mobile sources include the following classes of vehicles: light-duty gasoline
5 vehicles (LDGV), light-duty gasoline trucks, heavy-duty gasoline trucks, light-duty diesel
6 vehicles, light-duty diesel trucks, heavy-duty diesel trucks, and motorcycles. On-road mobile
7 sources account for 37.7% national butadiene emissions.
8 Although data on the butadiene content of motor vehicle exhaust were lacking until the
9 late 1980s, butadiene emissions from LDGV's are now reasonably well understood. As
10 mentioned previously, butadiene is not a component of gasoline and is not present in evaporative
11 or refueling emissions; thus, only exhaust butadiene emissions are included. Butadiene has been
12 found to be removed effectively from motor vehicle exhaust by catalytic converters (McCabe et
13 al., 1992). Thus, nearly all on-road motor vehicle butadiene emissions come from older,
14 noncatalyst vehicles, new vehicles with nonfunctional catalysts, the cold-start emissions from
15 catalyst vehicles, and diesel vehicles.
16 The emission factors calculated for all of the vehicles listed above range from 0.01 to
17 0.09 gm/mile (EPA, 1994b). A composite emission factor of 0.0156 gm/mile has been
18 calculated for the calendar year 1990 by the Office of Mobile Sources (OMS) using the MOBILE
19 model. The composite emission factor represents all vehicles classes and is based on the
20 percentage of total vehicle miles traveled (VMT) attributable to each vehicle class (EPA, 1993a).

21 2.3.1.2. Nonroad Mobile Sources
22 Nonroad mobile sources include mobile gasoline- and diesel-powered equipment and
23 vehicles and other equipment types. Types of equipment included in this category range from
24 construction, industrial, and agricultural equipment to small engines used in lawnmowers, chain
25 saws, and other gasoline-powered equipment. Nonroad vehicles include motorcycles,
26 snowmobiles, golf carts, and all-terrain vehicles (ATVs) used for off-road recreation and
27 recreational and commercial marine vessels. However, trains and aircraft are not generally
28 included in the nonroad vehicle category.
29 Generally, most nonroad engines are in use for many years and are noncatalyst engines.
30 The lack of a catalyst, in conjunction with the engine deterioration associated with increased
31 equipment age, may have profound effects on the amount of butadiene emitted. The emission
32 factors expected for the three major engines types in this category—gasoline-powered two-stroke
33 engines, gasoline-powered four-stroke engines, and diesel engines—are generally higher (by a

1/28/98 2-5 DRAFT-DO NOT CITE OR QUOTE

-------
minimum of a factor of 10) than the gasoline engines (EPA, 1991). The EPA 1994 draft denotes
that nonroad engines are expected to contribute 41% to the national butadiene emissions (EPA,
1994a).

4 2.3.1.3. Aircraft
5 Human exposure to aircraft emissions is considered to be limited to the emissions that
6 occur during aircraft landing and take-off (LTO). Airborne aircraft are assumed to fly at
7 sufficiently high altitudes that their emissions do not reach the surface. This assumption is likely
8 to be valid for butadiene because of its short atmospheric lifetime.
9 Butadiene has been reported in aircraft LTO emissions from military, commercial, and
10 general aviation. Based on the EPA SPECIATE database, the butadiene weight percents for
11 aircraft LTO hydrocarbon emissions range from 1.57% for general aviation (piston engines) to
1 2 1.89% from military aircraft (jet and piston engines). The 1994 EPA report estimates that 0.1 %
1 3 of the national butadiene emissions is attributable to aircraft LTO (EPA, 1994a).

14 2.3.2. Miscellaneous Sources
1 5 This section contains an overview of the miscellaneous sources of butadiene emissions.
16 These sources have been grouped as miscellaneous chemical production, secondary lead
smelters, petroleum refining, and combustion sources (especially biomass burning). Emissions
1 8 from these sources can account for 19.6% of the national butadiene emissions.
/ •
19 2,3.2,1. Miscellaneous Chemical Production
20 The 1994 EPA report notes that butadiene is used to produce other elastomers and
21 plastics not mentioned previously, as well as pesticides and fungicides at 19 separate facilities in
22 the United States (EPA, 1994a). This process accounts for 8% of the butadiene use, but only
23 contributes 0.1% to the national average butadiene emissions. The emission factors for process
24 vent butadiene released during miscellaneous chemical production range from 0.03 to 440 Ib
25 butadiene emitted/ton produced (product varies) measured at only four facilities.

26 2.3.2.2. Secondary Lead Smelters
27 Secondary lead smelting involves the reclamation of scrap automobile batteries to
28 produce elemental and lead alloys. There are 23 such facilities in the United States, most of
29 which are located near large population centers. The plastic and rubber components of the
30 battery are the source of the butadiene emissions, contributing 0.4% of the national butadiene
1/28/98 2-6 DRAFT-DO NOT CITE OR QUOTE

-------
1 emissions. The 1994 EPA report lists uncontrolled butadiene emissions measured from a blast
2 furnace yielding an average emission factor of 0.79 Ib/ton (EPA, 1994a).

3 2.3.2.3. Petroleum Refining
4 The 1992 Toxic Release Inventory contains the emission factor of 437,590 Ib/year for
5 petroleum refining. Using this emission factor would make this source the fifth largest emitter of
6 butadiene, contributing 0.3% to the national butadiene emissions. Data are currently being
7 collected to determine the actual contribution of petroleum refining to butadiene emissions.

8 2.3.3. Combustion Sources
9 Butadiene is, as mentioned previously, a product of incomplete combustion and has been
10 reported in the emissions from gasoline and diesel vehicles, as well as aircraft. Butadiene is also
11 released during the combustion of tobacco, biomass, and automobile tires, although only the
12 latter two will be discussed in this section due to the scarcity of data.

13 2.3.3.1. Tire Burning
14 There are approximately 240 million tires discarded annually, of which only 25% are
15 recycled. The remaining tires are discarded in landfills, stockpiles, or illegal dumps (Lemieux
16 and DeMarini, 1992). Tires are combusted through accidental fires at stockpiles, illegal burning,
17 tire-to-energy facilities, cement kilns, tire manufacturing facilities, and as a supplemental fuel in
18 boilers. Butadiene is a major constituent of the tire manufacturing process and therefore it is
19 present in emissions from tire burning. Emission factors have been calculated for the open
20 burning of tires (EPA, 1992; Lemieux and DeMarini, 1992). These emission factors range from
21 234.28 lb/1,000 tons of chunk tires to 277.95 lb/1,000 tons of shredded tires. No emission factor
22 for butadiene from tire incineration has been located.

23 2.3.3.2. Biomass Burning
24 Biomass burning includes residential wood combustion in both fireplaces and wood
25 stoves, open burning such as the backyard burning of yard waste, slash burning, land
26 clearing/burning, agricultural burning, forest fires/prescribed burning, structural fires, and other
27 wildfires. Although these fires differ in many important characteristics, the fuels in all cases are
28 primarily composed of wood. The relative contribution of biomass burning to the overall
29 national butadiene emissions was calculated at 18.8% in the 1994 EPA report (EPA, 1994a).
30 Emission factor models based on field and laboratory data were developed by the U.S.
31 Forest Service (Peterson and Ward, 1989). These models incorporate variables such as fuel type

1/28/98 2-7 DRAFT-DO NOT CITE OR QUOTE

-------
and combustion types (flaming or smoldering) and these models correlated butadiene emissions
with CO emissions to develop emission factors for biomass burning (Campbell and Mangino,
1994). The calculated emission factors range from 0.40 Ib/ton of yard waste burned to 0.90
4 Ib/ton for large wood burning in forest fires and prescribed burning.
5 Butadiene emissions have been reported from the combustion of wood (Sandberg et al.,
6 1975; Ward and Hao, 1992). The data of Ward and Hao (1992), in which both butadiene and
7 benzene were quantified from biomass burning, provides a butadienerbenzene ratio of 0.36 for
8 wood smoke.

9 2.4. AMBIENT CONCENTRATION OF 1,3-BUTADIENE
10 2.4.1. Air
11 In 1989, total emissions of 1,3-butadiene to the air in the United States were estimated at
12 approximately 2,512 tonnes from 158 locations; total land releases were estimated at 6.7 tonnes
13 (U.S. National Library of Medicine, 1991).

14 2.4.1.1. Ambient Monitoring Data
1 5 Several EPA databases exist that contain the results of various air toxics monitoring
programs. These programs have set up monitoring devices that are used to collect air samples all
over the United States over a period of months or years. Three of these programs/databases
contain data on 1,3-butadiene. This section summarizes the three monitoring programs and
1 9 presents annual average concentrations of 1,3-butadiene derived from these programs.
20 One of these programs is the Aerometric Information Retrieval System (AIRS), which
21 became operational in 1987 and uses a network of monitoring stations called the State and Local
22 Air Monitoring System (SLAMS) (EPA, 1989a). This network consists of monitoring stations
23 set up by every State in accordance with regulations promulgated in response to requirements of
24 the Clean Air Act. EPA's Office of Air Quality Planning and Standards (OAQPS) administers
25 the AIRS program.
26 The AIRS program allows State and local agencies to submit local air pollution data and
27 also have access to national air pollution data (EPA, 1989a). EPA uses data from AIRS in order
28 to monitor the States' progress in attaining air quality standards for ozone, carbon monoxide,
29 nitrogen oxides, sulfur oxides, and lead through the use of State Implementation Plans (SIPS). In
30 addition to containing information about each monitoring site, including the geographic location
31 of the site and who operates it, the AIRS program also contains extensive information on the
32 ambient levels of many toxic compounds. The AIRS database catalogs ambient air pollution
data from 18 to 55 monitors in 15 to 23 urban areas, depending on the pollutant. These monitors

1/28/98 2-8 DRAFT-DO NOT CITE OR QUOTE

-------
1 collect a 24-h sample every 12 days. However, in some cases not every target compound was
2 detected in every sample. Where this occurred, half the minimum detection limit was used in the
3 averaging of the data for this summary. The annual average ppb for each site was calculated
4 using only those sites that provided four quarters of monitoring data. The cities monitored and
5 the average concentrations determined can be found in Table 2-1.
6 Another air monitoring program is the Urban Air Toxic Monitoring Program (UATMP),
7 which the EPA developed in 1987 to assist State and local agencies in determining the nature and
8 extent of urban air toxic pollution (McAlister et aL, 1989,1990,1991; Wijnberg and Faoro,
9 1989). Data from the UATMP also is used in air toxic risk assessment models (EPA 1989b, c;
10 EPA 1990a, b). In 1989, the UATMP had 14 monitors in 12 urban areas, and in 1990, the
11 UATMP had 12 monitors in 11 urban areas, of which 9 also participated in the 1989 monitoring
12 program.
13 In 1989 and 1990, the UATMP network simultaneously monitored 37 nonmethane
14 organic compounds, selected metals, benzo(a)pyrene (1989 only), formaldehyde, acetaldehyde,
15 and acetone for a 24-h period once every 12 days. The UATMP database lists the data collected
16 from the monitoring network using two methods. In the first method, only the concentrations
17 above the detection limit of the compound are included in the data. In the second method, if the
18 concentration of a compound is below the detection limit, then one-half of the compound's
19 detection limit is incorporated into the data. The second method was used because it seemed
20 more reasonable and allowed a greater number of samples to be averaged. Data collected in
21 1989 and 1990 were used in this summary. The cities monitored and the average concentrations
22 determined can be found hi Table 2-2.
23 The monitoring data for the UATMP that were collected from 1991 to 1994 have not yet
24 been released as separate reports. The data collected in those 4 years were entered into and
25 reported as part of the 1991-1994 AIRS database.
26 The National Ambient Volatile Organic Compounds (NAVOC) Database contains
27 approximately 175,000 records on the concentrations of 320 volatile organic compounds (VOCs)
28 observed in 1-h air samples taken every 24 h between 1970 and 1987 (Shah and Heyerdahl,
29 1988; Hunt et al., 1988). However, only the most current NAVOC data, taken during 1987, is
30 used in this summary. In addition, samples that had nondetects of 1,3-butadiene were included
31 as one-half the detection limit in averaging the data for this summary. The NAVOC Database
32 includes air samples collected using indoor and outdoor monitoring devices. Personal monitors
33 were also used. The types of locations of outdoor monitoring sites included remote, rural,
34 suburban, and urban areas, as well as near specific point sources of VOCs. Indoor monitoring
1/28/98 2-9 DRAFT-DO NOT CITE OR QUOTE

-------



1-H
OS
Os
V*
o
-fcj
oo
00
Os
^
•-
,o
co^
£
•
e, o
E «*
« Os

*^«

•_
J= 00
s s
3

OO
oo
Os
1—1




s
'•C
u
_0
!M
_o
'S
o


"S °^
£ OS
•4^ VH
U
&
O Os
u oo
M i-l
2

oo
oo
Os














p-







^r-
u
(U
s
•e
5
.2
'o
Commei


o
Q
§
00
^










2
o


















Os
CN







.&>
O
s
§
u
1
D
12
'3
Commei

j
CS-
. Lauderda
4->










Os
CN
O


















en







^T1
O
o>
C
6
£2
CO
i
'3
Commei



1
S










^J-
o


















'O'OVOO'OOsOS'^t-^-'^-OsOO Os
-1 CN CN CN CN



VO



£" ^> ^> ^ ..^ ^> ^»
O U O O QUO
_gJ^J^J^J £j> . tnOJO
"S e c C o S S "S
(Dcjum •— ' eeSS
OUUUfe §UO
c^o^d^Jci^C^ ^0°^^
|||S|o|||||BS§
•£•£-£•£ gcWii'S'S'S g S-P-0
5555-§1^^^^-§^55
SisSSoi cosa^sooHD^s
.S .2 .2 -S ~ ^ ^ & co c/3 is — co co
o 'o 'o 'u .2 2 "* ~ — * — -2 .S '3 '3
D 3 a> <3 c >j -2 -2 .2 .S • a g s 5
CCCcSJiisisJaisSScc
5 5 S ST?-^ w coco WT-1T3 C C
1 . 1 1 1 . s | 1 1 1 1 s s 1 1
OUOOPiS'^^StSpiesicjO


d d d d >H -^
ijjijdcjdd^^oo
JJJi— < i— i -i— 'OUOUi^W aJ~3
•S -S 'Sill '§ "^ '§ '3 -^ -^ ^ ^
1 1 1 .§ .1 S u o o o -g -S •§• |
N5-<^N^^^^-4-»*J4-i*J^.^>». O CO
^^^UUCJ&O&Q&Q&OESESjeQ










Os ^1* en o Os en CN *— * C> *o en VO en
oooooooooooo o


^.
^sj"
0













•*







^.
o

U
3
"c
o>
U
c
Ui
is
'3
Commer


-3
co"
3
O
Pi
C2
'to










en
o





1/28/98
2-10
DRAFT-DO NOT CITE OR QUOTE

-------
I
••—
53
O

N""'
0>
«T\
*" *
2
00
OO
T-t
L*
cS
*••«
M^
CS
1— 4
•*-•

€3
S

•raation Rel
*^
(2
_.
i
E
o
o
P-,
W
1
S

o
!o
S
o
f*
.a
•a
^
ja
i
**».
*™^
o

>>

cs
S
E
to

<
i

a
S
c:
H
S
rH
ws

C< _,
Sv^
C7\
d O\
M f-t
<*-!
O
U
% ^
•Q CO
B*™ G\
^^
s **
Z

OO
OO
I °^
1 r"(

o

'es
u
0
Df monitor 1
0
e/i
1





1
*«
_C
"E.
S
c:
CO


s\
o\
1-1
s-
c.
e o
O G>
**•* n\
•^ US
_2 '""'
"S
U
U
gj
0 Ov
« 00
o o\
M

 3 3 3
i- X> XI X)
±) 3 3 3
S CO CO CO
co 5s i= CS
"3 2 .2 .2
<3 c S3 e
c 53 u 53
G -2 12 12
C "eft 'eft "eft
O CJ CD O
O p-1 Pi Pi





0
•2^22
•— cf cs' ef
3 0> CD CO
o -o -a -a
-J S | S
co CJ (3 O














CN cn ^o oo
— • CN — ' —
O O O O



















^t* O^ Os O*\ O^






* G "
r"\ "-1
O tH CD
CO •*-*
JB a §
c ,""0
lilli
11111
^3 ±3 ^i S CO
2 ~^ .2 .2 "o
c -2 c c3 -, > >
r, ^^ "Q *l2 *Tl
1=3 ^ "o ^2 £
TJ ^ ^ ^ ^
S £ -S £ £
rrt CO TO CJ CO
r ^ *y 7Z* *7" *^














>O VO O 1O CT\
— i CN CN CN CN
O O O O O



















en
CN





O

O
1
u
al/Urban &
"o
U
s
o
U






X
CO*
CO
*^
Q














^
o















"*



0
en










ySuburban
CO
d
'w
CD





X
P
Houston






CN
r-
o





o
VO
o



t— <


^H


OO
1— 1




en










O O
IH %-c

1 S
CD CD
U O
ayUrban &
al/Urban &
'o 'o
Ct> CL>
S S
o o
O O




5
f-,
> rS
1 §
.S °*
1 =5
03 <


—
0


o
CN
O










r~-
•3-

0


CN





Js















I/Suburban
CO
c
'eft
(S




^
>
8
o
ffi


CN
0


VO
1 — 1
o
















CN >0 CN
CN ^^ CN




2 S 2















S3 C S3
CO W CC
•£•£•£
333
x> x> x>
co co co
cC cQ cQ
a e c
CD CD CD
12 12 12
'co 'eft 'So





< ^ <
> -^ >
f 1 1
S OH §
CO O O
ffi ffi Pi


n g 2
0 O C3


CN cn •*
CN ^^ CN
o o o
















































LH
(D
'co
XI Cft
OH Cft
o b
"I <
0 e
o
n 1
o>
c^ o
CN S
0 °
0 0
-------



ON
ON
O
•*••>

ON
ON
^H
^
c2
/— V
(/)

I— I
S
g
It
"3
'C
&
e

•g
CS
>metric Inform
CJ
Bk
0>
s— »
S
£
CB
•w
2
cu
15
S
es
o>
•3
es
"3
t
fi
r- <
CM
O
^
i^
es
E
s
CO
.
w
1
«N
3
H


ON
ON

Cn
"H.
S
cs
en fO
O ON
i. »H

"1
S
z

ON
ON


_o

cs
u
ie of monitor lo
l«
•o
J





a>
Sampling sii



ON
^— N ON
0 T— <
C.
a
^^
e
_o
1 |
•w O\
e i-<
u
e
o
4)
;> C^
-< S
ON ON
c--






co
CX)















tial/Rural
tial/Rural
c e
CO CJ
•a -a
CO CO




J J
r r
Jefferson Co
Jefferson Co



CO O
CO — <
0 0





CN
o






—i. O
oo —<






CN
oo















tial/Suburban
tial/Suburban
§ §
•O T3
CO CO
CJ CJ




J J
<  tl>
^ "c
 O

•§ s
'i I
erf CJ




< O
0 Q
Simi Valley,
Washington,



r~ oo
—i CN
o o





oo ^o
— ' CO
0 0







1— t *— 1






CO VO











0 ^>
c3 CJ
•y
§ CJ
•cial/Urban & C
il/Urban & Cent
1 1
1 1




O
p
1 -
.s &
-s s
"2 ^
i> o



^o oo
0 0





CN ON
CO O
0 C5






1/28/98
2-12
DRAFT-DO NOT CITE OR QUOTE
-------
o
U
ss
a
_«


•o


.1
~
•3
C3

S
C/3


ca

S
bo
_o
"E.
S
a
•a ff>
<— 0<
0 OS
a>
1
s
2
t<«
o\
OS
_o
&
CI
Land use of monitor lo

CO
"to
ec
"S.
1
rr
OS
S^ S5
C.
a
5
1 s
*-• OS
u
CI
s
o
CI
o
M
2 ^

>. Pv


1—1





V")
»— 1











Residential/Suburban


•4-T
O
U
OO
O
o



VO
o











CN
*— t





O












St. Louis Metro Area

d
d
U
'co
o
4-«
oo
-
o



o
C5











en






,_«
,«^








O
^^
U
Industrial/Urban & Cent

oo
O
CO
vo
CN
O



£
O











00






^j.
_^








O
B

fl)
Commercial/Urban & C

•j.
1
c
o
to
CQ
en
O



CN
O











•o




Commercial/Rural


•cf
s
p
C5



OS
O








OO
O
O
CN
CN





! 	 1
VO





>0
*o




Commercial/Rural


1
^
g
o



o
o








t —
o
o

vo





! 	 [
vo





>0
"O




Industrial/Rural


S
•d
J
S
o
o



o
o








f-
o
o

VO





T—t
VO





>0
to




Agricultural/Rural


•a"
M
^
t-
0
o



0
o








f —
o
o

VO





0
vo





>o
1O




Industrial/Rural


S
•a"
S
r-
0
o



oo
o
o








r-
o
o

~





I.H
en





00
CN




Residential/Suburban


-a
CO
s
r-
0
CJ



o
o








t —
o
o

~





0
en





oo
CN




Residential/Rural


S
•o"
J
s
0
o



o
o








f —
o
o







*o
,-H





^J.
CN

O
'S

Commercial/Urban & C


1
d>
Q





oo
o
o








CN
f-
o
1/28/98
2-13
DRAFT-DO NOT CITE OR QUOTE
-------
^

*c
nformation Re
CJ
Aerometrii
cu
W
cu
5
o
-2
CQ
•a
"S
—
S
S

s
OS
"3
t

c
*
cs
S
3
cn
.
C3
,•4
i

U O 0
5 S § § •« « ««
i- JD ^ X5 C C fi
S M £3 fa ^ ^ ^
^O 737 jQ rQ _Q
3X)Xl^)5-' *-« «-« T i
^£££555 |
ca^a^^s ca.ca CCD^
'o .2 .S .2 'o '3 'o Sa
cSaacajcucS-S
Sa> cu a> e a c il
cJ2"S"2 S S § ™
S 'S3 '55 'to 5 S S £
O CO 0}  S3 C~^ w en *, «
OT ^S3 ^2 ^ ^«J2
^^S ca-S t- 8'^3'ra
JZ< f^ co O CQ CQ Q Q

o oo r-- ^^ r- • t-*
C^* C^ CO C*^ CO CO
o o o o o o



2 S 2 gj S §
o 000*00











\j"\ if\ c^ ^^ ^^ ^*J" ®® T^ *^i ON
r^" ^o v~j i—* iD







Q^ Q^ ^-*. Qv QQ -^J ,-. rs^. ^--
CO CO ,*— I> CN CO -H CN





»n *o





QUO - -^
o o3 fe
^ "c "S .S3
o> 53 S a
O O C) «
^^ ^ C3 Kt 03 ^ ^5 ^5 Z^; cC
_P? 3 S* iL y *j-H «^H »~< P-t
O bD O O O O Qi (U 
-------
f
s
e:
"a

"«

r^
O
-*«»
o\
T-t
«
CO*
C5
S
U
JO
CO
_CJ
!•
O
Information R
.0
2
<1>
fd
u
.a
S
P
e ambient data ft
e
•3
re
•*•*
A
m

o

CS
e
5=
3
CO

ca
*^*

s
0
w
o
C3
^-
>• ^*
-S CO^
"-^, -^ CO CO CO "-^ CO CO Q CO
.2 .2 S '•§ '•§ -2 '•§ '•§ 53 'i
i3i3Q>i3OC
c/3 co "O T3 T3 OT ^O *T3 5 'O
•S-fSScS-lcSSocS
^ t^ fv* |V* pyj ^ p^ p^{ r_^ p^





^s .,
KXpsceisB5B
rSrSf^t^ *•« w co*r" bn
HH^-r^-Ts °> w 2 -S C
gg||^111^S
22||<^^^|0
SSggtjtitj-e&'g
ooa>n





en









&
U
S

ial/Urban & Ce
1
1
u






f-
o
an
CQ

OS
•*
o




en
O












m m





en -^t









if
o
ca
e
ia!/Urban & Ce
ial/Suburban
B H
1 1
Sr-
C
O O
O U






Rutland, VT
Waterbury, VT

•<* Os
O O




m oo
0 0









































t-l
CO
X)
1
c»
CO
CO
2
ON o

°' 1
c
U
o
rs C
p, o
G, a
O jj
"* 00
o g

^;
^^
S3
-O ^
& 12
vo -S
_ 3
O ^7
— "

1/28/98
2-15
DRAFT-DO NOT CITE OR QUOTE
-------
I
 cs
 a
•s
 o
 in
 U

'i
H
 i.
 c
 R
 O

5



 O

£

 cs
s-»
 ft
t3
 

CO
S
cr
CJ

"aJ
- J
•4-*
M
"E,
E
S3
WD



S
.2
"S
U
0
-*^
'5
Land use of mo





Sampling sites


/•~s
•fi
a |
C i— i
^C
?
&M
"£
0)
U
C
o
U

05° °°
a3 ^^
^.
.^

Os OS
JN £N
OO •^f



rn JS
i — i i—t







& &
ca ca
Q Q
CN CN
53 

03 03
o o
1 1
CN CN



"c8
'g
1
— i O
2 O

33 •&*
Suburban-Resid
Urban/Center C:





1 1
CB CB
0 Q




O'
o









o oo
CN O
0 0



00
CN



rn rn
oo m







>. >>
c3 CO
Q Q
CN CN
iT ^
> >
03 03
o o
K K
t I
CN CS



"cB
H
. >>
cS c3
P Q
CN CN
£ sT

03 03
1 1
1 1
CN CN



"ca
'g
 b

O3 U3
3 3
O O
ffi E
4 4*
CN CN



^2
o
o
£
1
i "ca

Urban/Center Ci
Suburban-Indusi





0
S
_2f K
o t3
-i-: ca
OO GO




0
o









OS O
0 CN
0 0



o t-~
s ••§



i 1







CO CO
C? Co
Q Q
CN CN
£> &
o 53

03 03
. 3 3
i i
CN CN



CQ cO
Ho
LJ
u S3
a s
1 1
U U
1 1

Urban/Center Ci
Urban/Center Ci




1 1
Washington, DC
Washington, DC




O 10
0 0









<— ' Os
— < CN
O O



0
en



2 ^







to co
ca ca
Q D
CN CN
b b
• >•
03 03
z §
K ffi
4 4
CN CN



"c3
"*-*
•a
1 3

>> a
Urban/Center Ci
Suburban-Residi





CO GO
S .2
1 1




VO
O
O









\O OS
T-. O
0 0



00
CN











CO
ca
Q
CN
sT

03
z
'
CN






_

S
Suburban-Residi





Port Neches, TX




OS
o
rt














oo



§







« CO
Ctf CH
Q Q
CN CN
£> b

£9 03
Z §
4 4
CN CN



is
'o
CD
1 -
U .2

•&" ^
Urban/Center Ci
Suburban-Residf





f7> S'
O
-§ o
c "S
"C "o
0 H




o vo
•-< o
0 0



















































Q.
Q.
f~l
_0
—t
C
CN o
o a
^ o
0
o>
Ofi
2



"^
g
CO
^T* *^J
CN ca
O 3
CQ
1
2-
T3
3
o
0
C
.CO
•o
C
3
O
a.
S
0
o
s
43
43
U
1
S3 _;
•—  1
2" t'~'
Q J>
^ 1"
•- to

C m
o »
'minimum detectii
qually in 1989 anc
"ca M
43 0
oc -^
c w
1 2
in of all samples i
iples taken from
y c
1 0
C OS
O
e —
o ca
0 00
OJ C
ca° '5b
t- ^
> 
rt ^


*"*~* Q
S -g
J2
"C -^
a> .5^
4^ 13
f_ (J

    1/28/98
2-16
DRAFT-DO NOT CITE OR QUOTE
-------
 1      sites consisted of nonindustrial workplaces and residential environments. Personal monitors also
 2      are included in the indoor category.  This database was an interim precursor to the air toxics
 3      portion of AIRS. For this summary, only the outdoor urban data were used.  The cities
 4      monitored and the average concentrations determined can be found in Table 2-3.
 5            Table 2-4 summarizes the average concentrations (in ppb) of 1,3-butadiene found at the
 6      monitoring sites of each air monitoring program.  The table also shows the total number of
 7      observations for each average and the number of sites that monitored the compounds in each
 8      program. For AIRS, the average concentrations of 1,3-butadiene are listed separately for 1987
 9      through 1994.  Some of the highest averages in the AIRS database were from suburban
10      residential sites in Houston and Port Neches, TX. Both of these cities have high point-source
11      emissions that could be affecting the monitor. The AIRS and UATMP data from Houston and
12      Port Neches, TX, were excluded to create alternate annual averages (ppb and ug/m3) for the years
13      1988 through 1994 (where applicable) and are presented in Table 2-4. This alternate annual
14      average may be more representative of areas that are not near strong point sources.
15            Tables 2-5,2-6, and 2-7 regroup and summarize Tables 2-2, 2-3, and 2-4 according to the
16      sampling locations, i.e., rural, suburban, or urban settings. The data obtained from Port Neches,
17      TX, were not included in these averages because of the elevated levels due to industrial
18      emissions.
19            It should be noted that methods of averaging the data are not consistent between the air
20      monitoring databases. Also, in the NAVOC monitoring network, samples were taken for 1 h in a
21      24-h period while the other monitoring networks collected a 24-h air sample every 12 days.
22            It should also be noted that the ambient levels detected in these three databases are not
23      meant to be indicative of an individual's actual exposure to  1,3-butadiene. Times and
24      concentrations in microenvironments other than the outdoors need to be taken into consideration,
25      i.e., accounting for integrated exposure.
26            In addition, the ambient levels include contributions from a variety of source categories.
27      Typically, ambient monitoring data are adjusted to represent the amount attributed to a particular
28      source using emissions inventory apportionment. The derivation of an urban annual average
29      exposure estimate for all mobile sources will be used for illustration purposes (EPA, 1993a).
30      The range of ambient data from Table 2-4 (using alternate annual averages when available) is
31      0.22 to 1.02 (ng/rn3 (0.10 to 0.46 ppb). When this range is adjusted by the estimated proportion
32      of the inventory that is contributed by mobile sources (78.7%) and for integrated exposure to
33      account for time spent indoors and outdoors, the range becomes 0.11 to 0.50 ug/m3 (0.05 to 0.23
34      ppb).
        1/28/98                                   2-17       DRAFT-DO NOT CITE OR QUOTE
-------


tfi
T3
S
O
S1
o
U
'S
CS
bX)
1.
o
• M
CS
"3
"S
he National Ambii
S
o

«J5

"es
5*
•a
=1
O
O
f^^
•«•»
ys
o
o
es 2
S £
S 5
s ts
IX) Q
• /-N
fO I^J
fN O
4d ^
^M 1^^
-° 
oo
^p
0
^ — i


co
O
1







S3
cS
a

o
0"
co
o
1— 5
00





CN










10
CN
O





oo
t--
o
^
+->
c-~-
oo
ON
ON


co
1
•4-
CN







rj
1

Stockton, CA





•-










0
CO
o





00
ON
3
r--
oo
ON
ON





















'o
&;
^
O


S3
O
'•+3
_£j
; conceni
us
bo
S
o
z>
ID

.2
1
CQ

CO
, 	 1
cti
03
o




















>.
"re
cr
•1
"o

W


43
e
re
C/3
"Hi
CO
C
'5
"re
W)
averagin
>>
•a
•S3
-^5
§
"re
O
1/28/98
2-18
DRAFT-DO NOT CITE OR QUOTE
-------
      Table 2-4. Summary of air monitoring program results for 1,3-butadiene

AIRS
1988 level
(Ppb)
#Obs.
# Sites
1989
(ppb)
#0bs.
# Sites
1990
(ppb)
#Obs.
# Sites
1991
(Ppb)
#0bs.
# Sites
1992
(Ppb)
#Obs.
# Sites
1993
(ppb)
#Obs.
# Sites
1994
(ppb)
#Obs.
# Sites
Annual average
ppb Gig/m3)


0.67(1.48)
18
3

0.23(0.57)
399
30

0.29(0.64)
101
7

0.10(0.22)
117
6

0.16 (0.40)
656
20

0.40(0.88)
2069
64

0.59(1.30)
2666
70
Alternate annual average3
ppb (ug/m3)


0.46(1. 02)b
12
2

0.25(0.55)b
369
29

0.21(0.46)b
97
6

____c



0.16(40)b
650
19

0.32(0.71)d
1,931
59

0.42(0.92)d
2,401
64
1/28/98
2-19
DRAFT-DO NOT CITE OR QUOTE
-------
        Table 2-4.  Summary of air monitoring program results for 1,3-butadiene
         (continued)
                        Annual average
                          ppb (ug/m3)
                         Alternate annual average3
                                 ppb (ug/m3)
       UATMP
       1989
       (Ppb)
       #Obs.
       # Sites
0.21 (0.46)
   390
    13
       1990
       (ppb)
       #0bs.
       # Sites
1.02(2.25)
   349
    12
      0.12(0.27)b
         293
          10
       NAVOC

       1987
       (ppb)
       #Obs.
       # Sites
0.34(0.75)f
    9
    6
        no data
       "Alternate averages do not include data from Houston and Port Neches, TX, due to impacts from
       strong point sources.
       bAverage ppb from all 4-quarter data sites, excluding Houston, TX.
       cHouston, TX, was not monitored during this 4-quarter period.
       dAverage ppb from all sites, excluding Houston and Port Neches, TX.
       ePort Neches, Texas, was not monitored during this 4-quarter period.
       fAll urban California sites.
1/28/98
                2-20
DRAFT-DO NOT CITE OR QUOTE
-------







§
g.
s
o
en
&
>C3
^.
Ci
•«•<
•*-*
£
c
«»•*

5
,o
C2
HH
.«
-fc
o
o
•<

PM
W
o
.£!
•*••
S
o
cs
Is
T3

1
2
es

C
O
PQ
-S
S3

**i, a
»— < .2
'o "«
S-. w

OS "te£
S a'
i'-a
-I
"? C
cs o
O 13
**JS? ^
« en
H JS
Qi f. .
1 5 1

« "2 cs
— So
cs s _o
"S s ~"
_ H
cs
cs m
B M
es K
P ^



o

g-
CS
CS

"S
1 u S
S S 2
eg ^ J
W g CS
— Co
« S 0
0 c
H

3 •<
C/3





1


"S5
o <— ,
e. ° e
111
* S 1
Is 3 ^
"o e
H


°— >
CN




vo P- r- o\ os
TT CN CN O CN
O C5 O O O


t— en en — CN
rf r- en •— < r~
O O O C3 C5
i i i i i
•^ o o r-> os
•^T i— ; CS C> O
C5 O O O' C>


« n ^ t^-
\o CU wn ^^ •^~
i — i Os t — "O

•— 1





•— >n os o o
« CN CN •-; •-;
—I o o o o




O CN CN Os
vo C^- T— ; »--;
C5 O O O
i en vo vo tr)
' — ; — ; 0 0
O O O O




>0
CN







2
o




v~>
o
o
o

OO Os O ^H CN
OO OO Os Os Os
Os Os Os Os. Os


3
O
oo




en
o


oo
Os
CD
i
oo
o
o


en

00





o




cs
en
^f
o
o




o
OS







o
o




c,
cs
Cp
"O
o
o

en
Os
Os


5
o
oo




cs
VO
o


5i
^
1
VI
o
o


cs
en






c>




oo
VO
'-H
1
VO
o
cs




-
CN
cs







00
o




o
en
o
c~-
o
o

^J.
Os
OS













































j£
C.
ex
-S
C
•2
1
-------

















=
•1
(8
W

^
W>
"ft
S
08
O
•B
V
93
OS
JS


(•>•<
J2
os
H


cu
1
s
S
*eJi
QJ
"is
a o

•» C3
« 0
o-
H o





re
So
2
^





e
C!
OS



<»
.0
^. "S
S o
gv T*
C Q
re i.
CQ CU

5 s
H e





c 1u

*" 2
•° "
J™ ^C
\fj

0>
bJ)
i


es
4>
><









go
m
CN






oo
o




ON
O
1
oo
o
0











10
Io








CN
0

O
VO
o
1
ON
o
o


ON
00
ON
•"^









'C
5






IT)
o




VD
m
o
\0
o
0











JO
g








2
0

t-
o
vo
o
o


ON
ON
'~~l



T3
C
3
<£
o
c
CO
CS
•u
C
o
0,
s
o
o
o
•3
15
i
a,
CO
CO
;_<
o
M
>
_0
t>
0>

-a
1
g
"5
*2
CH
i!

•?
1
60
"co
3
M

"a,
S
CO
CO

& c2
.s °
O ®
"cfi ^
ic average concentr
:ic average concent)
§ g
11
£3 "S
CQ «
i O
v\ JS

ra^ _Q~*
&
_o
t^
OS
^o
be
"S.
08
CC
&
o
en
S3
I

M
08
H
S
o
<*4
08

ts
ft
CU
S
.2
^o
08
"S
^

^
'H
0
>j
08
S
S
s
02
a
^^,
i

an
2

CO
0
c
.2
S
CQ
cd
1/28/98
2-22
DRAFT-DO NOT CITE OR QUOTE
-------
 1      2.4.1.2. Ambient Source Apportionment
 2             There are three studies that attempt to apportion sources as to their contribution to
 3      ambient levels of 1,3-butadiene.  The studies assume that all emissions to the atmosphere
 4      contribute proportionally to ambient concentration. These three studies are summarized in Table
 5      2-8.
 6             As observed in Table 2-8, the source apportionment conducted by Systems Applications
 7      International for the American Automobile Manufacturers Association (Ligocki, 1993) contains
 8      the category of biomass burning as a large part of the inventory. The weight percentage of
 9      butadiene in TOG for emissions from residential wood combustion, open burning, forest fires,
10      and other burning that are used in this analysis are derived from a single estimate provided in
11      EPA's SPECIATE database.  An actual 1,3-butadiene TOG weight percentage for incineration of
12      wood was not found in the literature; therefore, the solid waste incineration TOG weight
13      percentage was used.  There is a great deal of uncertainty connected with the 1,3-butadiene
14      emission estimates that were developed for the biomass burning as well as for emissions from
15      aircraft. Many of the limitations revolve around the lack of real-world data on actual 1,3-
16      butadiene emissions and exposures for the scenarios mentioned above, as well as the allocation
17      of these scenarios nationwide. The emissions from residential wood combustion and forest fires
18      vary by season and region of the country. The mobile source and stationary source emissions
1 9      would, for the most part, remain constant throughout much of the year.

20      2.4.2. Indoor Exposure to 1,3-Butadiene
21             Information on 1,3-butadiene concentrations in homes or public buildings is limited at
22      this time. Indoor concentrations  of 1,3-butadiene are primarily dependent on the presence of
23      environmental tobacco smoke (ETS) (CARB, 1992). Several studies indicate that on the average
24      most individuals spend anywhere from about 60% to 70% (Robinson et al., 1989; EPA, 1993b)
25      of their time each day indoors at their residence. In addition, individuals also spend a lot of time
26      at indoor workplaces. This makes indoor air a major route of exposure to 1,3-butadiene for
27      individuals who are exposed to tobacco smoke. It is also apparent that the potential for indoor
28      exposure can exceed outdoor  exposure if ETS is taken into consideration. Lofroth et al. (1989)
29      and Brunnemann et al. (1990) measured 1,3-butadiene emissions in sidestream smoke ranging
30      from 200 to 400 ug/cigarette and 1,3-butadiene levels in smoke-filled bars ranging  from 2.7 to 19
31      (J-g/m3. Further research and measurements are needed to quantify typical indoor 1,3-butadiene
32      exposures.
        1/28/98                                  2-23       DRAFT-DO NOT CITE OR QUOTE
-------
                Table 2-8. Summary of the relative contributions to ambient 1,3-butadiene
                emissions given as percent of total mg/year
Study
EPA, 1994a
GARB, 1992

Ligocki, 1993
Mobile
sources"
78.7
96d

57
Stationary point
and area sources1"
2.4
4

.5
Aircraft
0.1
Included under
mobile sources
3
Biomass
burning0
18.8
NDe

35
          "Mobile sources included on-road and off-road vehicles and generally excluded trains and aircraft.
          bArea and point sources generally included all manufacturing and industrial process, oil and gas production
          facilities, commerce, residential fuel combustion, and other stationary fuel combustion.
          cBiomass burning includes residential wood combustion, incineration, and other biomass burning.
          dThe CARB off-road apportionment of mobile sources includes trains and aircraft.
          eND=not determined.
        2.4.3.  Water
               Although 1,3-butadiene has been detected in drinking water in the United States (U.S.
        EPA, 1978; Kraybill, 1980), it is not clear what happens to the chemical in the body (U.S.
        DHHS, 1992). Total releases to ambient water in 1989 were estimated to be 65 tonnes (U.S.
 5     National Library of Medicine, 1991).

 6     2.4.4.  Food
 7            Certain cooking oils release butadiene on heating.  For example,  1,3-butadiene emissions
 8     are approximately 22-fold higher from unrefined Chinese rapeseed oil than from heated peanut
 9     oil.  Of three fatty acids tested, heated linolenic acid produced the greatest amount of 1,3-
10     butadiene. Although cooking oils in the U.S. are refined for purity, U.S. rapeseed oil (canola)
11      also emitted 1,3-butadiene (Shields et al., 1995). Also, levels of <0.2 ug/kg 1,3-butadiene were
12     found in retail soft margarine; the plastic tubs containing the margarine contained < 5-310 fig/kg
1 3     (Startin and Gilbert, 1984).
14     2.5. PATHWAYS OF EXPOSURE
1 5            The 1992 U.S. DHHS report states that although 1,3-butadiene undergoes rapid
1 6     destruction in the atmosphere, it is almost always present at very low concentrations in urban and
17     suburban areas.  Automobile exhaust is a constant source of 1,3-butadiene release to the
        atmosphere.  Because of the compound's presence in the atmosphere, the general population is
        1/28/98
2-24
DRAFT-DO NOT CITE OR QUOTE
-------
1      exposed to ppb levels of 1,3-butadiene through inhalation. Exposure to 1,3 -butadiene may also
2      occur from the inhalation of cigarette smoke, or possibly the smoke from wood fires.  Possible
3      ingestion of contaminated drinking water may also lead to low levels of exposure, although the
4      concentration of this compound in drinking water has not been well characterized. The levels of
5      1,3-butadiene in soil are not known. Elevated levels of exposure for the general population may
6      occur for those near its site of manufacture or facilities where it is made into polymeric materials.
7            Occupational exposure to 1,3-butadiene is expected to be limited to those working at
8      facilities that manufacture 1,3-butadiene or convert it into commercial polymers.  Exposure by
9      inhalation is expected to be the dominant pathway for exposure.
       1/28/98                                   2-25       DRAFT-DO NOT CITE OR QUOTE
-------
                            3.  METABOLISM AND PHARMACOKINETICS

               The pharmacokinetics of 1,3-butadiene have been reviewed previously by the U.S.
  2     Environmental Protection Agency (U.S. EPA,  1985) and the International Agency for Research
  3     on Cancer (IARC, 1986). Data from both in vitro and in vivo studies on the toxic effects of 1,3-
  4     butadiene have established that 1,3-butadiene metabolites, not the parent compound, cause these
  5     toxic effects. Differences have been noted in the toxic responses to 1,3-butadiene among
  6     laboratory species, and understanding the pharmacokinetics of 1,3-butadiene and its metabolites
  7     is important in assessing the carcinogenic risk and evaluating other health effects associated with
  8     exposure to this chemical.  This chapter summarizes the recent research that has provided
  9     information on the pharmacokinetics of 1,3-butadiene in several animal species and elucidates
10     the metabolism of 1,3-butadiene, via both in vitro and in vivo studies.
1 1             The chemical terminology and units  used in the publications reviewed in this chapter
1 2     have been standardized for consistency. Epoxybutene (EB) is used for 1,3-butadiene
13     monoepoxide, 1,3-butadiene monoxide, l,2-epoxybutene-3,  l,2-epoxy-3-butene, vinyl oxirane,
14     and 3,4-epoxy-l-butene; diepoxybutane (DEB) is used for l,2:3,4-diepoxybutane; and butene
1 5     diol (BD) is used for l,2-dihydroxybut-3-ene and 3-butene-1,2-diol.   ,

        3.1.  OVERVIEW OF PHARMACOKINETIC STUDIES
               In recent years, considerable data have been generated regarding the pharmacokinetics of
1 8     1,3-butadiene in various laboratory species.  Although in vitro studies can elucidate possible
1 9     metabolic products and allow measurements of metabolic reaction kinetic constants under
20     controlled conditions, in vivo studies usually encompass several issues of pharmacokinetics and
21      provide an account of the total disposition of the  exposed dose. For 1,3-butadiene, because of
22     the toxicity of the 1,3-butadiene metabolites, in vivo pharmacokinetic studies validated the
2 3     existence of these metabolites  and their metabolic rates of activation and detoxification.
24     Absorption of the parent compound was often assessed either from its distribution in the tissue
25     organs or blood or from its excretion in urine, feces, and exhaled air.  Absorption and excretion
26     have also been measured from the presence  of 1,3-butadiene metabolites in blood, urine, feces,
27     and exhaled air. Species differences have been observed in the toxic effects of 1,3-butadiene in
28     mice, rats, and monkeys and are reflected in the in vitro metabolism and pharmacokinetics of
29     1,3-butadiene in these species.  This section summarizes the metabolic pathways of 1,3-butadiene
30     disposition and the species differences in 1,3-butadiene pharmacokinetics and metabolism from
31      in vitro and in vivo studies.
        1/28/98                                    3-1        DRAFT-DO NOT CITE OR QUOTE
-------
 1      3.1.1. Pathways Elucidation
 2             Several in vitro and in vivo studies have elucidated the metabolic pathways of 1,3-
 3      butadiene metabolism as shown in Figure 3-1 and summarized in Table 3-1 (Himmelstein et al.,
 4      1997). Results from in vitro studies show that 1,3-butadiene undergoes cytochrome P-450-
 5      mediated biotransformation to the reactive metabolite epoxybutene, which has also been
 6      validated from in vivo studies in rats, mice, and monkeys. Epoxybutene can be activated further
 7      to another reactive metabolite, diepoxybutane, or detoxified by epoxide hydrolase to butene diol,
 8      as shown by in vitro studies and detected in vivo via their glutathione (GSH) conjugates in rats,
 9      mice, hamsters, monkeys, and humans.  Further metabolism of these two metabolites can be
1 0      mediated by either the P-450 system or epoxide hydrolase, giving l,2-dihydroxy-3,4-
11      epoxybutane. The detoxification of epoxybutene occurs by hydrolysis and GSH conjugation and
12      is mediated by the enzymes epoxide hydrolase and glutathione S-transferase (GST), respectively;
1 3      these reactions have been supported by both in vitro and in vivo studies. Epoxybutene can also
14      form DNA and hemoglobin (Hb) adducts in both rats and mice. Of greater  significance is the
1 5      identification of crotonaldehyde, a DNA-reactive chemical and known mutagen, as a new
1 6      product of the oxidative metabolism of butadiene. Crotonaldehyde was formed by the
17      tautomerization of 3-butenal formed by chloroperoxidase-dependent oxidation of 1,3-butadiene
1 8      and was not a metabolic product of epoxybutene. 3-Butanal rapidly tautomerized to
19      crotonaldehyde at room temperature, which may explain its nondetection in in vitro studies.  A
20      possible pathway for the metabolism of 3 -butene-1,2-diol, a secondary metabolite of 1,3 -
21      butadiene, is oxidative dehydrogenation catalyzed by alcohol dehydrogenase.  The production of
22      GSH-epoxide conjugates, 5'-(2-hydroxy-3-buten-l-yl)glutathione (compound I) and S-(l-
23      hydroxy-3-buten-2-yl)glutathione (compound II), was confirmed using human placenta! GST.
24      While compound II is chemically stable, compound I tautomerizes to a stable sulfrane.  Because
25      these compounds are of low reactivity (including the stable sulfrane), this biotransformation
26      pathway may represent  a physiological protective mechanism against the DNA reactivity of
27      epoxybutene.

28      3.1.2. Species Differences
29      3.1.2.1. In Vitro Metabolism
30             Species differences for several reactions described in the previous section are shown by
31      measuring their in vitro reaction rates using microsomal and cytosolic preparations from several
32      organs.  Himmelstein et al. (1997) gives a comprehensive summary of the in vitro methodology
33      and the studies that measure the reaction rates of the reactions included in the metabolic
34     pathways shown in Figure 3-1.  Table 3-2 summarizes the reaction rates and rate constants
3 5      obtained from the main studies that compare these differences (modified from Himmelstein et
        1/28/98                                    3-2        DRAFT-DO NOT CITE OR QUOTE
-------
^
i
o
o

:c
x C
	 X 1
=C 2
\ 0
^ "^


A>,
. -o
	 	 	 3
S, 35
0
O
N\
\
\ -S
11-
CD
•o
<0
j §
/ 2
/ °




/
-^

ll-i*'
im
ico
:
(
; <


j , • • •
\S :
S\
/
! /
 5?
o * 	 """•
0
1
CO
CD
A *
/Im
P co/^
^— ' v_ -•/
/ ®
/— • "O
/DO 'x
, ui co o
( A=--o Q.
-^ r\ o 'x t
0Xf
=C 5
+ CO
d -'
0

I
o



D=



O
x


A
o~



31
o-

/
/
/
[


*^^ 	

CD
,j-tsuiaaiene monoxia
,2-Epoxybutene-3
««, ,1 „,,:„„_
r- T- ^








Y
D<










a}
c
o
c
X CO
\ a-
> e

&
\ e
> -o
/ ^
/ £
Q

CO

L


	 N.
§ |
x r 2
L0 =5 5
* k
I o 2
c^l.
cu tu ^
= 1?
[U ^ CM
	 y to na T-"
^ o .
\ m
^

f
i o-
<
p0 I1-
r o Q
/ o.^ -
/ 0 ^
/ 5g
4-2
p
*H. ^
^ i-^DQ


CD
"a
'x
ales:
Butadiene
Butadiene monoepc
Butadiene diepoxidi
$z
j§ -^ CN CO ^
3
W


CO
-2 •§
"o to
1 =
? 1
< o
0 U
yj
^ /^^^**\
o« X -
^ xco
^ WQ-
J: VZ.
Q
\ t-
\ 3 .1
..^ 	 ~. ts ®
/ X co co£=
( \ SS, CD 0
en V -Q
Ls < S
^ 	 Jl§§






C
c.
A
CD
1 £
to
•111 o-
•*. 3 _.
X «?•? «=
LO ^- -» /
s i, i^-r /
•p g- uiin Jt /
ta °-f
QLJ -j-
oi^. LO
y T-CO ^

\
\
s\ ~ • •
^-\ -2-
^ ^
Co
o>
3
§

5T :c
" CO
— o


E CD
 B «
4 | =5 g co i
•?• || Is , 1
S>CD o§ £2w5
*!« 11 Pi!
i§t
^4 gotiH-cSdl
s . ...
i- in J5>< CO O Q
c
LU
























CM
)
k.

-O J2
f
LU
'





















                                                                        c -S
                                                                        g C8 "S
                                                                                         •o
                                                                                         B
                           _c

                           •o
                                                                                I g
                                                                                3 5 '
                                                                                j3 ,g ^
                                                                       cc ••*  ^s v ^ * ^3
                                                                                    *f§
                                                                                    y  - «
                                                                                  g •«
                                                                       •2 O
                                                                       u
                                                                       es m
                                                                       ja
                                                                       «
                                                                         g TS  c

                                                                         ® 3 .2
                                                                               - 5 P

                                                                                St
                                                                                O) fiU
                                                                                en *^  _i
                                                                                d) o  w

                                                                                £ s ^
                                                                                e p >«
                                                                            =  c
                                                                            S  o
                                                                                   • y L« ^

                                                                                    ~ <2 g.
                                                                                      3 —
                                                                                      .Xj CA

                                                                                      c o
                     bi f|I
                     " *i o u e
                       o3 .2 «•? eu
                             CM
                                                                                  g '=
                                                                         « £ •=»
                                                                         I -T -S
                     & a i; = «
                     ^ « •" .2 £
                     •• _*^ T?*- *f **
                           S g>
                                                                       ' £ >> S
                                                                       dj ^Q o


                                                                       s flj ^5

                                                                       3 « S

                                                                       ZVB-3.
                     5 -2

                     L.
                                                                           « S


                                                                           «'l
   o  S o •o


^1^11
i  o  i| -g

^ -fii TT fii ••••
   W ^-f* *w cu
oT  W) *? u L.
C  3  >> O «

I I 2 -S s
•4^  U ~
6S   ^^
                                                                                s S^ ^
                                                                                o N£>J-
                                                                             ^;">, »
                         - cs •"

                         * 'I 1

                           ?-
                           o o
                           E C
                                                                                             C

                                                                                            •8
                                                                                             o
                                                                                                      «

                                                                                                     •s
                                                                                            c
                                                                                            a\
                                                                                      g .2 =
                                                                                       -C O
                                                                             2 &m -2
                                                                                            >3
                                                                         rt  - _!. ,i  U  3
                                                                                 £.§!§&_?
                                                                                     T> . U «
                                                                                      ' p g -g

                                                                                     ills
                                                                         .  .  '
                                                                       tL. CS
                                                                                     •w 3 0>
1/28/98
                                                      3-3
DRAFT-DO NOT CITE OR QUOTE
-------
o
A
«

•s
e
TJ



I

JS


C<>
to
A

£
Re
and tissue
Spec
CQ    co
Eb£}    r«*
t    5!

s   -I



«    s
0?
t-~.
CS
1 — t
'c3
^_i
CJ
.s
"co
I
C3
s
en
CJ
1
Wistar rat liver micros
CO
oo
C\
"re
•8
•4->
"o
m
8
SD rat liver microsom
e; SD and Wistar rat, rhesus (Ic) Schmidt and Looser (1985)
1) postmitochondrial lung and
(strain not stated), mice (strain (Id) Wistuba et al. (1989)
3 " -i
B6C3F,andNMRIm
monkey, and human (t
liver fractions.
Liver microsomes of r
rp
n
fH
not stated), and humai
T— I
C3
•s
ca
*— i
1
cn
|
B6C3F! mouse liver n
3 1
>. u
"tJ !-H
~C3 O
C3 *-"
co 2
t — 1 f-H
_W
! 1
S" M
n 3
^^c do
1 |
^ ^
CC C/3
^ P
ai O
SD rat, B6C3FJ mousi
lung microsomes.
SDratandB6C3F1m
microsomes.
cs
a\
OS
1
i — i
C3
-------





2 (_y{
• S ^
J^ &
K ^>
.2 &
"i ^
<*5t <~]
a; o
> (W
"5? o
*•* r— 4
^J" Vl ' r— <
^-^ /— x ON ON Is =8
re / — > S" c^ , — , ON ^j .-i^
vo c^- ^ ' — ' »o ' — ' ^
/— ^ °° °°2 ,—*. ^"'cN^^'tinSP
G^ ON ON ^ \& ,—>, — ; ON — < c g
oo r^- "—' • oo o « T-H cs -a -a
2 .^-^ s ' §; tj ^ ts s 1
c. . *a . -a -B °c,g "a .35 i
•a " ^ &. -a «• -§ « s ^ i
* - • a* ' 2 t * 5 1 ' 1 IS 1
*-> S IB -53 T-S^HiS PC! H
"3 -s :s'B ssg-liss
£ ' s £ s •'£ s a ! !• ss eg
S '^^ ^^^^ /^~^1 /*^s* s~-** ^"""N /"""N /^^\ O^ «^~N Q\
^ S O jg, c^ ^ w .<^> ^ > ON ^ c\
^^ ^3* di' dx t"~i T™~1 r~* i~K I™H i~^ *"~*
1/-1  OT
f-o ^.^^ 1J5 1>
XJ T3 O -* ;3
w o *o S cirtw
11 Is 4 ?| i i ! i *
11 1? 1 Jl ill 5 i
I! li 1 1|l t I i s s %
a -« ^S -i-^l o.^S ^S §8 ^
l! It 1 111 . 3 11 1 11 !
!| || s l||| s | a 3 s g| g
-M+2 oo ^^^S'^^mu « r' -
rere ^ oooo^ opq 0 g g o^S cnp
i->
1
o
1
1/28/98
3-5
DRAFT-DO NOT CITE OR QUOTE
-------




















/"—s
*^3
^5
2w
*^j

o
u
£
C/3
"o
.Q
«
•4*'
CU
S
CU
CU
1,3-butadi
o
!£>
«
Table 3-1. Metabolic pathw





Cl
g
i















3-butenal
In vitro experiments
EB and crotonaldehyde in reaction mis
Proposed 3-butenal as intermediate in
formation of crotonaldehyde.
3-Butenal detected at optimal pH of 6.



c^
ON

^
«
§
1

'c?
rn,

"§
o
4->

_-j-J
"o
42
CO

ON ON
ON ON
^^ CO v—- '
co g^ ca
r-^ rx^ S i — ' 3
" CN CO v 	 ^ ,CO
wN ON MH ^^ S-M
c, c, 3 "1 §
t3 *"' a3 ^ a3
2 S "S SP "S
C C en C co
«S 8 3^3
W CO Q O Q
'c? 2^ /cj' *O 'a?
S- S S^ S^ 2^
CO ^
'"« °
-S 5
_j 3
co (D
!>i 3
C* §
*§ S
' O 5™*
i— H • rH
Mn *^3
1 1
U' In" "S3 flj • (D
§ J J | § |
* "== J3* * * /S.

G g 2| O B G-
w ^^ p4 www
cs ' w o K! w KJ
«-* cc o c< fl cj
*G o w t-i ?— i i— t
U ^ Qj U O O
OH fi ^ Q^ QH OH
o m o 4-* -------





















(continued)
.,3-butadiene metabolism
T"™1
(«•
0
Cfl
es
•*•»
es
a.
u
"o
R
•s
^
TH
f)

V
2
H








cu
u
s
^
cu
Cu












en
*3
•O
«
.Si
'3
&
CO






c
u
R!










/^N
ON
ON
G
g
t~j
w
s
TO
1
C/3

D
Q
C^
^


































&


































cLU
^
>
.*-<
•s
•t-(
Q
&o
1
cT
1
S
frT
m
O
vo
ra
M3
II
^
(3
^
vJ-H


















c3
g
3
J3
1
1
•8
en

>o
m
CN
vs"
<;
CN

Cr-'T ^* -
I~, s
^D ^^



^_^
CN
ON
ON
v
O
>.
c

en

'c?
^
TS
C
c? S
' — ' >->
II ^^
G
53 n
1 '1
SD rat, B6C3FJ mouse, an«
lung (n - 5) microsomes. 1








































mouse liver microsomes.


















                                                   
-------














s~^
"O
OJ
s
.S
*L*I
,3-butadiene metabolism (con
T-l
O
C/J
J
Is
B.
f .
• M
"o
.0
C5
CU
s
*
*"?
CO
3

^




cu
CJ

.a
«
oa
Q
Same experiments as (Iv) above.





03
•>«
i
•g
&,
S

•S
^
Ji
•S
a.
'S
^











*o
C3N
^
liver, higher in mice
and lung, thymus, fat, spleen and
than in rats.

































§
f
-0
0
J3
3
.s
m
%
"3
"u
c5
'H
CS
^


"S
s / 	 ^

£ VI
^ ON
^0
iale only), higher in
Same experiments as (l\v) above
marrow, fat, lung, mammary (fen
females than in males.







































^
x— s ON
00 ^
ON _;
^— t cS

i— i "S
w S3
t3 N
in §
o H
ffl M
/ — s ' — v
& &
>
1— H
II
viS
Same experiment as (Ib) above.
SD rat, NMRI mouse, and liumaE
microsomes and eytosol.








CO
•fr*
wo -2
.5, §
§ .§
pa
ffi S,
o "
T I
a s
S "^


1 — I
ON
ON
r— t

'TS
•4-^

S3
S
r-<
OO

^
GSH S-transferase.
Purified human placental TC class





















CN
ON
ON
"— '
"3

0)
g
SS

oo
, — ^
^
•o
1
1
J
u
i
m
Q
oo

































d
C3
4-J
1
gn
1

































I
C?
i — i
1!
^S,
fe
"i
SD rat, 8603?! mouse, and hums




















CN
ON
ON
^
•s
tt
OJ
"O
^cs
§
CO
o

£
>sol. Same experiment
1
•s
CQ
CO
ll
co O
0 -g
b! ra
"S ^
B vo
£T^e
" g
^Sc?
(WCl-
§ CO
^ CB



















1/28/98
3-8
DRAFT-DO NOT CITE OR QUOTE
-------



















/— v

S
_a
,3-butadiene metabolism (conl
o
&
{?
1
Table 3-1. Metabolic p




0)
u

Qi
d
•C
§
x>
CO
&O
§

to °°

5 *0
s °

tin ^
(Tt t-*
Inhalation exposure of EB to SD i
B6C3FJ mice, Syrian hamsters, ai
monkeys. Urinary metabolites.

i .
^ i
cA ^
If
ti
^ j<;
CN g
i4^ -
Uj i1 fl>
*TD ^s] j^

c$

ON
¥
t5

pr;*]
.
yj
crt

g
^
r^
00
g
1
1

,, w
i s

s S
o ^^
i.p. injection of EB to SD rats. Is<
regioisomeric GSH conjugates in

o
a
|
i
CN,
I




ON
ON
^y

CO
_!_>
ID
CO
& .
tS

P3
g
&
S
S

|
+3
•8
is1
SD rats and B6C3Fj mice, by i.p.
metabolites.
t
. S
i i
*>L ^
CC "g
Jl.
rn V' .S
^T3 1"
o fi g1
^«x C/3 Q>
Co o co




VO
ON
ON
.
co

O

3
j:
p
CO

§
CO
co
•^
3
OT -

K5
r-^s
\S-l
8
Inhalation exposure of rats and m:
(3 a) above. Urinary metabolites.
BD
In vitro experiments
S£-



ro
oo
~|
•^^/

- j

ID
T— H

CQ
g
r-^
f^^ '^S
2 ^
-2 ^g?*
ie c^
^ en
_g co
^ g
SD rat liver microsomes treated w
epoxide hydrolase. Same experim
above.









T-H
ON
ON
" — '
, 	 ;
CO
-i->
O
53
N
CD

M '
oo


^
^
• T-4
T-H
II
3,
SD rat, NMRI mouse, and human
microsomes.






1/28/98
3-9
DRAFT-DO NOT CITE OR QUOTE
-------
      u

      g
          cN
          ON
          ON
            m
            ON
            ON
                 03
 •a
 
0
"re
g

re
•g
§
Same experi
conjugates'.
l,2-dihydroxy-4-(A^-acetylcysteinyl)-butane.
(BD necessary intermediate of this product.)


ON ON
OS ON
r— < rt
re "re
1o t>
13 co
1 1


m z
tlQ _ <~j
oo oo
§
re
o
0
o
.*
CO
§

^ i
"2 2

CO TO
O* • ^H"
^| ?

CO ^^ /-^
r \ ^O Kt
VO S, $S
ffl g CO
g™ >,
TO To
Ii Q &
1-7- l-t ^->
Z 3 co
1 s S
•^r 0^ c
en b? re
EJ— O OO
Human urinary clearance predominantly via
epoxide-hydrolase-mediated hydrolysis.





























) BD->GSH conjugates
In vivo inhalation experiments: BD-GSH
2^
CN
ON
ON
i — 1
13
•s
f^
'§
o
_o
tS
oo
'o?
ON,

^ -
« (D
_, CO
C 3
DI
§"§
jg re
CCt ^(
	 s
lT> 4_>
e 1
T3 cS
p? *^
TO ^_p
CH^ S
OT Q
TO V^H
^2 o
S S
S .S
Same experi
metabolites :
conjugates:
l,2-dihydroxy-4-(//-acetylcysteinyl)-butane.


ON .
ON
r—H
"re
ID
*^
1
o
0)
ffl
S"
Os,
B
*o
s
u
a
|
l~~^
i
o

03
"so
oo
CO
re
CO
1
Same experi
in humans.
l,2-dihydroxy-4-(Ar-acetylcysteinyl)-butane.


ON
ON
t — i
•re'
•s
CO
1
3
OS

•^
€^

>,
§
• «H
5
i
t'j"ij
oo

1^
A
5

^_n
{§,
CO
re
Same study
metabolites.
A^-aceryl-5'-(3,4-dihydroxybutyl)-L-cysteine
in mouse and rat, JV-acetyl-iS'-(l-
(hydroxymethyl)-3,4-dihydroxypropyl)-L-
cysteine in mouse but not in rat.

1/28/98
                                      3-10
DRAFT-DO NOT CITE OR QUOTE
-------























"O
CU
(^
adiene metabolism (contii
•^
,Q
CM
O
en
«
JS
-4-»
es
E.
Table 3-1. Metabolic]



tu
CJ
C
CU
sS

-3
P-I















0)
en
_cn
•o
CS
en
"D







Reaction


,—,
ON
ON
»— t

•3
"tv

S
CS
rf3
§
¥>
'cs'
O
I— H






^
O
•s
'c?
ON
^*^-S
1
S1
oo
g
'cSS
C-
cn
cs
1
CO
CD
1
CS
00




o>
C
(10) BD->l,3-dihydroxypropano
In vivo inhalation experiments
rs
O ^~s
 S
M JH
"^ O
'c? oo 2^
-—I ON r-*
^-s
VD,

en
CO
^j (L>
CJ S
S rs*>
'S
0, CO
vT 0)
^ g .
ar rat liver microsomes. Same
e.
strain not stated) liver microso
riments as (4d) and (8e) above.
II *&
^ "CO (X $




„,
o
(11) BD->3,4-epoxy-l,2-butanedi
In vitro experiments
Co*-"
ON
ON
i — i
1
OQ .
"S
CS
^O
CS
CS

O
PQ
'c?
CN
r— <

*73
C
es
.^

3 C3
K.3




!l
•o
(12) DEB-+3,4-epoxy-l,2-butane
In vitro experiments


G^
ON vST"
/ 	 ^ ON ON
ON X ' ^,
ON i^g
*-_ -* i * ^j
(D +.J
*c3 1"O ^
•*-* cc £2
o ^ 2
^* CuO _ci
-S S P
j^ O ca
E""1 p"l *2II
/c? S^ ^o^
mm m
C3 W

t/"""N
._ ,. ^
•rj t> o
[> Jr^ *""*
^O &•
o -P-» 'o
•g ^ d
«S ^
W |[ '^
"^* _^ Q .^-C
11 il li
!i i ti
^ S pciT ^O H fe^
^ i § -2 "g -1
1 ffi 5 °° g "
^cs tt? jr f?-i P s
3 °2 Q ^ S ji
Co CD oo S oo cs
y
••a
5J -^J
.a ^
^ 'If , ""
te ^ u "^ §
55 ^~> &J CO G
i i i|i 3 11
1 1 ..111 111
o t^t! * ^°? ^
T ej 's _&> -2 -0 y -i ^ H_>
a 't '|»^^ ° -| | % -S
r-l
1/28/98
3-11     DRAFT-DO NOT CITE OR QUOTE
-------























*^3
cu
_e

•*•>
a
o
^
IT
"o
td
es
•4^
CU
CU
•S
•*-«
S
i


^^
C.-I
O
e/j
C5
«i_i
Table 3-1. Metabolic pai







a>
1
0)
p5














S
en

^-*
•a
CS
.§
a,
cc














Reaction




^_^
vo
ON
C\
*ra
•s
CO
i
i
^
r— <




O
CO
T— *
-a'
C3
/cs
O
^
y^"^
o

3
CO
^
0
'c?
en
-&
9
CO
g >
« o
cc o
OQ CO








(14) DEB->l,3-dihydroxypropane
In vivo inhalation experiments
bo
e
'S
0
a,
n,
3
03
)-t
B
CS
-2
^—i
ublished
eaction.
&< M
% 3
































B
(15)3,4-Epoxy-l,2-butanedioI->GS
conjugates
M
Co C
ON '-g
o\ o

^^^ o i
•^ 3
S en
O 
w ' — ' ti
CC r^ •*•"<
i II
ffl •§ g
/c5s P-< >H
^2 ° is









to
-^
S
S
o
£L
Q
'a
1
CO
f^t
g
I
1
g1
fs
S
ffi


o
• •M
--
•s
£"
I •
(16) l,2-Dihydroxy-3,4-epoxybutan
7n vz/ro experiments
In vivo experiments
OD
a
••e
0
OH

pj
S
1
-S
'r=<
ublished
eaction.
a, ^
^ J

































(17) ErythritoI->CO2




n
m
w
Q
i !|
^.S^
"^ -H*
i g
If
s"l
.§ 2
1*0
•5: oo
tf b"
c c
o o
"3 S
•*? 43
°? 2
li S
rti ij_i ,.
<^ c| 3


S -r S
O O ^
CH C «J
"O "s D-
03 ^ O
•*—• ii t—i
3 cn OH
-? Ix OH
2s 5
iT ^ "o

•— • 3 ^
o "° tS
CL, &* -, s
S .C g
S ^ ciS
.1^ ?. s;
Abbreviations: EB = epoxybutene, 1,3-butad
diepoxybutane, l,2:3,4-diepoxybutane; BD = '.
intraperitoneal.
a Numbers and letters in parentheses preceding
Source: Modified from Himmelstein et al., 19
1/28/98
3-12
DRAFT-DO NOT CITE OR QUOTE
-------

to
a
, and hydrolysis reactio
a
CS
W>
3


O
cj
a
S
a"
o


es
."2
O
a.
.
s
««
 es
« "S
S xj
0 *t
a '—'
0 <<-
CB O
II
S "o
O ^2
v «i
M (U
« a
O **
G< -C
1X3 "^
• .S

r^ So
^
CJ — <
X! g
H fi
L ' •••*


g
0
Reference
(exposure concentrati


^•"\
S ^J
^^ **-<

^—^







«
03
!>'
ts .-
C8 0
4j




.C
'3
cc




tn
'S
Q-
!/3



CO
5
en
en
••*3
T3
g
es
C
o
u
C3



All studies with species
differences data








_



oo

cK •* i  o?" S
2 i a I;
•a-g °§
" & 2 °-
^> o u >o "
:§ o fe ^ ^
S °" 1 § S
0^ £§.;§
__ O x— x O CB
C^O bJOro -tt
T—^ \O, r-H CO ^



^^ I-?
. C*1- r— (
^ m' to"








vo ^ ! s -i- o ! !
cs 0- j ^ vo c^ j |







f£T p£f
CO ^^ CO ^^
ffl ffl




W r^l & {/} *v^ KJ
Is j J is!!
SpiiSffi Se^Sffi














^
ON
ON
(Ij) Duescher and Elfarra (J
(30,000-660,000 ppm 1,3-
butadiene)



0 0
VO (N
i— i i — i








ON CN 1







Et-T
m f^
• r i *— 1
vo °°
m




» &
111



















o
o

1
o
o
CN





OO
-------





e
o
'•§
«

O
S-i
TS
T3
e
ct

B
.2
ts
3
"s5
O
u
w
C/3
O
.2 -*-.
CS cU
"0 3
• M* _*
x .S
o ^£3
c. c
ou o
^* v^^
s2 CU
•*2 CU
e — •

•4-1 C3
CO *+*•*
S 3
O 43
O i
0) ^
« £
*" O
•§ s
« .2
•M A
*"* *cu
2 S
_" ^
.t; m
> Js
c
t— i c
CS "gj
•2J

l-s
u
'g^
o
^>^
rt
leference
e concentr
in
O
n,
1














«..
5
 "
<5 CM "s ' — ' ' — '
H §S $1 f|
5| 2^ £o S,o
•2 ro * S — ' S — ' S
fl" 4->P, C3 S +-> o" *J 0"
^ g ,§>0 ^ 0 « Vg
H 1| 111 111
rr^ ^ f "^ • C/5 ^^ i£4 C*O ^ i^*
txJ — . ^; i ^J <\ ^Q **-* #\ "^5
--- °« O x— v O c3 ^— x O cS
c> o CH^ o bo co *2 tiD m -a
T-H ro *"H ^«O ^"^ CO Er """^ co 5
v_,v_/ ^-^ ^^ ^^-


O C"*- •
^ (>• 01







®^ »— <
iri § ! t^. s !2 "-: ^ ! °°- *°.
"*. VO 1 Ol O O ^ ^^ 1 O)
*• o





H- 1
^^j |n-(
~*Z> K5 t , ,_|
r. *> ^^ Q ^^ Q *^ Q

CO Q £Q pT) pQ

vo ^^
pfi
,-M


(U S flj S flj C^ O
III 111 III ll

CO
O
CO
§
's
&•
1




X— N
C^
T-H
Ig
o ^
C4 ^
lo
o o
oo o^
*^3 i
VO *O

0
oo
i— < »O OO
S 2 4

CO






^;
i i — i i — I
__• "t oo
o m
' o







.-
o Q
vo ^

CQ






o d
co 52
ill


§3
S §
H co
Q a
.. o
« 1
a s
1/28/98
3-14
DRAFT-DO NOT CITE OR QUOTE
-------


s
o
"•3
es
.2
Kl
"o
•o
t3
S
cs
^
.0
CC

•1
o
u
w
5C
B"

"-C S^
CS cu
•o s
">5 S3
0 +S
c. c
 C
a •-
-M CS
CB **
S S
O ^2
CJ 1
«5
*" o
•§ s
es £
ee ^g
C^
1 !
•*" A

^^ ••*
Table 3-2. ]
involved ir
°^
o
1
4> "S
a S
1 s
S o
*S u
CH •_
VI
O
n<
X






t"""

- — •




"
I
fi
.2
"cj
CS
*j
^H




.S
"S
CO




*3
o>
a,
CO


0)
s
_23

Reaction and


01
ON
ON
i — i

"3 ^
'S PQ
S •?
I*
^0
^T
"— 'v— '




o o
0 O
1 t, ' IN
co co





§£









-«
PU
o 0
to °°
CQ




00 §
P ^ g
tn
CS
CUD
"I"
O
u
ffi -H
CO O
o s
A *
W fe


oT
S a
^ o\
& <^T^
t> w "« ^
>> g "S "
If Is
'ai'o ^3" T
x-'s-' "-"^



o o o
o o o o o
CO OO ^^"- f^ ^^
i^-j cY-p c^r * i, ' ^
co <— i i — i en co





o i— i '-; _ , !
<=> •* «D — CO
iri O4 ^j-









— —
fT , ^T l
CO Q CO /-\
OQ OQ




3 *j S '3 ^ g






'o
C/3
2
S4
S1
5


oT
1 I I
^H r— H
'cQ /vs^ ^~*^ ^""^
t> W "cS ,-v • 13 ^
>% fl t3 ® 'S OC
|g | g |1
^ o ^T1 ^T1
^CS £^£5 t-G



o o
o o o o o o
*O ^t" O C> ^3 O>
vo" t>r *-<„ ""i, •— *„ •— i
CO i— i CO CO CO^ CO*



?
o
rr) CN ' — <
c< _,: . \o ^ o—<
O4 -3- vn r™^ CO ^O
^O
04








{T t fr ! ^r !
CO Q CO Q CO Q
« ffl ffl




o S o d>
ill §^ ita






en *o
i. S
t : *••
6* w
H "*n
*t3 co
• r-< U
W H
1/28/98
3-15  DRAFT-DO NOT CITE OR QUOTE
-------



JO
C
.2
•M
1
.—
"t/3
!•••«
O
"1
e
0
'•§
W5
S
"c5
o
0
w
CO
O
.2 /-s
"es "2
*t3 S
*S jj
O ^£3
Q, C






J1





a .
on rate
'•§
rt
o>
^"<


C!
1
4^
CO
.2
"0
c.
CO




3
en
en
•4^
•0
Reaction an
o^
"g"
1
1
§
exposure c
^•w



JS
i



























c?
ON
e3N
^
"ra
•8
C3
OO

'T3
^<

O
0
O^
o"




^
o
o *"*
o
o
CN








CS
1



ffi
00
O
T— 1

1
« 1 °% Q
^1 | ?
"§ 1 J ca
S g g W
rj j-4 o
= ~ 0*0?
OH OO OO Ii>


1 — 1
ON
ON
^,
'a
Kreuzer et

'o'
oo

o
0






ON






S
*
(D
1





CO
S2
1
Liver micros


w
£
a
o
o
ro"
o"
o
m

O
m

o o
o o
t> V)






t— H 1— t






P
oo

a
+j S











ON
ON
T— 1

"3
Csanady et

rQ
OO

O
ON






ON






ftT
CO
O
ffl
(D
C/3
1













S"
w
1
a,
o
o
CN
t
o
CS
o
r— 1
O O
CN CN





'~|
oo oo"
CN CN
ON






P
OO

ro
CO H
OH ffi











CN
Ov
ON
v — '
3 m
s w
^S
•§ a,
C CM
CO ,_,
3|

'O ^^
ss ci





CO
o
C^ C^ *
VO CN !>."
oo m '
T-i rt ~J
tr!





co r>
^°

(D g
CO ro
J "8 3
2 P^ ffi





"~cn
O
S
Lung micros


CN
ON
ON
v — '
'cO
Csanady et

T3
OO






f
O









CO
g
sS







on
CO
1
•3
o
GO


























en
-s
T3
O
CU
*K
""S
!M
•8

S
T— (
ON
ON
r— I
t3
g
^.j pq
g§
*H 5
co i«S
CO P
00 «
o in
O r-l
« ^

CN "~i
•— < O

O O O
o ^o o
•— < t~- oo
oo" CN" •*'






O ON VQ
CN CN JQ






CO Oi
o 55
m
0 S3
en ^3





en
22
1
Liver micros
1/28/98
3-16  DRAFT-DO NOT CITE OR QUOTE
-------
\





•>
1
13

»S
T3
8
es
Q
^
^^
CS
bJD
S
's5

o
u
w
1/3
O
a"
•+•* >**^
Gt H3
s S
o-s
— "S
aj «
O
fa U
«2 ^
«*- a)
j£ a
^ «
3?
s 1
s|
« fO
•s i-j
fa CM
73 e
a g
es vi
a? ""*
* -s
i« "
sl
= |
to «
J8 •—
pfi ®

H.S





4)
U

i"*^







g
^u
p*n










4B
S
o
"•£3
es
&






g
*«



tn
1
a.
oo


 •<*•
i>T c-T CN" vo1







m m o- cvj
Os' o\ i— < ^^









^
fT j
m
O
VO
m




CO £3 W
^; P^ ffi tn <§
"S
6X1
'5s
o
u
OT ffi
1> ***
aW)
0 &
2 ' »
^3 • -PQ P
S g' ^
M) !rt
| £T ^
T ^H "*"3
>— ( S^' >-4







^^
CQ
pr"]
Q
. . .
S
«
cs
2




o
o
0
^








^
oo











Q
oo




1c
pi(







Cr^ Co^
Os Os
Os Os
i— i i— i
— J ^4
_2 ^-^ _2 x-v
(D pQ (D pQ
"O W 'O W
c3 f! cS Ci
cQ • t c3 v~i
0°'l o°e
O S O S
^(N ^ cQ
2 ^ 2 ^
^ S*~^' S"— '




O O O
o o o
i— ^ t>^ CN^
CS *~~* "^d"






en
0
vo ' =° ^ ^
__;















2 tn
S s ij
i! S OH


1 1

1 1




g
3
'S

_
§
c^3
Ui
,2,
^§
§ 1
11
O KX
^ bO
§1
11
*C *^
*S o
"o °

.c!
s ^
M x;

^^ ?j
° §
.2 g
CS -S
I ^
S ^
S §5
5P c
r J3
"s J
S ^


o c>
s ^
G o
^o "^L
M O
."t3 i— i
c o
3 • — i -
<-s « '
s s
-S 0
• g s
o •— *>
•^1 2 2
CO &0 g
t; .'g §
g< S g
S -3 §
^S S *x
on 'rates are either from
n rate (V,^ in units of n
entration atone-half ma

111









^ ^
i VH
CTJ ^—^
3 S
H "~1
•S 'i"
•O 3
U 0
•a S
§ VO
g ^
1 ^
1 S
^
<^ C3
CO
o 5?
s "S
.SP ^
fe «
• S &

fe fa"
o *S
•§:§
^-» fa^<
J 1
|-|R
^ '1 S
wp G rH
t> g5^"
«•? "S
"c3 -S ^>


O * nj
QH"O «
On g
G ^>
Hi
a> 3 o
*-< ta d<
a> M 
-------
  1      al., 1997).  In general, the range of reaction rates or maximal reaction velocity (Vmax) for different
  2      reactions in different tissues do not provide a clear pattern of species differences; in particular,
  3      the range of values for human tissues spans the range of values for both rats and mice. However,
  4      within any single study, for the oxidation of 1,3-butadiene to epoxybutene, the reaction rates of
  5      liver and lung microsomes are higher in rats than in mice. Multiple cytochrome P-450 enzymes
  6      are involved in the metabolism of 1,3-butadiene. For example, in human liver microsomes, the
  7      metabolic oxidation of 1,3-butadiene to epoxybutene is principally mediated by P-450
  8      isoenzymes 2A6 and 2E1. Biotransformation of 1,3-butadiene to the non-DNA-reactive butene
  9      diol is the predominant pathway observed in in vitro metabolism studies that used hepatic
10      microsomes from rats and humans, and formation of the DNA-reactive diepoxybutane is
11      relatively minor in these species. However, the latter pathway is significant in mouse hepatic
12      microsomes.  1,3-Butadiene can also be metabolized to epoxybutene by human myeloperoxidase
1 3      and by mouse and human bone marrow cells.
14            In the Csanady et al. (1992) study, the authors also extrapolated the kinetic constants
1 5      obtained from in vitro experiments to equivalent in vivo rates by adjusting the in situ protein
1 6      content  and organ weights across species, as shown in Table 3-3.  However, for GST, Kohn and
17      Melnick (1993) pointed out that the rate constants should be adjusted to the mg cytosolic
1 8      protein/g liver instead of to the mg microsomal protein/g liver as done by Csanady et al. (1992).
19      The corrected values are also included  in Table 3-3. These can all be used in pharmacokinetic
20      models as hepatic and lung metabolic clearance.

21      3.1.2.2. In Vivo PharmacoJdnetics
22             In vivo pharmacokinetic studies examine absorption, distribution, metabolism, and/or
23      elimination. Most studies report results on several of these four components.  Absorption is
24      often measured either by the distribution of 1,3-butadiene and/or its metabolites in tissue organs
25      or by the elimination of 1,3-butadiene metabolites in excreted urine, feces, and exhaled air.  In
26      vivo metabolism studies include measurements of concentration profiles of the various
27      metabolite pools after exposure to butadiene. Metabolic kinetic constants are usually calculated
28      from the rate of formation of the metabolites or from the clearance rate evaluated from excretion
29      data.  This  section summarizes the in vivo pharmacokinetic studies.  Because inhalation is the
30      principal route of exposure to 1,3-butadiene, most of the absorption data for the chemical have
31      been derived from inhalation exposure  studies. Based on the blood:air partition coefficient for
32      1,3-butadiene (0.603 in vitro;  0.645 in  vivo), the passage of 1,3-butadiene from the air into the
33      blood is by simple diffusion (Carpenter et al., 1944).
        1/28/98                                   3-18       DRAFT-DO NOT CITE OR QUOTE
-------






CS
m
jutadiene and E
•MM
AfK
«\
•^H
o
s
c
hi
OJ
13
u
• •N
•«^
c!
1
o
_c
i?
ea '>
tn C
C ""
S £
« 2
-t-* CS
tf "2
«?!
to Q.

11



02 wa
^ ^
O ^
s f
S *


Microsomal
conjugation of
EB with GSH"


•°c







OO CS Tf
c<( cs t-i
o o o
o o o
o o o


^ CS VO
o o o
o o o





"N!" C~*- ^sj"
•*' »o o

vo oo vo
i— i ^J- OO
0 O O
lO CS ON
ON 10 ON i— i
i^"j ^^ C*^* \o
>0

(D S3
I si
M 5 ^r(

~^a
fa Cd M
sH -
S E  •§
a s a ~ 1
•s-Boq^s s
§ £2 gl . g
§ §*>:'! || s
& § - & § 1 -5
S °  30
| I^oo^-B 3s
• B, 3 O O 1> "S ^ S
-B1 J3 « T-H .>; § 3«
ii^-sll 11
-^53°°^=^ «°
9 2 .§ 2 ' a ^ 1 §
__L - "^ '""^ O -j— * -+— • ' *~i
sil-f^ §1
|I|IM II
^ 5 "ea >>•= C
.^ g •§ 1 1 ^ ^8
~ i ra, s -^ s g «
en _3 -4-!1 fl t)-l E .£
n units of L/h/kg.
values were calculated from in vitro Vmaj[ (tables
[ and cytosolic protein concentrations and liver vo]
6.8, and 14.5 mg/g liver, respectively. Mouse, ra
lively. Mouse, rat, and human liver cytosolic com
mice, rats, and humans were 6.2, 5.0, and 3.1% o
i L/h/kg) were estimated by dividing the in vivo V,
cording to Kohn and Melnick, 1993.
rmic hydrolysis and reaction with glutathione, in v:
the in vivo clearance for reaction with glutathione
it for metabolism of EB to DEB.
anadyetal, 19 92; Kohn and Melnick, 1993.
•-1 a « H'-'.aoiS'uS  |s «-^O-«Ki(ul5SJpS
K>ow^!o!ujg:3""'3do 2
•^ 0 '§ 1 '& "0 §*•§ r^ t2 3 0
>£«^^>o^fr,H^ jg
1/28/98
3-19
DRAFT-DO NOT CITE OR QUOTE
-------
  1             Two main in vivo inhalation systems are used to conduct inhalation studies. The first one
  2     is the closed-system inhalation chamber, and the second one is the nose-only exposure inhalation
  3     system. These studies are reviewed by Himmelstein et al. (1997) and summarized in Tables 3-4
  4     to 3-6 for the closed inhalation chamber studies and Tables 3-7 to 3-10 for the nose-only
  5     inhalation studies.
  6            In the closed-system inhalation chamber study, rats or mice are placed in a desiccator jar
  7     chamber. Two rats or up to eight mice per experiment are exposed to different initial 1,3-
  8     butadiene chamber concentrations. Air samples from the desiccator are measured directly by gas
  9     chromatography-mass spectrometry (GC-MS) through an air valve. With the use of a two-
10     compartment pharmacokinetic model (Riser and Bolt, 1981),  shown in Figure 3-2, uptake and
11      clearance kinetic constants of 1,3-butadiene and epoxybutene can be evaluated, as shown in
12     Tables 3-5 and 3-6, which give the results of these studies. Because the metabolic elimination
13     rate constant (k^) cannot be determined accurately from the gas uptake studies, 1,3-butadiene and
14     epoxybutene were administered intraperitoneally to the mice and rats, and exhaled 1,3-butadiene
1 5     and epoxybutene concentrations were monitored in the chamber and used to evaluate kel (Bolt et
16     al., 1984). Tables 3-5 and 3-6 show that for both 1,3-butadiene and epoxybutene, uptake (k^Vj)
17     and clearance (Cltol) in mice are about twofold greater than in rats. Although the exhalation rate
1 8     constant (k21) and metabolic elimination rate constant  (kel) are comparable for 1,3-butadiene in
19     both mice and rats, mice exhaled epoxybutene about twice as much as rats (k21), whereas the
20     metabolic rate constant (k^ is about fivefold higher in rats than in mice (Laib et al., 1990).
21      Under these conditions, the steady-state epoxybutene  concentration in mice is about sixfold that
22     in rats (Melnick and Huff, 1992; Himmelstein et al., 1994).
23            A second inhalation experimental system is the nose-only exposure, where exhaled breath
24     is sampled by placing the animals in plethysmography tubes. Additional blood and tissue
25     samples can also be obtained by sacrifice of the animals after different exposure durations.
26     However, while the air samples are measured at real time, all blood  and tissue samples are
27     subjected to some time delay due to processing of the samples. These studies are summarized in
28     Table 3-7 (modified from Himmelstein et al., 1997). Table 3-8 summarizes the results of the
29     studies showing that 1,3-butadiene and its epoxide metabolites (epoxybutene and diepoxybutane)
30     have been found in blood at different inhalation exposure concentrations to 1,3-butadiene in rats,
31      mice, and monkeys.
32            Thornton-Manning et al. (1995a) also examined the disposition of epoxybutene and
33     diepoxybutane in various tissues following nose-only inhalation exposure of male Sprague-

        1/28/98                              3-20           DRAFT-DO NOT CITE OR QUOTE
-------





















t»
CJ
•3
1
e
o
1
"es
.=
JS
i.
 *r *2^ S l^-* (,-, i q^ o vjr ' ,
*"• t^ fH ""O 5^ Srt -.Vt PH ^5 ^ "*^ '"^ '"«%
r^ O * ^ ^^ ^ CC * «* C^ "c3 ^'^- OT
i^3 cj , 	 i ^ CD ccJ 'iri V ^" *™< *-!w J-^ '— * E3
dj (D (^ o o CD O C^-* hn - fliM >*^> ^ . ^C? ^_^
^"PP So^^^^^S o° -2 ^,1,^
111 sli^lijlll li ill
'S^°  c/3 ''^ Q^

^ 1 ^ g of '- 1 jf
L.J y LJ VO *~* (^ W T^j

? I *s i § j § i i-^
B*-^ 22 H2 P S !7i PI
^•* *C >^ "^ S3 "J-i 12 "tj ^5 cC
^ '^ ^ § w S w § "t^ ^^
tl tl II till







'cOS
oo

m oo r"^ "itf-
oo o\ . oo

^^ o ^_»
're ^j OB "o
•g u -S ffl
•3 2 ' 1 1
CQ ^^^ *^ j_^
x— s g 'O' 4>
- - ^ S
1/28/98
3-21     DRAFT-DO NOT CITE OR QUOTE
-------
     b£
    T3
     G

    S
 3



I
 O

^


.§
•3
 s
e
^o
*3
»
IS
,=







    I
    «5
    &4
    x
    o
    <*•
    o

    s
     u
     s
    Q
cc
erence
    •o

    a


    .2
    u
    es
        «
        a-
O
T3 CO
1 'V
\^i Cj
S &i
3 O
_o o
*O -a

C3 S *V
ll-?
"O '""' f
ill





H s^
• P-* VO
.S 0°
«aa
aS"* •„„?
^^ C3
"3 o t>
*^" ^^ W)
g 2 -1

Jf 2 t2
ure of B6C3Fi mice t(
Itration ranging from ]
ated from Figure 1 of
ll-l

W U vS











X™ %
CO
ON
r— 1
*S
"cj
.S

t2







.
"3
_o

*o
1

maximum
to

"w
CO
o
1
1
•S
en
g
o
o
1
2
1
i?
S
Vmjx = Michaelis-Menten i
g

1
3
CO
2
CL,
CO

il
Q
CO
c"
'-*

"s
ID
f"Li
•S
a
3-<
S
CA
§
•-C
R

O
1
<










'(
!e3-l.
n Table 3-
•° -a

r^ a>

•S ci
•o p
.8 1
i g
H
W r-<
T3 ro
§ e
r-i 3
' 00
^s
l-s
eo a
E ?
etabolic pathway shown in
ivolve several reactions sho
S-5
o 70
"« ^J
"C °
If

^ c
o a>
"H "3
o -o
^3 &
2 g
8 &
a <«
2 o
,O o
el
CD CL,
S 9
11 s
81
& 3
S2 S
4J n>
2 ^
•o o
i g
w S
0 0
11

a J>


































^
o\
a\
i — i
"o


u
"o
&
i
iS
a
<§
•a
u
G
"f
s

0
o
g
CO
1/28/98
                                     3-22
DRAFT-DO NOT CITE OR QUOTE
-------
   Table 3-5. Toxicokinetic parameters for uptake and elimination of
   1,3-butadiene in mice and rats
Parameter (units)
K12V, (mL/h)
K21 (h'1)
K81(NA)
K..CNA)
kel(h-')
Cltot"'b (mL/h)
Vmax(umol/h/kg)
Mouse
10,280
3.2
2.7
1
7.6
7,300
400
Rat
5,750
2.5
2.3
0.5
8.8
4,500
220
Definition of parameter
Equilibrium constant between chamber volume
and test animals; first order, Vj -• V2.
Equilibrium rate constant between chamber
volume and animals; first order, V2 - V^
Static equilibrium constant representing virtual
absence of metabolism.
Steady-state concentration; ratio of concentration
in animal to chamber concentration.
First-order metabolic elimination rate constant.
Total clearance of chemical from chamber.
Maximum rate of metabolism of chemical.
"Calculated for Y!-«,.
bValid for linear range of metabolism (up to 1,000 ppm for both species).
NA = not applicable.

Source: Filser and Bolt, 1981; Kreiling etal., 1990.
   Table 3-6. Toxicokinetic parameters for the uptake and elimination of
   epoxybutene in rats and mice
Parameter (units)
k12V! (mL/h)
K2I(h->)
Keq(NA)
^(NA)
k^h'1)
Clfc/-" (mL/br)
Vmax (umol/h/kg)
Metabolic saturation (ppm)
Mouse
33,500
0.79
42.5
10.2
2.3
24,900
350
500
Rat
13,800
0.37
37
1.16
11.5
13,400
>2,600
>5,000
Definition of parameter
Equilibrium constant between chamber volume
and test animals; first order, Vj — V2.
Equilibrium rate constant between chamber
volume and animals; first order, V2 - V\.
Static equilibrium constant representing virtual
absence of metabolism.
Steady-state concentration; ratio of concentration
in animal to chamber concentration.
First-order metabolic elimination rate constant.
Total clearance of chemical from chamber.
Maximum rate of metabolism of chemical.
Concentration resulting in saturated metabolism.
"Calculated for Vj-oo.
bValid for linear range of metabolism (up to 1,000 ppm for both species).
NA = not applicable.

Source: Filser and Bolt, 1981; Kreiling et al., 1987; Laib et al., 1990.
1/28/98
3-23
DRAFT-DO NOT CITE OR QUOTE
-------








CO

• M
•a
s
•4-»
to
O

**jj
"«
.13
B
*PK
_>>
"S
O
CO
0
e
o
£>
i
5
s
CO

{•4
CO


A
(5


b£
"3
S



*^*
tu
•s
n\
f*t
E2«
CU
y


C
_o
"-3
.£•

u
CJ
Q







B
1
w
's
•a
e
SS
e
o
"•C
u
s
P<

Epoxybutene tentatively identified in blood after 2, 4, or 6 h
of exposure. EB blood concentrations were 0.4 and 3.3 uM
in rats exposed to 70 and 1,000 ppm 1,3-butadiene and 0.5,
1.6, and 13 uM in mice exposed to 7, 70, and 1,000 ppm,
respectively. Concentration of EB in blood of mice 2 to 3
times > rats.
S
O >-i

lg
f£T g.
CO ^^
VQ ^^
P3 •*
rQ
C5 i^
c^ O
w ov
c3 C**
Q Sl>
OQ o
^-H £H

P ^ ^
U- C3 »^H
S 3 ^>
O ^D O
^rA •"
ass-







^.^
§ S
<*. T—t
I \«/
O -
C cc

"5 °
ctf ^t3
1 §
S ra
r* ^H
T-t
'•^-^
.6 d c^ * ^^co-2
fill
^ 1 «?' § 1 ^ 3 ! S 1 s 8 91^ § ^ ? i
••sl^-l^tj gmgS^s 6^0 a^^5^
l^^l-a l&ls§s ^Is £AJ£?
g^Wo.So COWTS-K-SO ^ r-, o ^g^m-^
Hi Ipti! ISI Iris*
SiSl
IP! It lillii ill Hi
lit!
fgfjs ;§0l^i)^ Ilm 1-slss
a^^illl 3:1 H!lS (&S11 dlllS
b o O O co
O i - ,_( [" -i\ 1*7
Cw ojgtM ain-i OJ^^1
g CJClnO CJo O '-J"*-' T3
•*-* pi C ^ «^H "S O "c o €
w S *^o xo ^3 ^0 M ^
li B-5 §5" . o^tl
2 co *o *~< -^-* *o v-4 ^cu vo ^ S r*> i
Sue cQOc8s cQtSs P3 o" « X ^<,
^ *^ *^~l W r^ rrt ,—t co »rt ^o CD f ,. ^
H. ^-3" scN*^ csno e!cN3o3
.SJ ;=; cavoHO re&i&i co 	 r J2 11 co
1° ^^-s^ ^^^ ^"--o^gL
§O" S CS 8 ^ 20^0 2c^iTl3cD
SO Q VS--0 & Q Cl"0 Q ^VO S In
cV(D OOl>O'«'a? &O°S OOlUI-|roS

OJ^ O *D i^ S3 O CD -*-i O CD ^"^ -*-i pj
CD "^3 CD *O 1_T Q^ o CD *"O _„- (D r^ $3 — 'S
t» t* re "^ ti X, iif re j-^ M re QH '^ fl
co"3 co'S^Sxl co"3*J cn'3cl|-«cN
°'?^. °^5o«> 0-98 °-^r>;8'-f

W -— <" CN W "-H" p 8 di pjr-To W i— T o 8 re





^ CT"
Cs C?s
O\ ^—^ CTs
^"^ o\
o^ "re r-H re
g; t> ^ t»
ri S3 "re CS
_• S o S
re ^o ^3 ^o
*j T3 -a • 3
i — 1 r— 1 i — 1 r-H
1/28/98
3-24
DRAFT-DO NOT CITE OR QUOTE
-------






^*^\
•o
V
S
• MM
•*•
O
u
en
.M
•5
-v>
in
a
o
i

cs
£
j>>
"3
O
i,
CO
O
C
^M
O
^^
^
es
S
s
5
CO
^
CO

M
2
SB
H



WD
"5
£





•w
^
«
c.
X
,CO ^ "^ " "« ^S^^c^aS
^ c^ ••! § J W 1 -^ w .g 2 Q | 8 .
•Js °1 § '^ 'S .g *£ 'i -•§ ^ | i *o *° '-^
ao-ls^ all aS^ISlS
5 S +j -a"
O O i ^ «M (L)

S Jo w 2 *ri. 'rt § ^03
ft ill 111! 111
'2Sc*=!^'a< T3r?ora ±*a3"-'/a?
gftOPS^ C«n,bo P-S^C
liiff Jtfl ||||
"s .§ 2 ^~ "s ^ •§ i *s -| 5 21

t---J «~ ~i O C^ ci-[ T"["^ prJ| rf*i ^j r*^ * r^ ^j1^ p^ \^ ^^



'cH^
UO VO^
0 0\
ON O\
^^/ *~H
^""^
cc cc
•S -1 S3
s § 's
s j g
1 . ji 1
o 3 § ,
I 1 1
-^ £ Co' -S
9~* 'r~< 10 E— '
l-H i— t T-H • r— 1
(
lary metabolite Ar-acetyl-S'-(l-hydroxy-:
in mouse urine using 13C-NMR. This
ted 3.7% of total urinary metabolites
it was not detected in rat urine.
•a s s M
I'fill
1 $$l
sail
It-ll
8 « -2 h
S t=S CD g
a£ a S
o
^"^ L ^"j ^
.§«o«S
5 s -§
tC & «
m 2 P
o a 3
MS &, «
m & a •
^§81?
03 2S- bO g
S3 
_g

W
03 T3
Q .1
T K
ffl ^
1/28/98
3-25
DRAFT-DO NOT CITE OR QUOTE
-------


















0
3
•***
O
^o,
V)
"2
4^1
CO
S
O
••G
*s
»••
*S
'>>
is
o
i
t/J
O
a
o
^^
ij*
a
5
|
CO

t~;
CO
«










M
_c
*S









•4-*
U
"i
a.
X

U
C
i-
1
•a
e
rt
_o
*5
CJ
§
P<
-«->
o
G
H|
111
o w o
S § Q
.S g d
•g.s g
lt|
O „ 3
•° S -?
«?-
0 a S
«-B &
0 g 0
^3 -2 O
^ C ^-c
o
111
f — 3
03 CQ M v^;
fg&l
-O • « O 1~<
III?
« to « |
00 S cs »

•c^
if
o w
*-* ?4
.§•1

's =|
"S §
S3 O
^ S
g 1
c_ co
o
8 g
S 2 •
o 1 S
fll
^ -









/-S
g
s*^-
.
*H
4— >
2
"o
o
                '









              t
43'o>«tM
ail !

ii i

           s '-a g
           til

                cgaa
ti
                «O

                 03



                 "S
                 C3


                 S


                 i




                  -
                       Ill

        S «

        Q S
        oo c
        PI O

       'g.g
ti

                                    « •   &= ^

       1



                    t
                                      5


fill
                                     g i
tl
            o
                            W
                               g

                               S
                    1
1/28/98
       3-26
       DRAFT-DO NOT CITE OR QUOTE
-------






/""S
continued
s- • '
.§
•3
s
%-*
P^
at
_a
c
o
i
O
e
0
^^
r^
C3
s
|
02
.
i
1





6J3
•3
1




periment
Hi
CM
O
e
•+•>
.&
£3
W5
ft





1u
o
C
GJ
«3
*s
T3
R
C
"•5
o
A
P5



.s
inant pathway for clearance of epoxybutene
1
1


1
CO
•8
1
S
'o
O
o
t









q^
ON
O\
N 	 '
13
•8
ft 1
M o
I I-Q
ft3 s~^*.
1 "
re
t-<
is by epoxide hydrolase-mediated hydrolysi
1
^

^J^j
1
T— H
I— 1
1
cA
^
mice to j




















ect conjugation with GSH. See (9b) below.
'£
£


o
*-H
to
1*
"If
O
0
CO
s
>>
,0

























to
n
5
o
to
a,
B
S
1
re























05
C
u
H
1
m
*\
r-H
S
-o
»
13
3
^->
it
1 1



















5!
ated BD in mouse-and rat urine using 13C-N1
1
a


2
8
1
•s
CO
«
**0

T3
7™1

l f*j
o
ffl
^

r-i->
human urine using isotope-dilution GC-MS. Metabolite
represents 97% of GSH conjugates derived directly from I




























































T3
H
(see (7g) above) or indirectly by epoxide hydrolase-mediat































hydrolysis and GSH conjugation of BD (see (8e) above).































Quantitated Af-acetyl-5'-(3,4-dihydroxybutyl)-L-cysteine in































mouse (7.1%) and rat (26.4%) urine using 13C-NMR. N-



1
"^^
t-H
3

to
C!
re
^2
ps
re
CTs

acetyl-SX 1 -(hydroxymethyl)-3 ,4-dihydroxypropyl)-L-
T3
C
C3
g
^
'S1
CO
S3
>^
TJ
_g
to
«a
g
re
CO

o
C
-3
re
>~t











cysteine occurred in mouse (7.1%) urine but was not














01
>
O
re
^"^
T™^











detected in rat urine. This latter metabolite was also






























CO
surmised to be a product of DEB as described in reaction 1































1
*s
,£)






























1/28/98
3-27
DRAFT-DO NOT CITE OR QUOTE
-------



/^^s
•o
o
JR
•*•*

O

S-*'
.§
•3
•*.»
CO
C
o

•*-»
MM
G5


J3
*]]*
.2?
C
o
i
CO
O

O
^
P
5
S

1/3
*
I


3





WO
•3
ts



S3
CJ
c
MM

Quantitated 1,3-dihydroxypropanone in rat urine as 5.3%
total 13C-metabolites excreted. It was not detected in mou
urine. Transformation of DEB to 1,3-dihydroxypropanon
may also contribute to excretion of this compound (see
reaction 14 below). This reaction would also be expected
contribute to the formation of C02.
r-j
OO

^_^
1 J^
	 p ^»^

0 m.

e> to
o ro
"s ^
e S
ro «
w 5^
•4-J fml •
C3 *g O

<£ cy5 o
O t o
^j gj ^
u< c3 /^N
ES O O
" iri o\
O ^3 vli'
H^ -*^ *T3
w ^ «





(U
=
o

OK
o (^
i c\
?• o\
5* '"^
p ^
"3 2
—
• M CO
"O P
i ra
CO J2
!? «
t &
Q '"^

W o
^ ^
 x_x
J a
S £? •
•S^ £
g «|
_W g 03
CO S 
^ ro ro
f . 1 ^u t^J
° d-
o ai _
w S3 ^3
s S rt
co **~i ro
o ro
&* ~t« <*>
pd .0 Si-








en ^ — ^
^^ ^^
cc ^^

j3 ^-^
's?
o "ro

on "
i^5 ro

gl
^^ /O*
/— s ^^
T-<
t*-t (L> r^*
O CO ^
Quantitated 1,3-dihydroxypropanone in rat urine as 5.3%
total 13C-metabolites excreted. It was not detected in moui
urine. Transformation of BD to 1,3-dihydroxypropanone
may also contribute to excretion of this compound (see stu
(lOa) above). This reaction would also be expected to
contribute to the formation of C02.
2s
oo

^^
rn ^--/
" x"\
ro
•2 C/ «->"

" cS _2
'g >-, 'ro
T3 S3 't?
PJ H-i ^rj
CQ 5g ^"^
^ n ^

r , t iy^ ro
O ^-T
o «i ro

r^ Q |

c^'ro x-%
^* "r^ ^^
w ._§ 2^





e
o
d
p
S, ^
>i O\

o ^
"a
J2> 'B
"-3 "
1 2
^ 5

? 1
CQ ^
Q 'c?
"vT
s ^

GSH conjugates of epoxybutanediol have not been
quantitated in either in vitro or in vivo studies.

CO
"I—!
"5
If
"B
S
cu
&
CO
B
I
+B
~_~i

y
r-j
C/3

fS C

p,'-Z^
O cc
t-^- Q_J
^ VH





a
o
T
.1
0)
0
•S
S
S
rH~
i,
r?
g
D- CQ
rvl 4^
4 "S
^ BC

<§ s
1/28/98
3-28
DRAFT-DO NOT CITE OR QUOTE
-------














T?
_g
•4^
e
g

VI

.«
•B
S
.2
"5
_s

^^
J"<*
a
0
£
0
e
s*-i
o
>>
es
S
|
cc
^
fO
«





c
•3
'I













4>


53
5"1

*T"i
g
OH
P
W
O
1

*0
^
•§

o

^









r*.
8
T
^
">>

^
T— (

nd B6C3FJ mice exposed to 1,3-[MC]-
C3

^J
^
t— <
i
p
^o
















°
o
CS
u
^


















i the chemical reactions leading to C02
town. Mass balance of "C02 from
CtD T

o C^
^ p3
*^"t ..X
II
T3 S
^ R
^O CM






































-s
residual 14C02 from carcass showed thi
T3

CC
_cl
^^
laled brea
t3







































&

3take of butadiene than rats. Assumptk
erives from erytkitol. Dahl et al. (1991
S"^

!— <
•£j
Kj
•a
1
^
2






































*T3
O



^^
J
-T3
^
-22







































)2 in exhaled breath of monkeys expose
sne. Uptake and retention of 14C in mic
r^ rTl

•T C^
""* "S ^-^
"^3 o Gi
ill
CO ^H
'cS -H A







































iromatography; GSH = glutathione; MS =
o

«J
00
II
8
b"^
O 2
*l?
^ «

OH 2
V oo
CN S
^" 0,
S Q
o1
"2 *c
Si
^1
o S3
JD pt
SCO
•c
II g.
0 «
Ji
« l>;
p -^3
."§> S
!-§

_o a>
"3 p.
^ ai
-------
   es
   •a

   o
   C
   o
   o

   •a
   o
   d   i

   IS JS
   5  c
 o. «
 -------
P
«s
cs
"S
•ts
B

"S

a
o
o
u
•a
o
o
— « B
.2
a "S
5 "«
s •§
S?^
O pQ

.« C

•o "3
SM "
°1 "^
'a *•?
£c ^
Q CJ
8-1
jT d
c
O W3
s 1
3 "S
T-T o
18 1
O A
•— ^
SCQ
.2
1 ^^
cu
« S,
r^ f^

H =5





















•3
O
O
2
•S 4"*
o g
>& S*
wj B
§ Oi
|t
1 §
IS
£ e
s ^
u
g
O


11
1^'

Butadiene e
concentratii













0)
u
§
 r?
O >n o
O, o, co,
ON os CO
-: o P
0 0





i — i * — ii — i in ON ^D
0 °. 0 P 0 P
^ £x ^"^ S ^"^ S
1-1 oo (-; vo ,_• •*
O C3 O


>-< "—i ^ -^- en in
o cs ^ S, d B
m ro ^t- oo . oo ^i
O i — ' ^h





co °° cN t— < O O
•* r-" ON _ o v-y
^ O CO VO
r--"


i>>
o
1
cd S
























1
"O
g


entration and blood cone
1
o





•o
. o
^o
zs
£
eg
.8
OJ
e
-2
"D
«
"3
J3
c^
i— t
13
^H

1
2
3
O
X*
O

CB
•§
°S
sure; animals continued 1
!; detection limit = 0.1 \i
lerefore no SE reported.
o S ^
headspace equilibriun
les collected at 4 h of i
id analyzed by GC-GC
butadiene concentrati
~S
1,3-butadiene was quantitated by a vi
C; « en
Values are means ± SE (n = 6) for sar
by vacuum-line cryogenic distillation
* One exposure was conducted at this 1,





"2
1
1~
a
§•
•73
g
a
o
1
jg
"3

o
DO
0
&
.1
|
1
£
>>
^

8
S
•g
§
••3
idwelling catheter; 1,3-t
^






























wing 2 h exposure us
ntillation counting.
—
Blood (n = 3) collected immediately ft
'3
as C-labeled equivalent using liquid :










CQ
S
Q
M
ttn
s
en

o
i
CQ
§
.2
t
en
o
0
II
&
J
'•§
O
'S
T3
io"
and quantitated by GC-1
3
an into methylene chic
«





























"1
S,
uM and 0.01 uM, res
g
s EB and DEB were recovered by extrai
4 Detection limits for EB and DEB - O.I
S
"I
T3
'2
s
fe
0 -J?-
.£> &•
"c3 ?

f 0.
O 
-------
    Table 3-9. Tissue levels of epoxybutene and diepoxybutane (pmol/g tissue)
    in male rats and male mice exposed by inhalation to 62.5 ppm 1,3-butadiene
    for4h
Tissue3
Blood
Heart
Lung
Liver
Fat
Spleen
Thvmus
Bonemarrowd
EB
Rats
36±7
40 ±16
NDb
ND
267 ± 14
7±6
12.5 ±3.2
0.2 ±0.1
Mice
295 ±27
120 ±15
33 ±9
8±4
1,302 ±213
40 ±19
104 ± 55
2.3 ±1.5
DEB3
Rats
5±1
3 ±0.4
0.7 ±0.2°
ND
2.6 ±0.4
1.7 ±0.5°
2.7 ± 0.7C
ND
Mice
204 ±-15
144 ±16
114±37
20±4
98 ±15
95 ±12
109 ±19
1.4 ±0.3
*Mean±SE;n =
bND = not detected; indicates that analyte was not detected or was not above control level.
'Includes at least one ND value.
dAs mean pmol/mg protein ± SE.

Source: Modified from Thornton-Manning etal., 1995a.
1/28/98
3-32
DRAFT-DO NOT CITE OR QUOTE
-------
   Table 3-10. Tissue levels of epoxybutene and diepoxybutane (pmol/g tissue)
   in male and female rats exposed by inhalation to 62.5 ppm 1,3-butadiene for 6 h
Tissue3
Blood
Femur
Lung
Fat
Mammary
EB
Males
25.9 ±2.9
9.7, 9.3
12.7 ±5.0
175 ±21
ND
Females
29.4 ± 2.0
10.4 ±1.0
2.7 ±4.3
203 ± 13
57.4 ±4
DEB
Males
2.4 ± 0.4
1.1,1.8
1.4 ±0.8"
l.liO.l
ND
Females
11.4±1.7C
7.1±1.3C
4.8±0.7C
7.7±1.3C
10.5 ±2.4°
an = 3, except for male femur, where n = 2.
bOne value was not detectable; instrument detection limit/2 was substituted to calculate the mean.
"Statistically greater than male tissue value, psO.05.

ND = not determined.


Source: Modified from Thornton-Manning et al., 1995b.
1/28/98
3-33
DRAFT-DO NOT CITE OR QUOTE
-------
                 Cp1
     v,
(atmosphere)
                                      Cp2
                          12
              ^k2i . .<>      (rats)
                              A


                           interface
                                                     kel
metabolic
elimination
      Figure 3-2.  Two-compartment pharmacokinetic model for inhalation
      chamber.

      Source: Filser and Bolt, 1981.
1/28/98
                            3-34       DRAFT-DO NOT CITE OR QUOTE
-------
         Dawley rats and male B6C3Fj mice to 62.5 ppm 1,3-butadiene for 4 h, as described in Table 3-7,
         with the results shown in Table 3-9.  The same group of investigators (Thornton-Manning et al,
         1995b) also examined gender differences in the production and disposition of epoxybutene and
  4      diepoxybutane by determining tissue concentrations of the two butadiene metabolites in male and
  5      female Sprague-Dawley rats, as described in Table 3-7, with the results shown in Table 3-10.
  6      The concentrations of epoxybutene did not differ significantly between male and female rats in
  7      any of the tissues examined.  The highest concentrations were observed in the fat tissues of both
  8      sexes. Tissue levels of the diepoxybutane, however, were consistently greater in females than in
  9      males. Blood diepoxybutane levels of female rats were 4.75-fold greater than those of male rats.
 1 0      The greatest gender difference was in the levels of the diepoxybutane in fat tissue, with females
 11      having a sevenfold greater tissue concentration than males.  The mammary tissue of females also
 1 2      contained relatively high levels of the diepoxybutane.  The authors suggest that the greater
 1 3      production of the highly mutagenic diepoxybutane in females may play a role in the increased
 14      incidence of mammary tumors observed in a chronic carcinogenicity study with rats (Owen et al.,
 15      1987).
 1 6            Dak! et al. (1991) exposed cynomolgus monkeys to nose-only inhalation of 1,3-butadiene
 1 7      and measured the levels of diepoxybutane and 3-butene-l,2-diol.  Results for diepoxybutane are
 1 8      included in Table 3-9 for comparison to 1,3-butadiene and epoxybutene levels measured in their
         previous study (Dahl et al., 1990).  Exhaled air and excreta were collected during exposure and
 TO      for 96 h  after exposure and are summarized in Table 3-11.
 21             Two in vivo studies provided data on the urinary excretion of butadine metabolites by
 22      humans.  In the first study (included in Table 3-7 and described in more detail here), Bechtold et
 23      al. (1994) identified and measured two metabolites of 1,3-butadiene, l,2-dihydroxy-4-(JV-
 24      acetylcysteinyl-^-butane (M-I) and 1 -hydroxy-2-(A'"-acetylcysteinyl-tS'-)-3 -butene (M-II) in the
 25      urine of workers employed at the Texaco Chemical Co. in Port Neches, Texas, a 1,3-butadiene
 2 6      extraction plant. The study population included (1) exposed employees who worked in two areas
 27      (described as low- and high-exposure areas) with time-weighted average concentrations of 3 to 4
 28      ppm 1,3-butadiene over the previous 6 months; (2) an intermediate exposure group spending
 29     variable time periods in low- and high-exposure areas; (3) nonexposed employees who worked in
 30      areas with historical time-weighted  average concentrations of less than 0.1  ppm 1,3-butadiene;
 31       and (4) outside controls who had no known exposure to 1,3-butadiene. Urine samples were
32     analyzed from 7, 3, 10,  and 9 subjects, respectively, from the above four groups. The assay was
33      based on isotope-dilution GC-MS.  After addition of deuterated internal standards, the
34     metabolites were isolated from urine samples by solid-phase extraction and selective
        precipitation. M-I but not M-II could be readily identified and quantitated in the urine samples


        1/28/98                                   3-35        DRAFT-DO NOT CITE OR QUOTE
-------
             Table 3-11. Excretion of 14C by monkeys exposed to l,3-[14C]-butadienea

Exposure
concentration
(ppm) CO,
10.1 1.5 ±0.2

310 0.21±0.04d

7760 0.08 ±0.02d>c


Exhalants

Other1" Urine Feces
0.45 ± 0.9 ±0.1 0.021 ±0.005
0.33
0.40 ± 0.8 ± 0.2 0.01 1 ± 0.003d
0.21
1.00 ± 0.58 ± 0.002 ±0.001d>e
0.35 0.06d
Uptake
Total metabolites
recovered'
2.88 ±0.22

1.40 ±0.42"

1.65±0.29d

        'Values are mean percentage of total inhaled ± SB measured for 96 h after 2-h exposure.
        "Includes all material (except CO^ exhaled during the 2-h exposure and 96-h postexposure.
        "Mean ± SE of the sums of CO2> other, urine, and feces values for individual monkeys; does not include
         residues, if any, in monkeys' bodies.
        dSignificantly different from low-level exposure (p<0.05).
        'Significantly different from mid-level exposure (p<0.05).
         Source: Dahletal., 1991.

 1       (limits of sensitivity for this assay, 100 ng/mL). The  average values of M-I for exposed,
 2       intermediately exposed, nonexposed, and outside control employees were 3,200 ± 1,600, 1,390±
 3       550, 630 ± 190, and 320 ± 70 ng/mL, respectively. Although the levels of exposure for each
 4       individual were not known, the urinary levels of M-I  for the exposed groups were significantly
 5       higher (p<0.05) compared with the outside control group. The implications of M-I in the urine
 6       from individuals with no known exposure to 1,3-butadiene are not known.
 7       In the second study that provided human data, Ward et al. (1996a)  reported increased levels of
 8       the urinary metabolite l,2-dihydroxy-4-(A^-acetyl-cysteinyl)-butane (a human urinary metabolite
 9       also identified by Bechtold et al., 1994) and somatic mutations in workers at a styrene-butadiene
10       rubber plant. Exposure was assessed in workers from areas of higher exposures (reactor,
11       recovery, tank farm, laboratory) and lower exposure (blend,coagulation, bailers, shipping,
12       utilities, shops) using badge dosimeters; the concentration of the metabolite was measured in
13       urine; and the frequency ofhprt mutant lymphocytes was determined
14       by autoradiography. The detection limit (0.25 ppm)  was exceeded in 20/40 dosimeter readings
15       in the high-exposure group and in 0/20 readings in the low-exposure group.  Sixteen high- and
1 6       nine low-exposure urine and blood samples were analyzed.  Expressed as ng/mg creatinine,
17       metabolite concentrations were 2,363 ± 1,880 and 937 ± 583 (p<0.05), respectively, for the
         1/28/98
3-36
DRAFT-DO NOT CITE OR QUOTE
-------
         high- and low-exposure groups. The respective mean mutant frequencies were 7.09 ± 5.2 x 10~6
         and 2.26 ± 1.34 x IQ'6 (p<0.05).
                Other in vivo studies that further confirm the pathways shown in Figure 3-1 but use
  4      different intermediate endpoints than those shown or with noninhalation exposure are described
  5      below. Deutschman and Laib (1989) studied the effects of 1,3-butadiene exposure on
  6      nonprotein sulfhydryl (NPSH) content of lung, heart, and liver tissue of rats and mice. In these
  7      experiments, male B6C3Fj mice and Sprague-Dawley rats were exposed to 1,3-butadiene at
  8      concentrations of 10, 50, 100, 250, 500, 1,000, or 2,000 ppm for 7 h. For rats, a reduction
  9      (»70%; significance level not stated) occurred in liver NPSH for animals  exposed to 1,000 to
1 0      2,000 ppm.  A reduction of approximately 20% was observed in lung NPSH of rats, and no
11       appreciable depletion of NPSH was observed for heart tissue of rats. For mice, depletion of
1 2      hepatic NPSH was observed at exposure concentrations of 100 to 250 ppm and declined to 20%
1 3      of control for the 2,000 ppm exposure group. Similarly, the NPSH content of mouse lung tissue
14      also declined by 80% to 90% at the two highest exposure levels.  For heart NPSH content in
1 5      mice, minor declines were noted for exposure levels up to 500 ppm, but a rapid decrease was
1 6      observed between 1,000 and 2,000 ppm that resulted in an «75% depletion.  Kreiling et al.
17      (1988) suggested that the greater susceptibility of mice to the carcinogenic effects of inhaled
18      1,3-butadiene might reasonably be explained by the  higher rate of formation of the epoxide  .
         intermediate and its limited detoxification and subsequent accumulation in mice. The authors
         applied the concentration-response data to the exposures used in earlier bioassays (HLE,  1981)
21       and noted that, for rats exposed chronically to 1,3-butadiene concentrations of 1,000 or 6,000
22      ppm, a daily hepatic NPSH depletion of about 25%  and 60% and lung NPSH content depletion
23      of 20% and 30% for the low and high exposures, respectively, was calculated.  However, these
24      values were based on assumptions that 1,3-butadiene metabolism and NPSH resynthesis
25      remained constant throughout the duration of the chronic exposure. Applying the same methods
26      and assumptions for mice, daily NPSH depletions for the low-exposure (625 ppm) and high-
27      exposure (1,250 ppm) levels, respectively,  were estimated for liver (50%  and 70%), lung (70%
28      and 90%), and heart (25% and 40%).  In studies assessing the effects of 1,3-butadiene exposure
29      on NPSH content of various tissues, Deutschmann and Laib (1989) reported that depletion of
3 0      cardiac NPSH content in mice after inhalation exposure was an indicator of systemically
31       available epoxide intermediates of 1,3-butadiene that reach the heart by efferent blood flow
32      from the lungs or liver.
33             The reduction and/or depletion of NPSH content in mice is also indicative of saturation
34      of conjugation of the epoxide metabolites of 1,3-butadiene by glutathione. Glutathione
3 5      conjugation of epoxybutene and metabolism by glutathione S-transferase was shown by
         Malvoisin et al. (1981) and Bolt et al. (1983), respectively, and a reduction in hepatic NPSH
        1/28/98                                   3-37       DRAFT-DO NOT CITE OR QUOTE
-------
 1       content in mice exposed to 1,3-butadiene was shown by Kreiling et al. (1987).  A more recent
 2       study by Kreiling et al. (1988) suggested that glutathione conjugation may be important in the
 3       detoxification of this reactive intermediate. Air-exposed control animals exhibited a moderate
 4       time-dependent decrease in hepatic NPSH content, whereas those animals exposed to 1,3-
 5       butadiene exhibited a significantly greater reduction in NPSH. After 7 h of exposure, the
 6       hepatic NPSH content in mice was reduced to approximately 20% and was reduced further to
 7       about 4% after 15 h of exposure.  For Sprague-Dawley and Wistar rats, hepatic NPSH content
 8       was initially reduced to 80% and 65%, respectively, but remained stable thereafter. Concurrent
 9       with the significant depletion of NPSH in the mice were signs of acute toxicity (specifics not
10       noted); no toxicity was observed in any of the rats tested. This experiment clearly demonstrates
11       species variability in the magnitude and time course of hepatic NPSH depletion after inhalation
12       exposure to high concentrations of 1,3-butadiene. Furthermore, the progressive decline in
13       hepatic NPSH content in mice correlates with a reduction in epoxide exhalation and a decline in
14       1,3-butadiene metabolism. The accumulation of epoxide intermediates (epoxybutene and
1 5       diepoxybutane) in mice (Bond et al., 1986) is consistent with the observed depletion of hepatic
1 6       NPSH in this species and the increased metabolism (i.e., production of epoxide intermediates)
17       observed for mice.
1 8              Nauhaus et al. (1996) indicated that metabolites were detected in mouse urine that are
19       also  seen following exposure to acrolein and acrylic acid, suggesting that these compounds may
20       arise directly from 1,3-butadiene oxidation or indirectly from further metabolism of
21       crotonaldehyde.  Rats excreted 1,3-dihydroxypropanone, a metabolite that may be derived from
22       hydrolysis of diepoxybutane. Metabolites derived from diepoxybutane were similar in rats and
23       mice when expressed as a percentage of total metabolites; however, when normalized to body
24       weight, the amount of diepoxybutane-derived metabolites was four times greater in mouse urine
25       than in rat urine. The greater body burden of diepoxybutane in the mouse and the greater ability
26       of rats to detoxify diepoxybutane through hydrolysis may be related to the greater toxicity of
27       1,3-butadiene in mice. The metabolites derived via reactive aldehyde intermediates in mice also
28       suggest a role of these aldehydes in the toxicity of 1,3-butadiene.
29              Following i.p. injection of 14.3 or 143 (imol/kg of epoxybutene, two glutathione
30       conjugates, S-(2-hydroxy-3-buten-l-yl)glutathione (I) and ^-(l-hydroxy-S-buten-l-
31       yl)glutathione (II), were detected in the bile of rats (Sharer and Elfarra, 1992). At either dose,
32       the amount of conjugates excreted in 30 min was at least 85% of that excreted in 120 min.
33       When the epoxybutene dose was varied between 14.3 and 286  umol/kg and the combined
34       amounts of conjugates I and II excreted in 60 min were determined,  an apparent linear dose-
35       relationship was obtained. Saturation was not observed at these dose levels. Total conjugates
36       excreted in 60 min averaged 7.6% ± 4.2% of the administered dose with approximately a 3:1
         1/28/98                                   3-38       DRAFT-DO NOT CITE OR QUOTE
-------
          ratio of conjugates 1:11. Although the study showed that epoxybutene GSH conjugates are
          formed in vivo after administration of epoxybutene, biliary excretion of GSH conjugates
          account for only a small portion of the administered dose.
  4             7V-Acetylcysteme derivatives of the two glutathione conjugates of epoxybutene identified
  5       in the bile of rats by Sharer and Elfarra (1992) were detected in the urine of rats and mice
  6       administered with epoxybutene intraperitoneally (Elfarra et al., 1995).  When rats were injected
  1       with epoxybutene at doses ranging from 71.5 to 285 junol/kg, the urinary excretion of S-(2-
  8       hydroxy-3-buten-l-yl)-^-acetyl-L-cysteine (I) and ^-(l-hydroxy-S-buten^-yO-JV-acetyl-L-
  9       cysteine (II) within 8 h of epoxybutene administration exhibited a linear dose-relationship; the
 10       total amount of the two mercapturic acids combined averaged 17% ± 4%.  No metabolites were
 11       detected in urine samples collected 8 to 24 h after dosing.  Mice excreted similar amounts of
 1 2       mercapturic acids (26% ± 13%) at 285 jimol/kg within 24 h of dosing. However, at 143  and
 13       71.5 [imol/kg, excretion accounted for only 7% ± 3% and 9% ± 3% of the dose, respectively.
 14       Rats preferentially excreted mercapturic acid II over I (approximate ratio 3:1), whereas mice
 1 5       preferentially excreted mercapturic acid I over II (approximate ratio 1.85:1).  The study showed
 1 6       that at low exposure levels, rats excrete higher levels of epoxybutene mercapturic acids than
 17       mice.
 1 8             In summary, the inhalation studies show the uptake of 1,3-butadiene exhibits first-order
          kinetics at exposure concentrations <1,000 ppm, but at higher concentrations, the process
 TO       becomes saturated and exhibits zero-order kinetics; mice exhibit saturation kinetics at lower
 21       exposure concentrations than do rats.  At exposure concentrations up to 1,800 ppm, the uptake
 22       of 1,3-butadiene is approximately fourfold greater in mice than in rats.  In addition, mice
 23       accumulate a greater amount of 1,3-butadiene or its metabolites or both than do rats exposed
 24       similarly.  Limited data on monkeys indicate that the metabolic uptake rate is less than that for
 25      rats or mice.
 26             After inhalation of 1,3-butadiene, mice appear to have greater levels of radioactivity (15-
 27      to 100-fold greater at all time points after exposure) in all tissues than do rats exposed similarly,
 28      but no significant qualitative differences have been observed regarding storage depots or target
 29      tissues.  However, immediately after a 2 h inhalation exposure, mice exhibited higher levels of
 30       1,3-butadiene metabolites (including the reactive epoxybutene) in the blood than did rats.  A
31       comparison of butadiene epoxide levels in target tissues (blood, bone marrow, heart, lung, fat,
 32      spleen, and thymus) of rats and mice following inhalation of low levels of 1,3-butadiene showed
3 3      consistently higher epoxide levels in mouse than in rat tissues. Other in vivo experiments
34      demonstrated gender differences in the production of butadiene metabolites in rats, with tissues
         from female rats containing higher concentrations of the diepoxybutane than tissues from male
         rats. The experiment also showed that the levels of epoxybutene were similar in males and

        1/28/98                                   3-39       DRAFT-DO NOT CITE OR QUOTE
-------
 1       females. In vivo experiments have confirmed the role of cytochrome P-450 in the metabolic
 2       activation of 1,3-butadiene observed in in vitro studies.  Epoxybutene, a reactive intermediate,
 3       may undergo epoxide hydrolase-mediated hydroxylation or conversion via P-450 to another
 4       reactive intermediate, diepoxybutane.  Conjugation with glutathione represents a detoxification
 5       process.
 6              Although 1,3-butadiene may be metabolized by microsomal cytochrome P-450 in both
 7       rats and mice, species-related quantitative differences in the fate of inhaled 1,3-butadiene are
 8       well documented. The greater susceptibility of mice to the carcinogenic effects of 1,3-butadiene
 9       may be related to the higher rate of epoxybutene formation and the limited detoxification and,
10       hence, the greater accumulation of this reactive intermediate in this species. At concentrations
11       >2,000 ppm, the metabolism of 1,3-butadiene follows saturation kinetics in both rats and mice,
12       but the rate of metabolism in mice is greater (about twice); furthermore, the metabolism of
13       epoxybutene is saturable in mice but not in rats. With increasing exposure concentration, the
14       metabolic capacity for epoxybutene becomes rate-limiting in mice but not in rats.  Data
15       available from studies with nonhuman primates show that at low-exposure concentrations (< 10
1 6       ppm), the steady-state tissue levels of reactive 1,3-butadiene metabolites are lower in monkeys
17       than in rats or mice. The lower uptake rate of inhaled 1,3-butadiene by monkeys suggests that,
1 8       for comparable exposures, monkeys will receive a lower internal dose of reactive butadiene
1 9       metabolites. The uptake and retention of 1,3-butadiene appears to be nonlinear in the
20       concentration ranges used in long-term exposure studies, and repeated exposures to 1,3-
21       butadiene do not appear to induce its metabolism.
22              1,3-Butadiene may be excreted via the respiratory tract, urine, or feces. The rate of 1,3-
23       butadiene excretion by rats and mice was shown to be unaffected by exposure concentration
24       (0.14 to 13,000 |ig/L). Half-lifes for urinary excretion of radioactivity were similar for both rats
25       and mice (5.6 and 4.6 h, respectively), but fecal excretion was somewhat greater in rats (22 h)
26       than in mice (8.6 h). A shift to excretion of 1,3-butadiene-derived [14C] via the lungs was noted
27       for rats but not mice at high (13,000 n.g/L) exposure concentrations. Approximately 2% of the
28       total inhaled dose was excreted as 14CO2 or in the urine of monkeys exposed for 2 h to 1,3-[14C]-
29       butadiene at concentrations ranging from 10 to 8,000 ppm.  At the higher concentrations, the
30       proportion of CO2 decreased, whereas exhaled metabolites (diepoxybutane and butene diol)
31       increased. Elimination of radioactivity from the blood and tissues of rats and mice after
32       inhalation exposure to l,3-[14C]-butadiene was biphasic; half-lifes for initial removal were 2 to
33       10 h and for slower elimination were 5 to 60 days. Excretion of epoxybutene via the lungs by
34      rats and mice also has been studied and notable differences between the species observed. For
35       rats, exhaled epoxybutene concentrations at 10 h attained a plateau  of about 4 ppm and

         1/28/98                                   3-40       DRAFT-DO NOT CITE OR QUOTE
-------
          remained at this level for >12 h. For mice, however, the plateau level was about 10 ppm but
          declined to 6 ppm at 15 h, a decline that coincided with signs of acute toxicity in the mice.
                 Studies on the urinary excretion of 1,3 -butadiene metabolites in mice, rats, hamsters,
  4       monkeys, and humans have shown that all these species predominantly produce two urinary
  5       metabolites, l,2-dihydroxy-4-(W-acetylcysteinyl-,S)-butane (M-I) and l-hydroxy-2-(7V-
  6       acetylcysteinyk$)-3-butene (M-II), but in different proportions. The M-II is a mercapturic acid
  7       formed by conjugation of GSH with epoxybutene, while M-I is a mercapturic acid that appears
  8       to form by GSH conjugation with butene diol, the hydrolysis product of diepoxybutane. M-I but
  9       not M-II was also found in the urine of workers exposed to low levels of 1,3-butadiene.

 10       3.2. MOLECULAR DOSIMETRY
 11              In addition to data on absorption, metabolism, and excretion, a complete dosimetry
 12       model for 1,3-butadiene should incorporate information on molecular dosimetry, which links
 1 3       exposure to some internal biomarkers of exposure.  This last component is best evaluated by
 14       assessing adduct formation.
 1 5              The use of Hb adducts as biomarkers of exposure to 1,3-butadiene was investigated by
 1 6       Sun et al. (1989a). In this study, male B6C3FJ  mice and male Sprague-Dawley rats were
J 7       injected intraperitoneally with l,3-[14C]-butadiene at doses of 1, 10, 100, or 1,000 umol/kg, and
          adduct formation was monitored. Hb adduct formation was linearly related to dose up to  100
 19       nmol/kg for both species. The Hb adducts accumulated linearly after repeated injections of 100
 20       |imol/kg for 3 days.  The 1,3-butadiene-derived Hb adducts showed lifetimes of -24 and -65
 21       days in mice and rats, respectively, which correlates with the lifetimes of red blood cells.
 22       Assuming that adduct formation is a function of the extent of 1,3-butadiene metabolism, the
 23       similarity in the degree of Hb adduct formation  between mice and rats does not reflect the
 24       species variability in toxicity of this compound.  Therefore, Hb adducts may not  serve as
 25       accurate indicators of levels of reactive metabolites in the blood and, thus, as indicators of
 26       toxicity. However, Hb adduct formation may be useful as an indicator of 1,3-butadiene
 27       exposure.
 28             Similar findings of exposure-dependent Hb adduct formation and stability of the adducts
 29       were reported by Osterman-Golkar et al. (1991) for Wistar rats exposed to 1,3-butadiene at
 30       concentrations of 250, 500,  or 1,000 ppm, 6 h/day, 5 days/week for 2 weeks.  In this study, the
 31        Hb adduct formation also increased linearly with exposure up to the highest exposure  level. The
 32      investigators also concluded that Hb adducts were useful for assessing dosimetry of long-term
 33       exposure to 1,3-butadiene.
                Osterman-Golkar et al. (1996) studied Hb adducts in 17 workers exposed to 1,3-
         butadiene in a petrochemical plant and nine referents employed at the same factory but not
        1/28/98                                   3-41      DRAFT-DO NOT CITE OR QUOTE
-------
  1       exposed to 1,3-butadiene. Using stationary and personal monitoring devices, the ambient 1,3-
  2       butadiene level for workers handling 1,3-butadiene containers was 11.2 ± 18.6 mg/m3 and < 1.2
  3       mg/m3 for maintenance and laboratory workers.  The Hb adduct measured was 2-hydroxy-3-
  4       butylvaline, formed by reaction of .TV-terminal valine with carbon 1 in epoxybutene. Higher
  5       concentrations of Hb adducts (0.16 ± 0.099 pmol/g) were recorded in the workers handling 1,3-
  6       butadiene containers compared with those in maintenance, laboratory workers, and nine
  7       unexposed controls («0.05 pmol/g).
  8              Citti et al. (1984) conducted an in vitro study that examined the reactivity of
  9       epoxybutene (referred to as epoxybutene by the authors) with isolated nucleosides and DNA.
1 0       They reported that two adducts were formed:  7-(2-hydroxy-3-buten-l-yl)guanine and 7-(l-
11       hydroxy-3-buten-2-yl)guanine.  The authors indicated that the epoxide reacted similarly with
12       either free DNA or DNA-bonded deoxyguanosine and that the half-life of these adducts under
1 3       physiological conditions was 50 h.
14              Kreiling (1987) reported the in vivo formation of the DNA adduct 7-(l-hydroxy-3-buten-
1 5       2-yl)guanine in the liver of mice exposed to l,3-[14C]-butadiene (exposure concentration and
1 6       duration not specified). No DNA adducts were detected in the livers of 1,3-butadiene-exposed
17       rats. Note that this adduct was one of two reported by Citti et al. (1984) for the in vitro reaction
1 8       of 3,4-epoxybutene with DNA and deoxyguanosine.  Additional details were not available in the
1 9       abstract by Kreiling nor was additional information reported in later publications by Kreiling
20       and coworkers.
21              Jelitto et al. (1989) reported species-dependent differences in the in vivo formation of
22       DNA adducts by male B6C3Fj( mice and male Sprague-Dawley rats exposed to 1,3-[14C]-
23       butadiene at concentrations of 250, 500, or 1,000 ppm for 7 h.  Analysis (alkaline elution and
24       comparison of HPLC profiles with synthesized adduct standards) of liver DNA from the mice
25       showed that two adducts had been formed: 7-A^-(l-hydroxy-3-buten-yl)guanine and 7-N-(2,3,4-
26       trihydroxybutyl)guanine, the latter being derived from diepoxybutane.  These products were not
27       detected in rat liver DNA. Alkaline elution curves showed that protein-DNA and DNA-DNA
28       cross-linking occurred in mice, but not in rats, after a 7 h exposure to 1,3-butadiene at
29       concentrations of 250 ppm and above. These findings provide additional evidence at the
30       molecular level for explaining the difference in the carcinogenic response between mice and
31       rats.

32       3.3. STRUCTURE-ACTIVITY RELATIONSHIPS
3 3              Studies by Del  Monte (1985)  and Dahl et al. (1987) have shown that the metabolism of
34       structurally related isoprene (2-methyl-butadiene) may be qualitatively similar to that of 1,3-

        1/28/98                                   3-42       DRAFT-DO NOT CITE OR QUOTE
-------
  1       butadiene. Although the diepoxybutane metabolite of isoprene has been shown to be genotoxic
  2       in Salmonella, data are unavailable regarding the carcinogenic potential of isoprene.
                Del Monte et al. (198 5) showed that mouse hepatic microsomal monooxygenases
  4       converted isoprene to epoxides and diepoxides and that the biotransformation was inhibited by
  5       cytochrome P-450 inhibitors such as CO, SKF 525-A, and metyrapone.  Specifically, 3,4-epoxy-
  6       3 -methyl-butene and 3,4-epoxy-2-methyl-1 -butene were major and minor metabolites,
  7       respectively, with the latter representing about 20% of the former.  The 3,4-epoxy-2-methyl-l-
  8       butene metabolite was metabolized further in microsomal incubations to the mutagenic isoprene
  9       dioxide (diepoxide).  Data from these in vitro metabolism  studies were used to calculate the KM
 1 0       and Vmax for the production of the diepoxide. The resulting KM (mM) and Vraax (nmol
 11       diepoxide/mg protein/min) values for diol production by microsomes from control,
 1 2       phenobarbital-induced, and 3-methylcholanthrene-induced mice were 0.24 and 1.7, 0.29 and
 13       5.1, and 0.22 and 2.0, respectively.  The Vmax for the formation of the diepoxide was
 14       significantly increased (p<0.01) in incubations using hepatic microsomes from phenobarbital-
 1 5       treated mice.
 1 6             Gervasi and Longo (1990) provided additional information on the metabolism of in vitro
 1 7       isoprene by hepatic microsomal preparations from rats, mice, rabbits, and hamsters.  Hepatic
 1 8       microsomal preparations from these species metabolized isoprene to epoxybutene, 3,4-epoxy-3-
          methyl-1 -butene, and 3,4-epoxy-2-methyl-l -butene.  The former was the major metabolite and
          was found to have a half-life of 85 min. Microsomal preparations from all species further
21        metabolized the 3,4-epoxy-2-methyl-l-butene to isoprene dioxide (2-methyl-l,2,3,4-
22       diepoxybutane), which was found to be mutagenic and to have alkylating ability.  The KM (mM)
23       and Vmax (nmol/mg/protein/h) for the rat, mouse, rabbit,  and hamster microsomal metabolism of
24       isoprene were 0.08 and 0.24, 0.09 and 1.79, 0.2 and 0.66,  and 0.06 and 1.20, respectively.
25       Unlike 1,3-butadiene, isoprene exhibited the same pattern of metabolism in all species tested
26       and did not result in mutagenic epoxybutene intermediates.
27             In the study by Dahl et al. (1987), groups of 30 male F344 rats were exposed  by nose-
28       only inhalation to [14C]isoprene at concentrations of 8.0, 266, 1,480, or 8,200 ppm for 6 h (5.5 h
29      for the highest exposure), and urine, feces, and exhalants were collected over a 66 h
30      postexposure period. During this period, >75% of the nonisoprene (metabolites) radioactivity
31       was excreted in the urine. Except for the highest exposure group where greater amounts of
32      radioactivity were excreted in the feces, a pattern of predominantly urinary excretion was
3 3      consistent among the various exposure groups. The half-life (mean ± SE) for urinary excretion
34      of 14C was 10.2 ± 1.0 h (range of 8.8 to 11.1  h). Generally, the concentration of metabolites in
         the blood increased with exposure concentration and duration of exposure. The authors  noted
         that 85%  of the radioactivity in the blood was associated with material of low volatility and that

        1/28/98                                   3-43      DRAFT-DO NOT CITE OR  QUOTE
-------
 1       it probably represented covalently bound metabolites, conjugates of isoprene metabolites, or
 2       tetrols.  Only at the two highest exposure concentrations were materials detected that possessed
 3       volatilities matching those of isoprene and isoprene monoepoxides. The percentage of inhaled
 4       isoprene-derived 14C present as diepoxide or diol in the blood remained fairly constant with time
 5       but decreased with exposure concentration. Assessing the distribution of isoprene and its
 6       metabolites in some  animals of the 1,480 ppm exposure group revealed that the liver and blood
 7       contained the majority of the radioactivity. Relatively large amounts of metabolites were
 8       present in respiratory tract tissues after 20 min of exposure.  The mutagenic metabolite, isoprene
 9       diepoxide, was identified in all tissues examined and, in the blood, represented between
10       0.0018% and 0.031% of the inhaled 14C label. Although exposure to high concentrations of 1,3-
11       butadiene result in CO2 as the major metabolite, this study suggested that the major route of
12       excretion for isoprene is in the urine. The authors noted, however, that this finding is tentative
13       and may be the result of a labeling artifact. Although no evidence for metabolic saturation was
14       detected for the isoprene concentrations used, the uptake and fate of inhaled isoprene are similar
15       to that of butadiene.
1 6              Peter et al. (1987) also studied the pharmacokinetics of isoprene in male Wistar rats and
17       male B6C3F! mice.  Animals were exposed in closed systems to concentrations as high as 4,000
18       ppm for up to 10 h.  At concentrations <300 ppm, the rate of metabolism was found to be
1 9       directly proportional to the isoprene concentration, but saturation of metabolism was detected at
20       higher concentrations. The Vmax for the metabolism of isoprene in rats and mice was 130 and
21       400 umol/h/kg, respectively.  Exhalation of the parent compound was approximately 15% and
22       25% in rats and mice, respectively.
23              Chloroprene (2-chloro-butadiene) is also structurally similar to 1,3-butadiene.  Studies
24       have shown that the biotransformation of chloroprene results in the formation of peroxides that
25       may interact with tissue thiols (Haley, 1978).  Furthermore, cytochrome P-450 mixed-function
26       oxygenases may form an epoxide intermediate similar to that formed during 1,3-butadiene
27       metabolism.
28              In summary, in vitro metabolism studies have shown that the structurally similar
29       isoprene is metabolized in a similar fashion by several different species and that epoxybutene
30       intermediates are formed, one of which may be epoxidized further to a genotoxic
31       diepoxybutane. In vivo inhalation studies that used rats and mice exposed to isoprene showed
32       that its uptake and fate are similar to that of 1,3-butadiene and that a genotoxic diepoxybutane
33       metabolite, but not a genotoxic epoxybutene intermediate, is formed.
34              Preliminary data indicate that Hb adducts may be useful as biomarkers of exposure for
35       1,3-butadiene exposure.  Research efforts are focusing on dosimetry modeling for extrapolating
36       from high- to low-dose exposures and for interspecies extrapolation.  Furthermore,  on
         1/28/98                                    3-44       DRAFT-DO NOT CITE OR QUOTE
-------
          validation, dosimetry models may be usefiil in predicting levels of 1,3-butadiene and its reactive
          metabolites in various tissues.

  3       3.4. DISCUSSION AND CONCLUSIONS
  4              Species variability in the metabolism and disposition of 1,3-butadiene may explain, in
  5       part, species variability in the toxicity of the compound. Current data indicate that the toxicity
  6       of 1,3-butadiene depends on the metabolic activation to reactive intermediates such as
  7       epoxybutene and diepoxybutane and that these biotransformation processes vary quantitatively
  8       and qualitatively among species.  The mutagenic epoxybutene and diepoxybutane metabolites
  9       have been shown to occur in the blood of rats and mice exposed to 1,3-butadiene, and their
 1 0       concentrations are two- to fivefold greater in the blood of mice. Limited data for humans have
 11       shown that liver microsomes have a higher capacity for the formation of epoxybutene than do
 1 2       rodent liver microsomes but that the metabolism of epoxybutene to 1,3 -butadiene epoxide by
 1 3       human liver microsomes was 20-fold greater than that observed in rat or mouse microsomes.
 14       These data suggest that levels of this reactive intermediate in humans may be substantially less
 1 5       than in the rodent species. The oxidation of epoxybutene to diepoxybutane (also a reactive
 1 6       metabolite) appears to be negligible in humans and rats (formation of the non-DNA-reactive
 1 7       butene diol l,2-dihydroxybut-3-ene is the preferred pathway) and is substantial in mice.  Study
          results have shown species-related differences in the uptake and retention of inhaled 1,3-
  9       butadiene. Uptake and retention by mice is greater than for rats, and saturation kinetics are
 20       observed in mice at exposure concentrations of 500 ppm but not in rats at exposures as high as
 21        5,000 ppm.  These differences may be used to support the hypothesis that the greater sensitivity
 22       of mice to the toxic effects of 1,3-butadiene may be a function of a greater internal dose, greater
 23       production of reactive metabolites, and lower detoxification potential.
 24             Although the previous findings provide considerable insight into the understanding of
 25       1,3-butadiene toxicity, some researchers have indicated the need for examining additional,
 26       although quantitatively minor, metabolic pathways (e.g., glutathione  S-transferase-mediated
 27       detoxification processes and formation of toxic metabolites such as butene diol and
 28       crotonaldehyde) and the possible effects of pulse exposures on the metabolism and disposition
 29       of 1,3-butadiene.
 30             Molecular dosimetry studies have also shown species-related differences in the
 31       formation of various adducts.  Additional work in this area will be useful in assessing these
32      adducts as either biomarkers of exposure or effects.
3 3             Dosimetry models are being developed or refined to extrapolate the relatively high
34      exposures and doses used in animal tests to the low exposure concentrations in human exposure


        1/28/98                                   3-45       DRAFT-DO NOT CITE OR QUOTE
-------
1      situations. These models will be especially useful in predicting blood and tissue concentrations
2      of butadiene metabolites.
       1/28/98
3-46
DRAFT-DO NOT CITE OR QUOTE
-------
                                          4.  MUTAGENICITY

        4.1. INTRODUCTION
  2            The mutagenic effects of 1,3-butadiene have been reviewed extensively (Rosenthal, 1985;
  3     de Meester, 1988; Arce et al., 1990; Norppa and Sorsa, 1993; Jacobson-Kram and Rosenthal,
  4     1995). The last of these reviewed publications through 1994 are on the genetic effects associated
  5     with butadiene (and metabolites). There is extensive evidence that butadiene and the two
  6     primary epoxide metabolites (epoxybutene and diepoxybutane) induce genotoxic effects in a
  7     variety of in vitro and in vivo test systems. Most of the in vivo studies discussed in the cited
  8     reviews were assays in mice and rats using cytogenetic endpoints, and the results generally
  9     support the dichotomy in carcinogenic response where mice are more responsive than rats. This
10     review will focus on recently published studies performed in vivo (both somatic  and germ cell
11      effects) with an emphasis on those studies providing information relative to the mode of action of
1 2     butadiene metabolites.
13
14     4.2. GENE MUTATIONS
1 5            Most of the earlier in vivo genotoxicity studies used cytogenetic endpoints (aberrations,
        micronuclei, or  sister chromatid exchange [SCE]).  It is recognized that this reflected the dearth
        of in vivo assays measuring gene mutations and limited the interpretation  of in vitro versus in
1 8     vivo findings. The ability to detect mutations at the hprt locus obtained from T lymphocytes
1 9     from exposed mammals including mice, rats, monkeys,  and humans provides an important step
20     in developing an understanding of chemically induced mutational processes.  Cochrane and
21      Skopek (1993, 1994a) used E6C3F1 mice and human TK6 cells to evaluate the mutagenic
22     potential of butadiene and the two major metabolites. In the  in vivo studies, mice were exposed
23     for  6 h/day, 5 days/week for 2 weeks to butadiene at 625 ppm.  The induced hprt mutant
24     frequency was 6.2 x icr6 compared with  1.2 x 10'6 from unexposed controls. For the
25      metabolites, mice received three daily intraperitoneal (i.p.) injections of 60, 80, or 100 mg/kg of
26      epoxybutene or 7, 14, or 21 mg/kg of diepoxybutane. Mutant frequencies in hprt from splenic T
27      cells were dose related for both metabolites, with maximal responses of 8.6 x icr6 and 13 x 1Q'6
28      for  epoxybutene and diepoxybutane, respectively. Similarly, they found diepoxybutane about
29      100 times more effective than epoxybutene when human lymphoblastoid TK6 cells were treated
30     in vitro.
31             In a recent meeting presentation,  Meng et al. (1996) reported on a study in which both
32      mice and rats were exposed by inhalation to 1,250 ppm butadiene for 2 weeks (6 h/day, 5
        days/week). Groups of animals were necropsied before exposure (controls) and weekly up to 10

        1/28/98                                   4-1        DRAFT-DO NOT CITE OR QUOTE
-------
  1      weeks after the last exposure.  The researchers measured hprt mutants in both spleen and thymus
  2     using the T-cell cloning assay.  Mutant frequencies in both tissues of both species increased for
  3     several weeks and then declined. Maximal frequencies were:  in thymus, 1.3 x 10~6 in mice (2
  4     weeks) and 4.9 x 10"6 in rats (3 weeks); in spleen, 19.7 x 10'6 in mice (5 weeks) and 8.4 x 10'6 in
  5     rats (4 weeks). They determined a relative mutagenic potency (RMP) as the ratio of cumulative
  Q     increase in mutant frequency in treated versus controls.  For the spleen the RMP was 7.18 for
  7     mice compared with 2.04 for rats.
  8            Several recent studies have measured in vivo mutations using the phage lad or lacZ
  3     genes incorporated into a rodent genome. Recio and Goldsworthy (1995) summarized several
10     experiments in which male B6C3F! lad transgenic mice were exposed to 62.5, 625, and  1,250
11      ppm butadiene (6 h/day, 5 days/week) for 4 weeks. Two weeks after the last exposure, animals
12     were euthanized and DNA was extracted from bone marrow to be examined for lad
13     mutagenesis. Mutations 'increased in a dose-response manner, reaching an apparent plateau at
14     625 ppm (about a fourfold increase above controls). Sequence analysis of lad mutant colonies
1 5     from the 625 and 1,250 ppm groups indicated an increased frequency of point mutations at A:T
16     base pairs. These findings are consistent with those observed in butadiene-induced hprt mutant
17     T lymphocytes from B6C3Fj mice (Cochrane and Skopek, 1994b).
1 8            Several studies of genetic effects in exposed workers have recently been reported. Ward
19     et al. (1994, 1996b) measured the frequency of hprt mutations in lymphocytes of workers in a
20     butadiene production plant (two studies) and in a styrene-butadiene rubber plant. In the first
21      study exposure estimates were based on 8 h  samples in two production areas and in a central
22     control area. Mean butadiene concentration in the production areas  was 3.5 ppm, but the
23     majority of samples showed concentrations below 1 ppm; mean butadiene concentration in the
24     control was 0.03 ppm. Variant frequencies at the hprt locus in PHA-stimulated peripheral blood
25     T cells of a high exposure group were increased more than threefold compared with the low-
26     exposure and nonexposed groups. The eight individuals in the high-exposure group had hprt
27     variant frequencies varying from 0.94 x 10'6  to 8.98 x i(r6 and the variant frequency generally
28     correlated with the level of the metabolite dihydroxybutane in the urine. Whether the difference
29     was due to differences in exposure or genetic differences in metabolism cannot be ascertained
30     from the data.  A second study was conducted in the same plant about 1 year later (Ward et al.,
31      1996b). Measured butadiene concentrations in personal samplers were markedly lower, 0.30 ±
32     0.59, 0.21 ± 0.21, and 0.12 ± 0.27 ppm in areas defined as high, medium, and low exposure (no
33     controls were reported for the second study). The corresponding hprt variant frequencies were
34     5.33 ± 3.76, 2.27 ± 0.99, and 2.14 ± 0.97 x  10'6, respectively.  Individual data were not reported
35     for this study, but again there is a high standard deviation in the highly exposed group. The

        1/28/98                                   4-2       DRAFT-DO NOT CITE OR QUOTE
-------
  1     Ward et al. (1996b) paper also reported preliminary results from workers in a styrene-butadiene
  2     rubber plant. Workers were assigned to high (20 of 40 personal samplers exceeded the 0.25 ppm
 *3     detection limit and 11 had a concentration over 1 ppm) and low (none of 26 exceeded the
  4     detection limit) exposure groups. In nonsmokers, the hprt variant frequencies were 7.47 ± 5.69
  5     and 1.68 ± 0.85 x 10'6 for the high and low groups, respectively.  While the variant frequency for
  Q     smokers in the high-exposure group (6.24 ± 4.37) was not different from nonsmokers, the
  7     frequency for smokers in the low exposure group was about twice the nonsmoker group (3.42 ±
  8     1.57).  These preliminary findings with small sample sizes and no  detail about smoking history
  9     or other confounding factors raise several unanswerable questions.  The autoradiographic
10     procedure for detecting hprt variants was used in these studies. The limitation of this method is
11     that it is not possible to distinguish between several independent mutations and a single mutation
12     giving rise to a clone of cells with the mutant phenotype.  The procedure using the T lymphocyte
1 3     cloning assay and subsequent DNA sequence analysis of clones as described by Albertini et al.
14     (1982), and Recio et al. (1990) provide sufficient data for ascertaining independent mutational
15     events.
1 6            Hayes et al. (1996) employed the T cell cloning assay to detect mutant frequencies in
1 7     lymphocytes of workers in a rubber production factory. Butadiene levels were measured using
  8     personal samplers during the 6-h work shift and expressed as 6-h time-weighted average. These
        were supplemented with several grab samples. Three different work areas were identified: initial
20     distillation and recovery from dimethyl fonnamide, polymerization, and recovery, with median
21     air levels of 3.5, 1.0, and 1.1, respectively. The T cell cloning assay was performed from
22     postshift blood samples. Unexposed subjects were age  and gender matched and a brief
23     questionnaire was administered.  Tabular hprt mutant frequencies were presented grouped only
24     by gender and exposed versus unexposed. Mean mutant frequencies were somewhat higher in
25     females than males.  Smoking (in males only) was not different in  either group, but mutant
26     frequency did significantly increase with age. Mean mutant frequencies, raw and adjusted for
27     age, sex, cloning efficiency, and exposure status, were similar in exposed and nonexposed
28     workers. Adjusted mean frequency for total exposed workers was 18.0 x 10'6 compared with
29     13.6 x  10~6 for nonexposed workers.
30            In a third study,  Tates et al. (1996) used the T cell cloning assay on blood samples
31      collected from workers in a butadiene plant in the Czech Republic. Workers were sampled in
32     1993 and 1994, but most of the blood samples from. 1993 were lost to technical errors.  A
33     detailed analysis was conducted on the later group of 19 exposed and 19 nonexposed workers
34     from other parts of the same plant.  Personal samplers indicated a mean butadiene concentration
        of 1.76 ppm, with individual samples ranging from 0.012 ppm to 19.77 ppm.  The geometric

        1/28/98                                   4-3       DRAFT-DO NOT CITE OR QUOTE
-------
  1      mean hprt mutant frequencies (adjusted for age, smoking, and cloning efficiency) were 7.10 x
  2     10"6 for exposed and 10.59 * 10"6 for the controls. The range of mutant frequencies among
  3     individuals was similar for both groups and individual mutant frequencies in the exposed group
  4     were not correlated with concentrations of butadiene detected in the personal samplers.
  5            The results in both of the T cell cloning assay groups are clearly in conflict with the Ward
  6     et al. (1994, 1996b) findings both for exposed versus nonexposed and for smokers versus
  7     nonsmokers. A simple explanation would be that the increase in the autoradiographic assay was
  8     due to clones of mutants having arisen from earlier mutations. Even if true, the increase is
  9     clearly exposure related because 7 of the 8 exposed workers exhibited higher variant frequencies
10     than the highest of the nonexposed controls. As indicated by Hayes et al. (1996), there are many
11      differences between the two studies and currently no basis for rejecting either finding.

12     4.3.  CYTOGENETIC EFFECTS—HUMAN
13            There have been four studies evaluating cytogenetic effects of exposed workers.  Au et al.
14     (1995) measured chromosome aberration frequencies in blood samples of 10 exposed workers
1 5     and 10 matched controls from the same population used in the Ward et al. (1996b) study cited
1 6     above.  They reported measurable, but not significant (p>0.1), increases in chromosome
17     aberrations  and chromatid breaks.  Also, cells were exposed to gamma-rays in Gl and
1 8     aberrations were measured in the subsequent metaphase. With this indirect measure of DNA
19     repair, chromatid breaks, deletions, and dicentrics were all significantly higher in cells from
20     butadiene-exposed workers.
21             Sorsa et al. (1994) investigated chromosomal damage in blood lymphocytes sampled in
22     1993 from workers in the factories described by Tates et al. (1996)  above. Chromosome
23     aberrations, micronuclei, and sister chromatid exchange (SCE) frequencies were not elevated
24     above samples from unexposed persons. They did note that smoking had a slight effect in
25     micronuclei and SCE but not chromosome aberrations. Preliminary data measuring chromosome
26     aberrations  and micronuclei in blood samples from the 1994 group of workers was reported by
27     Tates et al.  (1996).  The percentage of aberrant cells was significantly increased (/K0.01) in
28     exposed subjects; however, the frequency of micronuclei in lymphocytes was similar in exposed
29     and unexposed subjects. Evaluation of data for each subject would be required to determine the
30     basis for the apparent discrepancy of the results between the two years.
31             The role of glutathione S-transferase (GST) genes GSTM1 and GSTT1 enzymes in the
32     detoxification of butadiene metabolites has been evaluated by measuring the induction of SCE in
33     cultured human lymphocytes. Uuskula et al. (1995) found that SCE induction in lymphocytes
34     from GSTMl-null individuals was 31% higher than in lymphocytes from GSTM1-positive

        1/28/98                                  4-4       DRAFT-DO NOT CITE OR QUOTE
-------
        individuals when treated with 50 or 250 |iM l,2-epoxy-3-butene. The same group (Norppa et al.,
        1995) reported no difference in SCE induction among GSTM1 nulls and GSTM1-positive
        lymphocytes when treated in vitro with diepoxybutane; however, they observed a 60% increase
  4     in SCE in lymphocytes from GSTT1 -null individuals when treated with 2 or 5 uM
  5     diepoxybutane. Neither GSTM1 nor GSTT1 deficiency affected the induction of SCE by 250 or
  6     500 |iM of 3,4-epoxybutane-l,2-diol (Bernadini et al., 1996).  In a separate study, Kelsey et al.
  7     (1995) found that GSTT1 deficiency significantly increased the frequency of SCE induced by
  8     diepoxybutane in lymphocyte cultures of workers exposed to butadiene. Hence while all three
  9     epoxides of butadiene metabolism are effective inducers of SCE  in cultured human lymphocytes,
10     there  are differences in the role of at least two of the GST genes (GSTM1 and GSTT1) in the
11      detoxification of the three metabolites.

12     4.4. CYTOGENETIC EFFECTS—RODENT
1 3            Most of the rodent in vivo cytogenetic studies on butadiene—especially in somatic
14     cells—have been thoroughly treated in the reviews cited in the introduction of this chapter.  In
1 5     those  studies, positive results were reported for all cytogenetic endpoints studied in mice and
1 6     negative results were consistently reported in rats.  Recent efforts have focused on cytogenetic
V7     effects in germ cells of butadiene as well as effects of the two primary epoxides of butadiene.
               Butadiene induced dominant lethal effects in studies of male mice exposed by inhalation
1 9     (Adler and Anderson, 1994); the details are described in Chapter  4.  That study was followed by
20     an experiment measuring heritable translocations induced in exposed males (Adler et al., 1995).
21      Males were exposed by inhalation to butadiene at 1,300 ppm for 5 days for 6 h/day.   Offspring
22     were tested for translocations by both litter size and cytogenetic  analysis of meiotic and somatic
23     cells.  The translocation frequency from treated males was 2.7%  compared with 0.05% for
24     historical controls.
25            Xiao and Tates (1995) evaluated the cytogenetic effects of 1,2-epoxybutene  (EB) and
26     l,2:3,4-diepoxybutane (DEB) in both somatic and germ cells of mice and rats. Male animals of
27     both species received single i.p. injections of 40 or 80 mg/kg of EB. Animals were sacrificed at
28     various time intervals after treatment and spleen and testes were  processed for scoring of
29     micronuclei. In splenocytes, EB was almost four times more effective in the mouse as in the rat.
30     In mouse germ cells, the incidence of micronuclei was similar to  controls on days 1 and 3 after
31      exposure, but was significantly increased on day 14. In rats, EB  was equally effective on days 1
32     to 3 (late spermatocytes) and day 20 (early spermatocytes) and the  frequency of micronuclei at
33     80 mg/kg was slightly higher than that observed on day 14 in the mouse. For DEB, mice were
        injected with 15 or 30 mg/kg and rats received single i.p. injections of 20, 30, or 40 mg/kg.  In a

        1/28/98                                   4-5       DRAFT-DO NOT CITE OR QUOTE
-------
  1      separate experiment, rats received 3 daily injections of 10 mg/kg. The response in splenocytes
  2     was similar in both mice and rats at 30 mg/kg.  In mouse germ cells, DEB increased the
  3     frequency of micronuclei only on day 3 after treatment. Significant increases of micronuclei
  4     were observed in rat germ cells at  all doses and all time periods.  The results in somatic cells are
  5     consistent with all other reports of greater sensitivity in mice than in rats.  This difference is
  6     contradicted in germ cells with rats equally (or more) sensitive to micronuclei induction by both
  7     EB and DEB. The authors offered no explanation for this, stating that more research is needed to
  8     better understand the organ and species differences.  It is noted that the strains of both species are
  9     different from those used in most other endpoint measurements.  The mice were F! males of a
10     (102 x C3H) cross of parent stocks from Adler's laboratory in Germany.  The rats were Lewis
11      rats supplied by Harlan CPB, the Netherlands.
12            Sjoblom and Lahdetie (1996) used an in vitro meiotic micronucleus assay to examine the
13     effects of EB, DEB, and l,2-dihydroxy-3,4-epoxybutane (diolEB) in seminiferous tubule
14     sections of male Sprague-Dawley rats. Tissue sections were cultured for 4 days with EB  at 100,
1 5     500, or 1,000 mol/L; DEB at 5,  10, or 20 jj.mol/L; or diolEB at 10, 50, or  100  nmol/L. The
1 6     frequency of micronuclei was increased only by DEB and the increase was clearly dose-related.
17     That EB was not effective is contrasted with the findings of Xiao and Tates (1995) above. The
1 8     authors suggest that EB requires further metabolism by P450 enzymes, which they indicate does
19     not occur in rat testes microsomes.

20     4.5. SUMMARY
21             The studies cited here along with the many earlier genotoxicity studies discussed in the
22     cited reviews provide clear evidence that 1,3-butadiene is both mutagenic and clastogenic
23     through its metabolism, primarily due to the mono- and diepoxide. While the difunctional DEB
24     is clearly more effective than the monofunctional EB for most endpoints, it is not possible to
25     ascribe the effects observed to one or the other when animals are exposed to butadiene. Where
26     both have been studied, mice are more responsive than rats, except for the recent germ cell
27     studies. Whether this exception is strain specific (among or between species) can only be
28     answered with future work.
29            The role of GST is also clearly established for the genotoxic effects of butadiene in
30     human lymphocytes.  That the two glutathione S-transferases (GSTM1 and GSTT1) react
31      differently with the three epoxide metabolites suggests that the relative concentrations of these
3 2     metabolites will vary depending  on the individual's genotype.
        1/28/98                                    4-6        DRAFT-DO NOT CITE OR QUOTE
-------
                       5. REPRODUCTIVE AND DEVELOPMENTAL EFFECTS

         5.1. REPRODUCTIVE EFFECTS
  2            Several reproductive toxicity studies for 1,3-butadiene have been undertaken, starting
  3     with a study in rats, guinea pigs, rabbits, and dogs by Carpenter et al, 1944.  Two studies by
  4     Owen and coworkers were done in rats (Owen et al., 1987; Owen and Glaister, 1990). NTP
  5     conducted two chronic reproductive toxicity studies in mice (NTP, 1984; 1993). Hackett and co-
  6     workers undertook an acute "sperm head morphology" study in B6C3F1 mice (Hackett et al.,
  7     1988a) and a dominant lethal study in CD-I male mice (Hackett et al., 1988b). Dominant lethal
  8     studies, both acute and subchronic, have also been done in CD-I male mice (Anderson et al.,
  9     1993,1995) and in (102/ElxC3H/El)Fl mice (Adler and Anderson, 1994).

 10     5.1.1.  Carpenter et al., 1944
 11             Four groups, each consisting of 24 albino rats, 12 guinea pigs, 4 rabbits, and 1 dog, were
 12     exposed to 0, 600, 2,300, or 6,700 ppm 1,3-butadiene 7.5 h/day, 6 days/week for 8 months in
 13     546-L chambers. Except for the dogs, which were all female (only one in each group), the
 14     animals were divided equally between the two sexes. Body weights were measured weekly;
 1 5     blood was analyzed monthly; and urinalysis, blood chemistry, organ weights (kidney and liver),
         and gross and histopathologic examinations were performed at termination. Males and females
         were mated, but the authors did not indicate when this occurred relative to the treatment period.
 18     No deaths were noted in the exposed animals. Terminal body weights in rats were reduced to
 1 9     90.5%, 86.3%, and 81.2% in the 600, 2,300, and 6,700 ppm groups, respectively, relative to
 20     control body weights. A similar trend was noted for male guinea pigs, and weights for dogs and
 21      rabbits  fluctuated. No effects on organ weights that could be attributed to exposure to
 22     1,3-butadiene were observed. There were no abnormal findings for hematology values or blood
 23     chemistry. Microscopic lesions were not observed in the testes, ovaries, or other organs
 24     examined (heart, kidney, skeletal muscle, pancreas, or spleen) except for the liver, in which mild,
 25     cloudy  swelling was noted in 68% of the animals exposed to 6,700 ppm.
 26            Carpenter et al. (1944) provided a few results regarding fertility of rats, guinea pigs, and
 27     rabbits  exposed to 1,3-butadiene.  Fertility, defined as the number of litters produced within a
 28     given time, was reduced in rats, with 3.3, 2.7, 2.5, and 2.6 litters being produced by animals
 29     exposed to 0, 600, 2,300, or 6,700 ppm, respectively. Because the results were not analyzed
 30     statistically and other details regarding the duration of the mating periods were not presented, it is
 31      not possible to conclude that 1,3-butadiene had an effect on fertility in rats. Furthermore, fertility
^32     in rats was not affected by exposure to 1,3-butadiene  when litter size (8.4 pups/litter at 600 ppm,
         7.9 pups/litter at 2,300 ppm, and 7.8 pups/litter at 6,700 ppm) was used as the measure; the
         1/28/98                                  5-1        DRAFT-DO NOT CITE OR QUOTE
-------
 1      average litter size of the two higher exposure groups was similar to that of the control group.
 2      Two male and two female offspring from rats exposed to each concentration were exposed along
 3      with the parents. According to the authors, the F! controls and the 660-ppm group produced
 4      three times as many pups as did the F1 groups exposed to 2,300 or 6,700 ppm. Too few animals
 5      were used to adequately evaluate the fertility of the exposed offspring. Guinea pigs in each
 6      exposure group produced 16, 13,10, and 13 pups, respectively. Rabbits exposed to 600 or 2,300
 7      ppm produced no pups, whereas the controls produced 24 pups and the 6,700 ppm group
 8      produced 27 pups. Considering that the highest concentration had no effect on fertility in rabbits,
 9      it is doubtful that the lack of fertility at the lower concentrations was due to exposure to
10      1,3-butadiene.

11      5.1.2. Owen et al., 1987; Owen and Glaister, 1990
12             This 2-year toxicological and carcinogenicity study is the same as the Hazleton
13      Laboratories Europe, Ltd. (HLE, 1981), study discussed previously by EPA (U.S. EPA, 1985).
14      Male and female CD strain (Sprague-Dawley derived) rats (110 of each sex per group) were
15      exposed by inhalation to 1,3-butadiene (99.2%  purity) at target concentrations of 0, 1,000, or
16      8,000 ppm 6 h/day, 5 days/week for 105 (females) or 111 (males) weeks. Ten males and 10
17      females were killed at 52 weeks. The average weekly concentration of 4-vinyl-l-cyclohexene (a
18      1,3-butadiene dimer) was 413 ppm (v/v). A comprehensive postmortem examination, including
19      necropsy and histopathologic examination, was conducted of all gross lesions, all tissues from
20      control and high-exposure groups, and  selected tissues from low-exposure groups.
21      Nonneoplastic lesions were not induced in reproductive organs in either male or female rats,
22      although benign and malignant mammary tumors, uterine sarcomas, and Leydig cell tumors were
23      observed.

24      5.1.3. NTP, 1984
25             The first inhalation toxicological and carcinogenicity study conducted by the National
26      Toxicology Program (NTP, 1984) showed that, in addition to the numerous neoplasms induced
27      by high concentrations of 1,3-butadiene in male and female B6C3F! mice, nonneoplastic lesions
28      also were induced in reproductive organs. Male and female mice were exposed to 0, 625, or
29      1,250 ppm 1,3-butadiene 6 h/day, 5 days/week and then killed after 60 or 61 weeks of exposure.
30      Among female mice, ovarian atrophy was seen in 40/45 (89%) mice exposed to 625 ppm and in
31      40/48 (83%) mice exposed to 1,250 ppm, compared with an incidence of only 2/49 (4%) in
32      control mice. Involution of the uterus,  which was considered a manifestation of ovarian atrophy,
33      was seen in 7/46 (15%) and 14/49 (29%) mice  exposed to 625 and 1,250 ppm, respectively,
34      compared with 0/49 control mice.  Uterine involution was characterized by fewer and less
        1/28/98                                   5-2       DRAFT-DO NOT CITE OR QUOTE
-------
  1      prominent endometrial glands. A low incidence of mammary gland neoplasms (acinar cell and
  2      adenosquamous carcinomas) was induced by 1,3-butadiene; nonneoplastic mammary lesions
  3      were not induced.  Testicular atrophy was observed in 19/47 (40%) mice exposed to 625 ppm
  4      and in 11/48 (23%) mice exposed to 1,250 ppm compared with 0/50 control mice. Statistical
  5      analysis showed that the increased incidences of the lesions in male and female mice were
  6      significant (p<0.05) for all groups compared with their respective controls.

  7      5.1.4. NTP, 1993
  8            NTP (1993) conducted a second inhalation toxicological and carcinogenicity study in
  9      male and female B6C3F3 mice exposed to lower concentrations of 1,3-butadiene. Concentrations
 10      were 0, 6.25,20, 62.5, 200, or 625 ppm 1,3-butadiene for 6 h/day, 5 days/week for 103 weeks,
 11       with interim evaluations at 9 and 15 months. Additional male mice were exposed to 200 ppm of
 12      1,3-butadiene for 40 weeks, 312 ppm for 52 weeks, or 625 ppm for 13 or 26 weeks followed by
 1 3      observation for the remainder of the 2 years (stop-exposure protocol).  It should be emphasized
 14      that this study was designed to study neoplastic and general toxicological, rather than
 15      reproductive, endpoints. Further details are presented in Chapter 6.
 1 6            The effects of 1,3-butadiene  on reproductive organs in female mice are presented in Table
 17      5-1. Ovarian atrophy was seen in the 200 ppm and 625 ppm exposure groups sacrificed for the
 18      9-month interim evaluation. The atrophic ovaries were characterized by the absence of oocytes,
 1 9      follicles, and corpora lutea. No occurrences of this lesion were noted in the lower exposure
 20      groups. Hyperplasia of the germinal epithelium was observed in one animal exposed to 625 ppm
 21       for 9 months. Germinal epithelial hyperplasia was described as prominent down growth of the
 22       mesothelial surface into the parenchyma of the ovary, forming tubular and gland like structures.
 23      At the 15-month interim evaluation,  ovarian atrophy was observed in mice exposed to 20 ppm or
 24      higher; the incidence at 62.5 ppm or higher was significant compared with concurrent controls.
 25      Hyperplasia of the germinal epithelium was seen at 200 and 625 ppm at nonsignificant
 26      incidences. Angiectasis (dilation of blood vessels) was seen in one mouse in the control group,
 27      one exposed to 6.25 ppm, and two exposed to 200 ppm. The ovary, which was evaluated at 15
 28      months in only two female mice exposed to 625 ppm, was atrophic in both.  Among female mice
 29      exposed to  1,3-butadiene for 2 years, ovarian atrophy was observed in all exposure groups at
 30      incidences that were significantly elevated compared with controls. Therefore, using ovarian
 31      atrophy as an endpoint of reproductive toxicity, a no-observed-adverse-effect level (NOAEL)
3 2      could not be defined in this mouse study. The incidence of angiectasis was significantly elevated
33      only at 62.5 and 200 ppm, and the incidence of germinal epithelial hyperplasia was significantly
34      elevated at 20 to 625 ppm.  The occurrence of ovarian atrophy and germinal epithelial
 p 5      hyperplasia showed significant dose-related trends, whereas ovarian
        1/28/98                                   5-3        DRAFT-DO NOT CITE OR QUOTE
-------
a
•2
«
cs
•a
C
O
••a
"S
o
•*»•»
en
0
&

0

o
0
v^
cs

-A_^
o
o




^K^
o







o
o

i
o





n.
S





1
~


/— ^
m
s.
o





1
o







o
o

i
o



KJ
— i _ cn
KJ CO CS
•l^ &
§5 g.
S&^





o
•• — '
c->


/— x
0
i
o





|
o






f 	 s
o
V 	 '

o
	 .



VI
a
o
o
'So
<





cs




o
0





o







_^


o







°cn
P
0
3

o
o
N— '
0)


o
o
I
o





o
0







1'
o

1
o




0
1

















•a
•1

a
si
i
«


















ON




o
o





oo
••3-







ON


ON







a
£r*
>
O

f? 0
ON ^


SO" »-H
It
c>^ O
•* O
5So




Co* '•^
0s- O
c^ o
i£-S






''"a*' |
ON §
2K

g-S
^K





o,
o





^ 0
en o
£So
OO *


C^o
0s- *— "
o o
It
^ 0
0 0
C-o




y~ ^. t^~-
CO CX,






^
^i§
- Ik

^ o
^L



ci
— — w
C3 KJ CT
ll&
Iff





11
£"" II


C,p^ i— «
oo o
st
•o o




Vf>
C^o






^-v VQ
CM «^
^O Ci,

S?g
* t,



CO
&o
C3
8
•a
<





oo
t^




0
u-i
ON





o
in







ON


O
u-j







"co

b
P

en
T— 1


S—
1
OO
1




^_^
C4







i
o

^
—




">-
Atroph




                                                                                    -
                                                                                 *S w
                                                                                 Q O
                                                                                        a
                                                                                        ~ 2
                                                                                        .2 M


                                                                                        si
                                                                                        •rt ®
                                                                                        J5
                                                                                        «§*

                                                                                        >. i;

                                                                                        "2 §
                                                                                        to °
                                                                                        i- en

                                                                                        S:§
                                                                                   fe « QJ .2 '
                                                                                 .i:!.! :i § §•
                                                                                 1^1^11

                                                                                       S.|

                                                                                       o 3
                                                                                         >•!
                                                                                 •
                                                                                  £ g. o &.§ "s
                                                                                  t2 S „ S .« s
                                                                                  J2 S S 8 S 5
                                                                                  § « 5= S " •=
                                          ^ "« *^ '« ,O o
                                          O Q.-S o.'" S

                                          V-  n O  "
                                          u _ .= ^
                                          •= O en O w..

                                          c o '5 o g £
                                          3 w Py \/ >• .—
                                                                                       0 3 -
                                                                                  = o '-5 o
                                                                                  = v S v
                                                                                              o-
1/28/98
5-4
DRAFT-DO NOT CITE OR QUOTE
-------
         angiectasis did not. Although the functional integrity of the female reproductive system was not
         assessed, it can be assumed that animals without oocytes or follicles would be infertile and would
         express reduced estrogenic and progestin secretory capacities.
  4            Uterine atrophy was seen at the two highest concentrations at 9 months, but was seen only
  5      at the highest concentration at the 15-month evaluation. After 2 years, the incidence of uterine
  6      atrophy among mice exposed to 200 and 625 ppm did not increase relative to that observed at 9
  7      months.
  8            Data regarding the effect of 1,3-butadiene on the reproductive organs of male B6C3Fj
  9      mice are summarized in Table 5-2.  The testes of males exposed to the highest concentration of
 10      1,3-butadiene (625 ppm) were atrophic at the 9- and 15-month interim evaluations and at
 11      termination of the 2-year study.  Among male mice exposed to 1,3-butadiene in the stop-
 12      exposure studies, testicular atrophy was observed in only five mice exposed to 200 ppm (40
 1 3      weeks), five exposed to 625 ppm (26 weeks), three exposed to 312 ppm (52 weeks), and three
 14      exposed to 625 ppm (13 weeks). It is not possible to determine if the lack of a more prominent
 1 5      response in mice exposed to 625 ppm for 26 weeks was due to insufficient time for induction of
 1 6      testicular atrophy or if atrophy had been induced during exposure and the lesion repaired before
 17      termination of the stop-exposure study.

         5.1.5. Hackett et al., 1988a
               This sperm-head morphology study was conducted in B6C3F, mice at Pacific Northwest
 20      Laboratories for NTP as part of a series of studies to investigate the effects of 1,3-butadiene on
 21      reproductive function. Twenty male B6C3Fi mice (12 to 13 weeks old) per group were exposed
 22      to 1,3-butadiene (99.88% purity; 174 ± 13 ppm mean headspace dimer [4-vinyl-l-cyclohexene]
 23      concentration) at concentrations of 0 (filtered air),  200,1,000, or 5,000 ppm 6 h/day for 5
 24      successive days. Measured concentrations (mean ± standard deviation [SD]) were 199 ± 6.12,
 25      999 ± 22.6, and 4,980 ±130 ppm. The animals were exposed in a 2.3 m3 stainless steel chamber
 26      with a mixing volume of 1.7 m3.  Positive controls received intraperitoneal injections of 167
 27      mg/kg of ethyl  methane sulfonate daily for 5 consecutive days. After exposure, the mice were
 28      observed twice daily for mortality, morbidity, and  signs of toxicity; body weights were
 29      determined weekly.  The mice were killed 5 weeks after exposure, weighed, and examined for
 30      gross lesions, with particular emphasis on the reproductive tract.  Sperm collected from the right
 31      epididymis were examined for abnormal heads (blunt hook, banana, amorphous, pinhead, two
 32      heads/two tails, short) and other abnormalities (primarily midpiece abnormalities).
 33            Final body weights for the unexposed, treated, and positive control groups were similar,
J34      and net body weight gain over the period of the experiment was also similar for all groups.
         Piloerection and dyspnea were observed within the first 20 to 30 min after exposure in mice
         1/28/98                                    5-5        DRAFT-DO NOT CITE OR QUOTE
-------
 O
 o
  I
 to,
 T-K*
 O
 o
 CO
 o
 O
 • M
 O
  *-*

 0
 vo

 M

 i
 a
 o
 'S
 2
 2
 a,
 o
 «
 2
 cs














•§
c<
&

a
2
c
o
1





















3




1




ift







0




in

vo







o





CJ
.2
J






















g
1
"«
"C
.gj
_g
1


























0




o
— H





0






o





o








o









To

1
J3
CO
8
o
+1
OS
2.
o
,,
o
p
o
-H
s
T-M
O
^
O
O
+1
S
T-H
o"


8
o
-H
2
.
o
8
O
+1

r—
*""*
o


o
o
-H
r-
^^
o

3
•w
•f)
S3
—


CO

S
o
•H
S


0
C)
-H
in
oS

s
o
-H
oo
c^



OS
0
o
-H
^
c-i

OS
p
o
•H

S
CN


vo
o
o
-H
oo
oi

•2?
"Sb
^
•ob
1
u


1 ™


feS
i
vo



s.
o




1






1





"l







^
g.
o







•v.

a.
S


















§
1
1
s
1

5
1


























f-




o
1 — C





O






O





O








O










4)
CO

s
o
0
+1
§
o
JO
CO
8
o
-H
8
T~H
0
CO
8
o'
-H
2
.
o


o
o
o
+1
o
.
0
§
o
-H

CO
1 '
o
CO
0
o
o
-H
VD
1 '
O

s
^J
"5
'S
V
3

1

o
o
+1
2


0
o
-H
OS
CO
r-i

g
o
+1
vo
vo
oi



o
o
-H
oo
oi

oo
o
o
-H

OS
r-
oi


r-
0
o
-H
VO
c4

CD
"bb
4-*
•s
1
CD
>

CO ^
"S fQ


S?
En
*



e-
o




1






, — ^
o




1







o
g,
0







"^-^

c.
e




















1
i-

                          ta n\ •*
1/28/98
5-6
DRAFT-DO NOT CITE OR QUOTE
-------
  1      receiving 5,000 ppm; no signs of toxicity were noted for the other groups. Exposure-related
  2      gross toxicity was not observed in the reproductive tract.  The percentages of epididymal sperm
 *3      with normal morphology were 98.08%, 97.23% (p<0.05), and 96.34% (p<0.05) at 200, 1,000,
  4      and 5,000 ppm, respectively, compared with 98.40% for controls; these values also showed a
  5      significant exposure-related trend (psO.05). The percentage of the following abnormalities were
  6      significantly elevated compared with controls (p<0.05):  blunt hooks at 5,000 ppm, bananas at
  7      1,000 and 5,000 ppm,  and pinheads at 1,000 ppm.  Amorphous, two heads/two tails, and shorts
  8      were not significantly  elevated at any dose.  The predominant types of abnormalities were the
  9      banana followed by blunt hook and amorphous. The authors speculated that late spermatogonia
 10      or early primary spermatocytes were sensitive to 1,3-butadiene. The authors also stated that
 11      examining the sperm at only one time point following termination of exposure precluded a
 12      determination of the stage of spermatogenesis affected by the chemical.

 13      5.1.6. Hackett et aL, 1988b
 14            This dominant  lethal study was conducted using proven breeder male CD-I mice (20 per
 15      group) exposed to 0, 200, 1,000, or 5,000 ppm 1,3-butadiene 6 h/day for 5 successive days.
 16      Measured concentrations (mean± SD) were 200 ± 5.73, 1,010 ± 13.9, and 5,000 ± 85.4 ppm,
 17     respectively.  The purity of the 1,3-butadiene was 99.88%, and the headspace dimer
         concentration was 215  ± 49 ppm. For mating, one exposed or control male mouse was placed
        with two  unexposed female mice for 1 week for 8 successive weeks; the two females were
 20     replaced each week. Male mice were sacrificed at termination of matings, and female mice were
 21      sacrificed 12 days after the last cohabitation day. The reproductive status, total number, position
 22     and status of implantations, the number of early and late resorptions, and the number of live and
 23     dead fetuses were recorded.
 24           No animals died during the study, and body weights of the exposed groups were similar
 25     to those of the control group. All males exposed to 1,3-butadiene were fertile during the 8-week
 26     exposure  period. During the first week of mating (postexposure week), the total number of dead
 27     implants was significantly elevated for the group exposed only to 1,000 ppm (p<;0.05) compared
 28     with that  of controls. Early resorptions accounted for most of the dead implants. In addition, the
 29      percentage of dead implants relative to the total implants was significantly elevated in groups
 30      exposed to 1,000 ppm (p< 0.05), and the percentage of females with more than one intrauterine
31      death was significantly elevated in all exposed groups (p^O.05) relative to controls.  During the
32      second postexposure week, the total number of dead implants was also significantly elevated at
33      200 and 1,000 ppm relative to controls.  The percentage of dead implants and the percentage of
        females with more than one intrauterine death were elevated, but not significantly.  For
        postexposure weeks 3,  5, 6, 7, and 8, the number of dead implantations, percentage of dead

        1/28/98                                    5-7        DRAFT-DO NOT CITE OR QUOTE
-------
 1      implantations, and percentage of females with more than one intrauterine death in all exposed
 2      groups were similar to those of controls (i.e., not statistically significant). For postexposure
 3      week 4, the percentage of dead implants (5,000 ppm) and the percentage of females with more
 4      than one intrauterine death (200 and 5,000 ppm) were significantly reduced (prsO.05) relative to
 5      the control value. However, the control values for these parameters were unusually high
 6      compared with control values at other postexposure weeks.  Thus, the significantly reduced
 7      values for treated mice were probably not treatment related.
 8             The results indicate that exposure to 1,3-butadiene may affect mature spermatozoa and
 9      spermatids assessed by preimplantation deaths for postexposure weeks 1 and 2.  Interpretation of
10      these results is complicated because the effects occurred in the 200 and 1,000 ppm groups but not
11      in the 5,000 ppm group, which showed no indications of toxicity.
12
13      5.1.7. Anderson et aL, 1993
14             The ability of 1,3-butadiene to induce dominant lethal mutations in male mice following
15      acute and subchronic inhalation exposure was assessed by evaluating the number of dead
16      implants  in females mated to exposed males. For acute exposures, male CD-I mice were
17      exposed to 0,1,250, and 6,250 ppm 1,3-butadiene for 6 h; 5 days later, each male was mated to
18      two females.  Males used for subchronic exposures were treated with 0, 12.5, or 1,250 ppm, 6
19      h/day, 5 days/week for 10 weeks.  Following mating in both experiments, one female was killed
20      on gestation day (gd) 17 and the other was allowed to litter for evaluation of long-term effects on
21      the offspring. Results of long-term carcinogenic effects on the live offspring are not yet
22      available. The female killed on gd 17 was examined for number of live fetuses, number and type
23      of malformations in the fetuses, and number of postimplantation deaths. The only effect seen in
24      the acute study was a decrease (p<0.05) in the number of implantations in females mated to
25      males exposed to 1,250 ppm. In the subchronic study, females mated to males exposed to 12.5
26      ppro. had  an increase in the number of late postimplantation deaths (p^O.Ol; both fetal and
27      placental tissue were present); females mated to males exposed to 1,250 ppm had a decrease in
28      mean implantations per dam (p^O.Ol) and an increase in both early (p<0.001; resorption) and late
29      postimplantation deaths (p^O.OOl).  1,3-Butadiene appears to affect the male germ cell line,
30      resulting in late postimplantation death of the fetuses. It is unknown whether the
31      mutations/alterations of the germ cells resulting in a reduction in live fetuses are due to an effect
32      on reproductive ability or a teratogenic effect resulting in death.

33      5.1.8. AdLer et aL, 1994
34             To assess the stage at which male germ cells are affected by 1,3-butadiene, male (102/E1
35      * C3H/E1)F, mice were exposed by inhalation to 0 or 1,300 ppm, 6 h/day for 5  consecutive days.
        1/28/98                                   5-8        DRAFT-DO NOT CITE OR QUOTE
-------
        Four hours after the end of exposure, each male was mated at a ratio of 1:2 to untreated virgin
        females. Females judged bred by the presence of a vaginal plug were replaced with new females,
        and mating continued for 4 consecutive weeks.  Females were killed on gd 14 to 16 and
  4     examined for numbers of live and dead implants. Exposure of male mice to 1,300 ppm resulted
  5     in an increase in dead implants during the first to the third weeks of mating; however, statistical
  6     significance (p<0.01) was reached only in the second week. When expressed as a percentage of
  7     dominant lethals, a significant increase was seen in the second (12.4%,p<0.01) and third (5.5%,
  8     p<0.05) weeks. Because of the time course for dominant lethal mutations to manifest as  dead
  9     implantations,  1,3-butadiene appears to induce dominant lethality in spermatozoa and late
 10     spermatids.

 11     5.2.  DEVELOPMENTAL EFFECTS
 1 2            The developmental toxicity study sponsored by the International Institute of Synthetic
 1 3     Rubber Producers (IISRP, 1982) is the same as the Hazleton study discussed briefly in the 1985
 14     EPA document (U.S. EPA, 1985). Hackett and coworkers also conducted two developmental
 15     toxicity studies, one using rats (Hackett et al, 1987a) and one using mice (Hackett et al.,  1987b).
 16     The study using rats was conducted to confirm and extend the findings of the IISRP (1982) study
 17     in rats, and the mouse study was conducted for comparison of a rodent species more sensitive
        than the rat to the toxic effects of 1,3-butadiene.

 19     5.2.1. IISRP, 1982
20            Female Sprague-Dawley CD rats were mated with male rats of the  same strain (2f: 1m) to
21     produce 138 sperm-positive females. Groups of mated females (220 to 266 g) were exposed by
22     inhalation to 1,3-butadiene at target concentrations of 0, 200, 1,000, or 8,000 ppm 6 h/day on gd
23     6 to 15 and killed on gd 20. Measured concentrations (mean ± SD) were 2.8 ± 1.2, 202 ± 14, 990
24     ± 24, and 7,647 ± 375 ppm for 0, 200, 1,000, and 8,000 ppm, respectively. The animals were
25     exposed in stainless steel chambers. Twenty-four pregnant females were exposed to each
26     concentration of 1,3-butadiene, 40 were exposed to filtered air (controls), and 26 were given 250
27     mg acetylsalicylic acid/kg body weight by gavage on gd 6 to 15 (positive controls). The purity of
28     the 1,3-butadiene was not reported; the mean concentration of the dimer, 4-vinyl-l-cyclohexene,
29     was 108.6 ± 53.59, well below the target of 300. The rats were weighed on gd 0, 3, 6, 9,  12,15,
30     18, and 20. Various parameters of maternal and developmental toxicity were evaluated and
31      analyzed using the litter as the statistical unit.
32           Maternal effects of 1,3-butadiene are summarized in Table 5-3. No animals died of
        exposure to 1,3-butadiene. One animal exposed to 1,000 ppm was killed because of morbidity
        unrelated to treatment.  Clinical signs of toxicity were not observed in any group, and the
        1/28/98                                   5-9        DRAFT-DO NOT CITE OR QUOTE
-------
          Table 5-3.  Maternal toxicity in Sprague-Dawley CD rats exposed to
          1,3-butadiene by inhalation
Parameter
No. dams assigned
No. of deaths
No. pregnant (%)
Whole body weight (g)
DayO
Day 20
Body weight gainb (g)
Days 0-6
Days 6-9
Days 9-12
Days 12-15
Days 15-20
Gravid uterine weight (g)
Extragestational weight (g)
Extragestational weight gain6 (g)
/
Significant clinical signs
Concentration (ppm)
0
40
0
90

239
362

24
13
16
15
54
63.9
297.9
59
None
200
24
0
91.7

238
357

24
9
13
15
58
61.1
296.2
58
None
1,000
24
la
100

240
355

23
1°
14
16
61
66.5
280.8
49f
None
8,000
24
0
95.8

239
347

23
lc
ir
15
60
62.8
283.9
45f
None
This animal was killed in moribund state on day 19; morbidity was not related to exposure to butadiene.
bDetermined from differences in group mean body weights reported for specific days of gestation.
ep<0.01, compared with corresponding control; analysis of variance and t test.
dBody weight on gd 20 minus gravid uterine weight.
"Extragestational weight minus body weight on gd 0.
'p
-------
  1      pregnancy rates were similar in all groups.  Terminal body weights showed a dose-related
  2      decrease (no statistical analysis). Maternal body weight gain was markedly depressed in dams
  3      exposed to 1,000 and 8,000 ppm, especially during gd 6 to 9; a significant decrease was also
  4      noted during gd 9 to 12 in rats exposed to 8,000 ppm. During the later stages (gd 12 to 15 and 16
  5      to 20), body weight gain was similar to controls. The gravid uterus and extragestational weights
  6      were similar to controls, but extragestational weight gain was significantly depressed by 17%
  7      O<0.05) in  dams exposed to 1,000 ppm and by 24% in dams exposed to 8,000 ppm (p<0.05).
  8      No effects were observed on other measures of maternal toxicity. Developmental effects of 1,3-
  9      butadiene are summarized in Tables 5-4 and 5-5. Fetal body weight and crown/rump length were
 10      significantly reduced at 8,000 ppm (p<0.05). The percentage of fetuses with major skeletal
 11      defects was  significantly elevated at 1,000 and 8,000 ppm, and minor skeletal defects were
 12      significantly elevated only at the lowest concentration. The percentage of fetuses showing minor
 13      external/visceral defects, predominantly subcutaneous hematomas, was significantly elevated
 14      only at 1,000 ppm, but the percentage was similar in all three experimental groups. The
 15      incidence of bilateral lens opacity was elevated at all concentrations but was significantly
 16      elevated only at 8,000 ppm. The incidence of marked-to-severe wavy ribs and the total number
 17      of abnormal ossifications and irregular ossification of the ribs were elevated at 8,000 ppm. The
 18      incidence of thoracic bipartite centers was elevated hi all exposed groups; a dose-response
 19      relationship was not observed. Other malformations and variations occurred at incidences
 10      similar to those of controls or were not significantly elevated compared with controls.

 21      5.2.2. Hackett et aL, 1987a
 22            For the  experiment with rats, 208 female Sprague-Dawley CD rats and 108 male Sprague-
 23      Dawley CD rats (all 7 to 8 weeks old) were used. The rats were mated by placing two females
 24      with one male rat overnight for 5 consecutive nights or until  a sperm-positive vaginal smear was
 25      obtained; gd 0 was the day sperm were detected.  Thirty sperm-positive female rats per group
 26      were exposed to 0, 40, 200, or 1,000 ppm 1,3-butadiene (99.84% purity;  197 ±6 ppm mean
 27      headspace dimer concentration).. The measured concentrations (mean ± SD) were 40.1 ± 0.62
 28      (mean ± SD), 199.8 ± 2.61, and 1,005 ± 11.9 ppm, respectively. On gd 6 to 15, the females were
 29      exposed for  6 h/day in stainless-steel chambers having a total volume of 2.3 m3 and a mixing
 30      volume of 1.7 m3.  The females were weighed 1 week before mating and on gd 0, 6, 11, 16, and
 31      20 (the day of sacrifice). Various parameters of maternal and developmental toxicity were
 32      evaluated. The experimental design is summarized in the upper section of Table 5-6.
 33            The  effects of inhalation exposure to 1,3-butadiene on maternal endpoints in rats are
J34      summarized in  Table 5-7. All females survived to the end of the study. No clinical signs of
         toxicity were observed. Final body weights were similar to those of controls; body weight gain,
         1/28/98                                   5-11        DRAFT-DO NOT CITE OR QUOTE
-------
        Table 5-4.  Developmental toxicity in Sprague-Dawley CD rats exposed to
        1,3-butadiene by inhalation
Parameter
No. pregnant (%)
No. litters with live fetuses
No. implantations/dam"
Preimplantation loss (%)
Postimplantation loss (%)
Total no. resorptions
Early resorptions
Dead fetuses/litter
No. fetuses/no, litters examined
Fetal body weight" (g)
Females'
Males"
Crown/rump length (mm)
Sex ratio (% males)
Concentration (ppm)
0
90
36
13.0
15.4
3.6
17
16
0
450/36
3.3
3.2
3.4
37.8
49.8
200
91.7
22
12.8
17.1
6.0
17
13
0
265/22
3.2
3.1
3.3
37.2
54.7
1,000
100
23
14.1
11.5
4.9
16
16
0
308/23
3.2
3.1
3.3
37.2
51
8,000
95.8
23
13.8
12.4
7.3
23
20
0
294/23
3.1b
3.0
3.2 '
35.9°
50
"Mean values.
ycQ.05, Wilcoxon test.
e/?<0.01, Wilcoxon test.

Source:  IISRP, 1982.
1/28/98
5-12
DRAFT-DO NOT CITE OR QUOTE
-------
          Table 5-5.  Malformations and variations in Sprague-Dawley CD rats
          exposed to 1,3-butadiene by inhalation
Parameter
Total no. fetuses/no, litters
examined
External/visceral defects
Minor defects
Major defects
Unilateral lens opacity
Bilateral lens opacity
Bilateral ureter dilation
Skeletal defects
Minor defects
Major defects
Any thoracic center (10-13)
bipartite
Wavy ribs (marked to
severe)
Wavy ribs (slight to
moderate)
Variations (abnormal ossification)
Skull (occipital)
Skull (interparietal)
Stemebrae no. 6
Ribs
Metacarpals
Phalanges
Concentration (ppm)
0
450/36
76 (16.9)a
0
19 (4.2)
18 (4.0)
13 (2.9)
72 (23.2)
2(0.6)
4(1.29)
2(0.6)
3 (1.6)
267 (85.9)
79(25.4)
82 (26.4)
152 (48.9)
0
207 (66.6)
141 (45.3)
200
265/22
63 (23.8)
0
11(4.2)
24(9.1)
8 (3.0)
49 (26.9)c
4 (2.2)
ll(6.04)d
4 (2.2)
3 (1.7)
164(90.1)
58(31.9)
66 (36.3)
107(58.8)
1 (0.55)
140 (76.9)
1 14 (62.6)
1,000
308/23
75(24.4)b
0
8 (2.6)
30 (9.7)
3 (1.6)
45 (20.9)
6 (2.8)b
14(6.51)"
3 (2.3)
7(3.3)
185(84.0)
69 (32.1)
79 (36.7)
126 (58.6)
2 (0.93)
141 (65.6)
139 (64.6)
8,000
294/23
75 (23.3)
2 (0.7)
13 (4.4)
31(9.5)b
19 (6.5)
43(21.1)
12 (5.9)d
8 (3.92)d
7 (3.4)b
8 (3.9)
199 (97.5)c
71(34.8)
75 (36.7)
147(72.1)
6 (2.9)b
172 (84.3)
123 (60 3)
"Numbers of fetuses affected; numbers in parentheses denote the percentage of affected fetuses/fetuses examined.
bjD<0.05, Fisher's randomization test based on frequencies of affected litters.
      , Wilcoxon's test.
dp<0.01, Fisher's randomization test based on frequencies of affected litters.

Source: IISRP, 1982.
1/28/98
5-13
DRAFT-DO NOT CITE OR QUOTE
-------
        Table 5-6. Design of the developmental toxicity studies on 1,3-butadiene
Species/strain/route of
exposure
Rat/
Sprague-Dawley CD/
Inhalation
Mouse/
CD-I /Inhalation
Exposure
(ppm)
0
40
200
1,000
0
40
200
1,000
No. of animals/
group
30
30
30
30
32
33
31
33
Gestation days of
exposure
6-15
6-15
6-15
6-15
6-15
6-15
6-15
6-15
Gestation day of
sacrifice
20
20
20
20
18
18
18
18
Source: Hackettetal., 1987a, b.
1/28/98
5-14
DRAFT-DO NOT CITE OR QUOTE
-------
         Table 5-7. Maternal toxicity in Sprague-Dawley CD rats exposed to
         1,3-butadiene by inhalation
Parameter
No. dams assigned
No. of deaths
No. pregnant (%)
Whole body weight (g)
DayO
Day 20
Body weight gain (g)
Days 0-6
Days 6- 11
Days 11-16
Days 16-20
Gravid uterine weight (g)
Extragestational weight0 (g)
Extragestational weight
gain" (g)
Significant clinical signs
Concentration (ppm)
0
30
0
28 (93)

242±3.7a
362 ±7.1

21.4±1.6
25.5 ±1.3
29.2 ±1.4
44.5 ±1.8
73.0 ±2.9
289 ±5.7
47.6 ± 2.8
None reported
40
30
0
24 (80)

239 ±3.2
351 ±5.9

21.1 ±1.6
23.6 ±1.3
30.9 ±1.7
36.7 ± 2.5
69.5 ± 3.5
282 ±3.9
42.7 ± 2.2
None reported
200
30
0
26(87)

244 ±3.0
369 ±6.6

22.9 ±1.3
26.6 ±1.5
31.7±1.9
43.6 ±2.3
73.9 ±2.8
295 ± 5.8
50.9 ±3.0
None reported
1,000
30
0
28 (93)

242 ±4.0
354 ±6.1

20.1 ±1.5
17.5 ±1.9"
31.2 ±2.1
43 .2 ±2.9
71.2 ±4.1
283 ±3.5
39.9±3.5b
None reported
"Mean ± standard error.
bp<0.05, compared with corresponding control.
TBody weight on gd 20 minus gravid uterine weight.
dExtragestational weight minus body weight on gd 0.

Source: Hackett et al., 1987a.
1/28/98
5-15
DRAFT-DO NOT CITE OR QUOTE
-------
  1     however, was reduced by about 30% (p<0.05) in the 1,000 ppm group during the first 5 days of
  2     exposure (gd 6 to 11). From gd 11 to 20, body weight gain was not significantly different from
  3     that of controls.  The gravid uterine weights and extragestational weights (whole body weight
  4     minus gravid uterine weight) were similar to those of controls, but extragestational weight gain
  5     was significantly lower (16%;/K0.05) in dams exposed to 1,000 ppm than in control dams.
  6            The overall pregnancy rates were similar among all groups, ranging from 80% among
  7     dams exposed to 40 ppm to 93% among controls and dams exposed to 1,000 ppm (Table 5-7).
  8     Fetal measures, including the numbers of implantations/dam, resorptions/litter, dead
  9     fetuses/litter, fetal body weights, sex ratios, malformations, and variations, were not affected by
10     exposure to 1,3-butadiene (Tables 5-8 and 5-9). Overall, no developmental toxicity was
11     observed in rats exposed to 40 to 1,000 ppm during gd 6 to 15; a slight maternal toxicity,
12     manifested as reduced extragestational weight gain, was observed at the 1,000 ppm level.

13     5.2.3. Hackett et al., 1987b
14           Because 1,3-butadiene is more toxic in mice than in rats, a study was also conducted in
15     CD-I mice using a protocol similar to that used for the rats.  Groups of 31 to 33 sperm-positive
16     females were exposed to  0 (filtered air), 40,200, or 1,000 ppm 1,3-butadiene (99.88% purity;
17     338 ± 72 ppm mean headspace dimer concentration), 6 h/day on gd 6 to 15 (Table 5-6, bottom
18     section).  Measured concentrations were 39.9  ± 0.06,199.8 ± 3.0, and 1,000 ± 13.1 ppm (mean ±
19     SD). The dams were weighed on gd 0, 6, 11,  16, and 18 (day of sacrifice).
20           The effects of 1,3-butadiene on maternal toxicity in CD-I mice are summarized in Table
21     5-10. Three animals exposed to 1,000 ppm showed signs of dehydration: two died on gd 15, and
22     early parturition occurred in the third. No other clinical signs of toxicity were observed.
23     Exposure-related decreases in whole body weights on gd 18, body weight gain during gd 11  to
24     16, gravid uterine weight, extragestational weight, and extragestational weight gain were
25     significantly reduced in the 1,000 ppm exposure group compared with controls.  Whole body
26     weight gain during gd 11 to 16 and extragestational weight gain was also reduced in the 200 ppm
27     exposure group.  None of these parameters were significantly affected in dams exposed to 40
28     ppm.  The pregnancy rates in mice were uniformly low in all groups and unaffected by exposure
29     to 1,3-butadiene. The effects of 1,3-butadiene on various parameters of developmental toxicity
30     hi CD-I mice are summarized in Tables 5-11  and 5-12.  More resorptions per litter were
31     observed among control dams than among exposed dams. Fetal body weights were reduced in all
32     exposed groups compared with controls, and the reduction showed a significant exposure-related
33     trend. The overall fetal body weights (males and females combined) were reduced by 4.5% at 40
34     ppm (not significant), 15.7% at 200 ppm (p<;0.05), and 22.4% at 1,000 ppm 0<0.05).
35     Significant differences from controls were seen at all treatment concentrations for fetal males and
        1/28/98                                   5-16       DRAFT-DO NOT CITE OR QUOTE
-------
         Table 5-8.  Developmental toxicity in Sprague-Dawley CD rats exposed to
         1,3-butadiene by inhalation
Parameter
No. pregnant (%)
No. litters with live fetuses
No. implantations/dam
No. resorptions/litter
Early resorptions/litter
Dead fetuses/litter
No. fetuses/no, litters examined
Fetal body weight (g)
Females
Males
Sex ratio (% males)
Concentration (ppm)
0
28 (93)
28
14.4 ± 0.55b
0.46 ±0.17
0.39 ±0.15
0
389/28
3. 49 ±0.04
3.40 ± 0.05
3.59 ±0.05
50.2 ±2.281
40
24 (80)
24
14.0 ±0.71
0.58 ±0.17
0.54 ±0.16
0
321/24
3.44 ± 0.05
3.36 ±0.05
3. 52 ±0.05
52.5 ±2.95
200
26(87)
26
15.3 ± 0.45
0.96 ± 0.26
0.88 ± 0.25
0
372/26
3. 40 ±0.05
3.29 ±0.06
3.51 ± 0.06
50.5 ± 2.77
1,000
28 (93)
. 27a
14.8 ±0.63
0.67 ±0.14
0.63 ±0.14
0
382/27
3.50 ±0.06
3.38 ±0.06
3.59 ± 0.06
52 5 ± 2 58
"One rat had only one implant; this animal was excluded from statistical evaluations.
bMean ± standard error.

Source: Hackett et al., 1987a.
1/28/98
5-17
DRAFT-DO NOT CITE OR QUOTE
-------
         Table 5-9. Malformations and variations in Sprague-Dawley CD rats
         exposed to 1,3-butadiene by inhalation
Parameter
No. fetuses/no, litters examined
No. fetal heads examined
Malformations"
Generalized edema
Hydrocephalus
Meningoencephalocele
Missing rib
Variations
Low ear set
Hydroureter
Misaligned sternebrae
Extra rib
Reduced ossification
Skull
Stemefarae no. 1-4
Ribs
Thoracic vertebrae (centra)
Pelvis
Phalanges
Concentration (ppm)
0
389/28
196

1/1
__b
—
—

—
36/17
-
—

27/13
60/15
1/1
109/23
9/7
1/1
40
321/24
161

3/1
3/3
—
—

--
35/15
—
1/1

22/13
48/13
2/2
97/21
~
6/1
200
372/26
185

1/1
—
—
2/2

2/1
39/14
1/1
4/2

18/10
95/21
5/3
75/21
6/5
2/1
1,000
382/27
191

3/1
-
2/1
-

-
31/12
1/1
-

29/11
66/15
2/2
81/25
5/5
-
"Expressed as number of fetuses/number of litters; includes only those findings occurring in more than one fetus
 or at more than one concentration.
b—, no malformations observed.

Source: Hackettetal., 1987a.
1/28/98
5-18
DRAFT-DO NOT CITE OR QUOTE
-------
         Table 5-10. Maternal toxicity in pregnant CD-I mice exposed to
         1,3-butadiene by inhalation
Parameter
No. dams assigned
No. of deaths
No. pregnant (%)
Whole body weight (g)
Day 0
Day 18
Body weight gain (g)
Days 0-6
Days 6- 11
Days 11-16
Days 16-18
Gravid uterine weight (g)
Extragestational weightd (g)
Extragestational weight gaine
(g)
Significant clinical signs
Concentration (ppm)
0
32
0
18 (56)

28.4 ± 0.25a
54.9±1.21b

2.7 ± 0.3
5.5 ± 0.4
13.3 ± 0.6b
5.5 ± 0.3b
19.3 ± 1.00b
35.5 ± 0.48b
7.60 ± 0.48b
None
40
33
0
19 (57)

28.3 ± 0.32
55.4 ± 1.09

3.0 ± 0.3
5.8 ± 0.3
12.7 ± 0.4
5.7 ±0.3
203 ± 0.80
35.1 ±0.44
6.99 ± 0.38
None
200
31
0
21 (68)

28.2 ± 0.32
52.5 ±1.01

2.5 ± 0.2
5.6 ± 0.3
11.4±0.5C
4.7 ± 0.4
18.0 ±0.87
34.5 ± 0.46
6.20 ± 0.38°
None
1,000
33
*•»
3
22 (67)

28.4 ± 0.32
50.8 ± 0.86C

2.3 ± 0.2
4.8 ± 0.3
10.6 ± 0.4°
4.8 ± 0.3
16.8±0.67C
34.1±0.36C
5.91 ± 0.28C
Dehydration
"Mean ± standard error.
"^^0.05, significant linear trend.
c/>^0.05, pairwise comparison with corresponding control parameter.
dBody weight on gd 18 minus gravid uterine weight.
eExtragestational weight minus body weight on gd 0.

Source: Hackettetal., 1987b.
1/28/98
5-19
DRAFT-DO NOT CITE OR QUOTE
-------
         Table 5-11. Developmental toxicity in CD-I mice exposed to 1,3-lmtadiene
         by inhalation
Parameter
No. pregnant (%)
No. litters with live fetuses
No. implantations/dam
No. resorptions/litter
Early resorptions
Dead fetuses/litter
No. fetuses/no, litters
examined
Fetal body weight (g)
Females
Males
Placental weight (mg)
Females
Males
Sex ratio (% males)
Concentration (ppm)
0
18 (56)
18
12.7 ± 0.52
1.06 ±0.22
1.00 ± 0.23
0
11. 7 ±0.66
1.34±0.03b
1.30 ± 0.03b
1.38±0.03b
86.8 ± 2.99b
83.1 ± 3.03b
89.3 ± 3.03b
51.6 ± 3.91
40
19 (57)
19
13.3 ± 0.44
0.84 ±0.22
0.58 ±0.21
0
12.5 ± 0.52
1.28 ±0.01
1.25 ±0.01
1.31±0.02a
85.4 ± 2.29
80.9 ± 2.46
89.5 + 2.27
49.8 ± 3.06
200
21 (68)
21
13.0 ±0.64
0.67 ± 0.20
0.43±0.13a
0
12.3 ± 0.62
1.13±0.02a
1.10±0.02a
1.13±0.02a
78.6 ± 3.24a
74.7 + 3.52a
80.1±2.35a
51.5 + 3.68
1,000
22 (67)
20
13.1 ±0.43
0.90 ±0.19
0.75 ±0.16
0
12.2 ±0.51
1.04±0.03a
1.06±0.02a
1.06±0.02a
72.6 + 1.88a.
70.1 ±2.33a
74.5 ±1.81"
51.8 ± 3.29
    .05, pairwise comparison with corresponding control.
     5, significant linear trend.
Source: Hackett et al., 1987b.
1/28/98
5-20
DRAFT-DO NOT CITE OR QUOTE
-------
       Table 5-12. Malformations and variations in CD-I mice exposed to
       1,3-butadiene by inhalation
Parameter
No. fetuses/no, litters examined
No. fetal heads examined
Malformations2
Exencephalus
Open eye
Limb flexure
Fused sternebrae
Fused ribs
Variations
Pale
Hydroureter
Abnormal sternebraec>d
Misaligned
sternebrae
Ossification site between
sternebrae 5 and 6
Supernumerary ribsc'd
Supernumerary ribs
(total number)
Normal length
Rudimentary
Ossification site at
lumbar 1
Concentration (ppm)
0
211/18
105

1/1
1/1
2/1
— • -
—

2/2
2/2
0.6 ±0.9
10/6
—
1.7 ±2.3
30/11
6/5
13/6
11/5
40
237/19
120

__b
-
-
-
2/2.

-
6/3
0.4 ± 0.7
3/3
1/1
1.6 ±2.1
30/9
5/1
19/8
6/4
200
259/21
130

—
..
—
—
—

—
—
0.4 ± 0.8
9/8
1/1
6.0 ± 3.6e
127/20
29/9
81/20
17/10
1,000
244/20
120

2/2
1/1
-
2/2
—

—
—
o.8±i.3e:
10/8
3/3
9.9±3.0e
198/20
68/10
120/16
10/7
1/28/98
5-21
DRAFT-DO NOT CITE OR QUOTE
-------
         Table 5-12. Malformations and variations in CD-I mice exposed to
         1,3-butadiene by inhalation (continued)
Parameter
Reduced ossification (all sites)c
Skull
Stemebrae
Vertebrae
(centra)
Phalanges
Concentration (ppm)
0
1.7 ±1.7
-
31/13
—
—
40
1.2 ±1.5
—
20/9
1/1
—
200
2.7 ± 2.7
2/2
57/1 6f
—
-
1,000
3.9±2.6e
3/1
76/1 9f
1/1
2/16
"Expressed as number of fetuses/number of litters; includes only those findings occurring in more than one fetus
 or at more than one concentration.
b~, no malformations or variations noted.
"Mean percentage per litter (mean ± SD).
d/?<0.05, linear trend, orthogonal contrast test.
ep<0.05, Tukey's test.
jD<0.05, Fisher exact test (fetal incidence).

Source:  Hackettetal., 1987b.
1/28/98
5-22
DRAFT-DO NOT CITE OR QUOTE
-------
   1      at the two higher concentrations for females. Placental weights showed an effect similar to that
   2      of fetal body weights (Table 5-11). Malformations occurred sporadically and at low frequencies
  '3      in all exposure groups (Table 5-12). The frequency of supernumerary ribs was greatly elevated at
   4      200 and 1,000 ppm; 6% of the fetuses/litter were affected at 200 ppm (p<0.05) and 9.9% at 1,000
   5      ppm (p<0.05) compared with 1.7% in controls and 1.6% in the 40 ppm exposure group (not
   6      significant). There also was a marked increase in the total number of fetuses with supernumerary
   7      ribs at the 200 and 1,000 ppm exposure levels. The frequency of reduced ossification of the
   8      sternebrae was elevated at 200 (p<0.05) and 1,000 ppm (pO.OOl) (Fisher exact test); the litter
   9      incidence was elevated but not significantly.  The percentages of reduced ossifications at all sites
 10      and the percentages of abnormal sternebrae (misaligned, scrambled, or cleft) per litter were also
 11      significantly elevated at 1,000 ppm (p<0.05). The percentages Of supernumerary ribs and
 12      abnormal sternebrae also showed significant linear trends.
 13            These studies showed that inhalation exposure to 1,3-butadiene causes maternal toxicity,
 14      manifested as reduced body weight gain, in the mouse at 200 and 1,000 ppm; therefore, the
 1 5      NOAEL for maternal toxicity is 40 ppm. 1,3-Butadiene also caused developmental effects,
 16      manifested by reduced fetal body weight and increased frequency of skeletal variations at 200
 17      and 1,000 ppm. In addition, inhalation exposure to 1,3-butadiene during gestation caused a
 18      significant reduction in body weight of male fetuses at 40 ppm. Therefore, a NOAEL for
         developmental toxicity in CD-I mice could not be obtained. Although 1,3-butadiene did not
         induce gross malformations in the mouse fetus, the dose-related increases in supernumerary ribs
 21      and reduced ossifications, particularly of the sternebrae., may indicate delayed or altered
 22      development and should be cause for concern.

 23      5.2.4. Anderson et al., 1993
 24            The  acute and subchronic effects of inhalation exposure to 1,3-butadiene in male mice on
 25      fetal abnormalities were examined.  For acute exposures, male CD-I mice were exposed to 0,
 26      1,250, and 6,250 ppm 1,3-butadiene for 6 h; 5 days later each male was mated to two females.
 27      Males used for subchronic exposures were treated with 0,12.5, or 1,250 ppm 6 h/day, 5
 28      days/week for 10 weeks. Following mating in both experiments, one female was killed on gd 17
 29      and the other was allowed to litter.  The female killed on gd 17 was examined for number of live
 30      fetuses, number and type of malformations  in the fetuses, and number of postimplantation deaths.
 31      No treatment-related abnormalities were observed in offspring of males treated on the acute
 3 2      exposure protocol; one fetus from one control litter had a gastroschisis (fissure of abdominal
 33      cavity), and one fetus from each of the two  low-dose litters was a runt (body weight <67% of
_34      mean litter weight).  Following subchronic exposure of males to 1,3-butadiene, 7 of 306 fetuses
         sired by males exposed to 12.5 ppm (p
-------
 1      1,250 ppm were affected compared with 0 of 278 fetuses sired by control males. Abnormalities
 2      in low-dose fetuses included four exencephalies, two runts (<70% of mean litter weight), and one
 3      fetus with blood in the amniotic sac. At the high-dose level, one hydrocephaly and two runts
 4      (^75% of mean litter weight) were observed. The authors calculated statistical significance on a
 5      fetal incidence basis rather than on a litter incidence basis. Because litter incidence rates were
 6      not included in the data, it is not possible to discern whether the affected fetuses were only from
 7      one or two litters or whether a high percentage of litters sired by exposed males were affected.
 8      Therefore, this study is inadequate to assess the developmental toxicity of 1,3-butadiene
 9      following exposure of males prior to mating.

10      5.3. STRUCTURE-ACTIVITY RELATIONSHIPS
11             Data on structure-activity relationship are summarized in Table 5-13.

12      5.3.1. NTP, 1986
13             The 1,3-butadiene dimer, 4-vinylcyclohexene, and its diepoxide, 4-vinyl-l-cyclohexene
14      diepoxide, have been tested in long-term toxicological and carcinogenicity studies in rats and
15      mice. The NTP study (1986) on 4-vinylcyclohexene used male and female F344 rats and male
1 6      and female B6C3F, mice dosed by gavage with 0, 200, of 400 mg/kg 4-vinylcyclohexene in corn
17      oil 5 days/week for 105 weeks.  No nonneoplastic lesions attributed to exposure to 4-
18      vinylcyclohexene were observed in the reproductive organs of male or female mice or rats, and
19      hence data are not presented in Table 5-13. The incidences of granulosa cell neoplasms, mixed
20      benign tumors, granulosa cell hyperplasia, and tubular cell hyperplasia were increased in female
21      mice. Tubular cell hyperplasia is a proliferative lesion originating in the germinal epithelium; the
22      hyperplastic cells invade the ovarian stroma forming tubular structures. The granulosa cell
23      hyperplasia is a morphological continuum with granulosa cell neoplasms and tubular hyperplasia
24      with mixed benign tumors and therefore should not be included with nonneoplastic lesions.  The
25      authors noted that female mice treated with 1,200 mg/kg 4-vinylcyclohexene for 5 days/week for
26      13 weeks had reduced numbers of primary and mature Graafian follicles whether they survived
27      until termination (5/10) or died before termination (5/10).
28             In another NTP study (1989), male and female F344 rats and B6C3FJ mice were treated
29      topically with 4-vinyl-l-cyclohexene diepoxide 5 days/week for 13 weeks and 2 years. No
30      nonneoplastic  lesions occurred in reproductive organs of rats.  Female mice treated for 13 weeks,
31      however, showed evidence of diffuse ovarian atrophy in 10/10 animals that received 10
32      mg/mouse (highest dose) and in 4/10 receiving  5 mg/mouse.  Uterine atrophy was observed in
33      2/10 animals that received 10 mg/kg. In the 2-year study, ovarian atrophy occurred in almost

        1/28/98                                  5-24       DRAFT-DO NOT CITE OR QUOTE
-------
 1
 ti
 (ft
 •*•»
 *>m
 CA
 ^w
 U


 •-M
 09
 I
 o
 £
 "3
 'x


 i
 0>
  t

•I - 1-

Females
13 weeks: i5 mg/kg: ovarian atrophy;
uterine atrophy
105 weeks: ;>2.5 mg/kg: ovarian atropi
Males
105 weeks: k5 mg/kg: inflammation o)

eg
^
(U
•£
*?">L -3
5h erf J?
3 "° S
o v>
*— < ^ o

. s
S s
i*43


•* 0
s
,__,
„ „
KJ ON ^
O i— i O
li" 1





bo
1
1—1




a>
c
"C
1.43 mmol/kg: reduced ovarian and ute
weights; decreased follicular counts
none


^ o

° §
^ =? 1?
O bp ea •§
CN" ^ a o
o o o m
o E -c g
- C U eg
*° S« -^
0 m g .£»
o -: .S 1

_o M
S i 2
~ 4> -.^
go g 2 "5
(L> vo , o3 cu "3
fa m BL, oo Q

T3
U "X
go.
•3 g

3 O
CQ S
S
^J
~
"c3 ON "3
*j ON .«_>
li" 1





bD bo
:o, "o •
E S
E S
2 2
0 0


(U

1

a 0.1 4 mmol/kg: decreased ovarian and
weights; decreased follicular counts
^0. 14 mmol/kg: decreased ovarian wei§
0.29 mmol/kg: decreased uterine weighl

•<*"
0 u

^O C
m 0
S i> >.
„ bO"o .§

0 "o 0 m
o S -S S
. g S3 eg
o os £ J£?
§ 2 .5 1

0
s ,1
— w p*.
| 0 g 2 Is
D *«O u pi, ra
PL, CQ . tu OO Q


i>
g ^
•3 o
S Q,
3 4J
CQ =5

r r
"o *c3
IL> ^j
^ ^>
o o
"53 c\ '& o\
^ C7\ «^ 

0
E 'g
ex, a.
ex, a,
0 0
0 0
0 . 0_
r^- t*^


_§> 3
a. S
£ 1

13 and 26 weeks: 7,000 ppm: testicular
13 weeks: none
26 weeks: 7,000 ppm: hyperplasia of inl
cells
o
0 <"_
I--" it?
" "^ Vrt
l-3^
*> O
^^ o f^l
o" « -^
f^ •— ^
„ S $^
S .S ^ w
p ^ u
0" 0- c? g
r-~ CL*O ?
EO
fa" 2
en ^.
o 5
"O ro
-Q fT
Is 1



o
§
E.
o
C/3
 tS

 CS
 (U
 •a
 o
 I

1/28/98
5-25
DRAFT-DO NOT CITE OR QUOTE
-------
 1      all groups treated with 2.5, 5, and 10 mg/mouse (43/49,42/49, and 47/50, respectively, compared
 2      with 12/50 for controls). Ovarian atrophy such as that in animals exposed to 1,3-butadiene was
 3      characterized by a complete absence of follicles and corpora lutea. Tubular hyperplasia occurred
 4      at a high incidence in all dose groups (35/49, 38/49, and 34/50, respectively, compared with 5/50
 5      for controls). In male mice, subacute inflammation of the epididymis occurred in 0/50, 6/50, and
 6      13/49, respectively, compared with 0/50 for controls.

 7      5.3.2. Melnicketal., 1994
 8             Differences in susceptibility between rats  and mice were seen in inhalation studies with
 9      isoprene, the 2-methyl analogue of 1,3-butadiene. Male F344 rats and B6C3F, mice were
10      exposed to 0, 70,220, 700,2,200, and 7,000 ppm isoprene 6 h/day, 5 days/week for either 13
11      weeks or 26 weeks followed by a 26-week recovery period. After 13 weeks of exposure, no
12      effects were observed in rats at any concentration, but testicular atrophy occurred in mice at
13      7,000 ppm.  Following 26 weeks of exposure, all treated rats in the 7,000 ppm group had
14      hyperplasia of the interstitial cells of the testis (p<0.01; 10/10 vs. 1/10 controls); however,
15      following the 26-week recovery, there was only a marginal increase (not significant) in benign
16      testicular tumors: 9/30 compared with 3/30 for controls. Mice also had an increase in the
17      incidence of testicular atrophy following 26 weeks of exposure to 7,000 ppm (p^O.05; 5/10 vs.
18      0/10 controls). After 26 weeks of recovery, mice had a slight increase (not significant) in            •!&
19      testicular atrophy at 7,000 ppm (3/29 compared with 0/29 for controls).

20      5.3.3. Doerr et al., 1996
21             This study tested the ovarian toxicity of the mono- and diepoxide metabolites of 1,3-
22      butadiene in mice and rats.  Butadiene monoepoxide (0.005, 0.02, 0.09, 0.36, or 1.43 mmol/kg),
23      butadiene diepoxide (0.002, 0.009, 0.036, 0.14, or 0.29 mmol/kg), or vehicle (sesame oil) was
24      administered intraperitoneally once daily to female B6C3F, mice and Sprague-Dawley rats for 30
25      days. Following day 30, animals were sacrificed by CO2 inhalation, the ovaries and uteri were
26      weighed, and the ovaries processed for histologic examination of preantral follicles. At the high
27      dose, the monoepoxide resulted hi reduced ovarian and uterine weights (p<0.05) and decreased
28      follicular counts in mice; rats, however, were unaffected. The diepoxide resulted in decreased
29      ovarian weights  (p^O.05) in mice and rats at 2:0.14 mmol/kg and decreased uterine weights
30      (psO.05) in mice at kO.14 mmol/kg and in rats at 0.29 mmol/kg.  Because organ weights were
31      given in a histogram, the absolute differences were not available; all significant reductions
32      appeared to be approximately ^50% of the control values. The ED50 value was defined as the
33      effective dose that reduces the follicular number to 50% of control. In mice, ED50 values for the
34      monoepoxide were 0.29 and 0.40 mmol/kg and for the diepoxide were 0.1 and 0.14 mmol/kg for
        1/28/98                                   5-26       DRAFT-DO NOT CITE OR QUOTE
-------
         small and growing follicles, respectively. However, in rats, only 32% of the follicular population
         was depleted at the highest diepoxide concentration.  Therefore, mice were more sensitive than
         rats to the ovotoxic effects of the mono- and diepoxides of 1,3-butadiene, and the diepoxide was
  4      the more potent ovotoxicant in both species.
  5             Doerr et al. (1995) also studied the ovarian toxicity of 4-vinylcyclohexene and several
  6      related olefins, including butadiene mono- and diepoxide. Mice were administered 1.43
  7      mmol/kg of the monoepoxide or 0.14 mmol/kg of the diepoxide once daily for 30 days.
  8      Following day 30, the mice were killed and the ovaries removed and sectioned for histologic
  9      examination. Mean follicle counts in mice treated with the monoepoxide were depleted 98% and
 10      71% for small and growing follicles, respectively, compared with controls. In mice treated with
 11      the diepoxide, follicle counts were depleted 85% and 63% for small and growing follicles,
 1 2      respectively, compared with controls. Structural analogs of vinylcyclohexene that contain only a
 13      single unsaturated site (vinylcyclohexane, ethylcyclohexene, cyclohexene) and their
 14      monoepoxide metabolites were not ovotoxic to mice. On the other hand, butadiene
 1 5      monoepoxide, butadiene diepoxide, and isoprene were ovotoxic. The study showed a
 1 6      relationship between chemical reactivity, as assessed by nicotinamide alkylation, and ovotoxicity
 17      with vinyldiepoxide and butadiene diepoxide that was 3.5 to 10 times more reactive than their
 18      monoepoxide precursors and other structurally related monoepoxides. It can be concluded that
         those compounds that are metabolized to a diepoxide or are a diepoxide are ovotoxic.

 20      5.4. SUMMARY AND CONCLUSIONS
 21             Evidence has been presented showing that 1,3-butadiene induces reproductive and
 22      developmental effects in rodents. Although the studies conducted by Carpenter et al. (1944)
 23      examined reproductive toxicity in four different species (rat, guinea pig, rabbit, and dog), the
 24      experimental protocol and the results obtained are inadequate for evaluating reproductive
 25      toxicity. The three long-term toxicity studies conducted in Sprague-Dawley CD rats (Owen et
 26      al., 1987; Owen and Glaister, 1990) and B6C3Fj mice (NTP, 1984, 1993) suggest that mice are
 27      much more sensitive than rats to the reproductive effects of 1,3-butadiene.  Reproductive toxicity
 28      was not observed in either male or female rats exposed intermittently to 1,3-butadiene at
 29     concentrations up to 8,000 ppm for 2 years. However, ovarian atrophy was observed in female
 30     mice exposed to 6.25 to 625 ppm 1,3-butadiene.
31             Ovarian atrophy occurred in 39% of 49 mice at 6.25 ppm (the lowest concentration
32      tested) only after exposure for 2 years, a time at which this condition is expected to appear in
33      aged animals due to normal senescence mechanisms; however, it occurred in a significantly
34     greater number of mice exposed to 1,3-butadiene than in control animals. Furthermore, ovarian
        atrophy was observed as early as 9 months after exposure to 200 and 625 ppm and 15 months

         1/28/98                                   5-27       DRAFT-DO NOT CITE OR QUOTE
-------
 1      after exposure to 62.5 ppm.  Therefore, the dose-response relationship observed for ovarian
 2      atrophy and the significant increase at the lowest dose relative to that seen in control animals of a
 3      similar age is evidence for a causal relationship between ovarian atrophy and exposure to 1,3-
 4      butadiene at 6.25 ppm.
 5             Similar ovarian lesions have been observed in mice after exposure to the 1,3-butadiene
 6      dimer, 4-vinylcyclohexene, administered by gavage for 13 weeks at 1,200 mg/kg for 5 days/week
 7      (NTP, 1986) or its diepoxide, 4-vinyl-l-cyclohexene diepoxide, administered by topical
 8      application for 13 weeks or 2 years (NTP, 1989). Rats administered 4-vinylcyclohexene or 4-
 9      vinyl-1-cyclohexene diepoxide did not develop ovarian lesions, thus showing a species-specific
10      response similar to that after exposure to 1,3-butadiene.
11             Ovarian lesions induced by 1,3-butadiene, 4-vinylcyclohexene, or 4-vinyl-l-cyclohexene
12      diepoxide are characterized by the absence of oocytes, follicles, and corpora lutea.  The
13      functional integrity of the reproductive  system in animals exposed to 1,3-butadiene has not been
14      tested, but the severity of the ovarian lesion is indicative of reproductive dysfunction.
15      Furthermore, Maronpot (1987) compared the ovarian toxicity and carcinogenicity of eight
16      chemicals tested by NTP and concluded that the occurrence of ovarian lesions in a 90-day study
17      may also indicate that ovarian neoplasia would be  induced upon continued treatment.
18             Uterine atrophy is probably due to the indirect action of 1,3-butadiene metabolites and the
19      consequent interruption of ovarian sex steroid stimulation of the uterus.  Oocyte toxicity and
20      destruction of the follicular and subsequent luteal components  of the ovary result in reduced
21      steroidogenesis by the ovary. It is well known that ovarian estrogens and progestins have a
22      uterotropic function in laboratory rodents and humans.
23             Testicular atrophy, as reflected by reduced testis weight following 1,3-butadiene exposure
24      for 9 and 15 months in male mice (NTP, 1993) indicates  gonadal sensitivity in the male as well
25      as in the female. However, the ovary is more sensitive than the testis because ovarian atrophy
26      results at very low concentrations (6.25 ppm) of 1,3-butadiene compared with that seen in males
27      after 2 years of exposure. The sperm-head morphology study showed that male mice are affected
28      at concentrations sl,000 ppm (Hackett et al., 1988a), and the dominant lethal test showed that
29      male mice may be affected at 200 and 1,000 ppm (Hackett et al., 1988b), again indicating that
30      higher exposure concentrations are necessary to induce toxic effects in male mice than in female
31      mice. As observed for other effects of 1,3-butadiene, the reproductive organs of male rats are
32      more resistant than those of female mice exposed to 1,3-butadiene. This resistance in males may
33      be attributed, in part, to the blood-testis barrier. No homologous anatomical barrier has been
34      demonstrated in the Graafian follicle (Crisp, 1992).  No adverse reproductive effects have been
35      observed in male rats at  concentrations up to 8,000 ppm (Owen et al., 1987; Owen and Glaister,
36      1990).
        1/28/98                                   5-28       DRAFT-DO NOT CITE OR QUOTE
-------
                Several studies indicate that 1,3 -butadiene affects spermatozoa and spermatids as
         determined by postimplantation deaths during the first 3 weeks after exposure. The data from
         Hackett et al. (1988b) is equivocal because of a lack of the dose-response relationship, but the
  4     studies by Anderson et al. (1993) and Adler et al. (1994) confirm the hypothesis. Further
  5     evidence that 1,3-butadiene is a germ cell mutagen was presented in a comparison of the latter
  6     two studies (Adler and Anderson, 1994).  In the first experiment, the percentage of dominant
  7     lethality observed following 10 weeks of exposure of males to 1,250 ppm was 28.1%. During
  8     the acute exposure experiment, the sum of dominant lethality over the 3 weeks of mating was
  9     23.1%. The results are in close agreement despite differences in protocols such as exposure
 10     regimen, strains of mice, and mating scheme. It appears that 1,3-butadiene affects spermatozoa
 11      and spermatids because the effects observed after 10 weeks are representative of the last 3 weeks
 12     of treatment and increasing the length of exposure did not add to the response.
 13            The mechanism by which 1,3-butadiene induces ovarian lesions is not known, but  it is
 14     unlikely that 1,3-butadiene is a direct-acting reproductive toxicant.  Direct-acting compounds act
 15     as hormonal agonists or antagonists or chemically reactive compounds (such as alkylating
 16     agents), which directly interfere with hormone-receptor interactions or interact with
 17     macromolecules (Maronpot, 1987; Mattison et al., 1990). More likely, 1,3-butadiene is an
 18     indirect-acting reproductive toxicant. Indirect-acting toxicants require metabolic activation to
         exert their toxic effects, which then may proceed via mechanisms similar to those of direct-acting
         compounds, or they may interfere with endocrine homeostasis (Mattison et al., 1990).
 21             Developmental effects observed after exposure to 1,3-butadiene consisted primarily of
 22     reduced fetal body weight and minor skeletal defects such as abnormal ossifications, abnormal
 23     sternebrae, and supernumerary ribs.  No gross malformations were produced.  The pattern for
 24     developmental effects induced by 1,3-butadiene was similar to that of reproductive effects, with
 25     mice showing greater sensitivity than rats. This difference was also seen for maternal toxicity as
 26     manifested by decreased weight gain (whole body and extragestational) in both rats and mice.
 27     The NOAEL for maternal effects is 200 ppm in rats  (IISRP, 1982; Hackett et al.,  1987a) and 40
 28     ppm for mice (Hackett et al., 1987b). While the IISRP (1982) study showed increased
 29     frequencies for bipartite thoracic centers and minor skeletal defects combined at 200 ppm in rats,
 30     the response is not clearly dose-related. Several developmental effects occurred at significantly
 31      increased frequencies at 8,000 ppm and showed a dose-response relationship:  major skeletal
 32     defects combined, wavy ribs, and abnormal ossification of the ribs (IISRP, 1982). The results
 33     from the IISRP (1982) study were not confirmed in the more recent study by Hackett et  al.
 34     (1987a), which showed no developmental toxicity in the same rat strain similarly exposed to
J3 5     1,3 -butadiene at concentrations up to 1,000 ppm.  Therefore, the NOAEL for developmental
         effects in rats is 1,000 ppm (IISRP, 1982; Hackett et al.,  1987a). These independent observations
         1/28/98                                  5-29       DRAFT-DO NOT CITE OR QUOTE
-------
  1      strengthen the evidence that a decreased maternal weight gain during early development might
  2      contribute to subtle adverse effects (1) undetected by insensitive indices in the studies or (2) at a
  3      later point in time in the Fj or subsequent generations. An NOAEL for developmental effects
  4      could not be defined for the mouse, because male fetal body weight was decreased at 40 ppm in
  5      the Hackett et al. (1987b) study and postimplantation loss was observed at 12.5 ppm in the
  6      Anderson et al. (1993) study, the lowest concentrations tested.
  7             Reproductive and developmental toxicity studies show species specificity for exposure to
  8      1,3-butadiene in rats and mice. Like other toxicological effects induced by 1,3-butadiene, mice
  9      are more sensitive than rats to the induction of reproductive and developmental effects.
10      Pharmacokinetic studies show that uptake of 1,3-butadiene is about four times greater in mice
11      than in rats at concentrations up to 1,000 ppm (Dahl et al., 1990) and about two times greater at
12      8,000 ppm (Dahl et al., 1991). These data indicate that the availability of 1,3-butadiene is greater
13      in mice than in rats at comparable exposure concentrations. Nose-only exposure of mice and rats
14      to 1,3-[l4C]-butadiene resulted in greater or similar concentrations of radioactivity (expressed as
15      nM/g of tissue) in tissues of rats than those of mice under conditions in which the rats were
16      exposed to a 10-fold higher concentration of 1,3-butadiene (Bond et al., 1987). If tissue uptake
17      was expressed as 1,3-butadiene equivalents/uM inhaled 1,3-butadiene, however, radioactivity
18      levels were 15 to 100 times higher in mice. Mammary tissue, which had 4.6-fold higher
19      concentration in rats than in mice, was the only tissue analyzed that was relevant to evaluating
20      reproductive effects of 1,3-butadiene. Because male animals were used, subcutaneous fat, which
21      had similar levels of radioactivity as mammary tissue, probably contaminated the samples. The
22      ovary, uterus, and testis, which are targets for 1,3-butadiene, were not analyzed by Bond et al.
23      (1987). In a recent in vitro study, Sharer et al. (1992) showed that microsomes from the testes of
24      rats and mice are ineffective in forming butadiene monoxide, but the cytosol fraction was very
25      effective in forming glutathione conjugates. Therefore, it is unlikely that toxic effects on the
26      testes are due to metabolites formed within the testes but rather are due to metabolites formed
27      elsewhere, indicating that 1,3-butadiene is an indirect-acting reproductive toxicant in males.
28             Species-specific differences have also have been observed for the formation of
29      metabolites.  The monoxide hydrolase (detoxification) pathway is favored in rat microsomes,
30      whereas the monooxygenase pathway is favored in mouse microsomes. Although these data do
31      not fully explain the species differences, they show that the basis of the difference may be related
32      to the greater availability of 1,3-butadiene in mice, the greater production of toxic intermediates,
33      and a lower capacity for detoxification of these intermediates.
34             No data are available regarding the  reproductive or developmental effects of the
35      metabolites of 1,3-butadiene, 1,2-epoxybutene, and diepoxybutane. Mice are more sensitive than
36      rats to the ovotoxic effects of the mono- and diepoxides of 1,3-butadiene, and the diepoxide is
        1/28/98                                   5-30        DRAFT-DO NOT CITE OR QUOTE
-------
        the more potent ovotoxicant in both species.  Data regarding the 1,3-butadiene dimer 4-
        vinylcyclohexene and its diepoxide provide evidence that 4-vinylcyclohexene induces ovarian
        and uterine atrophy after treatment by gavage for 13 weeks with 10 mg/kg 5 days/week and that
 4     4-vinyl-l-cyclohexene diepoxide (2.5 to 10 mg/mouse) induces ovarian atrophy and tubular
 5     hyperplasia in mice after topical treatment for 2 years. Subacute inflammation of the epididymis
 6     is seen in male mice receiving 4-vinyl-l-cyclohexene (5 or 10 mg/mouse) topically for 2 years.
 7     Ovarian neoplasms are induced in mice by 4-vinylcyclohexene and its diepoxide. However,
 8     neither 4-vinylcyclohexene nor its diepoxide  induce either neoplastic or nonneoplastic lesions in
 9     the ovaries of rats.
10            In conclusion, the animal data show that there is a potential reproductive hazard to
11      humans upon exposure to 1,3-butadiene, with women being more sensitive than men.  The
12     quantitative aspects of this assessment will require application of pharmacokmetic parameters
13     because humans may be less sensitive than mice (Chapters 3, 8). The animal data also  show that
14     there is a potential for developmental effects in humans upon in utero exposure to 1,3-butadiene
1 5     and that these effects may occur at concentrations below those causing maternal effects (Section
16     9.3).
        1/28/98                                   5-31       DRAFT-DO NOT CITE OR QUOTE
-------
-------
                                     6. TOXICITY IN ANIMALS

               This chapter updates the evaluation of animal studies published from 1985 through
 2     January 1997.  The study by Owen et al. (1987) is not evaluated here because it was reviewed
 3     previously in U.S. EPA (1985) as the NTP study (1984), which was subsequently published by
 4     Owen etal. (1987).

 5     6.1. SUBCHRONIC TOXICITY
 6            Irons et al. (1986a, b) conducted studies to assess the potential of 1,3-butadiene to induce
 7     myelotoxicity by exposing male B6C3FJ mice and male NIH Swiss mice to 1,250 ppm 1,3-
 8     butadiene, 6 h/day, 5 days/week for 6 weeks. Treatment-related hematological changes included
 9     decreases in red blood cell counts, total hemoglobin, and hematocrit and increases in mean cell
10     volume and circulating micronuclei in both strains of mice. The observed anemia was not
11      accompanied by significant alterations in mean corpuscular hemoglobin concentration, increases
12     in reticulocyte counts, or increases in the frequency of nucleated erythrocytes in peripheral blood.
13     These hematologic changes were considered to represent a macrocytic-megaloblastic anemia,
14     because they were accompanied by mild megaloblastic changes in bone marrow cells.
1 5            Exposure of male B6C3Fj mice to 1,250 ppm 1,3-butadiene for 6 h/day, 5 days/week, for
 |6     6 or 12 weeks did not produce any persistent effects on humoral or cell-mediated immunity
1 7     (Thurmond et al., 1986).  Relative thymus weights were unaffected, but relative spleen weights
1 8     were decreased 20% and spleen cellularity was decreased 29% in exposed mice. Extramedullary
1 9     hematopoiesis and erythroid hyperplasia in exposed mice correlated with a twofold increase in
20     thymidine incorporation into spontaneously proliferating splenocytes. Although the number of
21      IgM antibody plaque-forming cells (PFC) per 106 splenocytes was unchanged, a 30% decrease in
22     PFC/spleen was observed. Proliferation of alloantigens was similar for 1,3-butadiene-exposed
23     splenocytes and controls. The mitogenic response of mature T lymphocytes to
24     phytohemagglutinin was significantly suppressed after exposure to 1,3-butadiene for 6 or 12
25     weeks.

26     6.2. CHRONIC TOXICITY
27            A 2-year chronic inhalation toxicity and carcinogenicity study on the effects of 1,3-
28     butadiene on B6C3Fj mice was conducted by NTP (1993).  In this study, groups of 70 male and
29     70 female mice were exposed by inhalation 6 hours/day, 5 days/week to 0, 6.25, 20, 62.5, or 200
30     ppm  1,3-butadiene for periods up to 103 weeks; groups of 90 male and 90 female mice were
 [1      similarly exposed to 625 ppm 1,3-butadiene, which was the lowest exposure level in the previous

        1/28/98                                  6-1        DRAFT-DO NOT CITE OR QUOTE
-------
  1      NTP (1984) study. The additional animals in the 625-ppm exposure group were included
  2     because high mortality rates, observed previously at this exposure concentration, might interfere
  3     with the scheduled interim evaluations. Interim evaluations were conducted at 9 and 15 months.
  4            Mean body weight gains of male and female mice exposed to 6.25-625 ppm 1,3-
  5     butadiene for 103 weeks were similar to those of controls. However, concentration-related
  6     decreases in survival were seen in male and female mice exposed to concentrations 5:20 ppm 1,3-
  7     butadiene (Table 6-1, Figures 6-1 and 6-2) primarily due to the development of malignant
  8     neoplasms. No female mice exposed to 200 or 625 ppm or male mice exposed to 625 ppm
  9     survived to the end of the  study. Statistical analysis for the probability of survival was estimated
10     using the Kaplan and Meyer (1958) procedure; the method of Cox (1972) and Tarone's (1975)
11      life table test was used to identify concentration-related trends.
12            At the 9- and 15-month interim evaluations, no clinical findings other than those
13     associated with lesion development and moribundity were observed. Some statistically
14     significant organ weight changes were observed at interim evaluations in male and female mice
15     exposed to  1,3-butadiene concentrations 2:62.5 ppm. Effects related to toxicity to reproductive
16     organs are discussed in Chapter 5.
17            Hematological indices measured at the interim evaluations showed significant (p<0.05)
18     decreases in erythrocyte counts, hemoglobin concentration, and packed cell volume in male mice
19     exposed to  ^62.5 ppm and in female mice exposed to 5:200 ppm at 9 months.  Mean cell volume
20     was significantly increased in male mice exposed to 625 ppm and in females exposed to 2:200
21      ppm at 9 months.  A similar profile of hematological changes was observed in male and female
22     mice exposed to 625 ppm for 15 months. Increases in the percentage of erythrocytes with
23     Howell-Jolly body inclusions and mean cell hemoglobin were seen at 9 and 15 months. At the
24     15-month interim evaluation, males exposed to 625 ppm 1,3-butadiene had a significantly
25     increased mean platelet value, a finding that correlated with the development of neoplasms.
26     Because these hematological changes were not associated with increases in reticulocyte counts or
27     in frequency of polychromatic erythrocytes in peripheral blood, they were attributed to a partial
28     or poorly regenerative bone marrow response to decreased levels of circulating erythrocytes.
29     There were no significant  changes in total serum enzyme activity of lactate dehydrogenase
30     (LDH) or creatine kinase in mice evaluated at 9 months. LDH values at the 15 month evaluation
31      were increased in males and females exposed to 5:200 ppm. At 625 ppm, LDH-1 and LDH-2
32     were decreased and LDH-5,  the principal enzyme in skeletal muscle and liver, was increased.
        1/28/98                                   6-2        DRAFT-DO NOT CITE OR QUOTE
-------
        Table 6-1. Survival of male and female B6C3F! mice exposed to
        1,3-butadiene by inhalation for 103 weeks

Concentration (ppm)
0
6.25
20
62.5
200
625
Male
Animals initially in study
9-Month interim evaluation3
15-Month interim evaluation1
Natural deaths
Moribund kills
Accidental deaths3
Missing3
Animals surviving until study termination
Percent survival at end of study0
Mean survival(days)d
Survival analysis6
70
10
10
6
9
0
0
35
70
597
/K0.001
70
10
10
5
6
0
0
39
78
611
/>=0.430N
70
10
10
11
15
0
0
24
49
575
^=0.044
70
10
10
12
15
0
1
22
46
558
p=0.02l
70
10
10
23
23
0
0
' 4b
8
502
/K0.001
90
10
7
39
33
1
0
0
0
280
pO.OOl
Female
Animals initially in study
9-Month interim evaluation3
15-Month interim evaluation3
Natural deaths
Moribund kills
Accidental deaths3
Animals surviving until study termination
Percent survival at end of study0
Mean survival (days)d
Survival analysis'1
70
10
10
3
10
0
37
74
608
pO.OOl
70
10
10
7
10
0
33
66
597
p=0.510
70
10
10
11
14
1
24
50
573
p=0.013
70
10
10
8
31
0
11
23
548
pO.OOl
70
10
10
12
37
1
0
0
441
pO.OOl
90
8
2
33
46
1
0
0
320
pO.OOl
a Censored from survival analyses.
b Includes one animal that died during the last week of the study.
c Kaplan-Meier determinations. Survival rates adjusted for interim evaluations, accidental deaths and missing animals.
d Mean of all deaths (uncensored, censored, terminal sacrifice).
c The result of the life table trend test (Tarone, 1975) is in the control column, and the results of the life pairwise comparisons
 (Cox, 1972) with the controls are in the dosed columns. A negative trend or lower mortality in a dose group is indicated by N.
 Source: NTP, 1993.
1/28/98
6-3
DRAFT-DO NOT CITE OR QUOTE
-------
 1             Histopathological effects observed at the 9-month evaluations included bone marrow
 2      atrophy (depletion of cells) in 50% of males and in 13% of females at the highest concentration
 3      (625 ppm). The atrophy increased in severity from mild depletion of hematopoietic cells at 9
 4      months to marked depletion in mice that died or were killed at or before 15 months. An
 5      increased incidence of bone marrow hyperplasia and an increased incidence or severity of
 6      hematopoiesis of the spleen, liver, and lung occurred in females exposed to the three highest
 7      concentrations (^62.5 ppm).  Thymic necrosis (atrophy) and decreased thymus weights were
 8      seen at the 9-month evaluation in males and females exposed to 625 ppm.  Thymic necrosis also
 9      occurred in females exposed to 62.5 or 200 ppm.
10             In mice exposed to 1,3-butadiene for 103 weeks, nonneoplastic effects were observed in
11      the bone marrow, liver, testes, ovary, heart, upper respiratory tract, and various other organs.
                         MALE MICE
                        • CONTROL
                        O S.25PFU
                        A ZOffVI
                        D «2.5 PPU
                        • 200 PPM
                        O 62S ff U
              0.0
                                           -i	r
                                           -15       «0
                                             WEEKS ON STUDY
                                     120
       Figure 6-1.  Kaplan-Meier survival curves for male B6C3FJ mice exposed to 1,3-butadiene by
       inhalation for 103 weeks.
       Source:  NTP, 1993.
        1/28/98
6-4
DRAFT-DO NOT CITE OR QUOTE
-------
           1.0
                     FEMALE MICE
                     • COMTROL
                     O 8.25 PPM
                     A 20 PPM
                     D 62.5 PPM
                     • 200 PPM
                     O S2S PPM
                                                                                     120
                                          WEEKS ON STUDY

      Figure 6-2. Kaplan-Meier survival curves for female B6C3Fj mice exposed to 1,3-
      butadiene by inhalation for 103 weeks.
      Source:  NTP, 1993.
1     Effects in reproductive organs are discussed in Chapter 5.  Organ weights, hematological indices,
2     and serum chemistry were not evaluated at 103 weeks.
3            In the 2-year study, bone marrow atrophy was recorded in 50% of males and 14% of
4     females exposed to 625 ppm.
5            The incidence of liver necrosis was increased at the higher exposure concentrations in
6     males and females, occurring in 8%, 10%, 16%, 27%, 29%, and 26% of males and in 4%, 4%,
7     14%, 10%, 38%, and 21% of females exposed to 0 (controls), 6.25, 20, 62.5, 200, and 625 ppm,
8     respectively. Centrilobular hepatocellular necrosis of the liver was seen in 4% and 8% of males
9     exposed to 62.5 and 625 ppm, respectively, and in 2%, 8%, and 9% of females exposed to 62.5,
      200, and 625 ppm, respectively. Hepatocellular necrosis was not seen in any of the concurrent
       1/28/98
6-5
DRAFT-DO NOT CITE OR QUOTE
-------
  1      controls. Liver necrosis with no particular lobular distribution was found primarily in animals
  2      with malignant lymphoma and hemangiosarcoma; centrilobular hepatocellular necrosis was often
  3      found in animals described as anemic and in animals with atrial thrombi.
  4             Myocardial mineralization, a lesion of unknown pathogenesis, occurred with increased
  5      frequency in both sexes at 625 ppm (males, 27%; females, 14%), but was not observed in
  6      controls. A low incidence was observed at the lower concentrations. Myocardial mineralization
  7      was also observed in a separate stop-exposure study in which male mice were exposed to 312
  8      ppm 1,3-butadiene for 52 weeks or 625 ppm for 13 or 26 weeks, and observed for periods up to
  9      103 weeks. The incidence of myocardial mineralization for these three exposure groups was
10      12%, 18%, and 28%, respectively. Details of the stop-exposure study are presented in Section
11      3.3.
12             Minimal to mild olfactory epithelial atrophy occurred in females exposed to 625 ppm and
13      in males exposed to concentrations ^20 ppm. However, the incidence in males exposed to 625
14      ppm was lower than that seen in females. The olfactory epithelial lesions were unilateral at the
15      lower concentrations and bilateral at the higher concentrations.  The lesions were similar to those
16      seen in the NTP (1984) study, but osseous or cartilaginous metaplasia was not observed.  The
17      investigators considered olfactory nasal atrophy a possibly compound-related lesion.
18             Compared with controls, mice exposed to 1,3-butadiene exhibited increased incidences of
19      proliferative lesions (hyperplasia) in several organs, including the heart, lungs, forestomach,
20      ovaries, mammary gland, and Harderian gland (Table 6-2). Hyperplasia of the endothelium
21      (cardiac blood vessels), alveolar epithelium, forestomach epithelium (focal), germinal epithelium
22      and granulosa cells of the ovaries, mammary gland, and Harderian gland were all considered
23      preneoplastic lesions.  Other preneoplastic lesions identified in the 2-year study were
24      hepatocellular foci (basophilic, clear cell, mixed cell, and eosinophilic) in female mice exposed
25      to 1,3-butadiene. Hepatocellular foci were observed in 16% of controls and in 29%, 38%, 24%,
26      10%, and 5% of females exposed to 6.25,20, 62.5, 200, and 625 ppm, respectively. Hyperplastic
27      lesions were also observed in separate studies with male B6C3F, mice using variable exposure
28      and durations (stop-exposure experiments). Hyperplasia in these studies occurred primarily in
29      the endothelium (cardiac blood vessels), alveolar epithelium, forestomach epithelium, and
30      Harderian gland (Table 6-3).
        1/28/98                                   6-6        DRAFT-DO NOT CITE OR QUOTE
-------
   Cfl
   O
   o
   U
  G

  IS
   CS  en
  .2  |
   «! JO

    '^
   1  «
  -5
  •« -0
   c  «
  M  §
   cs w
  H ^








?.
a.
s
"•C
i
B
o
U












m
vo

o



in
VO

o




vo





1/3
Organ/tissue

c^ ^?
>r o^
?~ °
CO ,— j
C; oo
•" So

g o^
^ <=
"* ft


oo ON
CN ^

O CN
O O
>n vt
o —

CN M-
ON O
— CN

0 O
0 0
0

Heart, endothelium

vl 1
CO 00
-~^ • —

Tf CN
CO CN
o o
« Zl

CN °°
| |

« v%
o o
vo "

00 O
0 0
in 
° °
- ^
S &
g
_3
f
"o
u
c
3

VO ON
CN ON
^ t-~
0 t^

°° CN
00 0
•^ «

o m
00 OO

^ 1;
o r-
co ^

1 °
° §;
m >n

00 OO
o o
u-t tn
* *

Forestomach, epithelium

CO
CN
ON
cS

0
CO
o
«

8
o

r-
00
oo

vo
5?
en

s?
ON
-------
               Table 6-3.  Incidence of hyperplasia in male B6C3FJ mice exposed to
               1,3-butadiene by inhalation in the stop-exposure study
Organ/tissue
Heart, endothelium
Lung, alveolar epithelium
Forestomach, epithelium
Harderian gland
Concentration (duration of exposure)
Oppm
0/50 (0%)
2/50 (4%)
4/50 (8%)
1/50 (2%)
200 ppm
(40 weeks)
6/50 (12%)
18/50 (36%)
10/48 (21%)
4/48 (8%)
312 ppm
(52 weeks)
3/50 (6%)
14/50 (28%)
20/48 (42%)
6/48 (13%)
625 ppm
(13 weeks)
7/50 (14%)
10/50 (20%)
8/50 (16%)
3/42 (7%)
625 ppm
(26 weeks)
7/50 (14%)
11/50(22%)
15/50 (30%)
7/36 (19%)
        Source: NTP, 1993.

  1      6.3.  CARCINOGENICITY
  2            The first NTP mouse inhalation study of 1,3-butadiene was terminated early due to
  3     induction of fatal neoplasms (NTP, 1984); therefore, a second study (NTP, 1993) was conducted
  4     to better characterize the exposure-response relationship for neoplasms and nonneoplastic lesions
  5     induced in mice by exposure to 1,3-butadiene for 2 years. The concentrations ranged from
  6     100-fold  lower (6.25 ppm) up to the lowest concentration (625 ppm) used in the first study. In
  7     addition, stop-exposure studies were conducted to assess the relationship between concentration
  8     and duration of exposure on the induction of neoplasms by 1,3-butadiene.  Results of this study
  9     have also been published by Melnick et al. (1990a, b, c) and Melnick and Huff (1992). Miller
10     and Boorman (1990) provided morphological descriptions of the neoplastic lesions induced in
11      B6C3F, mice by 1,3-butadiene. The results are presented here in two parts, 2-year study and
12     stop-exposure study.

13     6.3.1. 2-Year Study (NTP, 1993)
14            The details of the study design are described in Section 6.2.  For neoplasms that were
15     considered to be lethal tumors, the tumor incidence was analyzed using the life table test, a
1 6     survival-adjusted procedure appropriate for rapidly lethal tumors (Cox, 1972; Tarone, 1975). For
17     incidental tumors (tumors discovered as a result of death from an unrelated cause), the primary
18     statistical method used was the logistic regression test. Alternate statistical methods included the
19     Fisher exact test and the Cochran-Armitage trend test (Armitage, 1971; Gait et al., 1979),
20     analyses based on the overall proportion of tumor-bearing animals.  Tests of significance
        1/28/98
6-8
DRAFT-DO NOT CITE OR QUOTE
-------
  1      included pairwise comparisons of each dose group and a test for an overall concentration-
  2     response trend.
 *3            Exposure of male and female mice to 1,3-butadiene induced a variety of common and
  4     uncommon tumors at multiple sites. The incidences of primary neoplasms associated with
  5     exposure to 1,3-butadiene (for the 2-year study) are presented in Tables 6-4 and 6-5. The
  6     percentage of animals bearing malignant tumors increased from about 30% in the controls to
  7     nearly 90% in the highest exposure group, 625 ppm. The results of interim evaluations for 9
  8     months and 15 months are presented in Tables 6-6 and 6-7.
  9           As in the previous study (NTP, 1984), exposure of mice to 1,3-butadiene was associated
10     with the development of malignant lymphocytic lymphomas and to a lesser extent with
11      histiocytic sarcomas.  The incidence of malignant lymphomas, particularly lymphocytic
12     lymphomas, was significantly increased in males and females exposed to 625 ppm and in
1 3     females exposed to 20 and 200 ppm (survival-adjusted) compared with controls. In addition,
14     there were significant exposure-response trends (pO.OOl) in both sexes.  The lymphocytic
1 5     lymphomas were well differentiated and occurred as early as week 23, peaking before the 15-
1 6     month interim evaluation. Many organs, particularly the spleen, lymph nodes, liver, lung, and
17     kidney, were affected  in mice with lymphocytic lymphoma; however, the thymus was involved
18     in most mice and was  the primary organ affected in some.  The lymphocytic lymphomas
        consisted of uniform populations of small- to medium-sized lymphocytes, whereas the mixed and
20     undifferentiated lymphomas generally consisted of more heterogeneous populations of
21      lymphocytes with pleomorphism and atypia.  Other histological types of malignant lymphomas
22     (mixed and undifferentiated), commonly associated with the spontaneous lymphomas in aging
23      B6C3Fj mice, were seen at low incidence in some groups. The incidences of histiocytic sarcoma
24     were significantly increased in males and females exposed to 200 and 625 ppm and in males
25      exposed to 62.5 ppm.  The histiocytic sarcomas (previously referred to as reticulum cell
26      sarcomas or type A sarcomas) were large and monomorphic, with dark basophilic nuclei and
27      relatively abundant eosinophilic cytoplasm.
28            Hemangiosarcomas of the heart were observed in male (at >20 ppm) and female (at >62.5
29      ppm) mice exposed to 1,3-butadiene for 2 years.  The incidences of hemangiosarcomas of the
30      heart were significantly increased in male mice exposed to 2:62.5 ppm and in female mice
31      exposed to ^200 ppm. There was a significant exposure-response trend in both sexes. The
32      cardiac hemangiosarcomas were observed in all ventricular locations, but were more frequent in
        1/28/98                                  6-9        DRAFT-DO NOT CITE OR QUOTE
-------
o
c
.H
-3


I

I
5









'i'
c.
B
5
1
<3

























V)
vo

g



in
S


o
in
vo

o







c
"ert
O

"E,
o
2
c
M
O

&£

es
H


*"
^1? P
P "^ ?
^ ^
NP ("•••
O^ ,j« f\)
O v ^
to  P
en =>•
l|i


CS
p"1* v^
ON *«J* S?
CN
£ *
C3 vP CD
0-Sg
|"l
20/50 (40%)
47.3%
p=0.517
3?
C2- o\ cs
o vo" o
.



CO
s
o
c
"O
cd

Alveolar/bronchiola







e°
3
J
^-^
S^ p
en o\ ??
^ ""  o\ V
s ^
l*i
o\ oi o
1-«
g
PI
6/50 (12%)
15.4%
. p=0.577
,-.
C,en §
o ^ ?

o
CO
g
o
e
"o
CO
o
:-!
Alveolar/bronchiola
adenocarcinoma










,-,
S? P
r*"i i*iC ^5
en °-
|||
o £? o
l/-j ^3 v
>?> "" a-
en jc S
s; ^ ?
^5 °° fc-

^ - VO
egR
Btii
at *•
!*i
c?
9^0
O •* O
!Q >o V



co" cd
S E
0 0
a _e
cd co
J~ 0
Alveolar/bronchiola
adenocarcinoma, or










^.^
en ^- 2
C/o •*

S* t
!«i
^_, o
w^ *^ ^
F=
i~^g
^^ .' ^^
*o ^^ 11

-? z
PI
o sx
	 	
eg §
1^1,






cd


Squamous cell papil


jz
0
r^

o
S3
0
fe
^
C-lo o
m ~ o
^ ^
vf 5/ S
^ o
^? *o V
So
l^vS
l^?

^ ^
Ss!
o sa.
/ 	 	
G^ 8
° oi *•*







E

Squamous cell papil
carcinoma











b-P S

cs o P
S"^
1^, o o
oo ^ o
«^- c^ \/
-------
I
t3
I
 O
 •*•»
 -a

 o
 CU
 s
0
 S
en

"£*§"
O  S
s .S
f^ "S
a  s
ii
o
4)
a  §
 . '-C
T «
vo "3



a,
•St

BJ
S
'c
1
o
U






s
o
o

tfj



o
^
vo
o



e
_o
^o



P)I
e





c
cc
OB
o

DC
CC



5" s? O
^ 0 3
I- 1
lil
S '§ t
CO
Iss
00 °< S?
50"
?5 <*•
CN

?>* p
S- — o
1 ^ t
m
^ cN ^
o en O
«2 VN «
m ^
CN
^ K
3 ;£ s
•S/ ON 0
o t-^ o

•^
i. i I
m <3 o
P 0 ||
S ~" ^
^^ Np J— s
s s 1
0 °-

1 1 1
i ^ ^


^, oo m.
? * 1
1 - I
CS ^ o
*•»—'•' ^^ ^^
o ^ o
~*^ ^H
VD









O
C
u
•o
<

•a
c

"3)
c
S

•

0* — ' 0
•n — II
en ^*


o 3 2
t| |
| n ?
i" ^ «
^ ° ^
o ^










i
§
ca
U









5~ vp CS
« £ §
R § ?
^ — o.
~~s V~t <^>
u?> "^ "v
m
!*§
0 •* 0
c5 ^
CN

9/50(18%)
29.5%
p=0.217
§ 2? S
g-t
-------
0)
u


2
a
o
O
c

1






















tn
5o

O
«*«

V)
vo


o
ts

vo

o



c
0
1
tn
cs
o
o


c
cs
M
o

BO
u

H



*— ' ob °
8*1
CO
1*8
5\
i? o «•>
2£S
° s 1.
^

§SP°J
cspS
§s!
—"
1*8
1*1
„
1-1
VO


o
00 _
O 'w%
>\ ,;ii ^cd
ll|
i& ^1* 3






TO
00
O

"*
f #%
goo
?- Hx

*. J° g
t>
^
&,* 2
o t~*- ^


III
f1
!*i
s*t

« o.




a
o
C3
CO
&
s











III; §
O CO P
go t— V
vo O,
CO
c^ * 0
o vo o
IQ WO V
VO °H
1— 1
cs-
G-cs o
§ ^X
—

1*1
'—'
14/50 (28%)
37.0%
p=0.136
C? 0
^^5
C*""1 o
S°'|
ON ^<
O
0

0
cd
o
JZ
a.
I
c
^ 1












3N ^ S
» o °
CS
S'O 8
o o o
in o v
i: a.
cs
,.£,
ci* m
§^ ?


||<

^

sil
§°' v_




cd
O
B
«
O
'i
cd
u
S








•e
e
K
rS ^ 2
|2°,
I— H
Ci-o o
-* o °
^- o.
g>
G.TC §
§ s ^
-H

CS^ o
-^
o o o

* ••? o
l"t

CO
e
0
g
•s
—
o
1
g
J3
^








00

J
f^s
C3 ^^

G-o o
|§i
C5- ]_|
C-» o
§ ° 1.
ON

CJ-ON o
sg?
— '
02?

**~* o
o ^«


P
§
'g
"o o
is
J2 g
S °
> 
< CO











lii
?2 8 9
cs
2 ^ s
cs
«? »-
S-v^ o
g g2 ?
cs

ro o^ ^
*— '
I^i
S °^ "?
^ m a.

}l|

co" co
e G
§|
fc- O
«* tH
•o °
i!
^CO g
"s °
_> g
< Cd











1^^
° o 8
lit
»— i
f*-5
i2* ^ o
|§?
c^ °
CS* cs
§ ON" ?


o • o
V) °° 1!

fis

o _• o"
«o o V



cd
o
'E.
Cd
0,
u
g
i
cr
01



o
rt





e
gsg
0 0 0

*^ 0 '*
o°^o
•0 CO ||
^
ci* cs
S 2 ?


fs|

l*<
vO i-^
CD — ^ /&H
•n o
0

|||
O O«



cd
c
§
0
o
en
i
K>











i"*s
CS " °H
cs
^^s
SS* *•— ? o
^2^
^,
S^ 0
O vo O


fit

fis

III



0
cd
o
"3.
cd
O.
u
to cd
g S
l|
CO 0











O o "~>
§ 0 ?

Ci-o §
0 ON 0
 "i
^ o

*•— ' CO *~]
ON o O
•* — II
-3= o.



CO
1
2
ca
t-
o
td
ex
K











    1/28/98
6-12     DRAFT-DO NOT CITE OR QUOTE
-------
I
*8
 O
 I
 OJ
 C3
 O 0
 « a
 a •-
sZ^ w
•^ ^
« tT)

a S
"I
"? ^S
vo "«
^5 -S
S -5









If
CL
e
_o
Concentra









!Q
NO



O
o



V)

0


s
NO


o




s
o
'to
_o
En
JS
O
&

s
03
bfi
o
"*•*
o
ff
S



III
ill


Co1
CS sP £§
|sf-
2


19/50(38%)
74.8%
p=0.027

o in P
§ "* a'

g> y.
^-"t 
J

=£ £? 0
ON f~I O


^~,
o^ sPO
2^8
!§t


CN g?f|
^,!

^^5
C-° o
g§?
2^ ^


••£-
S^g
Pi


|1|

^ CN

^^
^^ S
l^t


o'^ §
5 ci,







Carcinoma









C^SNO
Pi



s^<:
go2


0

III
in °
o

^^
0,2? <
o


III
^ c^




!_'
1
•g
Malignant mix









-*£
ON


/^.
°^0
O ON P
VI oo 8


o"-; o
^^ ^.

6/50 (12%)
20.7%
p=0.511N

? vo
0 vO JS
fNj ^N. W '
Ci-o) «
0 ON P
in  o
•n CN II

^ vo
§,^ £
O ON P
>n c^
£•§
2«o
Sit
C*^ >••
g'is
3™ oft
SS-c
aj 5*J>
|--o ra
M_
5 C oj-a G
a*— s 2 M
=!"!!' f
Iflll 1
S 2^^= >^o. o
3 C ,y*r3 o ^
£ W «^ C w §
3 « w u u S- u
'if-l--i ™ 'I
" o Siz: &o QJ ~
^•55 o ,^t t-H
^|Sg.» i-g-a
s &g sp £c §•
' iS W)»J o 1.°


3 *"" § O ™ —.23 o c*~)
— "O(^"O ® -^ S*^ ^^
i^ -gl 15.5 2
iT-i-l-l Ss| ^~
g?«22 2-s| S
—"COS'S 2 o cS
•g-SSg-K cllS g
>i:~j'l? <§! s
Orrt. 1~. 1 Jiii, 	 O

    1/28/98
6-13    DRAFT-DO NOT CITE OR QUOTE
-------
o
c
.—
'O
c:
to
in
O


I

O
u
0
«
*«
s

.s
VI
s  s
J- rrj

   a
'S  2
a-S
<*-<  (-
o  §
o  a
o  B
S3 CN
"1
 • *l^
vp js
vi "K
« -s
3 .S


c.
c.
c
o
C3
O
C
o
U



V)
8
|
in
8





S
V)
\o
o




e
o
CO

CO
CS
"5,
o
C2
e
cs
BC

O
•s
ff
o














c
o
O*v

r-












.c
c
o

^«


O oj
. ,0 co" .2
•s I "M "f> i I
g S & o z. |
f>> ,t!i»* ^Zi' _*^* C 3




CO
C
cs
I














J3
G
0
o>

c;
en
0










CO

•^

^
»-*

ed
Hemangiosarcom




t:
1
0
en
0
o
0





^


o



c
o
Os

s
0
o









CO

"fi

s
1—"

g
Alveolar/
bronchiolar
adenoma,
adenocarcinoma,
carcinoma




1
0













x:
c
o
OS

r?
0










CO

•g

S
^^


Squamous cell
papilloma or
carcinoma


.c

a
o
CO
a
O
0
0
o





o
0
s
o


•s

o
OS

t—
0
O
CO





o
§
o

o
CO
_C

O

^-.


Hepatocellular
adenoma or
carcinoma




Ui














J=
c
1
Os

S
o
0





0




_c


S
m



Adenoma or
carcinoma
1



|| Harderian
.§
OJ
O.
E
S
2-='
*> "«
C w
•a "°
s-s
S  o
0->-
. 2 o
£••$ 2
Hi 8s,
| 8 S
£ SI
S2t
l-ll
o a.13
1 1-3
.S S) c
<§!§
g-§S
E « "
~ •£ o
.2,5 'S.
•f s SP
u, C .= '
 > o
.£ -S x) g
s « a S
S S -s -i •
iJlN
S ^ CO w C
So « .2 o
i -s S s. S
£ E •« T3 «
2JL^l
O CU -. « to
imt
n *o *-• {W u
ilili
.£ -B O t_^ rn
C to *O ^^ -O
PJ (L> *™ . lg
W3 ^ C o, bO
Illli
IM K Q U CA
g a w o 15
I^^^H
"Overall rate: number of
'Survival-adjusted rate.
'Life table test. Beneatt
between the control an
'Logistic regression test























































o.
3
O
00
~5
c
1 1
2 -S
00 C
fll SO
CO J"

' tumors in tl
ecified.
roup is lowei
NA = not applicable; no
NOS = not othenvise sp
N = incidence in dose gi































































m
O\
OS
3
0
00
     1/28/98
6-14     DRAFT-DO NOT CITE OR QUOTE
-------
    cu
    o>
    t»
    O
   NO

   CQ
^BP^ i^^
   CM O
   2  S
   a in
   l
   ="  fi
   o  o
   g <*
   il
   3 «M

   -1
    • -^^
   T ^
   vo "«
   «* -



IT
n.
Or
O
C
e
0
u
c
o
u







O
c
V)






o
pj

IT)
VO
o

E

tu
-J
CO
"a.
o
e
L.
o
&£



















9 months

C>
o
0












o
1


•c

o
















•si
3
OS




O













to
3



c
Adenoacanthoma,
adenocarcinoma,
carcinoma, or maligna
mixed tumor
3
"ob
3 -

CO
"^
















•sj
3



O
O
o







-»^



|N
t/
c




Adenoma or carcinom
•o
"3>
3
tu
•o
C3

                                                                    a.
                      3 °

                      §>3
                      =5 S
                                                                   II
                                                                  c 12 *°

                                                                  fc J3 TO
                                                                  fe
                                                                  ttt
                       .
                      (M >>
                                                                 f I
                                                                  3 -O g
                                                                 *- *s a
                                                                 rt M *"*
                                                                 •o-l

                                                                iSli
                                                                I-H
                                                                .£ o
                        8"!
"sills
Hs-s i
IHif

11 I'll
'c '•= 4J H ">
i g 2 -g
on .. y co d
.g .2 -S §• g> .
« ^ ^ 2 i
.S'? S °og
j? c S -o "
o ^ 8 « .a
Pi!5
iill!
li«ii
c g „• 8 g
« ^ S g eh
g ? u ~ £
— "a 3 g .2
"5 •- - g ts
S | .-a 1 '§>
o M j * 2
si
1
.DO
1
C
1 8'
s .s
00 c
o =S
s r
•- o
CO - — •
h; *a co
2 u —
f S §•
f 8 1
§ g-M-
j?8 1
11 =
til
s is
?••" s
1! cy-J —
-------
 1      the left ventricular wall. Typical hemangiosarcomas had solid foci of anaplastic, pleomorphic
 2      spindle cells at the center with a loose arrangement at the periphery.  They were occasionally
 3      multifocal and frequently coexisted with foci of endothelial hyperplasia distant and separate from
 4      the main neoplasm.  Hemangiosarcomas of the heart are considered uncommon in untreated
 5      B6C3F[ mice (none were observed in 573 male and in 558 female historical controls in NTP
 6      inhalation studies).  In male mice, the lower incidence of cardiac hemangiosarcoma at 625 ppm
 7      compared with that at 200 ppm, was attributed to the early mortality  due to induction of lethal
 8      lymphocytic lymphoma at 625 ppm.  The time-to-tumor detection for all hemangiosarcomas of
 9      the heart ranged from 682 days at 20 ppm to 289 days at 625 ppm for males and from 649 days at
10      20 ppm to 307 days at 625 ppm for females. When hemangiosarcomas occurred in multiple
11      organs, the cardiac neoplasms were usually designated as primary, because the incidence of
12      hemangiosarcomas was highest in the heart and the earliest lesions occurred in the heart.
13      However, it could not be determined with certainty if the hemangiosarcomas observed in other
14      organs were metastases or primary neoplasms. Subcutaneous, splenic, and hepatic
1 5      hemangiosarcomas that were found hi the absence of cardiac hemangiosarcomas may reflect the
16      development of spontaneous vascular neoplasms known to occur in B6C3FJ mice.
17            Exposure of mice to 1,3-butadiene was also associated with an increased incidence of
18      pulmonary neoplasms in male and female mice. Although the incidence of alveolar/bronchiolar
19      adenomas was not significantly increased in male mice in the 2-year study, the combined
20      incidences of alveolar/bronchiolar adenocarcinomas and carcinomas and the combined
21      incidences of the benign and malignant pulmonary neoplasms were significantly increased at
22      62.5,200, and 625 ppm. In female mice, the incidences of the benign and malignant neoplasms
23      analyzed separately or together were  significantly increased in all exposure groups compared
24      with controls.  Thus, even at 6.25 ppm, 1,3-butadiene was carcinogenic to female B6C3FJ mice.
25      The lower incidence of lung neoplasms at 625 ppm compared with the incidence at 200 ppm was
26      attributed to the high rate of early deaths due to the competing risks of lymphocytic lymphoma in
27      female mice exposed to 625 ppm. There was a significant exposure-response trend for combined
28      adenomas and carcinomas in both sexes. The time-to-tumor detection for lung tumors combined
29      ranged from 587 days at 6.25 ppm to 251  days at 625 ppm for males, and from 519 days at 6.25
30      ppm to 275 days at 625 ppm for females.  The spectrum of lung lesions ranged from alveolar
31      epithelial hyperplasia (Section 3.2 of this chapter) to adenomas, carcinomas, and
32      adenocarcinomas. Histologically, the alveolar/bronchiolar adenomas exhibited distortion of the
33      alveolar structure due to the formation of complex, irregular papillary patterns; the
34      alveolar/bronchiolar carcinomas were similar, but consisted of heterogeneous cell populations
        1/28/98                                   6-16         DRAFT-DO NOT CITE OR QUOTE
-------
        with various degrees of cellular pleomorphism and atypia. The adenocarcinomas were larger,
        highly anaplastic neoplasms, often accompanied by hemorrhage or necrosis.
               In the forestomach, significant increases in squamous cell papillomas and carcinomas
  4     combined were observed in male mice exposed to >200 ppm and in female mice exposed to
  5     £62.5 ppm compared with controls.  There was a significant exposure-response trend for
  6     papillomas and carcinomas combined in both sexes.  The combined incidence of squamous cell
  7     papillomas and carcinomas of the forestomach (males, 4/575 [0.7%]; females, 9/561  [1.6%]) for
  8     historical controls suggests that these lesions are relatively uncommon in B6C3F! mice.
  9            Increased incidences of hepatocellular adenomas and carcinomas were also seen in 1,3-
 10     butadiene-exposed mice (Tables 6-4 and 6-5). The hepatocellular adenomas were discrete,
 11     expansile masses; the carcinomas were larger than the adenomas and consisted of markedly
 12     disorganized hepatocytes. The low incidence of liver neoplasms observed in males and females
 13     at 625 ppm probably reflects increased early deaths from malignant lymphoma. Hepatocellular
 14     adenomas and carcinomas are common neoplasms in B6C3FJ mice, occurring in 196/572 (34%)
 15     of male and 87/558 (15.6%) of female historical controls in NTP inhalation studies. The data
 16     suggest that 1,3-butadiene has only a weak tumorigenic effect in the livers of male and female
 17     mice. However, a chemical-related effect is supported by the detection of an activated K-ras
 1 8     oncogene in liver neoplasms obtained from mice exposed to 1,3-butadiene (Goodrow et al.,
        1990).  According to Reynolds et al.  (1987), activated K-ras oncogene had never been detected in
 20     liver neoplasms from untreated B6C3FJ mice.
 21            Although a variety of neoplasms were seen in the ovaries of female mice, only benign
 22     and malignant granulosa cell tumors  were definitely attributed to exposure to 1,3-butadiene
 23     (Table 6-5). The ovarian granulosa cell tumors varied from small benign tumors to large cystic
 24     tumors with thick trabeculae and spaces filled with blood or clear fluid. The overall historical
25     control incidence at NTP for benign and malignant granulosa cell rumors each was 1/548 (0.2%).
26            Increased incidences of mammary gland neoplasms were seen in female mice exposed to
27     s62.5 ppm 1,3-butadiene. Mammary tumors included adenoacanthomas, adenocarcinomas, and
28     malignant mixed tumors, the latter occurring only at 625 ppm. The mammary gland tumors
29     combined exhibited a significant exposure-response relationship.  The adenoacanthomas were
30     considered variants of adenocarcinomas that have prominent squamous differentiation. The
31      malignant mixed tumors consisted of epithelial components arranged in glandlike structures and
32     anaplastic spindle-cell components. Mammary gland adenocarcinomas and adenoacanthomas
33     were considered uncommon in female B6C3Fj mice; the overall historical incidence at NTP was
34     21/561  (3.7%) for carcinomas and 1/561 (0.2%) for adenoacanthomas in female control mice.
        1/28/98                                  6-17         DRAFT-DO NOT CITE OR QUOTE
-------
 1            The Harderian gland was identified as another site of 1,3 -butadiene-induced neoplasms in
 2      male and female mice (Tables 6-4 and 6-5), with significant exposure-related increases in
 3      adenomas at 62.5 and 200 ppm and a low incidence of carcinomas in males exposed to 6.25, 20,
 4      62.5, or 200 ppm. The low incidence of Harderian gland tumors at 625 ppm was attributed to
 5      early deaths due to lymphocytic lymphoma which precluded the development of Harderian gland
 6      tumors.  The investigators noted that the occurrence of Harderian gland carcinomas in mice,
 7      particularly males, is unusual. The overall incidence of Harderian gland carcinomas was 2/575
 8      (0.3%) in male and 3/561 (0.5%) in female historical controls at NTP. The 2-year historical
 9      incidence of adenomas and carcinomas (combined) of the Harderian gland for control groups in
10      NTP inhalation studies was 25/575 (4.3%) for males and 13/561 (2.3%) for females.
11            Preputial gland carcinomas, also considered to be rare neoplasms in B6C3F! mice, were
12      seen in five males (p<0.05) exposed to 200 ppm (none were reported in one survey of NTP
13      historic control data). These tumors were also thought to be exposure-related lesions. Some
14      preputial carcinomas were composed of large eosinophilic epithelial cells that were well
15      differentiated; more frequently, the carcinomas had necrotic cores and a thin layer of very
1 6      anaplastic basophilic epithelial cells that aggressively invaded surrounding tissue and blood
17      vessels.
18            Renal tubule adenomas were seen in 2/50 females exposed to 200 ppm 1,3-butadiene and
1 9      in 1/50, 3/48, and 1/49 of males exposed to 6.25, 62.5, and 200 ppm, respectively.  At the 15-
20      month evaluation, renal tubular adenoma occurred in 1/7 males exposed to 625 ppm.  The
21      historical incidence of spontaneous renal tubule adenomas in untreated control groups in NTP
22      inhalation studies was 1/571  (0.2%) for males and 0/559 (0.0%) for females.  Histologically, the
23      renal tubule adenomas contained multiple dilated tubules separated by thin connective tissue
24      septa. These renal lesions were probably related to exposure to 1,3-butadiene in males and
25      possibly related to exposure in females.
26            One neurofibrosarcoma of the subcutaneous tissue was observed in two females exposed
27      to 625 ppm at the 15-month evaluation. In the 2-year study, the combined incidences of
28      neurofibrosarcomas and sarcomas of the subcutaneous tissue were significantly increased in
29      female mice exposed to 62.5 ppm (p=0.017), 200 ppm (p=0.002), and 625 ppm (p=0.013) by the
30      life table test.  Subcutaneous tissue sarcomas (all types) were considered uncommon  spontaneous
31      neoplasms; the historical incidence was 2/561 (0.4%) for female controls at NTP, suggesting that
32      these subcutaneous tissue neoplasms may have been exposure-related.  The historical incidence
33      for NTP inhalation studies was not reported.
34            One adenoma and one carcinoma of the Zymbal's gland were seen in females exposed to
35      625 ppm; one adenoma also occurred in a concurrent control male mouse, but none were

        1/28/98                                   6-18        DRAFT-DO NOT CITE OR QUOTE
-------
         reported in historical controls. The report indicated that these Zymbal's gland neoplasms may be
         related to 1,3-butadiene exposure.
                Carcinomas of the small intestine, another uncommon tumor in the B6C3F! mouse, were
  4     seen in two females exposed to 6.25 ppm and in one female exposed to 62.5 ppm.  One
  5     carcinoma each was seen in one male each exposed to 6.25,20, or 62.5 ppm, and in two males
  6     exposed to 200 ppm. The relationship of these neoplasms to exposure to 1,3-butadiene could not
  7     be determined; however, controls did not exhibit proliferative lesions of the intestine.
  8            In supplemental analyses, the authors performed a "Poly-3" quantal response test (Bailer
  9     and Portier, 1988; Portier and Bailer, 1989) as an alternative to the logistic regression analyses,
 10     whose sensitivity was reduced by the decreased survival in the higher exposure groups.  For
 11      tumor sites related to butadiene exposure, the "Poly-3" test detected significant responses in
 1 2     some of the exposure groups that had not been detected by the logistic regression analyses. The
 13     overall  results were consistent with those already presented.
 14            The authors also fitted a modified Weibull model (Portier et al., 1986) to the "Poly-3"
 1 5     survival-adjusted tumor rates to determine the shape parameters for the exposure-response
 16     relationships.  About half of the tumor sites associated with butadiene exposure had exposure-
 17     response relationships consistent with a linear model (i.e., shape parameter of 1). Most of the
 1 8     other tumor sites exhibited supralinear exposure-response relationships (i.e., steep slope in low-
         exposure region; shape parameter significantly <1). These sites were the liver in male mice, the
•f^
 20     mammary gland in females, and the Harderian gland and lung in both sexes.  Only the malignant
 21      lymphoma in males and heart hemangiosarcoma in females had a shape parameter significantly
 22     greater than 1, suggestive of a sublinear exposure-response relationship.

 23      6.3.2. 2-Year Stop-Exposure Study (NTP, 1993)
 24           An additional study with B6C3FJ mice, referred to as "stop-exposure study" was
 25      conducted to assess the relationship between exposure level and duration of exposure to outcome
 26      of 1,3-butadiene carcinogenicity. Groups of 50 male mice were exposed 6 hours/day, 5
 2 7      days/week at concentrations of (a) 200 ppm for 40 weeks, (b)  625 ppm for 13 weeks, (c) 312
 28      ppm for 52 weeks, or (d) 625 ppm for 26 weeks.  After the exposures were stopped, the animals
 29      were placed in control chambers for the remainder of the 103-week studies. The total exposures
 30      to 1,3-butadiene (concentration x duration of exposure) were approximately 8,000 ppm-weeks
 31      for groups exposed to 200 ppm for 40 weeks or 625 ppm for 13 weeks; the total exposures were
 32      approximately 16,000 ppm-week for groups exposed to 312 ppm for 52 weeks or 625 ppm for 26
 33      weeks.  No additional controls were included for these studies, because they were run
         concurrently with the 2-year studies.

         1/28/98                                    6-19         DRAFT-DO NOT CITE OR QUOTE
-------
 1            Using the stop-exposure protocol, inhalation exposure to 1,3-butadiene had no effect on
 2      mean body weights. However, exposure to 1,3-butadiene markedly reduced survival in all stop-
 3      exposure groups as a result of the development of neoplasms, particularly malignant lymphomas
 4      and hemangiosarcomas of the heart (Figure 6-3). A comparison of the two groups receiving total
 5      exposures of 8,000 ppm-weeks showed that the survival of mice exposed to 625 ppm (13 weeks)
 Q      was similar to that of mice exposed to 200 ppm (40 weeks). By contrast, the groups exposed to
 7      16,000 ppnrweeks, survival of mice exposed to 625 ppm (26 weeks) was significantly lower
 8      than that of mice exposed to 312 ppm (52 weeks).
 9            Neoplasms induced in the stop-exposure studies are summarized in Table 6-8. Overall,
10      the data show that exposure of male mice to 1,3-butadiene using the stop-exposure protocol
11      induced neoplasms at the same sites as those observed in the 2-year  study.
         s

         CD
         §
         0.
            1.0»
            0.9-
    MALE MICE
  • CONTROL
  O 200PPMSE40
1  A 312PPMSE52
  D 625PPMSE13
  • 625PPMSE26 !
                       15
            30
45      60       75
  WEEKS ON STUDY
90
105
120
        Figure 6-3. Kaplan-Meier survival curves for male mice in the stop-exposure inhalation
        study of 1,3-butadiene.
        Source:  NTP, 1993.
        1/28/98
                             6-20
                     DRAFT-DO NOT CITE OR QUOTE
-------

 S
 •o
 o
 &
 S
fa"
u
 cs
 E
 A
 tfl
 CS
 i =
1 1
 & &
•8
 1 "
a .a
o  §




I
VI
&,
S
o
s
i
s
s
_o
1
&
o
o

























s
e.
 xd o
|N Ik

^
So? ,
S "xt
CS


CS
S
o
•a
s
-g
oi
.SP
13
S
_o
^
t
^

0
"S
'S
0.
ff
cS
(U
a
/— \
|g|
0 co" 
ro ^

"N° _ V>
fei§
o oo o

fgl
s « X

^ 0
feg s
l^t

^
^^ :
S "^
c5




o


o
T3
.1
g
I
O
•a
i









lis
® iQ ?
CS ^

$o?§
OP ^
fi

OX N? ^
sl; p
O 00 °

NO o
c,^ P
s « ^

^
^o !
^ o
o







sarcoma
_o
o
o
ffi









NO
0 *~"
O I-H P
in ft"
ro
o° 0 ^
$3° o
o ° o
«n ° V
In ^ ^
^— <
Isi
s^?
^ ft,
CN)
^
\O NP HI*
CNl OX O
N_/ OO O
CO xO CO
1
^
xS^ j
s ^
^
eS
1
O
cS
en
^
'43
O
o
"1
la
S

lymphoma
-g
.£?
1









NP
^ 0 ""
xS^ WO O
0 OS ^
CO ^
CO
o^ vP ^
\O ON CO
^ O O
S~' O C5
0 0 V
OO

so
Igs
0 00 P
cs
Co" „
c,^ §
OO

y-"— -,
ssg ,
S OS
^
-cT
tu
X
"s
^o1

0
o
1*
^

iyraphoma
-^
1 5?
bpO
>=S t-,
^; o









O^ co o
CO O O
° 0 v
2

NO
III
PI
m
xp
O^ i— 1
T^- X^ ^^
I'S?
C"-
s.O
OS 0
^^ "s^O ^~*
C5 ^O C5
V^ £*•"- \/
1—1
^
eg ,
S o
o







arcoma
en
O
5b
P
S
ffi






•e
C3

cs, o o
o § o
CS ^
T— <
f^ o^ 0
5C-0 0
pt
cs
o?
CO o? "^"
o in" <=>
1 — 1
ox "^
OO N° J^J
x— ' CO O
o ^ o
la^lk
CS
0?
CO =£•
x^ OS i
O \O '
OO
T— i




08
O •







60
c
3
^^, ^ ^
^§?
^d ^ ^
T— H
CS o° S
C,P p

•— "
^ _
XO NP , — ,
C2-^ §
|SS?

^
I o? 0
^ ^ l/"i f*^
O OS O
10 oo V
cs ft,
cs
x?
2 ^
o 2





cS
1
"o
I*
b "*
cti
.2 °
,_r^ ^
S s
o 2
i- S
*8 "o
^ (-<
cS ^
0 |
_> CD
< "c3









    1/28/98
6-21     DRAFT-DO NOT CITE OR QUOTE
-------
S)
a
a>
VI
O
a,
0
   3
  .S
  <^rf
   S9
   in
 o
 §
S
   &
.a js
oS g

vi|
 ... e?
 0> «M
"S es



exposure)
CM
O
a
.2
«**•
2
3
gj
O
2
•*-*

a
o
O































S
CM

VO
CM
CM
T-C
fO



S
CM
CM
S

s
IM
^
«•••»
(S

£M
CM
CM
O











£

oo



o









3
1
£
CM
>&
¥
3

0
I
"3
o





































s
o
1
to
si
1
g





03
a
^
o
"S
a
i^
CS
H

fill
<= § °
^- a,
"xt" ox o
VO co co
o o o"
CO

/~s
!£,o o
S§?
OO ^^
cs
CS ^ ~
Bog
O ^ O
l/^ ^^ \/
vo ^*
m
/— s
^-> ON J
C5 ^J*
io in
»— t
CJ






s s
o o
c .g
CS 03
1_ O
CS !_
.2 ^
,.tr1 03
0 g
1.1

"« cs
"3 0
tt> C
> o
<•§












Sii
o o o
S 2 ii
^~ ^
G/ ^^ o
,__!

, — ^
x3- S^i °
ON r-^ cp'
^^
oX. -o
tO sO T~H
(H ox. o
x_^ t— < CO
ON ^ O
5; ON V

cs
ox
o cs" '
CO









ca
S
o
§
1
~5



CS
O
o
CS
S
o
1
"3

 cs II
co a.

^
s cQ
T— *









.2
'SL
ca
n.
"S
o
CO
O
1
cr
oo



J3
O
03
a
o

o
o

c^^g
^ON °.

vo
0s- 0 •— <
*>*^ ^"^>
S 8 v
10


SS VO O
O r-^ C>
10 wo V

^
S-^- <3
o • JZ
10 °
^

^
xB'o ,
o








CS
s
o
CS
o
"o
o

o
g
03
§•
OO












|||S
o cs ??
>o 10 V
o
•x?
OX sP ' — '
OO <=x O
,-< C3 O
" ' o o
o o v
^ -1 a,
ON


|§§
|^ t

^p- g^
^ — ' ^" O
o ,-; o
;o cs II
co a<

&
sp !
»— 1

OS
s
o
.S
"o

03
O
u_
o
cS
_0
1
Q*
1
CO
O
cr












35 xo VO
xS P. P
o o o
CO
T-H
^ 0 ^
CS

, — ^
S-c^ §
0 ^J-' 0
§0, ^
cs
N?
0s- T.
fSl N.O T""t
^ ^ S
v — ' ON O
o r-^ o
>o oo V
v^ a
cs
"x?
o 2;
v^"5













I
O
i
-a


"O
CS
"M
j2
ca
*^
(D
"2
CS
ffi
    1/28/98
6-22    DRAFT-DO NOT CITE OR QUOTE
-------



£
4>
1
-3
3
s
^
i
TH~
o
S"*
-a
cu
l«
O
c-

^
M
JO
C O
• »4 53
1 1
"S^
2 •§
S3 -3
R S

•s |<
'o ?
ible6-8. Incidence
halation in the stop
H -5












^
9
O
a,
«
Q

a
_o
1
2-
s
^3
2
+->
0
w
0
U























S
a,
a,
in
g
a,
a,
TH
ro



S
a,
a,
rq
^0


1
o
0


g
a
a.
0








Parameters











VO


^




fO
^



o



fO
o






-
1
1
I
i
a
X
0
§,0 0
o o o
^2 ^
^ 0 -
o ^ S
cT o o
S o v
o ^ ^
CO


^ 0 §
o § °
CO ^
cs

NO
" — ' CO O
o od o
in oo V
C~- ^i
cs

O^ !
in ^^







si
Adenoma or carcino










S® 1 — 1
CSSN c^
o> • o
,—,
S'OO P
CO ^



£co S
S 2; °
^ ^


Co^ "°
o\ x<= CO
oo °> E;
s 	 ^ -xj- O
oo c-~ o
^F ^


Ig ,
in O
O






ca
o
 1
in O
O






I
Adenoma or carcinoi










s-~<
0s- N® ^~1
^J* 0s* — .
o'P o
iO CO "
cs ^
^ 2
ell
o a



^ oo
o^o
?3 a


^ rf
O _,; O
in ^r ||
^ a


|g ,
in ^
1 — '






03
Adenoma or carcinoi

T3
C
^
bO
CO

^rt
^
S
N
1/28/98
6-23    DRAFT-DO NOT CITE OR QUOTE
-------
The
gro
o
3
 t
•8
en
Q
g
O
U
0
CJ
'"* ;£
p s
jl
"S. s>->
O T3
« a
ss •*••
~ t«
:e of primary
op-exposure j
c "£
0> o
15 js
o *•'
5.s
Table 6-8.
inhalation


•SSg
•3 S51
•S ^ '5 1
15 al
cs » 3 ns
8 o o J* _^
istil
S -a o 'H c
c S w p o
•E e g, g c
(5 Sc *^ ?3
«t- S S o ^
O (« CH *»
fe^ o o E
•5 -o g .S «
lis^S
e E 8^2 -
1 1 8 1 f
= .b 1 1 I
1 ^ g.s ^
S "r w "> 8
•g g,o E S
ijlP
c > e -
£ 1 »l S 4
£ 2 o]2 e &
| | S |-2 |
*1 = tn .S3 S r=
t? i=n t> M GO ?J
« * « >* £> •-
£• -i-^- .2 ci . In
— «5 .O C= .Si 'S
1 > a 2 ts I
issJts
ps$ -I a jz






in control group.
1
g
O
•2 la
applicable.
t othenvise specify
;nce in dose group
— ° -0
o c .H
f " S
II c/3
-------
               Lymphocytic lymphomas of thymic origin occurred at a markedly increased incidence in
         mice exposed to 625 ppm for 13 or 26 weeks. According to the life table test, the incidence of
         lymphocytic lymphoma was also significantly increased in mice exposed to 200 ppm for 40
  4      weeks or 312 ppm for 52 weeks. The incidence of histiocytic sarcomas was significantly
  5      increased (life table test) in mice in all stop-exposure groups as well.
  6            The lower incidences of lymphocytic lymphomas at 200 ppm (40 weeks) and 312 ppm
  7      (52 weeks) compared to 625 ppm for 13 and 26 weeks, respectively, demonstrate that the
  8      concentration of 1,3-butadiene is a greater contributing factor in the development of this lesion
  9      than the duration of exposure, i.e., a high concentration for a short duration is more effective than
 10      a lower concentration of longer duration. A comparison of the 200 ppm (40 weeks) versus the
 11      625 ppm (13 weeks) and of the 312 ppm (52 weeks) versus the 625 ppm (26 weeks) lymphocytic
 12      lymphoma results using a life table test confirms that the higher concentration/shorter duration
 13      regimen is significantly more effective than the lower concentration/longer duration regimen
 14      within each cumulative exposure grouping (p=0.005 for 8,000 ppnvweeks and_p<0.001 for
 15      16,000 ppm-weeks) after survival differences are taken into account.
 16            As observed in the 2-year study, lymphocytic lymphomas occurred very early after
 17      exposure started: as early as 23 weeks in the group exposed to 625 ppm for 26 weeks and as early
_18      as 24 weeks in the group exposed to 625 ppm for 13 weeks. This lesion accounted for 24 and 17,
         respectively, of the first 25 deaths occurring in these groups by weeks 45 and 79, respectively.
 20      Therefore, early deaths due to lymphocytic lymphoma would have a tremendous negative effect
 21      on the  incidence of late-developing lesions.
 22            Hemangiosarcomas of the heart, which also accounted for some of the early deaths, were
 23      significantly increased in most stop-exposure groups compared with the controls. The highest
 24      incidence, which was about twice as high as that of other groups, occurred in the group exposed
 25      to 312  ppm, followed by the groups exposed to 200 ppm and 625 ppm (26 weeks).  The lowest
 26      incidence occurred in the group exposed to 625 ppm for 13 weeks. Hemangiosarcomas appeared
 27      at about 9 months in the 200, 312, and 625 ppm (26-week) stop-exposure groups. Comparison
 28      (life table test) of groups having the same total exposures showed that the incidences of
 29      hemangiosarcomas in mice exposed to 625 ppm were significantly lower than that of the
 30      corresponding group exposed to 312 ppm (/?=0.032) but not 200 ppm. The incidences of
 31      hemangiosarcomas in both 625-ppm stop-exposure groups were higher than that in the 625-ppm
 32      2-year  exposure group, probably due to longer survival of the stop-exposure groups.
 33            The incidences of neoplastic lesions of the lung (alveolar/bronchiolar adenoma,
 34      adenocarcinoma, or carcinoma) were significantly elevated in each exposure group. The highest
         incidence occurred in the 200-ppm stop-exposure group, followed by the 312-, 625- (13 weeks),

         1/28/98                                  6-25          DRAFT-DO NOT CITE OR QUOTE
-------
  1      and 625-ppm (26 weeks) groups. The adenomas developed after week 47 and the
  2     adenocarcinomas and carcinomas developed after week 53; the late appearance of these lesions
  3     relative to lymphocytic lymphomas probably accounted for the lowest incidence of lung
  4     neoplasms occurring in 625 ppm (26 weeks) group. A life table analysis suggested the incidence
  5     of lung lesions in the 625 ppm (26 weeks) group was significantly greater than in the 312 ppm
  6     (52 weeks) group (p=0.013), but no difference was detected between the 200 ppm (40 weeks)
  7     and 625 ppm (13 weeks) groups.
  8            Mice exposed to 200 ppm 1,3-butadiene for 40 weeks had significantly increased
  9     incidences of hepatocellular adenomas and adenomas/carcinomas combined; the incidences of
10     hepatocellular carcinomas analyzed alone were not significantly increased.  Exposure to 1,3-
11      butadiene at 312 ppm or 625 ppm (13 or 26 weeks) did not increase the incidence of
12     hepatocellular neoplasms of any type. The earliest detection of these neoplasms was 67 weeks
13     for the 625 ppm (13 weeks), 57  weeks for the 200 ppm, 47 weeks for the 312 ppm, and 45 weeks
14     for the 625 ppm (26 weeks) stop-exposure groups. A logistic regression analysis found no
15     differences between the 200 ppm and 625 ppm (13 weeks) or the 312 ppm and 625 ppm (26
16     weeks) groups.
17            A low incidence of squamous cell papillomas of the forestomach occurred in each of the
18     groups, and squamous cell carcinomas were seen in mice exposed to 312 ppm or 625 ppm for 13
1 9     and 26 weeks.  The incidences of squamous cell papillomas were not significantly greater than
20     controls for any group, but the incidences of squamous cell carcinomas were significantly greater
21      by the life table test, which is considered to be the appropriate test (NTP, 1993) for these fatal
22     neoplasms.  A life table analysis also revealed a statistically significant exposure-rate effect for
23     the squamous cell carcinomas in both of the total exposure groupings (p=0.019 for 8,000
24     ppm-weeks and p=0.015 for 16,000 ppm-weeks), suggesting that the higher concentration/shorter
25     duration exposures were more potent.
26            The incidence of adenomas of the Harderian gland was significantly greater in each
27     exposure group than in the controls by a logistic regression test. A low,incidence of Harderian
28     gland carcinomas occurred in mice exposed to 200 ppm for 40 weeks (not significant), 312 ppm
29     for 52 weeks (p=0.006), and 625 ppm for 13 weeks (not significant). No Harderian gland
30     carcinomas were observed in the controls or in mice exposed to 625 ppm for 26 weeks. A
31      logistic regression analysis did not detect any exposure-rate effects.
32            Other neoplasms occurred at low incidence in the stop-exposure studies; they were
33     considered to be related to exposure because of their low spontaneous incidences in NTP
34     historical control male mice.  These neoplasms occurred in the kidney, brain, Zymbal's gland,
35     and preputial gland.  The incidences of these neoplasms are also summarized in Table 6-8.

        1/28/98                                  6-26         DRAFT-DO NOT CITE OR QUOTE
-------
               Renal tubule neoplasms occurred in historical male control mice; the range was 0 to 1%.
        The small number of these neoplasms in each of the exposure groups are considered to be related
        to administration of 1,3-butadiene because the incidences were greater than the upper range for
  4     historical controls.
  5            Brain neoplasms, including two neuroblastomas and two malignant gliomas, observed in
  6     male mice exposed to 625 ppm for 13 weeks and one malignant glioma observed in male mice
  7     exposed to 625 ppm for 26 weeks may have been related to 1,3-butadiene exposure. Brain
  8     neoplasms are rare in untreated B6C3F! mice; none have been reported in 574 NTP historical
  9     control male mice. Furthermore, a low incidence of gliomas was also reported in the previous
 10     NTP (1984) study. For these reasons, the brain neoplasms are considered exposure-related
 11     lesions.
 12            A low incidence of preputial gland carcinomas occurred in the exposed groups in the
 1 3     stop-exposure studies, and none were seen in controls. Compared with the incidence in
 14     concurrent controls, the combined incidences of preputial gland tumors (adenoma and
 1 5     carcinoma) were significant in male mice exposed to 312 ppm (52 weeks) and to 625 ppm (13
 16     and 26 weeks) by the life table test. Preputial gland carcinomas were not reported in a survey of
 17     NTP historical control mice, further indicating that these neoplasms are probably related to
 18     exposure to 1,3-butadiene.
               One male exposed to 200 ppm for 40 weeks, two males exposed to 625 ppm for 13
 20     weeks, and two males exposed to 625 ppm for 26 weeks developed Zymbal's gland carcinomas.
 21     This lesion did not occur in male mice exposed to 312 ppm for 52 weeks; one control male,
 22     however, developed an adenoma. The combined incidence of Zymbal's gland adenomas and
 23     carcinomas in animals exposed to 625 ppm for 26 weeks was significantly increased compared
 24     with controls by the life table test. Zymbal's gland neoplasms are rare spontaneous neoplasms
 2 5     that had not been observed in any NTP historical controls before the only occurrence of this
 2 6     adenoma in the control male mice for these studies.
 27            To summarize the results of the stop-exposure study pertaining to the relationship
 28     between exposure  level and duration of exposure:  For lymphocytic lymphomas, there is strong
 29     evidence that higher concentration/shorter duration exposures are more potent than the lower
30     concentration/longer duration exposures for both the 8,000 ppm-weeks and 16,000 ppm-weeks
31     total exposure groupings. There is also some evidence for a similar exposure-rate effect for
32     forestomach squamous cell carcinomas in both total exposure groupings. Any exposure-rate
33     effects at other sites are less clear, especially because it is difficult to distinguish a small apparent
34     increased potency  effect of higher concentration/shorter duration exposures from an effect of
        longer potential postexposure follow-up times following the shorter-duration exposures.

        1/28/98                                   6-27         DRAFT-DO NOT CITE OR QUOTE
-------
 1      6.3.3. Summary of NTP (1993) Study
 2            The 2-year inhalation study showed that 1,3-butadiene is a potent carcinogen in mice at
 3      all concentrations evaluated. It also demonstrated that exposure to lower concentrations of 1,3-
 4      butadiene than those used in the previous NTP (1984) study allowed expression of neoplasms at
 5      other sites and provided clearer exposure-response relationships because of increased survival.
 6      Statistically significant increases in the incidences of malignant tumors at one or more sites
 7      occurred in male mice exposed to ^20 ppm and in females exposed to 2:6.25 ppm (the lowest
 8      exposure concentration used) 1,3-butadiene for periods up to 103 weeks.  The possibility,
 9      therefore, exists that lower exposure concentrations would also cause cancer in B6C3F, mice.
10      The percentage of animals bearing malignant tumors increased from about 30% in the controls to
11      nearly 90% in the highest exposure group, 625 ppm.  Lymphocytic lymphomas,
12      hemangiosarcomas of the heart,  lung neoplasms, and neoplastic lesions of the forestomach,
13      mammary gland, ovary, and liver, lesions identified in the NTP (1984) study, were again
14      increased in this study. In addition, the Harderian gland and preputial gland were identified as
15      sites of 1,3-butadiene-induced neoplasms.  Tumors observed in the kidneys, skin, Zymbal's
16      gland, and intestine may also have been related to 1,3-butadiene exposure.
17            The stop-exposure study demonstrated that limited exposure to 1,3-butadiene also
18      induces neoplasms at multiple organ sites in male B6C3F, mice.  Incidences of lymphocytic
19      lymphomas, hemangiosarcomas of the heart, alveolar-bronchiolar neoplasms, forestomach
20      squamous cell neoplasms, Harderian gland neoplasms, and preputial gland neoplasms were
21      increased compared with controls after exposure to 625 ppm 1,3-butadiene for only 13 weeks.
22      The stop-exposure study also demonstrated an apparent exposure-rate effect for the induction of
23      lymphocytic lymphomas by 1,3-butadiene. At equivalent total exposures, the induction of
24      lymphocytic lymphomas was greater with exposure to a higher concentration of 1,3-butadiene
25      for a shorter time than for exposure to a lower concentration for a longer duration.
26            Overall, the NTP (1993) was a very well conducted study with a precise and
27      comprehensive presentation of the data. Adequate numbers of animals of both sexes were
28      exposed to multiple concentration levels of 1,3-butadiene for a major portion of their life span.
29      Comprehensive histopathological evaluations were performed and mortality and tumor
30      incidences were analyzed statistically using multiple methods.

31      6.3.4. 1-Year Study (Irons et aL, 1989; Irons, 1990)
32            To elucidate the mechanism of murine leukemogenesis, Irons and coworkers compared
33      the induction of thymic lymphomas and expression of murine leukemia virus in NIH Swiss male
34      mice and B6C3F! male mice by exposing them to 1,250 ppm 1,3-butadiene, 6 h/day, 5

        1/28/98                                   6-28         DRAFT-DO NOT CITE OR QUOTE
-------
         days/week for 52 weeks. Activation of an endogenous esotropic retro virus has been associated
         with spontaneous lymphomas in the B6C3F, mouse.  The NIH mouse strain was used because it
         does not express the esotropic murine leukemia viruses expressed in B6C3Fj mice. The
  4      background rate for thymic lymphoma in NIH mice is nearly zero. Although there was a marked
  5      difference between the incidence of thymic lymphoma/leukemia in B6C3Fl mice (57%) and the
  6      incidence in similarly exposed NIH mice (14%), the study showed that 1,3 -butadiene can induce
  7      thymic lymphomas independently of an activated retrovirus. In addition, because these studies
  8      were for only 52 weeks, they did not necessarily allow for a full response for induction of
  9      lymphomas by 1,3-butadiene.

 10      6.4. RELATED COMPOUNDS
 11            The draft report on the toxicology and carcinogenicity of 4-vinyl-l-cyclohexene, a dimer
 12      of 1,3-butadiene, was reviewed in U.S. EPA (1985). The final report (NTP, 1986) contains the
 13      same information; therefore, the data are not summarized in this update. The basic conclusion
 14      was that there was clear evidence of carcinogenicity of 4-vinyl-1 -cyclohexene (by gavage) in
 1 5      female mice based on increased ovarian neoplasms and equivocal evidence in male mice based
 16      on marginal increases of malignant lymphomas and alveolar/bronchiolar adenomas. In rats, there
 17      was inadequate evidence in males, at least in part because  of excessive mortality, and equivocal
         evidence in females based on increased neoplasms of the clitoral gland.
 1 9           The 1,3-butadiene metabolites l,2-epoxy-3-butene and l,2:3,4-diepoxybutane have been
 20      shown to be carcinogenic in rats when administered by skin application or subcutaneous
 21     injection (van Duuren et al., 1963,1966). In addition,  1,2-epoxybutane, a related compound that
 22     is used as a stabilizer for chlorinated hydrocarbon solvents, was administered by inhalation 6
23     h/day, 5 days/week for 24 months at exposure concentrations of 0, 200, or 400 ppm to F344/N
24     rats and 0, 50, or 200 ppm to B6C3F, mice (Dunnick et al., 1988). The treatment and control
25     groups consisted of 50 male and 50 female animals of each species. Exposure-related
26     inflammatory, degenerative, and proliferative lesions occurred in the nasal cavity of both rats and
27     mice.  Neoplastic lesions were restricted to the respiratory  tract in rats. At 400 ppm, nasal
28     papillary adenomas were observed in seven male rats and in two female rats; none were  observed
29     in controls.  In male rats exposed to 400 ppm, there was also an increased incidence of
30     alveolar/bronchiolar adenomas or carcinomas (combined) (5/50) compared with controls (0/50).
31      No exposure-related neoplastic lesions were seen hi male or female mice.
        1/28/98                                   6-29         DRAFT-DO NOT CITE OR QUOTE
-------
 1      6.5. DISCUSSION AND CONCLUSIONS
 2            Previous long-term inhalation studies have shown that 1,3-butadiene is carcinogenic in
 3      rats and mice, inducing tumors at multiple organ sites (NTP, 1984; Owen et al., 1987).  Results
 4      of the 2-year inhalation study (NTP, 1993) presented in this report confirmed the carcinogenicity
 5      for 1,3-butadiene in male and female B6C3F, mice as demonstrated in an earlier study (NTP,
 6      1984).
 7            Of particular interest in this study were the large number of primary organ sites of tumor
 8      induction by 1,3-butadiene; the early and extensive development of lymphomas; the induction of
 9      uncommon tumors, such as hemangiosarcomas of the heart and squamous cell neoplasms of the
10      forestomach; and the development of malignant lung rumors at exposure concentrations as low as
11      6.25 ppm. Because there were no exposure levels of 1,3-butadiene at which a carcinogenic
12      response was not induced, it is likely that exposure to concentrations below 6.25 ppm would also
13      cause cancer in mice.
14            Exposure to 1,3-butadiene at concentrations ranging from 6.25 to 625 ppm for 2 years
15      caused increased incidences of neoplasms in the hematopoietic system, heart, lung, forestomach,
16      mammary gland, ovary, and liver, all lesions identified in the NTP (1984) study. The Harderian
17      gland and preputial glands were identified as additional sites, and tumors in the kidneys, skin,
18      ZymbaTs gland and intestine were marginally associated with 1,3-butadiene. Because  of
19      increased survival, the study also established clearer concentration-response relationships than
20      the 1984 study.  Competing risks of early-developing lethal lymphocytic lymphomas at high
21      concentrations preempted the appearance of late-developing neoplasms at some organ sites.
22            Separate experiments with reduced exposure durations (stop-exposure study) showed that
23      continued exposure is not necessary for development of neoplasms. The incidences of
24      lymphocytic lymphomas, hemangiosarcomas of the heart, and tumors of the lung, forestomach,
25      Harderian gland, and preputial gland were increased in mice exposed for only 13 weeks to 625
26      ppm 1,3-butadiene and it is likely that even shorter exposure durations would have produced a
27      carcinogenic response. The stop-exposure study also showed that the concentration is a greater
28      contributing factor in the development of lymphocytic lymphomas than the duration of exposure.
29      At comparable total exposures, the incidence of lymphocytic lymphomas was greater with
30      exposure to a high concentration of 1,3-butadiene for a short time compared with a lower
31      concentration for a longer duration.
32            A morphological continuum of 1,3-butadiene-induced proliferative lesions to neoplasia or
33      the progression of benign to malignant neoplasms was evident for a number of sites in both the
34     2-year and the stop-exposure study (NTP, 1993). Increased incidences of proliferative,
35      nonneoplastic lesions (hyperplasia) of the cardiac endothelium, alveolar epithelium, forestomach

        1/28/98                                   6-30          DRAFT-DO NOT CITE OR QUOTE
-------
  1      epithelium, germinal epithelium and granulosa cells of the ovaries, mammary gland, and
  2      Harderian gland probably represent treatment-related preneoplastic changes at these target sites.
 F3      The distinction between adenoma and carcinoma further reveal the biological progression of the
  4      benign lesions to malignant neoplasia.  For example, in the lungs of male mice, progression from
  5      alveolar-bronchiolar adenoma to carcinoma was evident in the 200-ppm exposure group and in
  6      all of the stop-exposure groups.                                                     .
  7            The mechanism of 1,3-butadiene-induced carcinogenicity is not known; however,
  8      metabolism likely involving two reactive metabolites, 1,2-epoxy-3-butene and 1,2:3,4-
  9      diepoxybutane, is thought to be an important factor (Chapters 3 and 4).
 10            Results of previous carcinogenicity studies reviewed in U.S. EPA (1985) have shown
 11      different effects of exposure to 1,3-butadiene in rats and mice, with mice being more sensitive to
 12      the induction of carcinogenic effects than rats.  The carcinogenic activity in Sprague-Dawley rats
 1 3      exposed to 1000 or 8000 ppm 1,3-butadiene was largely limited to endocrine tissues or hormonal
 14      responsive tissues, such as pancreas, Ley dig cells of the testis, uterus, Zymbal gland, mammary
 1 5      gland, and thyroid (Owen et al., 1987), whereas exposure of B6C3F, mice to much lower
 16      concentrations of 1,3-butadiene caused significantly increased incidences of mammary gland
 17      neoplasms and granuloma cell neoplasms of the ovary as well as malignant lymphomas,
J 8      hemangiosarcomas of the heart, alveolar-bronchiolar neoplasms, squamous cell neoplasms of the
         forestomach, and hepatocellular neoplasms. The reason for the species difference is not known,
 20      but may in part be due to differences in toxicokinetics.
 21            Toxicokinetic studies have shown species-related quantitative and qualitative differences
 22      in the metabolism and disposition of inhaled 1,3-butadiene that may, in part, account for the
 23      observed species variability in the toxicity (Chapter 3). For example, metabolism studies have
 24      shown that blood concentrations of 1,3-butadiene are higher in mice than in rats, and are lower in
 25      monkeys than in either rodent species.  In vitro studies using liver microsomes have shown that
 26      the metabolism of the reactive intermediate, l,2-epoxy-3-butene, to the non-DNA-reactive 1,2-
 27      dihydroxybut-3-ene is the prevalent pathway in human and rat preparations, whereas mouse liver
 28      microsomes convert l,2-epoxy-3-butene to DNA-reactive l,2:3,4-diepoxybutane in addition to
 29      the nonreactive l,2-dihydroxybut-3-ene (Csanaday and Bond, 1991).
 30            Investigations by Irons and coworkers (Irons et al., 1989; Irons, 1990) to explain the
 31      species differences of 1,3-butadiene-induced carcinogenicity have focused on the possibility that
 3 2      activation of an endogenous leukemia retrovirus may play a critical role in 1,3-butadiene-induced
 33      lymphoma in B6C3F, mice. The incidence of thymic lymphomas was greater in B6C3F, mice
 34      (57%) than in NIH Swiss mice (14%) exposed to 1,250 ppm 1,3-butadiene for 1 year. However,
         the NIH Swiss mouse does not express the endogenous leukemia retrovirus and has a very low

         1/28/98                                  6-31         DRAFT-DO NOT CITE OR QUOTE
-------
  1      background rate for thymic lymphomas. Thus, the finding that exposure to 1,3 -butadiene caused
  2      a 14% incidence of thymic lymphomas in NIH Swiss mice suggests that 1,3 -butadiene can
  3      induce thymic lymphomas independently of an activated retrovirus.
  4            Identification of activated oncogenes in chemically induced tumors also may provide
  5      information regarding the mechanism of tumor induction by butadiene.  For example, because K-
  6      ras is the most commonly detected oncogene in human cancers, tumors from the NTP (1993)
  7      study were evaluated for the presence of K-ras oncogenes (Goodrow et al., 1990). Activated K-
  8      ras oncogenes were detected in 6/9 lung tumors, 3/12 hepatocellular carcinomas,  and in 2/11
  9      lymphomas obtained from B6C3F, mice exposed to 1,3-butadiene at concentrations ranging from
10      62.5 to 625 ppm.  A specific codon 13 mutation was found in most of the activated K-ras
11      oncogenes, suggesting a chemical-specific effect. Activated K-ras genes have not been found in
12      spontaneously occurring liver tumors or lymphomas (Goodrow et al., 1990) and were observed
13      only in 1/10 of spontaneous lung tumors in B6C3Fj mice (Goodrow et al., 1990;  Reynolds et  al.,
14      1987). Furthermore, it was shown that tumor suppressor genes are inactivated during 1,3-
15      butadiene carcinogenesis.  Soderkvist et al. (1992) identified allelic losses in the p53 tumor
16      suppressor gene in lung and mammary carcinomas and lymphomas of B6C3Fj mice  exposed to
17      1,3-butadiene, that were analogous to those observed in a variety of human cancers.
18            Immune-function assays  conducted by Thurmond et al. (1986) in which B6C3FJ mice
19      were exposed by inhalation to 1,250 ppm 1,3-butadiene for 6 or 12 weeks showed that 1,3-
20      butadiene exerts no significant immunosuppressive effects, suggesting that 1,3-butadiene causes
21      neoplasia by mechanisms other than by compromise of immune function.
22            In addition to the carcinogenic effects noted in the NTP (1993) study, exposure to 1,3-
23      butadiene caused hematological  changes indicative of a partially regenerative anemia in mice
24      exposed to s62.5 ppm 1,3-butadiene.  Mice exposed to 625 ppm exhibited bone marrow atrophy
25      and splenic and hepatic extramedullary hematopoiesis.  Increases in mean cell volume and mean
26      cell hemoglobin at 625 ppm 1,3-butadiene suggested that although 1,3-butadiene caused
27      suppression of hematopoiesis in  the bone marrow, younger larger cells may have been released
28      into the blood from extramedullary sites. A macrocytic-megaloblastic anemia was reported in
29      B6C3F, mice exposed to 1,250 ppm 1,3-butadiene for 6 weeks (Irons et al., 1986a, b).
        1/28/98                                   6-32         DRAFT-DO NOT CITE OR QUOTE
-------
                      7. EPIDEMIOLOGIC STUDIES OF CARCINOGENICITY

               This updated review presents the evaluation of studies published from 1985 through
  2     January 1997.  The follow-up proposed by Lemen et al. (1990) of the cohort studied by
  3     Meinhardt et al. (1982) and Downs et al. (1992), an abstract submitted for the International
  4     Symposium are not reviewed in this evaluation. Lemen et al. (1990) did not present any results,
  5     while no details of study design and analysis were available for Downs et al. (1992). Since 1985,
  6     investigators have conducted studies of workers who produce 1,3-butadiene as a raw material
  7     (monomer production) or who use 1,3-butadiene in styrene-butadiene rubber (SBR) production
  8     (polymer production).

  9     7.1. MONOMER PRODUCTION
10     7.1.1. Texaco Cohort
11      7.1.1.1. Downs et al., 1987: Mortality Among Workers at a Butadiene Facility
12            Investigators  examined a cohort of 2,586 permanent male employees who worked a
13     minimum of 6 months in a Texaco butadiene manufacturing plant (monomer production) that
14     supplied the raw material to two adjacent SBR plants studied by Meinhardt et al. (1982) and for
15     which an update has been proposed by Lemen et al. (1990). Data were available for the 37-year
        period from January  1, 1943, through December 31,1979. Vital status of the cohort was
1 7     determined through the Social Security Administration (SSA). Individuals whose vital status
1 8     was unverifiable through SSA were traced through the Texas Department of Public Safety.
1 9     Death certificates were obtained from the health departments of the states where the individual
20     resided at the time of death. When this effort was unsuccessful, the individual's name was
21      placed on a list, which was submitted to the health departments of Texas and Louisiana, to obtain
22     the death certificates. A trained nosologist coded the death certificates using the eighth revision
23     of the International Classification of Diseases (ICD).
24            Because quantitative exposure data had not been accumulated for individual workers, the
25     investigators used department codes to construct a qualitative exposure scale composed of four
26     groups:  Group I, low exposure (included utility, office, and management workers, N = 432);
27     Group II, routine exposure (included process, laboratory, storage, and transport workers, N =
28     710); Group III, nonroutine exposure (included skilled maintenance workers, N = 993); and
29     Group IV, unknown exposures (N = 451). The investigators postulated that Group III workers
30     may have had exposure to higher concentrations with a lesser frequency than Group II workers.
31             Of 2,586 employees in the cohort, 175 (6.8%) were black. Scrutiny of death certificates
        uncovered that 45 blacks (7.5% of total deaths) were improperly coded as whites. At this point,

        1/28/98                                   7-1        DRAFT-DO NOT CITE OR QUOTE
-------
  1      investigators conducted a preliminary analysis on the total cohort, using both black and white
  2      national death rates. The standard mortality ratios (SMRs) were higher based on black rates as
  3      compared with white rates for four cause-specific deaths only (i.e., all lymphohematopoietic
  4      cancers) (SMR = 169 vs. 138), lymphosarcoma (SMR = 336 vs. 220), Hodgkin's disease (SMR
  5      = 135 vs. 102), and leukemia (SMR = 155 vs. 119).  Most of the other SMRs for both cancers
  6      and noncancers were decreased based on black rates. Therefore, using black rates would have
  7      underestimated the risks. Thus, the entire cohort was treated as white, and all further analyses
  8      were conducted using white death rates.
  9            Expected deaths were calculated using two referent populations: U.S. white males
 10      (national comparison) and white males in a seven-county area surrounding the plants (local
 11      comparison). The rates were standardized for age, race, sex, and calendar year.  SMRs (labeled
 12      NSMR for national comparisons and LSMR for local comparisons) were calculated in the
 13      customary manner by dividing the observed deaths by the expected deaths and multiplying the
 14      ratio by 100. Under the null hypothesis, the significance of the ratios of observed to expected
 15      deaths was tested assuming that the observed (O) deaths followed a Poisson distribution using a
 1 6      two-sided test and assuming ap value of <0.05 to be significant. Comparisons between Groups
 17      I, II, and III were done by using the Mantel-Haenzel procedure for computation of relative risks
 18      in follow-up studies  with stratified data (Rothman and Boice, 1982),  and power calculations were
 19      performed using the normal approximation to the Poisson distribution (Beaumont and Breslow,
20      1981). The person-years at risk were not accrued until after the sixth month of employment.
21            A total of 64,800 person-years were accrued for the follow-up period.  There were 603
22      deaths from 1943 through 1979; death certificates were obtained for 579 (96%) individuals. The
23     vital status was unknown for 73 individuals (2.8% of the total cohort).
24            Results of this investigation indicated lower than expected mortality for these workers
25     from all causes (NSMR =80, ^<0.05 and LSMR = 96, p>0.05, O = 603) and from all cancers
26      (NSMR = 84, ^0.05 and LSMR = 76, p<0.05, O = 122). However,  a site-specific comparison
27     indicated a statistically significant increase in mortality from lymphosarcoma and
28     reticulosarcoma (LCD code 200, NSMR = 235, 95% confidence intervals [CI] = 101-463, O = 8)
29     compared with national rates and a nonsignificant excess (LSMR = 182, p>0.05) compared with
30     local rates.
31             A comparison of wartime workers (N = 1,061; 452 deaths) who had worked for at least 6
32     months prior to 1945 and postwar workers (N = 1,525; 151 deaths) found an increase for all
33     lymphohematopoietic cancers among wartime workers (NSMR = 150, 95% CI = 84-247, O = 15)
34     and among postwar workers (NSMR = 134, p>0.05, O = 6). However, stratification reduced
        1/28/98                                   7-2        DRAFT-DO NOT CITE OR QUOTE
-------
         sample sizes considerably. The rationale for this comparison was based on the assumption that
         wartime exposures may have been higher than in postwar periods.
               The analyses by duration of employment on mortality showed an increase among those
  4      who worked <5 years for all lymphohematopoietic cancers (NSMR = 167, _p>0,05, O = 11), with
  5      most of the increase attributed to leukemia (NSMR = 187, _p>0.05, O = 5) and residual
  6      lymphohematopoietic cancers1 (i.e., non-Hodgkin's lymphoma, multiple myeloma, and other
  7      lymphohematopoietic cancers) (NSMR = 172,/>>0.05, O = 5). Among those who worked >5
  8      years, a nonsignificant increase was found for all lymphohematopoietic cancers (NSMR = 127,
  9      O = 10), mainly due to an increase in residual lymphohematopoietic cancers (NSMR = 200, O = 7).
 10            Further analyses were conducted for the four groups identified on the qualitative exposure
 11      scale. For those with routine exposure (Group II), increases were noted for all
 1 2      lymphohematopoietic cancers (NSMR = 187,^>0.05, O = 6), Hodgkin's disease (NSMR = 197,
 13     p>0.05, O = 1), and residual lymphohematopoietic cancers (NSMR = 282,£>>0.05, O = 4).  An
 14      excess of kidney cancer (NSMR = 254, p>0.05) was also observed in this group based on one
 15      case.  Similarly, in those with nonroutine exposure (Group III), excesses were observed for all
 1 6      lymphohematopoietic cancers (NSMR = 167, p>Q.Q5, O = 10), Hodgkin's disease (NSMR = 130,
 17     p>Q.Q5, O = 1), leukemia (NSMR = 20l,p>0.05, O = 5), and residual lymphohematopoietic
J8      cancers (NSMR =150, ^>0.05,O = 4).
               For those in the low-exposure group (Group I), excess mortality was seen for the same
 20      cancers (excluding Hodgkin's disease): all lymphohematopoietic cancers (NSMR = I2S,p>Q.05,
 21      O = 3), leukemia (NSMR = 105, £>>0.05,0 = 1), and residual lymphohematopoietic cancers
 22      (NSMR = 190, p>O.Q5, O = 2). In general,  use of local southeast Texas coastal rates resulted in
 23      lower SMRs for the above three groups except for Hodgkin's disease in routine and nonroutine
 24      exposure groups, which showed slight increases over national rates.  Both of these SMRs were
 25      based on one observed case in each group.  None of the  excess found in these three groups was
 26      statistically significant
 27            The comparison of Groups II, III, and IV with the low-exposure group (Group I) resulted
 28      in inconsistent findings due to a small number of cause-specific deaths and could not be reliably
 29      interpreted.
 30            Analyses were also done by latency and number of years worked using national rates.
 31     Although the results for number of years worked were inconsistent for total cancers, the SMRs
 32      increased from 80 to 93, with increasing latency for this category. Similarly, excess SMRs for
 33     all lymphohematopoietic deaths were observed in all latency periods (0 to 9, 20 to 29, 30 to 39)
          'Residual lymphohematopoietic cancers include ICD codes 200, 202,203, 208, and 209.

        1/28/98                                  7-3        DRAFT-DO NOT CITE OR QUOTE
-------
 1      except for 10 to 19 years.  The number of years of employment results showed an inverse
 2     relationship for these cause-specific deaths.  For cause-specific deaths due to lymphosarcoma and
 3     reticulosarcoma (ICD code 200), both the latency as well as number of years employed showed
 4     an inverse relationship.  The notable finding in this analysis was for workers who had a latency
 5     of 0 to 9 years and had worked for less than  10 years (NSMR =1,198, p<0.01, O = 4). This
 6     increase was statistically highly significant (tested by the author of this document using the
 7     Poisson distribution).
 8           This is an extensively analyzed cohort mortality study. As correctly acknowledged by the
 9     investigators, there are a few methodological limitations to this study, the major ones being a
10     lack of industrial hygiene (IH) data and a lack of personal work histories. In addition, half of the
11      total cohort worked less than 5 years in the plant. Some of the workers from this cohort had also
12     worked in two neighboring SBR plants. The exposures to other chemicals in the SBR plants and
13     in their prior jobs are the confounders that were not adjusted for in this study. The cohort is
14     relatively small to start with, but stratification in several subgroups further reduced the power.
15           The major strength of the study is that it is conducted in a butadiene (monomer)
16     production facility in a cohort where confounding exposure from styrene is absent. The excesses
17     observed are in cancers of the lymphohematopoietic system, which are consistent with cancer
18     findings of the SBR plant workers. Most of the cases of malignancy are concentrated in workers
19     employed for less than 10  years, which may be due to the occurrence of higher exposures during
20     wartime years. The exposures during subsequent periods were lower. Thus, the finding of
21      excess cancer mortality in short-term employees is not evidence against dose-response
22     relationship.

23     7.1.1.2.  Divine, 1990:  An Update on Mortality Among Workers at a 1,3-Butadiene
24             Facility—Preliminary Results
25           In  1990, Divine reported an updated analysis of the same Texaco plant (monomer
26     production) cohort. The follow-up on the original cohort was extended through 1985 by
27     updating the information on workers from company data and the SSA. Death certificates were
28     obtained from the health departments of Texas, Louisiana, Ohio, and Mississippi and were coded
29     by a trained nosologist according to the eighth revision of the ICD.  The National Death Index
30     records were searched for workers for whom the  SSA failed to provide the vital status.
31            Mortality analyses were performed using Monson's computer program (Monson, 1974).
32     Again, the white male death rates of the U.S. population were used due to uncertainties about
33     race information in the company files and because there were few blacks in the cohort.  Person-
34     years were accrued similarly to the Downs etal. (1987) study.                                     Jttk

        1/28/98                                    7-4        DRAFT-DO NOT CITE OR QUOTE
-------
               The qualitative exposure categories remained the same. IH sampling data at the time of
 .2      this study supported the exposure categories developed earlier. For this study, lymphosarcoma
 *3      (ICD code 200) was reported separately from the cancers of other lymphatic tissues (ICD codes
  4      202, 203, and 208).
  5            A total of 74,219 person-years had accrued through 1985.  The number of deaths had
  6      increased to 826, and death certificates were not available for 49 (6%) individuals.  Of 2,5822
  7      employees in the cohort, 1,708 individuals were still alive and 48 (1.9%) were lost to follow-up.
  8      Overall, the pattern of results was unchanged from the report by Downs et al. (1987) for this
  9      cohort. For the total cohort, the SMRs for all lymphohematopoietic cancers and Hodgkin's
 10      disease were increased but not significantly; however, for lymphosarcoma and reticulosarcoma,
 11      the excess was significantly larger (SMR = 229, 95% CI  = 104-435, O = 9) and accounted almost
 12      entirely for the increase in overall lymphohematopoietic cancers. Analyses by various
 1 3      subcohorts also yielded results similar to those observed  in the earlier study (Downs et al., 1987).
 14      The highest increase was observed in lymphosarcoma and reticulosarcoma among workers who
 1 5      had worked more than 5 years but less than 10 years (SMR = 245, 95% CI = 79-572, O = 5).
 1 6      Prewar and postwar subcohort analyses demonstrated a statistically significant increase among
 1 7      the prewar subcohort for the same cause-specific deaths (SMR = 269, 95% CI = 108-555, O = 7),
J 8      while an excess in the postwar subcohort was not statistically significant (SMR = 155, 95% CI =
         17-558,0 = 2).
 20            Among the subcohorts based on exposure levels, the only statistically significant excess
 21      was observed for lymphosarcoma and reticulosarcoma among workers who were ever employed
 22      in routine exposure category (SMR = 561, 95% CI = 181-1,310, O = 5).  Among workers who
 23      were ever employed in nonroutine exposure category, the excess was observed for all
 24      lymphohematopoietic cancers (SMR =141, 95% CI = 70-253, O = 11) due to an increase in
 25      leukemia (SMR = 185, 95% CI = 68-403,  O = 6). The lymphosarcoma in this group was slightly
 26      increased (SMR =126, 95% CI = 14-454, O = 2).
 27            For the total cohort, no pattern with latency or duration of years worked  was observed for
 28      either all deaths or total cancer deaths. For all lymphohematopoietic cancers, excesses were
 29      observed in the latency groups of 30+ years (SMR = 205, O = 8) and 0 to 9 years (SMR = 200,
 30      O = 4). Both of these groups had worked  less than 10 years. Deaths from lymphosarcoma were
 31      also increased in the same duration and latency groups. For 30+ year and 0 to 9 year groups, the
 32      SMRs were 3,333 (O  = 2) and 1,333 (O =  4), respectively. No statistical test results were
 33      presented for this analysis. Similar analyses by different  exposure groups failed to show any
          2It was not explained in the paper how the cohort was reduced to 2,582 from 2,586.
        1/28/98                                   7-5        DRAFT-DO NOT CITE OR QUOTE
-------
  1      pattern for all lymphohematopoietic deaths and lymphosarcoma deaths among low-exposure and
  2      unknown exposure groups. Among routinely exposed groups, the excesses were observed for the
  3      same two latency and duration groups as for the total cohort, whereas for nonroutine exposure
  4      the excesses were observed only for 20 to 29 and 30+ years' latency groups who had worked for
  5      less than 10 years. All of these excesses were based on <3 deaths in each group, making
  6      interpretation of these findings by exposure levels very difficult.
  7             This also is a well-conducted study; unfortunately, the same methodological limitations
  8      that were present in the Downs et al. (1987) study are applicable to this study.  However, the
  9      findings of this study are consistent with the earlier study, as well as with other SBR plant
10      studies,

11      7.1.1.3. Divine et al., 1993: Cancer Mortality Among Workers at a Butadiene Production
12              Facility
13             This update added another 5 years of follow-up to the earlier cohort of monomer workers
14      (Divine, 1990).  Cohort inclusion criteria remained the same but were extended from December
15      31,1979, to December 31,1990. This yielded additional workers resulting in a total cohort of
16      2,749 individuals. The four exposure groups were similar to those used in earlier studies with
17      slight changes as follows: (1) The background exposure group (included office utility,
18      warehouse, and transportation workers, N = 347). This group was called the low-exposure group
19      in the previous two studies (Downs et al., 1987; Divine, 1990). (2) The low-exposure group
20      (included workers from operating units, planners and engineers, welders, carpenters, and workers
21      from brick masons, N = 958). This group was a combination of some of the low-exposure and
22      all of the unknown exposure group from the previous two studies. (3) The nonroutine exposure
23      group (included skilled maintenance workers such as pipefitters, tinsmiths, instrument and
24      electrical workers, and insulators, N = 865). (4) The routine exposure group (included process,
25      lab, storage, and transport workers, N = 1056). Although the last two categories appeared to be
26      the same as in the earlier two studies, the change in the number of individuals in these categories
27      was not explained in the paper.  For this study, the investigators reviewed the results of the IH
28      data and information obtained from the plant personnel and found that the main difference
29      between the routine and nonroutine exposure groups was in the frequency and not the intensity of
30      exposure.
31             Monson's computer program (Monson, 1974) was used for the analysis of this study also.
32      All the analytical methods included use of white male death rates of the U.S. population (since
33      there were very few blacks in the study, they were assumed to be white for the analysis) and
        1/28/98                                   7-6       DRAFT-DO NOT CITE OR QUOTE
-------
         calculation of person-years.  The follow-up procedures and acquisition of death certificates were
         the same as in an earlier study by Divine (1990).
                A total of 83,591 person-years was accrued. At the end of the follow-up period, 1,660
  4      individuals were still alive, 38 were lost to follow-up, and 1,051 were deceased (death certificates
  5      were obtained for 1,036 individuals).
  6             The overall results observed in this study were similar to the earlier two studies. The
  7 '     only statistically significant elevated SMR observed was for lymphosarcoma and reticulosarcoma
  8      for workers employed for less than 5 years (SMR = 286, 95% CI = 104-622, O - 6). Again, this
  9      increase probably came entirely from the prewar employees (SMR = 254, 95% CI = 102-523, O
 10      =7). The analysis by exposure group showed an increase for the same cause in the routine
 11      exposure group (SMR - 452, 95% CI - 165-984, O = 6). The analysis by latency and duration
 1 2      of employment yielded the largest increase in 0 to 9 years latency for the individuals employed
 13      for less than 5 years (prewar individuals?). The SMR was 3,333 based on two observed cases.
 14      No statistical test results were presented for this analysis.

 1 5      7.1.1.4.  Divine andHartman, 1996: Mortality Update of Butadiene Production  Workers
 1 6            This recent follow-up of the same cohort added 46 more individuals to the cohort (2,795)
J7      by extending the inclusion criteria and the follow-up period through December 31,1994. The
         person-years accrued increased to 85,581. Of 2,795 individuals, 999 were still alive, 574 were
 1 9      lost to follow-up (28 known to be alive), and 1,222 were deceased (death certificates were
 20      obtained for 1,202 individuals).  The follow-up procedures and analytical.techniques (for SMR
 21      analysis) were the same as for earlier studies.  The exposure categories also remained the same
 22      for this follow-up.
 23            Based on IH data available since 1980, each employee's potential exposure to butadiene
 24      was estimated by separating the employee's work history by job categories into 1-year segments.
 25      Two variables were used to calculate the estimated exposure (job categories and calendar time
 26      periods). There were six exposure classes based on job categories: 0, 1, 2, 3,4, and 5 with 0,
 27      0.1, 0.2, 0.3, 0.4, and 0.5 weights (wt), respectively, and five calendar time periods: <1946 (wt =
 28      10), 1946-59 (wt = 8), 1960-76 (wt = 4), 1977-85 (wt = 2), and 1986-94 (wt = 1). The
 29      cumulative exposure was obtained for each individual by summing up the scores for all the years
 30      of employment.  These exposure estimates were used to conduct survival analyses for:  (1) total
 31       lymphohematopoietic cancer, (2) lymphosarcoma, (3) non-Hodgkin's lymphoma, (4) multiple
 32      myeloma, and (5) leukemias.
 33            Three different models were used for the survival analysis, i.e., a Cox proportional hazard
         model with a time-dependent estimate of cumulative exposure, a person-time logistic regression

         1/28/98                                   7-7       DRAFT-DO NOT CITE OR QUOTE
-------
 1      model with a time-dependent estimate of cumulative exposure, and a nested case-control model
 2      using conditional logistic regression. Each case had 10 matched controls by date of birth
 3      (+2 years). The selection of controls without replacement was from noncases at the time of the
 4      occurrence of each case.
 5            The results of the SMR analyses were very similar, to the earlier two follow-up studies of
 6      this cohort (Divine, 1990; Divine et al, 1993).  The survival analyses failed to show any
 7      significant increase hi the risk ratios, in any cause-specific cancer, by any of the three methods.
 8            Although the investigators have done a good job of estimating the exposure and have
 9      conducted various analyses, the increase observed in the prewar subcohort for lympho-
10      reticulosarcoma, when exposures were probably the highest, still persists. Upon completion of
11      this study, this cohort has 52 years of follow-up but has failed to show any increase in leukemias
12      which were observed hi SBR production workers.

13      7.1.2.  Shell Oil Refinery Cohort
14      7.1.2.1.  Cowles et al, 1994:  Mortality, Morbidity, and Hematological Results From a Cohort
15              of Long-Term Workers Involved in 1,3-Butadiene Monomer Production
16            Shell Oil's Deer Park Refinery produced a butadiene monomer from 1941 to 1948 and
17      1970 to the present.  The cohort consisted of male workers who had a minimum of 5 years
18      employment in the jobs with potential exposure to butadiene or at least 50% of their total
19      duration of employment (minimum of 3 months) in these jobs. This facility also had several
20      other refinery operations and chemical production units. Three different analyses were
21      performed on this cohort: (1) mortality, (2) morbidity, and (3) hematological.

22      1.  Mortality Analysis:
23            A total of 614 employees comprised the cohort. The follow-up period was from 1948 to
24      December 31, 1989. Vital status was assessed from company records, SSA, master beneficiary
25      files, and the National Death Index (NDI). Death certificates were obtained for all the deceased
26      workers and coded by a trained nosologist according to the revision of the ICD in effect at the
27      time of death. Mortality rates of Harris County, TX, were used to compute the age-, race-, and
28      calendar year-adjusted SMRs, using the Occupational Cohort Mortality Analysis Program
29      (OCMAP) from the University of Pittsburgh.
30            A total of 7,232 person-years were accrued. Of 614 employees, 589 were still alive, 1
31      was lost to follow-up, and 24 were dead.  No excess mortality, either for total deaths or total
32      cancers (including cause-specific cancers), was observed.
        1/28/98                                  7-8        DRAFT-DO NOT CITE OR QUOTE
-------
         2.  Morbidity Analysis:
                Original cohort members who were active at some time between January 1, 1982, and
         December 31,1989, qualified for the morbidity study. Morbidity data were obtained from the
  4      Shell Health Surveillance System. The follow-up period was from 1982 to 1991. Causes of
  5      morbidity were coded according to the 9th revision clinical modification of the ICD. Morbidity
  6      ratios (SMbRs) were calculated by using the internal comparison group of employees who were
  7      active during the same time period and had no exposure to butadiene.
  8             A total of 43 8 employees were included in this analysis.  No excess morbidity by any
  9      cause was observed.

 10      3. Hematological Data Analysis:
 11             Of 43 8 individuals included in the morbidity study, periodic hematological data were
 1 2      available for 429 individuals.  These hematological data reveal that seven hematological
 1 3      outcomes were measured (between 1985 and 1991). The most recent laboratory test results were
 14      used for the analysis. Comparisons were done with similar results from 2,600 nonexposed
 1 5      employees.  No differences were observed between butadiene-exposed vs. nonexposed groups.
 1 6            This study has quite a few methodological limitations.  The cohort is small, and deaths
 17      are few. The number of employees selected for this study from the time period 1941 -1948, when
         exposure was probably higher, is unclear.  Over 50% of the cohort was hired in 1970 or later,
 1 9      with an average follow-up of 12 years.: This means that the cohort was still young, showing
 20      "healthy worker" effect, and enough latent period had not elapsed to show increases in cancers,
 21       which usually have a long latent period. Thus, despite the absence of any positive results, this
 22      study fails to provide any negative evidence towards the causal association between butadiene
 23      and occurrence of cancer.

 24      7.1.3. Union Carbide Cohort
 2 5      7.1.3.1.   Ward et al, 1995:  Mortality Study of Workers in 1,3-Butadiene Production  Units
 26             Identified From a Chemical Workers Cohort
 2 7              Ward et al,, 1996c: Mortality Study of Workers Employed in 1,3-Butadiene
2 8             Production Units Identified From a Large Chemical Workers  Cohort
2 9           The study cohort was selected from 29,13 9 workers at three Union Carbide Corporation
30     facilities in the Kanawha Valley, West Virginia. A total of 527 male workers who had worked
31      between 1940 and 1979 were identified from the work history records as having ever worked in
32      the departments where there was a potential for butadiene exposure. Only the individuals who
        worked in these departments during the butadiene production period (during World War II) were

        1/28/98                                   7-9        DRAFT-DO NOT CITE OR QUOTE
-------
  1      selected for the study (i.e., 364 individuals). The vital status was determined through December
  2      31, 1990, using the National Death Index. Death certificates were obtained for decedents and
  3      coded according to the revision of the ICD codes in effect at the time of death. Both U.S. and
  4      Kanavvha County mortality rates were used for comparison. A modified life table analysis
  5      developed by the National Institute for Occupational Safety and Health (NIOSH) was used to
  Q      compute the SMRs.
  7            Of 364 workers, 176 were alive, 3 were lost to follow-up, and 185 were dead at the end of
  8      1990. The SMR for all causes was 91, while for all cancers it was 105. Neither of them were
  9      statistically significant. The only statistically significant increase was observed for
10      lymphosarcoma and reticulosarcoma, which was based on four cases (SMR = 577, 95% CI =
11      157-148). A county-based comparison also resulted in a similar result.  By duration of
12      employment and latency, a statistically significant excess of the SMR was observed among
13      workers who were employed for more than 2 years and with more than 30 years of latency (SMR
14      - 1980,95% CI = 408-5,780, O = 3).
15            The investigators stated that except for butadiene exposure, there were no common
16      exposures to other chemicals in the four individuals who had died of lymphosarcoma and
17      reticulosarcoma, although two of them had been assigned to an acetaldehyde unit for some time.
18            This study has a few methodological limitations.  The cohort is very small, no
19      adjustments for confounding exposures to other chemicals were done, and no exposure
20      information is available. The qualitative exposure is assumed based on the job coded for
21      butadiene exposure. It is still interesting to note that the exposure in these plants was to
22      butadiene monomer alone either in the production process or the recovery from the olefin
23      cracking process and not to styrene-butadiene polymer. The only other cohort exposed to
24      butadiene monomer (Downs et al., 1987; Divine, 1990; Divine et al., 1993; Divine and Hartman,
25      1996) also found excess in lymphosarcoma and reticulosarcoma in the prewar subcohort.
26            Studies in monomer production workers are summarized in Table  7-1.

27      7.2. POLYMER PRODUCTION
28      7.2.1. Cohort Identified by Johns Hopkins University (JHU) Investigators
29      7.2.1.1.  Matanoski and Schwartz, 1987: Mortality of Workers in Styrene-Butadiene Polymer
30              Production
31            This cohort mortality study of SBR polymer production workers from eight plants (seven
32      U.S. and one Canadian) was reviewed in a 1985  document (U.S. EPA, 1985).  At that time, this
33      study was submitted to the U.S. Environmental Protection Agency but was not published.
        1/28/98                                   7-10       DRAFT-DO NOT CITE OR QUOTE
-------

 o
 o


 I
.1
13
 C3
+*

I
 ce
 O
 C-
§
"3
 4>
A
 o
 x
_0

"o

|

T3
M
2
es
H






^3
Strengths an
limitations




42
1

B
wj
Butadiene expo
assessment
•
^*\
y^



Population studied

o
"3
**



 O ;*
3 _, -" JS  *•" co ? o
8 S> jo if 03 ^^>o
a § z; o •=  °>--j3 <3 ^ TV
§ £ | | S =3 * g s £ £
1 1 * | g- g &| i^g
«g « S |« ^.| 5 •g'.S.g
^g 'S • S.s ^ o 2 ^>> p-J^atri _3o£
0
t/3 O 22
^y »~^ !— i CH CC
«* 03 O Qi 4?
•§ •«? Si-ails
^^gcS - 1 ra eg 7
rrs O C^ CO '"""' CO CO *^3
S ^ S S ii S S g
iiC^OO n/ 0 0 +*
Hoooo f^ooge
Di7c3ca «!' «3 cs S B
»5l|coco ^cocoS^C
f*-t j O O <5 O O ^ >-i
g||i| sfis-i,
^^S-MO COS'-ScwS
ooJ^Co &o^So§
ll on
_g " p ^
» § ^ II "aT S
ll I ? 1 i ' '
o^ S -s I .S
- ^ > 3 0 8 ,
ll \ S 5 >•
a> g o, a, • a, a
!-S< 3 3 3^^, 3,—^
3 2 8 S^ 8"""1
t2o O O Oc> O^
0)
|, | j- §
"°» § -"* S t~- "S. '
1 -i &s 1 .-§1
0 S§<^ *" OOcS-S
! ifs- 1 III
It 111 II III
i ^i r& ii L
^i -^ScA J » §-3S^
S?JS o^>^f =r~ 80.30
"n, o > «*- os o os o o o ^
CNO f>o^^ ti-i— t_)ao ^I>
oo
OS
"-f
"co
CO
O
Q
CO
CO
CO
8
_O

•1— J
Same methodologic lim:
the earlier study



~o
e
CO
CO
S
For lymphosarco
reticulosarcoma:
I
CO
xposure groups
stlidy
i— c
<1> O
p i^H
W cO
OO <0


Update of the cohort from
Downs etal. (1987)


o*
OS
CD
_c
^
Q








^
tH "<3
co 8
— % -So
S S 3 ~*
s • s. s §'
O O O i~^
OS OS T-H M
C^ *O *O i
CN CS ""> «_
II II ^ 11 eg
erf o5 . o od TS
oo o co S oo o S2
oo -g oo •§ oo & g
00 0 00 co 00 S >-.









bO
CM P
oo S oo
""i, '— i
CM" ~a i—
s •» &
•a g 2
-,
eo n
No pattern with 1
duration of empli

















1/28/98
7-11
DRAFT-DO NOT CITE OR QUOTE
-------
C
O
'•O
cj

«
O
s-
C.

o

o

o
S

1
8
.2
'O
•9
I

CO
O
O
w
•s

S
o

^*
"es
u
bfi
u



•3
C (/)
C3 C
« .2
*£! •*-•
wj.5
l.§
^-» *"•












w
"a
%
(A



1,3-Butadiene exposure
assessment


**3
.SJ
•a
Population stu



CO
M
O

?
•->
e ts
I*
t«
Oi _CJ
C "t^

00 S


•o
c
C3
C3

§ S
« i
B 8
•§,«
So
>>13
"Z! .H
£^

Similar exposure groups as
earlier with some


|
.2
Update of the cohort f
Divine (1990)



en
c\


^
"eo
CJ
u
_£
Q















CO


O ^
^ v>
eg §

si
^ **"*
^ O
oo "H.
oo S
oo S

II
as ° Z
||l
1 2£ 1 o
O O oo O — "

"on
0

Follow-up extended tl
1990
Comparison group U.!
population


















0,
_o

cS
2


?>
CN



•E2
y j_
°cO ^3

1 S
S CO
00 CO
tn ~
11
3 (U
S £
04 £

Based on IH data and work
histories using


o
."
Update of the cohort!
Divine etal. (1993)
OS
OS
c
CO

•c
CO
ffi
•a
CO
CJ
_c
5


1
CO
3
S
o
'i
CO

1
C/3
o
a-
^
w








S
•S
3
CO

• 6 exposure classes
• 5 calendar periods
JS
-t-»
CO
CO
OS
r-
Cohort increased to 2,













1 *

||
o .S

•a |
c 43
•2 '«
.2 co

1 -2
kn CO

'c? -~
*^3 CO
f£, cu

s 1"

"S .c
'3 M

CO O
•« g
"re oo
I ^
3 °
00 "OT

Individual exposures were

s;
"S 2
"O
inclusion period exten
through December 3 1














00
00
•S
CO
CO

^

i


j^
3
CO




c3
c
CO
0
1
CA
i
o

estimated for each worker





















CO
&>
o
CO
S
o

CO
en
0
^5

g

-J-I*










"O
Three different models use

43
00
1
J-H
Follow-up extended tl
1994






































for the survival analysis















































CO
Comparison group U.!
population
Internal comparison













1/28/98
7-12
DRAFT-DO NOT CITE OR QUOTE
-------
-a
 Ci
 o
 u
T5

 2
 o
 &
 o


 !
 I
 OS
 O
 G.
 C5
o

to


-3
s


o

'Sb


"o


1
CD
^






•o
a x
03 S
en 0
fi '-S
M-2
S S
t, £5
55 ""





Cfi
1
OS





s
V3
O
C, +-
f
£ 3
 -fi
is S
w C
^3 G
O £
« °
O ^*
0 S
 ?^
>> -^

Deaths are v<
50% of coho
S -o «
^ W ^

2 &,£
q~^ di rj
^ ^ 0>
til
0 S fi-
"S3 ••§ S
|^1
*^ C
5 o c
,^( C4_( CC













5 years or 50% of total duration
worked in jobs with potential to
1,3-butadiene exposure


















C/3
CO
> *5
•-^ ^3
O Q
£ hosarcc
arcoma
ffj
>-, 3
T! -2

PH £
•o
S
8
i 0
^ o
T~~i 2
>> g
1 g-
o 
JjH ^
Jj- J5
^ •*

>S M
Cohort of 364 male employees
who had worked between 1940

^
VO
OS
os
*— '
^ =
OS
OS
^
"3
^3
•a
w
CO
^

&0
.3 ,
*O w"

|1
c i
o .*~i
o
^0 Jj
=2 "S.
r^ O
No adjustmel
exposures to

 >,
OS O
5 t
§ w















JO
X)
.s
'c3
^k-
c
_o
ra
^
g
'S

No exposure
•s
S

o* o"
-^ c
S ts
•g o
S2 "
£ §
^o
c
133 "c
•£ ra













! monomer production during
'World War II




















































Follow-up through December
31, 1990
U.S. population
Kanawha County population




















































y significant.
y nonsignificant.
"ca -jg
o o
'•£2 '£3

'•£2 *-£2
CO ^
OO &O
II It
11 It

00
oo "Z
00 00






































tandard mortality ratios.
dard mortality ratios.
hygiene.
M g — <
"ca -K -S
S3 *3
o *S OT
'•£3 o p^
CO o t^J
Z >-4 hS
11 11 11

Dd ei
^ ^
OO OO Jj-i
^^ ^ >JH

1/28/98
7-13
DRAFT-DO NOT CITE OR QUOTE
-------
  1      Because the findings of the published study are essentially the same, it will not be reviewed
  2     again.

  3     7.2.1.2. Matanoski et al, 1989:  Epidemiologic Data Related to Health Effects of
  4             1,3-Butadiene
  5             Matanoski et aL, 1990:  Mortality of a Cohort of Workers in the Styrene-Butadiene
  6             Polymer Manufacturing Industry (1943-1982)
  7            These two publications essentially reported the same updated reanalysis of the earlier
  8     cohort.  In addition, Matanoski et al. (1989) also presented the results of the nested case-control
  9     study in this population.  Three methodological differences in the original analysis (Matanoski et
10     al., 1987) and the reanalysis presented in these two publications should be noted: extension of
11      follow-up through 1982,  fewer workers whose vital status was unknown (3.4% vs.
1 2     6.6% in the earlier report), and deletion of workers from the Canadian plant who had relatively
13     short-term exposure (i.e., workers who had worked for less than 10 years or who had not reached
14     the age of 45 during employment). Analytical methods were essentially unchanged from the
15     earlier analysis.
16            In addition to information received from the SSA and the Motor Vehicle Administration,
17     follow-up through local plant beneficiary records and the National Death Index was done to
18     assess the vital status of the study cohort. Follow-up procedures for Canadian workers were
19     similar to the earlier study. Death certificates were obtained from the local health departments.
20     The total cohort was reduced from 13,920 to 13,422 in this study. Of 12,113 workers for whom
21      the vital status was successfully traced, 23% (2,784) were still working in the plants, 53.4%
22     (6,472) were alive but not working in the plants, 20.2% (2,441) had died, and vital  status was
23     unknown for 3.4% (416). The racial distribution was 75% whites, 10% blacks, 15% unknown
24     (presumed to be white for the analysis), and less than 1% other.3  Death certificates were
25     obtained for 97.2% of the deceased individuals and were coded by a trained senior nosologist,
26     using the eighth revision of the ICD.
27            Data analyses were done by using age, race, calendar time, and cause-specific U.S.
28     population rates. A modified life-table program by Monson (1974) was used. The person-years
29     were calculated through December 31,1982.  The first-year work experience was omitted from
30     person-years because one of the inclusion criteria was that an individual had to have worked for
          ^he percentages, which are quoted from the paper, add up to 101.  This is due to the rounding
        of the numbers by the authors of the paper.
        1/28/98                                  7-14       DRAFT-DO NOT CITE OR QUOTE
-------
         at least 1 year. A total of 251,431 person-years were accrued, of which 226,475 were contributed
         by whites.
               Statistically significant lower SMRs for all causes of deaths (81) and for all cancers (85)
  4      were virtually the same as in earlier studies. The SMRs for all causes of deaths by 5-year
  5      calendar period demonstrated increasing SMRs with increasing time period, indicating a "healthy
  6      worker" effect in earlier calendar years. Blacks showed higher  SMRs than whites in later years.
  7      A statistically significant excess for all causes of deaths was observed for blacks in the last 3
  8      years of follow-up (SMR = 134, 95% CI = 101-175, O = 54). Most of the cause-specific cancer
  9      SMRs showed deficits in both races.  A few cancer sites demonstrated excess mortality in both
 10      races.  Among whites, excesses were observed for esophageal cancer, kidney cancer, Hodgkin's
 11      disease, and other lymphohematopoietic cancers. Among blacks, excesses were observed for
 12      stomach, liver, and prostate cancer; all lymphohematopoietic cancers; lymphosarcoma; leukemia;
 13      and other lymphohematopoietic cancers. None of the excesses were statistically significant.
 14            Because the risks for kidney, digestive, and lymphohematopoietic system cancers
 1 5      approached those of the reference population, which was unusual for an occupational cohort with
 1 6      low overall risks, investigators further analyzed the data by work areas.  For production workers,
 17      deaths from lymphohematopoietic cancers, Hodgkin's disease, and leukemia were
         nonsignificantly increased for the total cohort and among whites (except for leukemia). The only
         significant excess observed for the total cohort was for other lymphohematopoietic cancers,
20      which included non-Hodgkin's lymphoma and multiple myeloma (SMR = 260, 95% CI = 119-
21      494, O = 9). Among blacks, however, statistically significant excesses were observed for all
22      lymphohematopoietic cancers (SMR = 507, 95% CI = 187-1,107, O = 6) and leukemia (SMR =
23      655, 95% CI = 135-1,906, O = 3). The other two excesses observed among blacks for
24      lymphosarcoma and other lymphohematopoietic cancers (including non-Hodgkin's lymphoma
25      and multiple myeloma) were based on one and two cases, respectively, none being statistically
26      significant (p>0.05).
27            Among white maintenance workers, no excesses of lymphohematopoietic cancers were
28     found with the exception of Hodgkin's disease (SMR = 170, 95% CI = 35-495), based on only
29     three deaths. However, rates were nonsignificantly increased for digestive tract malignancies
30     (i.e., esophagus, stomach, and large intestine).  Among black maintenance workers,
31      nonsignificant excesses were observed for cancer of the rectum and stomach. For utility
32     workers, the numbers were reported to be too small to reach firm conclusions about risks.  For
33     the "other" category of workers (including laboratory workers, management, and administrative
34     workers), excesses were observed for Hodgkin's (SMR = 130, 95% CI = 16-472, O = 2) and
        leukemia (SMR =116, 95% CI = 43-252, O = 6) among whites and for leukemia (SMR = 246,

         1/28/98                                  7-15        DRAFT-DO NOT CITE OR QUOTE
-------
  1      95% CI not given, 0 = 1) among blacks. Nonsignificant increased SMRs for the digestive
  2     system among blacks were also observed for the stomach, liver, and pancreas, all of which were
  3     based on fewer than five cases.
  4            Analysis by duration of work or latency for the total cohort did not show an increase in
  5     the hematopoietic cancers.
  6            This is still the largest cohort of SBR workers. The increased follow-up, better tracing,
  7     and exclusion of short-term workers from the Canadian plant have resulted in demonstrating the
  8     excess mortality from malignancies of the lymphohematopoietic system, digestive system, and
  9     kidney. However, the limitations of the earlier study of this cohort (i.e., the lack of exposure data
10     and inclusion of less than 50% of the population in the follow-up cohort) still exist. The
11      magnitude of the bias introduced by exclusion of workers (2,391) due to missing information on
12     total work history or crucial information such as date of birth could be substantial. Although an
13     attempt was made to correct the race, the race was unknown for 15% of the eligible cohort, and
14     this segment was assumed to be white for the analysis. This would result in an overestimation of
15     rates in blacks and an underestimation of rates in whites. No explanation was given as to how
16     the total eligible population of 13,422 was reduced to 12,113. No data were presented by
17     individual plants, but as indicated in the earlier study, only four plants had follow-up starting
18     from 1943, whereas in the other four plants the starting dates of the follow-up ranged from 1957
19     to 1970; thus, these latter four plants may not have had long enough follow-up for the
20     malignancies to develop.

21      7.2.1.3. Matanoski et al, 1989: Epidemiologic Data Related to Health Effects of
22             1,3-Butadiene
23             Santos-Burgoa et aL, 1992: Lymphohematopoietic Cancer in Styrene-Butadiene
24             Polymerization Workers
25            To elucidate the separate contributions of 1,3-butadiene and styrene to the cancers
26     identified hi the updated cohort study, a nested case-control study of this cohort of SBR workers
27     was conducted using estimates of exposure to 1,3-butadiene and to styrene for each job.
28     Fifty-nine cases and 193 controls (matched for duration of work) were included in the analysis.
29     Among the case group were 26 cases of leukemia; 18 of other lymphatic cancers, which included
30     10 multiple myelomas and 7 non-Hodgkin's lymphomas; 8 Hodgkin's lymphomas; and 6
31      lymphosarcomas.
32            Cases (workers who had lymphohematopoietic cancer as either the underlying or
33     contributory cause of death on death certificates) arose from the original eight plants with the
34     same selection criteria for the eligibility of that cohort (13,422), with the exception of the

        1/28/98                                  7-16       DRAFT-DO NOT CITE OR QUOTE
-------
        Canadian plant. For the Canadian plant, the restriction of either 10 years of work or those who
        had reached age 45 during employment was dropped from the selection of cases, which added
 "3     two more cases to lymphohematopoietic cancers. Another four cases were added in which
  4     individuals had died of another cause of death but had a lymphohematopoietic cancer at the time
  5     of their death. Two cases were deleted from the final analysis, one lymphosarcoma due to lack
  6     of any controls and one non-Hodgkin's lymphoma due to lack of job records from which
  7     exposure could be identified.
  8            Controls included workers from the same cohort who were alive or had died of any cause
  9     other than malignant neoplasms. Controls were individually matched to cases by plant; age; hire
 10     year; employment as long as or longer than the case; and, if the control was dead, then survival to
 11      the death of the case. Based on these criteria, an average of 3.3 controls per case were selected
 12     instead of 4 controls per case as intended by the investigators.  This average of 3.3 controls per
 13     one case had more than a 90% chance of detecting the twofold risk from exposure to  1,3-
 14     butadiene. Both cases and controls had about 15 years of employment and were hired at 36 to 37
 1 5     years of age, somewhat older than usually seen in occupational populations.
 16            Exposures to 1,3-butadiene and styrene were calculated from the job records of each
 17     subject, the number of months that each job was held, and an estimate of the 1,3-butadiene and
 18     styrene exposure levels associated with that job. Both the job identification and exposure
        estimation were done independently and without knowledge of case or control status of the
20     subjects.  To estimate 1,3-butadiene and styrene exposures, all jobs within the rubber industry
21      were ranked from 0 to  10 by a group of senior engineers with many years of experience in the
22     industry.  One-third of the jobs were determined to have no routine exposure, but almost all jobs
23     were thought to have intermittent exposure. Cumulative dose for both styrene and 1,3-butadiene
24     was calculated using the score and duration for each job in the  participants' work history.
25     Because the distribution of exposure scores was skewed to the  right, a log transformation of the
26     scores was used in the analyses. As  the logarithmic transformation approached normal
27     distribution, only the transformed exposure variables were used for the analyses.
28            Analyses were done by using "ever/never exposed" categories to both butadiene and
29     styrene and using high-exposure vs.  low-exposure groups (based on mean log exposure
30     cumulative rank for each substance determined by combining cases and controls). Both
31      conditional (matched) and unconditional (unmatched) logistic regression analyses were
32     performed.  Odds ratios (OR) for matched sets were then calculated based on maximum
33      likelihood estimates of the OR, and test-based confidence limits around the OR were calculated.
34            Unadjusted for the presence of the other chemicals and unmatched, analyses by
        "ever/never exposed" to butadiene and styrene found significantly increased relative odds  for

        1/28/98                                 7-17       DRAFT-DO NOT CITE OR QUOTE
-------
 1      leukemia for both high and low exposures. Relative odds for butadiene were 6.82 (95% CI =
 2      1.10-42.23) and for styrene were 4.26 (95% CI = 1.02-17.78).
 3             Nonsignificant excesses were also observed for all lymphohematopoietic, other
 4      lymphohematopoietic cancers for exposures to both butadiene and styrene. Other excesses were
 5      for Hodgkin's disease among workers exposed to butadiene and lymphosarcoma among workers
 6      exposed to styrene.
 7             Matched analyses demonstrated that risk for all lymphohematopoietic neoplasms was
 8      significantly increased among workers exposed to butadiene (OR = 2.30, 95% CI = 1.13-4.71).
 9      Separate evaluation of these neoplasms revealed that most of the association could be explained
10      by a significant excess risk for leukemia (OR = 9.36, 95% CI = 2.05-22.94), but other cancers in
11      this group were not significantly elevated.  Leukemia also showed a threefold increase associated
12      with styrene exposure (OR = 3.13, 95% CI =  84-112).
13             Conditional logistic regression was used to separate the  risks associated with each of
14      these substances. Again, there was a significant excess of leukemia associated with butadiene
15      (OR = 7.61, 95% CI = 1.62-35.64) and a nonsignificant excess of leukemia associated with
16      styrene exposures (OR = 2.92, 95% CI = 0.83-10.27).  When exposures to both chemicals were
17      evaluated in the model as dichotomous variables, only butadiene was found to be associated  with
18      leukemia (OR = 7.39, 95% CI = 1.32-41.33).
19             To determine if specific j obs within the SBR industry might explain some of the risk of
20      leukemia, the investigators categorized each worker according to the longest job held. A
21      mixed-job category that combined utilities, operation services, and laboratory jobs was
22      associated with a relative odds of 3.78 (95% CI = 1.2-11.9). When butadiene was added to the
23      model, the OR increased to 6.08 for the mixed-job category (95% CI = 1.56-23.72). The relative
24      odds were 13.3 (95% CI = 2.24-78.55) for association between  butadiene exposure and risk of
25      leukemia adjusted for mixed jobs in this model. Thus, both the mixed-job  category and exposure
26      to butadiene seem to contribute to the risk of leukemia.
27             The trend test for increasing risk of leukemia with increasing exposure levels of
28      butadiene (0 through 8) was statistically significant (trend = 3.76,p = 0.05).  A similar trend was
29      not found for styrene. The higher risk of leukemia seen in the original cohort for black workers
30      could not be evaluated adequately because race was partially controlled in this nested
31      case-control study.
32             Unlike the mortality study of this cohort, the case-control study did not show other
33      lymphoma to be associated with production jobs, but the number of cases was small.
34      Interestingly, when each chemical was analyzed by stratification, there was an excess risk for
35      butadiene exposure when exposure to styrene was low (OR = 6.67, 95% CI = 1.06-42.7). A

        1/28/98                                   7-18        DRAFT-DO NOT CITE OR QUOTE
-------
        similar nonsignificant increase also was observed for styrene when butadiene exposure was low.
        This might have resulted from small numbers of non-Hodgkin's lymphoma or multiple myeloma
        included together with potentially different etiologies or correlated exposure data. Thus,
  4     investigators suggest further evaluation of each cancer in this other lymphoma category should
  5     be performed separately.
  6            Investigators also caution that estimated exposures in this study were crude and were not
  7     substantiated by monitoring data. As correctly pointed out by them, the original ordinal rank
  8     does not create a perfect exposure scheme. The distribution of ranks was skewed to the right and
  9     had to be log-transformed to differentiate between no exposure and low exposure. Matching on
10     duration of work may have overmatched the dose and resulted in underestimation of the risk.
11      Validation of diagnosis of lymphohematopoietic malignancies was not done in this study, which
12     is an important methodologic limitation of the study given the fact that lymphohematopoietic
13     cancer recording on death certificates is unreliable (Percy et al., 1981). The panel ranked 71% of
14     the jobs in ranks of two or less; thus misclassification of exposure based on the estimated
15     exposure by job as judged by the panel members is quite possible. Because the panel members
1 6     were blind concerning the status of the individual being the case or control, the distribution of
17     misclassification should be the same in cases and controls.

        7.2.1.4. Matanoski et al, 1993: Cancer Epidemiology Among Styrene-Butadiene Rubber
19             Workers
20            This was an effort by the investigators to verify the findings of their earlier nested case-
21      control study among styrene-butadiene production workers (Santos-Burgoa et al., 1992).  This
22     study had shown statistically significant elevated relative odds for leukemias. The results from
23     the analysis conducted with a new set of three controls per case were similar to the results from
24     the earlier study. The new controls were matched to all the variables except duration of work
25     with the case.  Comparability between the previous and new controls was checked by reviewing
26     the information on cases and controls from the earlier study. To verify that the cause of death
27     was correctly coded on the death certificates, hospital records for cases were obtained. Of the 55
28     records reviewed, two cases had been incorrectly coded on the death certificates as
29     lymphohematopoietic cancers. Records were obtained for 25 out of 26 leukemia cases and were
30     found to be correctly coded on the death certificates.
31             Exposure estimation was done based on measurements provided by seven rubber plants,
32     the International Institute of Synthetic Rubber Producers, and NIOSH. Although there was
33     variability among plants, a significant correlation was observed between the log transformed data
        provided by the company and the ranks of 464 job and area specific titles.  Of the seven plants

        1/28/98                                  7-19.        DRAFT-DO NOT CITE OR QUOTE
-------
  1      that provided exposure measurements for butadiene, three had geometric means. Thus, using the
  2     geometric means, the cohort data were reanalyzed for these three plants.  The workers who were
  3     hired before 1960 and had 10 or more years of service showed excesses for all
  4     lymphohematopoietic cancers (SMR =163, 95% CI = 113-227, O = 34) and leukemia and
  5     aleukemia (SMR = 181, 95% CI = 101 -299, O = 15).
  6            This reanalysis of earlier data with new information on exposure estimation validates the
  7     earlier results found by these investigators.

  8     7.2.2. Cohort Identified by University of Alabama (UAB) Investigators
  9     7.2.2.1. Delzell et al, 1996: A Follow-Up Study of Synthetic Rubber Workers
10            A retrospective cohort mortality study was conducted by Delzell et al. (1996) of synthetic
11      rubber workers employed hi seven U.S. and one Canadian plant. Of the eight plants, seven
12     plants (including the Canadian plant) were studied by JHU (Matanoski and Schwartz, 1987;
13     Matanoski et al.,  1989,1990,1993; Santos-Burgoa et al., 1992) and one (two initial plants
14     combined into one) by Meinhardt et al. (1982).  Of seven plants studied by JHU, one located in
15     Texas that had a starting time of 1970 was not included in UAB study. The cohort comprised all
16     the male workers who had worked for at least 1 year between January 1, 1943,  and January 1,
17     1992 (49 years), which was the end of the follow-up period. The follow-up period was shorter
18     for plants 1, 2, and 6 because the complete records of the employees from these plants were
19     available much later than 1943. The Canadian plant (plant 8) also had a shorter follow-up period
20     because follow-up of men who had left employment before 1950 was not feasible.
21             Since the inclusion criteria for this study were different, there were some additions and
22     deletions to the earlier study cohort. The vital status was assessed by using plant records; the
23     SSA's death master file; the NDI; DMV records of Texas, Louisiana, and Kentucky for the U.S.
24     plants; and plant records and record linkage with the Canadian Mortality Data Base for the
25     Canadian plant.
26            Death certificates were acquired from plant and corporate offices and from state vital
27     records. The underlying cause of death was coded by a trained nosologist using the ninth
28     revision of the ICD.  Any cancer was coded as a contributory cause of death. For the Canadian
29     decedents, the underlying cause of death was used from Canadian death registration and coded
30     according to the ICD revision in effect at the time of death. All ICD codes were converted to
31      eighth revision codes for analysis. The Ontario Cancer Registry provided the information on
32     incident cancer cases (including the date of diagnosis, primary site, ninth revision ICD code, and
33     histologic classification) for the study period.
        1/28/98                                  7-20       DRAFT-DO NOT CITE OR QUOTE
-------
               Mortality analysis included computation of SMRs using the U.S. male general and state
        population rates and Ontario male rates; SMRs by quantitative exposure (cumulative ppm-years
        and peak ppm-years) to 1,3-butadiene, styrene, and benzene; and stratified internal comparisons.
 4     Various within-cohort analyses were conducted using Poisson regression models.
 5            This study included exposure estimation for each individual. A detailed description of
 6     this estimation appears in Section 7.2.2.2, Macaluso et al., 1996. Complete work histories were
 7     available for 97% of the cohort. Analysis for process group was conducted on the workers from
 8     all the plants. Subgroup analyses were restricted to 6 plants (1,354 workers from 2 plants were
 9     excluded from the analyses due to the lack of information on specific work areas).
10            Of 15,649 males who had worked in SBR and related processes, 13,586 were white and
11      2,063 were black. Vital status assessment indicated that 10,939 (70%) workers were alive, 3,976
12     (25%) were dead, and 734 (5%) were lost to follow-up. Death certificates were acquired for
13     3,853 (97%) individuals. A total of 386,172 person-years (336,532 for whites and 49,640 for
14     blacks) was accrued.
15            Total cohort analysis found SMRs of 87 and 93 for all causes and all cancers,
16     respectively. The SMR for leukemia was 131 based on 48 observed deaths (95% CI = 97-174).
17     The SMRs for lymphosarcoma and other lymphopoietic cancers were close to null.
18            Subcohorts of whites, blacks, ever hourly, and never hourly showed a similar pattern of
        below null results for both all causes and all cancer deaths. Ever hourly was the only subcohort
20     in which statistically significant excesses were found for leukemia. The SMR was  143 ( 95% CI
21      = 104-191, O = 3 6) for this subcohort.  For white ever hourly workers, the SMR was 13 0 (95%
22     CI = 91-181, O = 36), while for blacks the SMR was 227 (95% CI = 104-431, O = 9).  The
23     lymphosarcoma SMR for this subcohort was 102 based on 4 cases, while the SMR for other
24     lymphopoietic cancer was 106 based on 17 cases. Neither of these excesses was statistically
25     significant. The further analyses  of this ever hourly subcohort by year of death (<1975,1975-84,
26     1985+), year of hire (<1950, 1950-59,  I960), and age at death (<55 years, 55-64 years, 65+
27     years) showed statistically significant SMRs for 1985+ year of death (SMR = 187, 95% CI =
28     111-296,0 = 18), 1950-59 year of hire (SMR = 200, 95% CI- 122-310, O = 20), and <55 years
29     at death (SMR = 179, 95% CI = 104-287, O = 17).
30            When this subcohort was further restricted to >10 years of employment and >20 years
31      since hire, the SMRs of 224 (95% CI = 149-323,  O = 28) for all workers, 192 (95% CI = 119-
32     294, O = 21) for whites, and 436 (95% CI = 176-901, O = 7) for blacks were observed.
33     Furthermore, in this restricted subcohort, the SMRs for leukemia were 209 (95% CI =100-385)
34     and 228 (95% CI = 135-160) for the workers from plants with the solution polymerization
        process and workers from plants without such a process, respectively.

        1/28/98                                  7-21       DRAFT-DO NOT CITE OR QUOTE
-------
  1             When analyses were done by various process groups, more than twofold increases were
  2      observed for leukemia in polymerization process SMR = 251 (95% CI = 140-414, O = 15),
  3      coagulation process SMR = 248 (95% CI = 100-511, O = 7), maintenance labor SMR = 265
  4      (95% CI = 141-453, O = 13), and laboratory workers SMR = 431 (95% CI = 207-793, O = 10).
  5      Analysis by further restricting the process groups by 5+ years of employment and 20+ years
  6      since hire hi each group showed the excesses in leukemia SMRs in the same processes as above.
  7             Analyses by mutually exclusive process groups showed excesses for ever in
  8      polymerization and never hi maintenance labor or laboratories (O/E = 8/4.7), ever in maintenance
  9      labor and never in polymerization or laboratories (O/E = 6/3.7), and ever in laboratories and
10      never in polymerization or maintenance labor (O/E = 8/1.6). Within the labor group, leukemia
11      increase was observed for workers ever in maintenance labor and never in production labor (O/E
12      =11/3.8). On the other hand, for workers in production labor and never in maintenance labor,
13      the leukemia excess was negligible (O/E = 2/1.4). No  excess mortality from leukemia was
14      observed among ever in finishing and never in polymerization process workers (O/E = 4/4.5).
15             An unpublished report by the same authors (Delzell et al., 1996) submitted to the
16      International Institute of Synthetic Rubber Producers (IISRP) in October 1995 (Delzell et al.,
17      1995) included many more results of the analyses of this cohort that are relevant to this
18      assessment. A review of the unpublished results is presented in the folio wing paragraphs.
19             Various analyses by estimated 1,3-butadiene and styrene exposures were conducted. The
20      RRs calculated by Poisson regression for 1,3-butadiene ppm-years adjusted for styrene ppm-
21      years, age, years since hire, calendar period, and race for 0, >0-19, 20-99, 100-199, and 200+
22      ppm-years were 1,1.1,1.8,2.1, and 3.6, respectively.  When analysis was restricted to leukemia
23      as the underlying cause of death and person-years 20+ years since hire, the results were similar.
24      Analysis restricted to ever hourly also showed positive results for butadiene. Various analyses
25      were conducted by using alternate ppm-years categories of exposure. All the analyses
26      consistently showed similar results, strengthening the association between 1,3-butadiene and
27      occurrence of leukemias. It is interesting to note that all the leukemia subjects who were
28      exposed to 1,3-butadiene were also exposed to styrene. There were only two leukemia cases who
29      had exposure to styrene but none to 1,3-butadiene.
30             Analysis by 1,3-butadiene peak-years and styrene peak-years  demonstrated an association
31      with 1,3-butadiene peak-years and occurrence of leukemia when adjusted for styrene peak-years,
32      1,3-butadiene and styrene ppm-years, and other covariates. The association, however, was
33      irregular. A similar analysis for styrene peak-years was weak and imprecise.
34             The investigators also conducted a cancer incidence study in the Canadian plant.
35      Information was obtained from the Ontario Cancer Registry from 1965 to 1992. Standard             JMk

        1/28/98                                  7-22       DRAFT-DO NOT CITE OR QUOTE
-------
        incidence ratios (SIRs) were calculated by using the male general population of Ontario.  No
        increased incidence was found for any cancer in this study.
               This is a well-designed, -conducted, and -analyzed study. The main strengths of the
  4     study are large cohort size; long follow-up period (49 years); availability of exposure estimations
  5     on each individual, processes, and tasks; and in-depth analyses using both general population as
  6     well as internal comparison groups.
  7            There are a few limitations as correctly pointed out by the investigators. The cause of
  8     death on death certificates was not confirmed by medical records. Histologic typing was  not
  9     available for leukemias.  These limitations may have led to misclassification. Furthermore, as
10     pointed out in the Macaluso et al. (1996) study, there may have been misclassification of
11      exposure, but this was thought to be nondifferential. Two plants were eliminated from the final
12     analysis due to the lack of detailed work histories. Although this may have resulted in fewer
1 3     uncertainties, valuable data may have been lost due to this elimination. Nevertheless, the
14     association between exposure to butadiene and occurrence of leukemia was present among both
1 5     white and black workers and was fairly consistent across plants.

1 6     7.2.2.2.  Macaluso et aL, 1996: Leukemia and Cumulative Exposure to Butadiene, Styrene,
                 and Benzene Among Workers in the Synthetic Rubber Industry
               A cohort mortality study conducted in synthetic rubber workers by Delzell et al. (1996)
1 9     (Section 7.2.2.1) had a component of exposure estimation. The exposures to 1,3-butadiene,
20     styrene, and benzene were estimated by Macaluso et al. (1996).
21             An exposure estimation was conducted on each and every worker based on detailed work
22     histories, work area/job specification, IH monitoring survey records, IH recommendations,
23     various records from the plants, historical aerial pictures, use of protective and safety equipment,
24     walk-through surveys, and interviews with plant management as well as long-term employees in
25     specific areas/jobs. The quantitative exposure estimation was based on process analysis, job
26     analysis, and exposure estimation.  The job-exposure matrices (JEMs) were computed for 1,3-
27     butadiene, styrene, and benzene, which were linked to work histories of each employee.
28            Quantitative estimates of exposure to 1,3-butadiene and styrene were based on
29     background exposure plus task-specific exposure, using multiple exposure and point source
30     models, respectively.  Input variables for these models were derived from several information
31      sources described earlier. Limited validation of exposure estimates was attempted by comparing
32     the available IH data from the 1970s and 1980s as well as actually measuring the air
33     concentrations of 1,3-butadiene and styrene under controlled conditions. The latter method
        1/28/98                                   7-23       DRAFT-DO NOT CITE OR QUOTE
-------
  1      showed a good agreement among the methods of sampling, while the comparison of IH data
  2      indicated overestimations of 1,3-butadiene exposure.
  3             For each job, 8-h time-weighted average (TWA) intensities and the number of peak
  4      exposures (15-min exposures over 100 ppm) were calculated. Based on job exposures, a JEM
  5      database was developed that was linked with individual work histories to develop individual
  6      quantitative work exposure estimates. For each individual, the exposure indices were multiplied
  7      by the length of employment in that particular process or job and were added up for the total
  8      employment period in various jobs to estimate the cumulative exposure.
  9             Mortality analysis was done by calculating the SMRs and risk ratios (RR) using
10      estimated quantitative exposures to 1,3-butadiene, styrene, and benzene. Both cumulative ppm-
11      years and peak-years were calculated for each individual in the study. Person-year data were
12      grouped by 1,3-butadiene, styrene, and benzene ppm-years for both SMR analyses as well as RR
13      analyses.  Comparability between cohort mortality rates and general population reference
14      mortality rates was assured by limiting the SMR analysis to the individuals whose underlying
15      cause of death was listed as leukemia (51 people).  Risk ratios were computed by using the
16      Mantel-Haenszel method and 95% CI were computed by the Breslow method. Poisson
17      regression models were used for adjustment of multiple confounders and to compute within-
18      cohort mortality rates, and the X2 test for linear trend was used to examine the dose response.
19             Work histories were available for 97% of the population. Fifty-two in-depth interviews
20      with plant management and long-term employees identified 446 specific tasks/work areas with
21      potential for 1,3-butadiene, styrene, and benzene (3 plants only) exposure. Eight-hour TWAs for
22      1,3-butadiene, styrene, and benzene were 0-64 ppm, 0-7.7 ppm, and <1 ppm, respectively, the
23      median exposures being <2 ppm for 1,3-butadiene and 0.5-1.1 ppm for styrene.
24             Exposure analysis found that 75% of the cohort was exposed to 1,3-butadiene, 83% was
25      exposed to styrene, while only 25% was exposed to benzene. The median cumulative exposure
26      to 1,3-butadiene, styrene, and benzene was 11.2, 7.4, and 2.9 ppm-years, respectively. The
27      exposure prevalence as well as median cumulative exposure was higher in individuals who had
28      died of leukemia. Among the leukemia decedents, 85% had exposure to 1,3-butadiene, with their
29      median cumulative exposure being 36.4 ppm-years. This exposure was two times higher as
30      compared with all decedents and three times higher as compared with all the other employees.
31      The exposure to styrene was present in 90% of leukemia decedents, with median cumulative
32      exposure hi them being 22.4 ppm-years, two times and three times higher as compared  with all
33      the decedents and all other employees, respectively. Benzene exposure was found to be less
34      frequent among leukemia decedents as  compared with all the other employees. Analysis by
        1/28/98                                   7-24       DRAFT-DO NOT CITE OR QUOTE
-------
         benzene exposure showed no association with the occurrence of leukemia after adjustment for
         1,3-butadiene and styrene.
                Leukemia SMRs increased with increasing cumulative exposure to 1,3-butadiene as well
  4      as styrene. Mortality RRs computed for cumulative 1,3-butadiene exposure adjusted for race,
  5      age, and cumulative styrene exposure also showed increasing RRs for increasing cumulative
  6      exposure to 1,3-butadiene.  The adjusted RRs for cumulative exposures of butadiene of 0, <1,1-
  7      19, 20-79, and 80+ ppm-years were 1,2.0,2.1, 2.4, and 4.5, respectively. The linear X2 test for
  8      trend was statistically significant (p = 0.01). When similar RRs were computed for styrene
  9      exposure, neither showed a consistent pattern nor a trend of increasing risk with increasing
 10      exposure. A similar trend test was statistically not significant.
 11             Analysis by exclusion of the nonexposed population resulted in RRs of 1,1.5, and 1.7 for
 12      0.1-19, 20-79, and 80+ ppm-years of the cumulative exposures of 1,3-butadiene. The linear
 13      trend test was statistically significant (p = 0.03), substantiating the earlier finding of increasing
 14      risk of leukemia with increasing cumulative exposure to 1,3-butadiene. Although the same
 1 5      analysis suggested increasing risk of leukemia with increasing cumulative exposure to styrene
 16      after adjustment for 1,3-butadiene and other eovariates, the results were imprecise and
 1 7      statistically nonsignificant.
J 8             There was neither any positive or negative interaction found between the cumulative
         exposures to 1,3-butadiene and styrene.
 20             For the last decade or so, epidemiologists have been including exposure estimation in
 21      their studies. The methods used and efforts made to do exposure  estimations are improving but
 22      variable. This study is one of the best efforts of exposure estimations to date. The investigators
 23      have used many available methods to come up with best estimates of exposures of 1,3-butadiene,
 24      styrene, and benzene. They also have validated these estimates on a smaller scale. Although this
 25      is considered as the best effort, it should be noted that these are estimates and not actual
 26      measurements. Two plants were eliminated from the analysis because detailed work histories
 27      were lacking. Thus it is possible that individuals may have been misclassified with respect to
 28      process or job, resulting in either over- or underestimations of exposure. However, there is no
 29      reason to believe that the misclassification of exposure occurred only in individuals who had died
 30      of leukemia.
 31             Studies in polymer production workers are summarized in Table 7-2.

 32      7.3. SUMMARY AND DISCUSSION
 33             1,3-Butadiene has been shown to be both mutagenic as well as carcinogenic in animals
 |4      and humans. Data in animals, particularly hi mice, show that butadiene is a multisite carcinogen

         1/28/98                                   7-25       DRAFT-DO NOT CITE OR QUOTE
-------
  a
  o
  ts
  2
  c,
  o

  r
  u
  -3
  «
  •+•»
  Or



VI
c
_o
03
•*•>
J
•a
03
w
H
ob
o
u
-**
V)





"3
ca
1,3-Butadiene exposure
assessment

Population studied

L*
o
£
3
"*



^r^
m
CO

**-<
O
t
W3
Largest cohort mortality
workers


1


Among production wo
Divided in four major
areas based on the
e
Update of the cohort fror
Matanoski and Schwartz
C*J C\
o 2
^ ""O
O C3
Is
C3 0
*5JJ >•"- 1






Lack of exposure data
e
CO
8
w g
Ci O
SSSMR = 260foroth
lymphohematopoietic
whites
longest job held:
Production workers
erf
(1987)
Cohort mortality of 8 SB








C
e
_o
Js
1
o.
Exclusion of 50% of the
the follow-up





ui
o
II
CO
CO
CO
Utility workers

polymer production plani
workers







S
o

M)
_C
"5b
1
ex
Four plants had follow-i
ca
« "i
g j§
g J
o ,
lymphohematopoietic
andSSSMR = 655foi
Maintenance workers
CN
CM
•^
cn~
tl-l
o
•e
o
_c
o
u
•o
s
-§
£







•a
3
0
CD

51 -8
12 years to 25 years; ma
time for malignancies to



•S
'S
in blacks
No relation observed v
All other work sites
,_,
followed through 1982
Worked for at least 1 yea



















latency or duration of






















employment


Comparison group U.S.
population









CO

^ S
II
S £
One of the strengths is ai
made to estimate actual i





(L)
•3
i
i
Exposure to 1,3-
butadiene and styrene
-^
Nested case-control studj
•a
13
CO
O ^^
_« g
03 C^
S c^


^

"S
0
C3
d>
O
>
"S
1
1 0
o j;
«*r< O
2 -a



T3 O

• Ever/never exposu
SS OR = 6.82 (high) a
4.26 (low) were found
was done by job
identification and levels
associated with that job

Cases:
- of leukemia 26
- of other lymphatic
cancers 18
1
1
CQ r?
03 ^

G
co "3
rt-. c

> 0
S -o

° ca" «
III 1
•2 ti ca G,
^ cd o S<
C+.4 C ! 1£J 0>
0 .£P'S5 ^
in i
n1 i-i
? '3 .S Stt o
«2 §--0 J3 S3
0 § S JS ta
•g §3 ^ «
a l s s t

u
 UTS s a -s is
jg *f •§ -| § > 1 S
§ Jsls S«g
C3 CXOco-S oO?
O




















§
1
•o
-a
CO O
e'^
8 g
CO D









  o

  JS
  •»-*


  o
  •s
  a
  "o
  • ^•4

  S
  O
  r?
  t-~
1/28/98
7-26
DRAFT-DO NOT CITE OR QUOTE
-------
    &
   -o
    o

    a
    1.
    
    o


    r
   I
    I
    as
    O
    o


   g

   "e3

   ^

    4>
   en

   •S

   "3
   CU
CA
S
o
•-B


S
2
T3
fi
C3
«j
J3
"So
C
g
co




"3
CA
S
1,3-Butadiene expos
assessment
"O
o>
•3
•*•»
CA
Population

g
o
y




.2
CD
1
1-4
vq
II
O
OO
00
a
A log transformatio
of the scores was
t*
t-
II
00
oo

CO
CD
CO
"cO
c
CO
CD
•S
.£
T)
ID
CO
lymphoma when styrene
exposure was low

CD
1
1
CO
CD
CO
"co
c
SS trend of 3.76 was found
for increased risk of

- by ever/never
exposed
leukemia with increasing

1
.£P
exposure levels of butadiene

exposure
- both matched
(conditional) and
unmatched
(unconditional)

ID
2 g^
o oo
C ON
.cO
S 13,
CO
O
P3
t
1
^
oo

If

~^'~
a o
" ^


















-


•S
CO
CD
"O
O
CD
CO
CO
O
<*-.
o
G
_O
to
o
S
S3

*^
Similar results with new
controls

Exposure estimation
done based on


CD
CO
CO
o
•o
It
CO OT
CO §
00 0
"co :
i
o
la
CO O\
S 0
1
CO
CD
CO
CD

T3
"c3
>•
CO
"o

»-H
s

O
'ca
™ *S
CD CO
>->• CD
r^ hH
Reanalysis of cohort data for

measurements
provided by seven
WJ
o
1
o
o
0
ID
CO
1
<



























three plants

T3
§
00
s- x
||



CD
CO
CO
O
CD



























SSSMR=163forallLHC
SS SMR = 181 for leukemia



CD
«
use of deaths
hospital reco
o ^



























and aleukemia



CA
'£>
13
hort data rear
o
U






1/28/98
7-27
DRAFT-DO NOT CITE OR QUOTE
-------
  si
  o
 •J3
  u
  a


 1
  o
  a
 *o

  r
  c
  o
 'O
 •9
  cn
  O
  B,

  g
  u

  o
  o
 -a

 S
 cn
 u


 f
 "S

 1
 u
 3
 'S.
 N
    0


    C!
 .a
  «





















en
.0

cs
*w
1

3

T3i
g;
O

55




1

1,3-Butadiene

exposure
assessment
•a
.2
•5
s
*»-*
VI
to
%-•
—
"3
c,
o
Pi








en

o
"
4jj


g
*£
-o
CS
s
CO
1
T3
"O

en

en
•u
O
•£
0>

*=
ca
_>
ta
3

Q
j_
Q
S
K
"S
In
|p
"*

Exposure estimation

Cohort of seven U.S. anc
CS
4-*
 cT

conducted based on
several information

one Canadian SBR
workers mortality study





if
c\
s»^

o
's

o
o.
g
.g
'1
_0
ca
o
C^4
'en
ra
"o
en

^

.— r
•
CS
cf
C4

^^
rti
}_
O
^^
a ^
S T3
^ g
I,"*"
Q. CS

sources including IH
._
en
CS
L-i
-a
1








0
o


73
"3 M

^^
o .2
3 G
'55 K
CO •*
° S
d« ,-"
^ c
~^
K** 1*^1*
5 C
£ o






00
oo
en
CS
4-»
cn
S
•a
1
H
ai
Quantitative exposun
estimates on

between January 1, 1943
and January 1, 1992


























•a
ID
en
o
d,
g
C
O
c
CD
£
<4-4
0
c
.2
'cn
_3
"o
w

background, task-





























is

1 1

"S c
S .£
*c i*
1?
c «
'43 !^
•i ^
§* %
£X S

S
ja "S
.2* ^
11
« Q>
«| 1
"" S
0) CU
£X X
en 
J2
Follow-up period throng!
January 1, 1992









































sources models for










































 O
co "r- CU
£ S S
" S g
O 1> en >
ft, A cu '43
8 S S —
0 .S en 3
•S ^ § s
& 3% 5

U.S. population
Respective State
populations where the
plants were located
Ontario male rates for
Canadian plant









































Exposures estimated
for each individual
00
Internal comparison usin
Poisson regression







1/28/98
7-28
DRAFT-DO NOT CITE OR QUOTE
-------
c
o
•*•<
u
"O
0
SN
cu
i«
1
•^
r
c
T3
S3
-*J
ja
^
i-T
of exposure to
•52
«S
0>
 a> a
« i- cu
111
« li
.j- s




"cj
•3
5
SS
_o
—
"3
o
0.




en
O
S
"^

^
Co" cs
os .S
2 -g §
- S ^ 2
2 cu Ji "^

o o ° ii
S § g5 S c
rt J3 « 2
IS T3 o *^ ia
co o 60 ^H co
i i HI
C s en Kt w
^ ;u S g S
"O „ c/3 w
> i 0s CB CO T3 J3
g JD co+ "^ >,>, g^
I ~ li 1 ^2^ 1 |
§ 2 •?" T: I £^ :§!
^ 0 00 0 0 0« ^: —
£H Cj_( Ct_!CM t4_( t4_, C4_(Ca ^X3
§ mj2r-r~- o o\ •* 'S S .° fe
?_^I-SCNOO__,O r- cNg-a^cS
>,.2~-.^CN~-'£cN >"_r-C^cJ=lCN
•S g ii o ii H § ii .g H ^ ii 1, 8 2 ^
o_S(^ ?et.'Si5CH~'^C(~l^^"H.'« II
£ .2 OT- T= M S £co Soo^co | ^|o!_
> oCO OOOCO o>CO ScO S)COci_ S^S O?
W eg 00 -J3 00 00 ^00 >,00 C3COO>»5£,COCO

S CO
0 , < c5
J iu 53 g
o ca « a.
1 £-5 £
"^ ^o £
„ vo co S co
cs gs „ "55 M
rt ^^ 'jQ ^1 *i *^Q 3
CO "cS < JS JS < 00


13
"55
0
3
"cs
g

.s ^
d> *O
S OS
CO d^



'
CS
=3 x~>
JS Os
Q C!-













o =-
ll J _ . ^ III
jlii
h™ r-< ^0 . O ^J 5 ^0 *~T KJ 03 JJ»«* ^7*
^^ OT t^ O.i CO C CO .^ C^ ^^ 03 03



























1/28/98
7-29
DRAFT-DO NOT CITE OR QUOTE
-------
 o
 3

1
 a
 o
 S
 >?
 o

 r
 a
 . «j

M^ .H
•»B  IS
^ £•«
—    ea
        S
V3
     CO
     3
    •a
               w
     o
     o

     I
    _o

     o3

    1
    Q
     a

     I
8 .2 _

lit
•sit
        w oo £
        II II II  II  II  0

          oi        U
        oo S a 03  ^ ffi
        GO 00 K CS  O J
1/28/98
                        7-30
                                    DRAFT-DO NOT CITE OR QUOTE
-------
         even at the lowest dose of 6.25 ppm (NTP, 1993). Occupational populations are exposed to
         butadiene in the production/recovery of butadiene monomer and production of resins and
         plastics.  Exposure to this colorless, odorless gas is entirely via inhalation due to its extremely
  4      volatile nature. The general population is exposed to butadiene in ambient air, the major sources
  5      of its release in ambient air being automotive exhaust and cigarette smoke. Its potential to cause
  6      cancer in humans has become an important public health issue.
  7            Butadiene becomes diluted in ambient air and is eliminated by photooxidation.  Thus it is
  8      difficult to study Hie health effects of exposure to butadiene in the general population. Since
  9      exposure to butadiene is ubiquitous in the general population, "unexposed" reference populations
 10      used in occupational cohort studies are likely to contain a substantial number of individuals who
 11      are exposed to butadiene nonoccupationally. Furthermore, the issue of health measurement is
 12      complicated by the fact that occupational cohorts tend to be healthier than the overall general
 13      population and have below average mortality, which is referred to as the "healthy worker effect."
 14      Thus the standard mortality ratios observed in occupational cohorts, computed using the general
 15      population as the reference group, are underestimations of real risk.

 16      7.3.1. Monomer Production
J}_7            To evaluate the carcinogenicity of 1,3-butadiene, cohorts from monomer and polymer
         production were studied by several investigators. The largest cohort of monomer production
 19      workers was initially studied by Downs et al. (1987) and had three follow-ups by Divine (1990),
 20      Divine et al. (1993), and Divine and Hartman (1996). The cohort included 2,586 workers
 21      initially and had 2,795 individuals in the last follow-up due to an extended time period for the
 22      inclusion criteria. The four exposure groups were identified by Downs et al. (1987) based on a
 23      qualitative exposure scale. They remained the same in Divine's (1990) follow-up and were
 24      similar but slightly changed in Divine et al. (1993). In their last follow-up, based on IH data, the
 25      investigators (Divine and Hartman, 1996) estimated the potential exposure to butadiene for each
 26      employee by their work histories (in 1-year segments), using job categories and calender time
 27      periods.  Cumulative exposures were obtained by summing the scores of all the years of
 28      employment.
 29            The findings of all four investigations were essentially the same even after 52 years of
 30      follow-up.  There were deficits observed for mortality from all causes and all cancers.  The only
 31      statistically significant excess observed was for lymphosarcoma (ICD code 200).  Downs et al.
 32      (1987) observed this excess for the total cohort and for the subcohort of workers who had worked
 33      for less than 10 years and latency of 0-9 years.  This excess was seen in the prewar subcohort in
         all three follow-up studies (SMR = 269, and SMR = 254 in both Divine,  1990, and in Divine et

         1/28/98                                   7-31        DRAFT-DO NOT CITE OR QUOTE
-------
 1      al., 1993; Divine and Hartman, 1996). No information on exposure levels was available for this
 2      period, but it was believed that the exposures were high during the prewar period.  When
 3      analyses were done by years of employment and latency excess for lymphosarcoma, mortality
 4      was always found to be in individuals employed for less than 10 years and with latency of 0-9
 5      years. It should be noted that after 52 years of follow-up, no elevated mortality was observed for
 6      leukemia, which was the main finding in SBR workers.'
 7            A small cohort of 364 individuals was identified from 29,139 workers at three Union
 8      Carbide Corporation plants who had potential exposure to butadiene during World War II (Ward
 9      et al., 1995,1996c). The exposure to butadiene was assumed based on job categories, and no
10      adjustments for confounding by other chemicals were done. As observed in the Divine Studies
11      (1990,1993,1996), a statistically significant excess for lymphosarcoma (SMR = 577) also was
12      observed in this cohort.
13            A third cohort of 614 workers exposed to monomer was studied by Cowles et al. (1994)
14      and the study failed to show any excess mortality or morbidity. Due to several methodologic
15      limitations, this study failed to provide any negative evidence towards the causal association
16      between exposure to butadiene and occurrence of lymphosarcoma that was observed in the other
17      two cohorts.

18      7.3.2. Polymer Production
19            A further follow-up and reanalysis of a large SBR polymer production workers' cohort
20      (Matanoski and Schwartz, 1987) was conducted by Matanoski et al. (1989,  1990). This follow-
21      up added 3 years to the earlier study. The findings of this follow-up were essentially the same as
22      the earlier study. The only statistically significant excesses were found among production
23      workers.  Among whites the excess was for other lymphohematopoietic cancers (SMR = 260)
24      and among blacks the excesses were for all lymphohematopoietic cancers (SMR = 507) and
25      leukemia (SMR = 655). Analyses by duration of work and latency did not show any increases in
26      hematopoietic cancers. There were no exposure measurements or estimations done in this study.
27            A nested case-control study from this cohort (Matanoski et al., 1989, 1990) was
28      conducted by the same investigators and reported in Matanoski et al. (1989) and Santos-Burgoa
29      et al. (1992). Fifty-nine cases of lymphohematopoietic cancers and 193 matched controls were
30      identified. Exposures to 1,3-butadiene and styrene were estimated in these individuals using the
31      job records and levels of exposures to 1,3-butadiene and styrene associated with those jobs
32      independently of the case or control status. The jobs were ranked and cumulative dose was
33      calculated for each case and control. Analyses were conducted using log transformed scores.
34      The relative odds were increased for high (OR = 6.82) and low (OR = 4.26) exposures in the

        1/28/98                                  7-32       DRAFT-DO NOT CITE OR QUOTE
-------
  1      ever/never exposed analysis, matched analysis (OR = 9.36), and conditional analysis (OR = 7.61)
  2      for leukemia.  All the increases were statistically significant. A statistically significant trend was
 *3      also observed for increasing risk of leukemia with increasing exposure levels of butadiene.
  4            Because the findings of the nested case-control study were questioned by Acquavella
  5      (1989) and Cole et al. (1993), as they were in disagreement with the base cohort study,
  6      Matanoski et al. (1993) reevaluated the analysis of the nested case-control study by choosing a
  7      new set of three controls per case. The investigators also verified the cause of death by obtaining
  8      the hospital records. The findings of the new analysis were similar to the earlier analysis,
  9            Furthermore, they estimated the exposures to the cohort based on measurements provided
10      by seven rubber plants, IISRP, and NIOSH.  In an analysis of the subcohort from three plants
11      who had the geometric means of exposure, statistically significant excesses were observed for all
12      lymphohematopoietic cancers (SMR = 163) as well as for leukemia and aleukemia (SMR = 181).
13            Delzell et al. (1996) and Macaluso et al. (1996) reported separately the two components
14      of the follow-up study of synthetic rubber workers. These investigators studied the seven plants
1 5      studied by Matanoski and Schwartz (1987), Matanoski et al. (1989,1990, 1993), and Santos-
1 6      Burgoa et al. (1992) and one plant (two initial plants combined into one) by Meinhardt et al.
17      (1982). The follow-up period was 49 years.  Investigators estimated the  exposures to  1,3-
1 8      butadiene, sryrene, and benzene for each worker. This was done by using various means such as
        job histories, work areas, IH data, historical plant data, aerial pictures, interviews with long-term
20      employees and managers, walk-through surveys, etc.  Quantitative exposures were calculated and
21      limited validation of exposure estimates were attempted using available 1970's and 1980's IH
22      data. Cumulative and peak exposures were calculated for each worker. Comparison with the
23      U.S. population resulted in statistically significant excesses for leukemia in ever-hourly workers
24      (SMR = 143) and its subcohort of blacks (SMR = 227). The excesses were also found hi the
25      ever-hourly cohort for year of death (SMR,= 187 for 1985+), year of hire (SMR = 200 for 1950-
26      59), age at death (SMR = 179 for <55 years), and for more than 10 years employment and more
27      than 20 years since hire (SMR = 192 for whites and SMR = 436 for blacks). Laboratory workers,
2 8     maintenance workers, and polymerization workers also showed increased SMRs of 431, 265, and
29      251, respectively. All these analyses were done adjusting for styrene and benzene.  When
30      internal comparison was done using the estimated ppm-years exposure data, relative ratios
31       increased with increasing exposures.  The trend test was statistically significant.
32            The incidence study conducted in the Canadian plant employees did not show any
33     increases in any cause-specific cancers.
        1/28/98                                   7-33       DRAFT-DO NOT CITE OR QUOTE
-------
 1      7.3.3. Relevant Methodologic Issues and Discussion
 2            Throughout this chapter, various methodologic issues including strengths and limitations
 3      are discussed. The major concerns are lack of exposure information and short follow-up periods
 4      in earlier studies, small cohort size, lack of data on confounding variables, and lack of latency
 5      analysis hi one study. Furthermore, death certificates were used by all the investigators, which
 6      could lead to misclassification bias. Validation of diagnosis of lymphohematopoeitic cancer was
 7      not done in any of the studies except in Matanoski et al. (1993). This is a methodologic concern
 8      given the fact that lymphohematopoeitic cancer recording on death certificates is unreliable
 B      (Percy etal., 1981).
10            Lack of exposure information is another major limitation in Cowles et al. (1994) and
11      Ward et al. (1995,1996c). Cowles et al. (1994) made no attempt to even do job classification.
12      This cohort was very small, there were very few deaths, and more than 50% of the cohort had an
13      average follow-up  of 12 years.
14            Ward et al.  (1995,1996c) also did not attempt any exposure estimation.  This cohort also
15      was very small but was restricted to workers who had worked in the 1,3-butadiene production
16      period (during World War II). The high SMR for lymphosarcoma and reticulosarcoma observed
17      in this study was based on only four cases. They used employment of 2 years+ as surrogate for
18      exposure and stated that there were no other common exposures to other chemicals. Considering
19      that the cohort was small and only four deaths occurred from lymphosarcoma and
20      reticulosarcoma, it should be noted that this finding is consistent with the finding of the other
21      monomer facility studied by Divine (1990), Divine et al. (1993), and Divine and Hartman (1996).
22            A monomer cohort study conducted by Downs et al. (1987) and followed by Divine
23      (1990) and Divine et al. (1993) also lacked exposure information, although the surrogate
24      exposure grouping was done by qualitative exposure information based on job descriptions/work
25      areas. The investigators attempted the exposure estimation in their last follow-up (Divine and
26      Hartman, 1996) and found that except for an excess observed for lymphosarcoma and
27      reticulosarcoma in the prewar subcohort, there were no excesses in any cause-specific cancer
28      mortality.  However, investigators did not have any information on work histories or levels of
29      1,3-butadiene exposure during the prewar period, which made exposure estimation in the prewar
30      workers impossible. Even after 52 years of follow-up and extensive analyses, this cohort has not
31      observed any excess hi mortality from leukemia that was observed in SBR workers.
32      Nonetheless, the finding of excess mortality from lymphosarcoma and reticulosarcoma is
33      consistent with findings of Meinhardt et al. (1982) and Ward et al. (1995,1996c). In addition,
34      the excess of lymphosarcoma and reticulosarcoma in short-term workers but not in long-term
35      workers was consistent with the similar findings of Meinhardt et al. (1982).

        1/28/98                                  7-34       DRAFT-DO NOT CITE OR QUOTE
-------
               Matanoski and Schwartz (1987) and Matanoski et al. (1989, 1990) did not have any
         exposure information available. The cohort was distributed in four major areas based on longest
         jobs held and the qualitative exposure information used as surrogate.  When the nested case-
  4      control study was undertaken by these investigators (Matanoski et al., 1989; Santos-Burgoa et
  5      al., 1992), exposure estimation was done by using various sources only for the selected cases and
  6      controls. They observed a statistically significant high excess from leukemia mortality, which
  7      the authors concluded as being causally associated with exposure to 1,3-butadiene.
  8            Matanoski et al. (1993) validated their earlier results of the nested case-control study by
  9      using a new set of three controls per case. They also verified the cause of death noted on the
 10      death certificates and  diagnosis noted on the hospital charts. They found that the diagnosis noted
 11      on 25 out of 26 charts agreed with the cause of death noted on the death certificates. The results
 1 2      of this study were similar to the earlier nested case-control study.
 1 3            This finding of a high excess of leukemia mortality in the case-control study was
 14      questioned by Acquavella (1989) and Cole et al. (1993) because no excess leukemia mortality
 1 5      was found in the base cohort study from which the cases and controls were selected. Their
 1 6      argument that the results of the case-control study were statistically incompatible with the results
 17      of the cohort study was based on the calculations of number of leukemias that should have been
J 8      seen in the cohort study, based on the relative odds observed in the case-control study. The Cole
         et al. (1993) calculations resulted in approximately 104 leukemia cases if relative odds of 7.6
 20      were applicable to 60% of the cohort that was exposed to 1,3-butadiene and an additional 9.2
 21      expected leukemias for the remaining 40% cohort that was not exposed, resulting in an observed
 22      113 leukemias for the cohort as against 22 leukemias actually observed in  the cohort study.
 23      Variability in both the prevalence of exposure and the relative odds were looked at by these
 24      authors (Cole et al., 1993), and they concluded that there was no reasonable combination that
 25      resolved the incompatibility between the findings of the cohort and case-control studies.
 26            Matanoski and Santos-Burgoa (1994) disagreed with this criticism. They asserted that the
 27      60% exposure observed among the controls in the case-control study overestimated the
 28      prevalence of exposure for the cohort population and that the matching criteria may have skewed
 29      the control selection and produced controls who were not representative of the base cohort.
 30            The main limitations of the cohort study were that more than 50% of the population was
 31      excluded due to lack of work histories or start date and lack of exposure data.  The follow-up for
 32      four plants where the starting date was 1957 to 1970 may not have been long enough for
 33      malignancies to develop.  As far as the nested case-control study is concerned, as pointed out by
 34      the authors, the estimated exposures were crude and were not substantiated by IH data. The
 |5      exposure misclassification may have occurred based on the estimated exposure by job if the jobs

         1/28/98                                   7-35       DRAFT-DO NOT CITE OR QUOTE
-------
  1      were incorrectly identified for higher or lower exposure. However, the panel members were
  2     blind towards the status of cases and controls, thus the distribution of misclassification should be
  3     the same in cases and controls.
  4           Although the controversy about the cohort and case-control study is still not resolved, the
  5     nested case-control study was the first one to demonstrate a strong association between exposure
  6     to 1,3-butadiene and occurrence of leukemias.
  7           The Delzell et al. (1996) and Macaluso et al. (1996) cohort study is one of the best efforts
  8     of exposure estimation to date. Some misclassification of exposure may have occurred with
  9     respect to certain jobs, but it is unlikely to have occurred only in leukemia cases. The
10     investigators also did some validation of exposure estimates based on IH data. They pointed out
11      correctly that the excess mortality observed for leukemia was based on death certificates and was
12     not verified by medical records. Histologic typing of leukemia was also not available. This may
13     have resulted in misclassification.  Two plants were eliminated from the final analysis due to the
14     lack of work histories, which may have resulted  in the loss of valuable data.
1 5           Based on these monomer and polymer production worker cohorts, it is obvious that an
1 6     increased number of lymphohematopoietic cancers is observed in these populations. A clear
17     difference is becoming apparent though. Increased lymphosarcomas  develop in workers exposed
18     to monomer (Downs et al., 1987; Divine, 1990; Divine  et al., 1993; Divine and Hartman, 1996;
19     Ward et al., 1995,1996c), while excess leukemias occur in workers exposed to polymer
20     (Matanoski et al., 1990,1993; Santos-Burgoa et  al., 1992; Delzell et al., 1996; Macaluso et al.,
21      1996). Furthermore, the lymphosarcomas were observed in the monomer workers, who were
22     probably exposed to higher levels of 1,3-butadiene for shorter periods of time (wartime workers)
23     and not in long-term workers with low levels of  exposures. A confirmation of this observation
24     comes from the stop-exposure studies conducted by Melnick et al. (1990a). They observed that
25     at a similar total exposure, the incidence of lymphoma was greater among mice exposed to
26     higher concentrations of butadiene for a shorter period of time (625 ppm for 26 weeks) than
27     among mice exposed to a lower concentration for a longer period of time (312 ppm for 52
28     weeks).  Consequently, this suggests that it is the concentration of 1,3-butadiene rather than the
29     duration of exposure that is important in the occurrence of lymphomas. There is a null
30     relationship between exposure to 1,3-butadiene monomer and occurrence of leukemias that is
31      observed in polymer workers. This may be due to very low exposures to 1,3-butadiene in
32     monomer production workers or exposure to a necessary co/modifying factor or a confounding
33     factor in SBR production workers. Data are currently lacking to confirm or refute any of these
34     possibilities. The findings of Delzell et al. (1996) and Macaluso et al. (1996) are inconsistent
        1/28/98                                   7-36       DRAFT-DO NOT CITE OR QUOTE
-------
  1      with confounding by exposure to other chemicals. The findings of excess leukemias in SBR
  2      production workers are consistent with a causal association with exposure to 1,3 butadiene.


  3      7.3.4.  Criteria of Causal Inference
  4            In most situations, epidemiologic data are used to delineate the causality of certain health
  5      effects. Several cancers have been causally associated with exposure to agents for which there is
  6      no direct biological evidence. Insufficient knowledge about the biological bases for diseases in
  7      humans makes it difficult to identify exposure to an agent as causal, particularly for malignant
  8      diseases when the exposure was in the distant past.  Consequently, epidemiologists and
  9      biologists have provided a set of criteria that define a causal relationship between exposure and
 10      health outcome. A causal interpretation is enhanced for studies that meet these criteria. None of
 11       these criteria actually proves causality; actual proof is rarely attainable when dealing with
 12      environmental carcinogens. None of these criteria should be considered either necessary (except
 13      temporality of exposure) or sufficient in itself.  The absence of any one or even several of these
 14      criteria does not prevent a causal interpretation. However, if more criteria apply, it provides
 1 5      credible evidence for causality.
 16            Thus, applying the criteria of causal inference to the monomer and polymer cohort
 17      mortality studies and one nested case-control study in which risk of lymphohematopoietic
 } 8      cancers were assessed resulted in the following:

 19          •      Temporality. There is temporality of exposure to 1,3-butadiene prior to the
 20                 occurrence of lymphosarcoma in monomer workers and leukemias in SBR workers.

 21           •      Strength of association.  Strength of association between exposure and the
 22                 occurrence of lymphosarcoma in the prewar period ranged from 154% to 477%
 23                 higher risk among workers exposed to monomer as compared with the nonexposed
 24                 general population (Divine, 1990; Divine et al., 1993; Divine and Hartman, 1996;
 25                  Ward et al., 1995,1996e). The excess risk of leukemia ranged from 43% to 127%
 26                  higher among workers exposed to SBR in ever-hourly workers as compared with
 2 7                  the general population (Delzell et al., 1996).  Internal comparison of SBR worker
 2§                  population resulted in a 4.5-fold increased leukemia risk among the highest
 29                  exposure group in the same cohort (Macaluso et al., 1996). The nested case-control
 30                  study from the SBR cohort showed a 7.6-fold increase in the risk of leukemia
31                  (Matanoski et al., 1989, 1993; Santos-Burgoa et al., 1992).

32           •      Consistency. Two cohort studies in monomer workers showed an increased risk of
33                  lymphosarcoma (Divine, 1990; Divine et al., 1993; Divine and Hartman, 1996;
34                  Ward et al., 1995, 1996c), while one cohort study (Delzell et al., 1996; Macaluso et
                    al., 1996) (with a cohort derived from seven U.S. plants and one Canadian plant)
                    and one nested case-control study (Matanoski et al., 1989,1993; Santos-Burgoa et

        1/28/98                                  7-37       DRAFT-DO NOT CITE OR QUOTE
-------
  1                  al., 1995) showed an excess risk of leukemia in SBR workers. The SBR workers
  2                  cohort defined by Delzell et al. (1996) showed a fairly consistent association
  3                  between exposure to butadiene and occurrence of leukemia across plants.  Excesses
  4                  for both lymphosarcoma as well as leukemia were observed by McMichael et al.
  5                  (1974,1976) and Meinhardt et al. (1982).

  6           •      Specificity. All monomer studies showed an increased risk of lymphosarcoma
  7                  while SBR studies showed an increased risk of leukemia.  Overall, they show
  8                  increased risks of lymphohematopoietic system cancer among populations exposed
  9                  to 1,3-butadiene. It should be noted that exposure to a particular chemical (or drug
10                  or radiation) may cause more than one type of leukemia or another type of
11                  hematopoietic cancer (Linet, 1985).

12           •      Biological gradient. The biological gradient, which refers to the dose-response
13                  relationship, was observed only in SBR workers. Both the nested case-control
14                  study and the cohort study showed increasing risk of leukemia with increasing
15                  exposures. Such a relationship was not observed in monomer workers.  The reason
16                  may be because a very small number of people were exposed to high levels of 1,3-
17                  butadiene for a shorter period of time who showed the occurrence of
18                  lymphosarcoma. They could not be further stratified to evaluate the dose response.

19           •      Biological plausibility.  As described in Chapter 4, hemoglobin adducts have been
20                  detected in humans exposed to 1,3-butadiene (Osterman-Golkar et al., 1993; Sorsa
21                  et al., 1996). Significantly increased frequencies ofhprt mutant lymphocytes were
22                  observed in high-exposure groups by Legator et al. (1993) and Ward et al. (1994).
23                  Mutations, chromosomal aberrations, and cell transformations, all well-established
24                  steps in the process of carcinogenesis, were observed in human and animal studies.
25                  This makes a convincing argument for the biological plausibility of occurrence of
26                  leukemia in SBR workers and lymphosarcoma in monomer workers.

27             In conclusion, some of the causality criteria apply to monomer workers and occurrence of
28      lymphosarcoma while all the criteria apply well for leukemia among SBR workers.  Based on
29      strength of association, dose-response relationship,  specificity of cancer (leukemia—specific cell
30      type is not known at this tune), and biological plausibility, there is sufficient evidence to consider
31      1,3 -butadiene a known human carcinogen.
        1/28/98                                  7-38       DRAFT-DO NOT CITE OR QUOTE
-------
                                 8. PHARMACOKINETIC MODELING

         8.1. INTRODUCTION
  2             Several physiologically based pharmacokinetic (PBPK) models of 1,3-butadiene
  3      metabolism and disposition have been developed to attempt to explain the interspecies
  4      differences in the potency and site specificity of the carcinogenic response between mice and rats
  5      and to provide a corresponding dosimetric basis for quantitatively extrapolating carcinogenic
  6      potency from rodents to humans (Hattis and Wasson, 1987; Hallenbeck, 1992; Kohn and
  7      Melnick, 1993; Johanson and Filser, 1993; Evelo et al., 1993; Medinsky et al., 1994).  PBPK
  8      models use species-specific physiological parameters such as alveolar ventilation rates and blood
  9      flow rates, chemical-specific distribution parameters such as bloodrair and tissue:blood partition
 10      coefficients, and species- and chemical-specific metabolic rates to elucidate the pharmacokinetics
 11      (i.e., the uptake, distribution, metabolism, and excretion) of a chemical.
 1 2             Ideally, such models provide species-specific target tissue doses of the toxicologically
 1 3      active form(s) of the chemical.  Carcinogenic risks from bioassay data can then be extrapolated
 14      to humans on the basis of equivalent effective doses, reducing some of the uncertainties that
 1 5      occur when interspecies extrapolation is based simply on exposure to the parent compound,
 1 6      especially when nonlinear physiological processes are involved.  Assumptions must still be made
         to the effect that the mechanisms of action of the active form(s) of the compound at the target
 1 8      tissue(s) are the same across  species and that the tissues of different species are equally sensitive.
 1 9      If these assumptions are not valid, pharmacodynamic data and modeling are required for more
 20      precise risk assessment.
 21            PBPK models that fall short of describing target tissue doses of the active form(s) of a
 22      chemical may still be useful for improving the dosimetric basis of interspecies extrapolation for
 23      quantitative risk assessment.  For example, it is well established that metabolic activation of
 24      1,3-butadiene is probably necessary for its carcinogenic action (Chapter 4). Therefore, a PBPK
 25      model describing the production and disposition of l,2-epoxy-3-butene (EB), the first product of
 26      metabolic activation of 1,3-butadiene, may be able to provide a better dose metric than the
 27      default methodology of using exposure to 1,3-butadiene itself.
 28            This chapter reviews and analyzes the six PBPK models for 1,3-butadiene that are
29      currently available and assesses their usefulness for quantitative risk assessment of 1,3-butadiene
30     based on interspecies extrapolation. Each of these PBPK models assumes, for simplicity, that the
31      transfer of 1,3-butadiene to tissues is blood flow-limited, that each tissue compartment is "well
3 2     mixed," and that tissue concentrations are hi equilibrium with the venous blood concentration
        leaving the tissue.

         1/28/98                                   8-1        DRAFT-DO NOT CITE OR QUOTE
-------
  1      8.2.  PBPK MODELS FOR 1,3-BUTADIENE
  2      8.2.1. Hattis and Wasson (1987)
  3             The first PBPK model for 1,3-butadiene was that of Hattis and Wasson (1987).  They
  4      defined the effective dose of 1,3-butadiene as the amount that is metabolically converted to EB
  5      and used this dose as a basis for a risk assessment of occupational 1,3-butadiene exposure. Their
  6      model consists of three compartments: a fat compartment; a muscle compartment; and a liver
  7      and vessel-rich compartment, which includes the brain, heart, kidneys, and other small visceral
  8      organs.  The transfer of 1,3-butadiene between blood and tissues is assumed to be blood flow-
  9      limited.  Metabolism to the monoepoxide is ascribed to the entire liver and vessel-rich
10      compartment and is assumed to follow simple Michaelis-Menten kinetics. No further
11      metabolism of EB is considered.
12             The only chemical-specific parameter values then available were whole-body maximal
13      metabolic rates for mice and rats inferred from the chamber study data of Kreiling et al.  (1986b).
14      These data provided the KM and preliminary Vmax estimates for the liver and vessel-rich
1 5      compartment. Tissuerblood and blood:air partition coefficients were estimated from chemical
1 6      structure and solubility data using empirical relationships (e.g., Fiserova-Bergerova and Diaz,
17      1986).  Model simulations were then run, adjusting KM and the partition coefficients to fit the
18      blood 1,3-butadiene concentration data of Bond et al. (1986), to derive "best estimates" for these
19      parameters. Human metabolic rates were estimated by allometric scaling of the mouse and rat
20      rates because no PBPK data were available for human metabolism of 1,3-butadiene. The
21      parameter values used by Hattis and Wasson (1987) are summarized in Table 8-1.
22             No additional data were available at that time for an independent validation of this model.
23      A minimal sensitivity analysis was conducted by varying KM and the blood:air partition
24      coefficient among a few values and observing the effect on the ultimate risk estimates. Hattis
25      and Wasson (1987) claimed that their model is not very sensitive to reasonable differences in
26      partition coefficients. Similarly, the model is insensitive to the precise value of the metabolic
27      parameters because, given the blood:air partition coefficient values that were used, metabolic
28      conversion in their model is limited by blood flow to the liver and vessel-rich compartment.
29      Hattis and Wasson concluded that differences in pharmacokinetics fail to account for differences
30      in carcinogenesis between mice and rats and that, with respect to risk assessment, uncertainties in
31      the PBPK modeling are trivial compared with the differences in apparent sensitivities between
32      these species.
        1/28/98                                   8-2        DRAFT-DO NOT CITE OR QUOTE
-------
        Table 8-1. Parameter values used in the Hattis and Wasson (1987) PBPK model
Parameter
Alveolar ventilation
(L/min)
Weight (kg)
Qf (L/min)
Qm (L/min)
Qlvr (L/min)
Vf(L)
Vm (L)
V,,,(L)
Blood'air partition
coefficient0

Pf

Pm

Plvr
Vmax (mol/min)
KM (mol/L)
Rat
0.15
0.40
0.0136
0.0226
0.1042
0.028
0.300
0.036








1.47E-6d

Mouse
0.0233
0.028
0.00192
0.00319
0.01617
0.0028
0.0196
0.00308
03^

	 11S9

	 5 7/c

	 « A

1.87E-7d
	 5E-6f - 	 	
TTutndn
11.38a
4.8
70
0.69a
0.35b
2.61a
l.lb
5.09a
4.35b
14.024
34.756
8.513








8.0E-56

aAwake.
"Asleep.
The blood:air partition coefficient of 0.35 is the "best estimate" value from "fitting" the model. The
 tissue:blood partition coefficients (P) are from functions of the bloodrair partition coefficient for which the
 "best estimate" value of 0.35 was used.  Partition coefficients are assumed to be the same across species.
dFrom Kreiling et al. (1986b).
eFrom allometric scaling of the rodent values.
f"Best estimate" from "fitting" the model.

Subscripts f, m, and Ivr designate the fat, muscle, and liver and vessel-rich compartments (tissues), respectively.
Q: tissue blood flow rate.
V: tissue volume.                             .                                  •   .    •.
P: tissue:blood partition coefficient
1/28/98
8-3
DRAFT-DO NOT CITE OR QUOTE
-------
 1             The Hattis and Wasson (1987) model is not discussed further here because it has been
 2      superseded by new data and other modeling efforts.

 3      8.2.2. Hallenbeck (1992)
 4             Hallenbeck (1992) reported having done a PBPK-based cancer risk assessment for
 5      1,3-butadiene; however, he provided no details of the PBPK model that he used. Furthermore,
 6      he used the area under the 1,3-butadiene concentration-versus-time curve for the lung as his
 7      tissue-dose surrogate, taking no account of metabolic activation. As presented, this model
 8      contributes nothing to the current state of knowledge regarding the pharmacokinetic modeling of
 9      1,3-butadiene.

10      8.2.3. Kohn and Melnick (1993)
11             The PBPK model of Kohn and Melnick (1993) focuses on the disposition of EB in the
12      mouse, rat, and human.  This model incorporates additional tissues (compartments) and
1 3      metabolic reactions based on experimental data that were not available at the time of the Hattis
14      and Wasson (1987) model;  however, it also relies on theoretically derived partition coefficients.
1 5      The Kohn and Melnick model is blood flow-limited and consists of six compartments: lung,
1 6      blood, fat, liver, other rapidly perfused tissues (viscera), and slowly perfused tissues (muscle).
17      Metabolism occurs in the liver, lung, and viscera compartments.  The metabolic reactions include
1 8      conversion of 1,3-butadiene to EB, the conversion of EB to l,2:3,4-diepoxybutane (DEB), the
1 9      enzymatic hydrolysis of EB, and the enzymatic conjugation of EB with glutathione.
20             With the exception of the partition coefficients, which were derived in advance from
21      published methodologies, all of the mouse, rat, and human parameter estimates were from the
22      literature; none of them were adjusted to obtain a fit to experimental data. The parameter values
23      used by Kohn and Melnick  (1993) are summarized in Table 8-2.  Blood:tissue partition
24      coefficients for 1,3-butadiene were from Hattis and Wasson (1987). The blood:air partition
25      coefficients reported by Csanady et al. (1992) for 1,3-butadiene and EB were used as lung:air
26      partition  coefficients. The fatblood partition coefficient for EB was calculated using an
27      empirical relationship from Lyman et al. (1990), whereas the tissue:blood partition coefficients
28      of EB for the other tissues were derived using the method of Fiserova-Bergerova and Diaz
29      (1986). These are essentially the same procedures used by Hattis and Wasson (1987).
30             Michaelis-Menten kinetics were used to describe the oxidation of 1,3-butadiene and EB
31      by the cytochrome P-450 isozyme CYP2E1, the hydrolysis of EB by epoxide hydrolase, and the
32      glutathione S-transferase-catalyzed conjugation of EB with glutathione.  KM and Vmax values for
33      each of these reactions in the liver and lung of the mouse, rat, and human were taken from the in

        1/28/98                                   8-4       DRAFT-DO NOT CITE OR QUOTE
-------
      Table 8-2. Parameter values used in the Kohn and Melnick (1993)
      PBPK model
Parameter
Physiological parameters"
Body weight (kg)
Cardiac output (L/h)
Ventilation rate (L/h)
Fraction blood
Fraction fat
Fraction Ever
Fraction viscera
Fraction muscle
Fat flow fraction
Liver flow fraction
Viscera flow fraction
Muscle flow fraction
Partition coefficients0
Air partition BD
Fat partition BD
Liver partition BD
Viscera partition BD
Muscle partition BD
Air partition EB
Fat partition EB
Liver partition EB
Viscera partition EB
Muscle partition EB
Mouse

0.028
1.044
2.64
0.05
0.04
0.062
0.05
0.78
0.05
0.16
0.52
0.19
Rat

0.4
7.32
15.6
0.054
0.08
0.05
0.083
0.59
0.07
0.16
0.40
0.36
Human

70
660b
l,200b
0.077
0.144
0.025
0.037
0.547
0.036
0.16
0.446
0.361

1.5
118.2
5.49
5.34
5.26
60
1.8083
0.6545
0.6348
0.6533
1/28/98
8-5
DRAFT-DO NOT CITE OR QUOTE
-------
        Table 8-2. Parameter values used in the Kohn and Melnick (1993)
        PBPK model (continued)
Parameter
Biochemical parameters'1
Liver V cytl (nmol/h/mg)
Liver Km cytl (mM)
Liver V cyt2 (nmol/h/mg)
Liver Km cyt2 (mM)
Liver V EH (nmol/h/mg)
Liver Km EH (mM)
Liver V GST (nmolMng)
Liver Km GST (mM)
Liver micro prot (mg/L)
Liver cyto prot (mg/L)
Lung V cytl (nmol/h/mg)
Lung Km cytl (mM)
Lung k hydr (h'Vmg)
Lung V GST (nmol/h/mg)
Lung Km GST (mM)
Lung k GST (hrVrng)
Lung micro prot (mg/L)
Lung cyto prot (mg/L)
Mouse

155.4
0.002
12
0.0156
347.4
1.59
30,000
35.3
11,600
82,800
138.6
0.00501
0.1116
6,380
36.5

3,000
82,800
Rat

35.4
0.00375


148.8
0.26
14,460
13.8
16,800
108,000
9.6
0.00775
0.0792
2,652
17.4

3,000
108,000
Human

70.8
0.00514


1,110
0.58
2,706
10.4
14,500
58,000
9
0.002
0.1914


0.1536
3,000
58,000
"Compartment volumes are given as fractions of body weight; compartment blood flow rates are given as
 fractions of cardiac output.
bHuman cardiac output at rest: 336 L/h; human ventilation rate at rest: 240 L/h.
lAingrair and tissue:blood; assumed same for all species.
dData from Csanady et al. (1992).

BD: l,3-butadiene;EB: l,2-epoxy-3-butene.
V: Vmnx;Km: KM.
cytl denotes oxidative metabolism of butadiene to EB; cyt2 denotes oxidative metabolism of EB.
EH: epoxide hydrolase.
GST:  glutathione S-transferase.
micro prot:  microsomal protein; cyto prot:  cytoplasmic protein.
k hydr: apparent first-order rate constant for EB hydrolysis; k gst: apparent first-order rate constant
for glutathione conjugation.
1/28/98
8-6
DRAFT-DO NOT CITE OR QUOTE
-------
        vitro data of Csanady et al. (1992). The lung values were also assumed to apply to the viscera
        compartment. Csanady et al. detected DEB formation only in mouse liver preparations.
        Therefore, Kohn and Melnick (1993) included this reaction only in the mouse liver compartment
  4     and only as a disappearance route for EB; the distribution of DEB was not further modeled. 1,3-
  5     butadiene and EB were treated as competitive inhibitors of each other in the rate equations for
  6     mouse liver CYP2E1. Finally, although glutathione was treated as saturating for glutathione S-
  7     transferase in the mouse, rat, and human liver, glutathione conjugation with EB in human lung
  8     and viscera was assumed to be first order.
  9            To validate their model, Kohn and Melnick (1993) compared predicted 1,3-butadiene
 1 0     absorption and blood concentrations for mice and rats with the measurements of Bond et al.
 11     (1986). They also modified the model to include a chamber compartment and compared
 1 2     predicted EB concentrations in the chamber and maximum metabolic elimination rates with the
 1 3     Laib et al. (1990) results for mice and rats.  Kohn and Melnick claimed that their model
 14     predictions are comparable to the experimental results except for overestimates in the blood 1,3-
 1 5     butadiene concentrations, which they ascribed to inadequacies in the model or experimental
 1 6     sources of error in the blood concentration measurements.
 17            To assess the sensitivity of the model to the values of various parameters,  relative
J 8     sensitivity coefficients for different model variables were estimated by finite differences,  as
        given by Frank (1978). The physiological parameters to which the model was the most sensitive
 20     were the lung:air partition coefficient and the cardiac output. Because the ventilation rate is
 21     greater than the rate of 1,3-butadiene absorption, the lungiair partition coefficient and the cardiac
 22     output are the major parameters governing  1,3-butadiene uptake.  Predicted 1,3-butadiene
 23     concentrations were not very sensitive to variations in the biochemical parameters; however,
 24     monoepoxide levels were somewhat more sensitive to the parameters describing hepatic
 25     glutathione S-transferase and epoxide hydrolase kinetics.
 26            Based on their model simulations, Kohn and Melnick (1993) reported that 1,3-butadiene
 27     uptake and the disposition of EB are controlled to a greater extent by physiological parameters
 28     than by biochemical parameters.  The model further suggests that storage in fat is a significant
 29     fraction of retained  1,3-butadiene, especially in rats and humans. Kohn and Melnick also found
 30     that predicted EB tissue concentrations do not  correlate with tumor incidences in mice and rats,
 31     and they concluded  that other factors are crucial in  1,3-butadiene-induced carcinogenesis:  These
 32     other factors may include pharmacokinetic variables that were not part of the model, such as
 3 3     accumulation of the diepoxide or formation of other metabolites or mechanistic
 34     (pharmacodynamic) phenomena,  such as formation of DNA adducts or efficiency of DNA repair.
         1/28/98                                    8-7       DRAFT-DO NOT CITE OR QUOTE
-------
  1             The Kohn and Melnick (1993) model appears to have a reasonable basic structure, in
  2      terms of the compartments and metabolic reactions included, given the biochemical parameters
  3      that are currently available. A major strength of their model is that none of the parameter
  4      estimates is adjusted to fit experimental data. Two important drawbacks of the model are the use
  5      of empirically derived partition coefficients and the lumping of various tissues with different
  6      metabolic capabilities (Chapter 3) into a viscera compartment, which is assumed to have the
  7      same metabolic activity as the lung. Partition coefficients for 1,3-butadiene and EB have
  8      recently been measured by Johanson and Filser (1993) and Medinsky et al. (1994), and
  9      experimental values for the 1,3-butadiene partition coefficients are substantially less than the
 10      empirically derived estimates, which suggests that the specific results reported by Kohn and
 11      Melnick may not be relevant.  For example, the role of physiological parameters in controlling
 12      1,3-butadiene  uptake and the amount  of 1,3-butadiene storage in fat may not, in fact, be as great
 13      as the Kohn and Melnick model predicts (Medinsky et al., 1994).

 14      8.2.4. Johanson and Filser (1993)
 1 5            Johanson and Filser (1993) developed a PBPK model for 1,3-butadiene and EB
 1 6      disposition in rats and mice. Their model is blood flow-limited and consists of four main
 17      physiological compartments—lungs and arterial blood, muscle and vessel-rich tissues, fat, and
 1 8      liver—as well  as a chamber compartment and an intrahepatic subcompartment.  Metabolism is
 19      assumed to take place exclusively in the liver. The metabolic reactions include oxidation of 1,3-
 20      butadiene to EB; hydrolysis of EB; intrahepatic first-pass hydrolysis of EB; conjugation of EB
 21      with glutathione, which is described by a "ping-pong" mechanism; and the turnover and
 22      depletion of hepatic glutathione.
 23            In contrast with the previous PBPK modeling efforts for 1,3-butadiene, Johanson and
 24      Filser (1993) conducted in vitro studies of rat homogenates to obtain empirical values for the
 25      tissuerair partition coefficients for 1,3-butadiene and EB.  All physiological parameters were
 26      taken from Arms and Travis (1988), except the alveolar ventilation rates, which were reduced to
 27      60% of those suggested by Arms and Travis on the basis of generalized observations of uptake
 28      rates of various gases in closed-chamber experiments (Johanson and Filser, 1992).  For the
 29      oxidative metabolism of 1,3-butadiene, the model uses the Vmax values from the in vitro studies
 30      of Filser et al.  (1992). A KM value was derived by fitting the model to the in vivo data of Lieser
 31       (1983) for the  rat and Kreiling (1986b) for the mouse because the model could not reproduce the
 32      results observed in these closed-chamber studies the KM values of either Filser et al. (1992) or
33      Csanady et al.  (1992). Values for the  metabolic parameters pertaining to the conjugation of EB
34     with glutathione and to the hydrolysis  of EB were taken from the in vitro data of Kreuzer et al.

         1/28/98                                    8-8       DRAFT-DO NOT CITE OR QUOTE
-------
         (1991).  The value of the "intrinsic KM" for the intrahepatic hydrolysis of EB (see below) was set
         to 20% of the "apparent KM" value of Kreuzer et al. because the model then fit various in vivo
         data. The flow rate between the hepatic and intrahepatic compartments was estimated from the
  4      kinetic parameters.  The physiological and biochemical parameter values used by Johanson and
  5      Filser (1993) are summarized in Table 8-3.
  6            In terms of the metabolic reactions involved, the Johanson and Filser (1993) model
  7      differs from the Kohn and Melnick (1993) model in that further oxidation of EB to DEB is not
  8      included, conjugation of EB with glutathione is described by the two-substrate ordered sequential
  9      ping-pong mechanism (reviewed by Mannervik, 1985) rather than by Michaelis-Menten kinetics,
 1 0      and glutathione turnover and the intrahepatic first-pass hydrolysis of EB are incorporated.  Given
 11      the KM values for glutathione conjugation used in the model, the conjugation of EB becomes
 1 2      rate-limited by glutathione only when glutathione is almost completely depleted.  Cytosolic
 1 3      glutathione turnover is depicted by zero-order production and first-order elimination.
 14      Intrahepatic first-pass hydrolysis of EB is hypothesized to occur, based on the observations of
 1 5      Filser and Bolt (1984), because of proximity of the monooxygenase to the epoxide hydrolase in
 1 6      the endoplasmic reticulum. Newly formed EB within this intrahepatic compartment will be more
 1 7      readily hydrolyzed than EB that must diffuse in from outside the compartment, as reflected by a
J 8      lower KM in the intrahepatic compartment.
               To attempt to validate the model,  Johanson and Filser (1993) compared simulated results
 20      with the data from various in vivo experiments.  In addition to the 1,3-butadiene kinetics data
 21      used to fit the KM for 1,3 -butadiene oxidation and the  EB kinetics data of Filser and Bolt (1984)
 22      for the rat and Kreiling (1987) for the mouse that were used to fit the intrinsic KM for intrahepatic
 23      first-pass hydrolysis, the model apparently reproduces the EB concentrations appearing in
 24      chamber air as a result of 1,3-butadiene exposure in the experiments of Rolzhauser (1985) for the
 2 5      rat and Kreiling (1987) for the mouse.  However, it is  not clear from the text whether these
 26      experimental data were also used to fit the intrinsic KM.  The model also reproduces the
 27      glutathione concentrations observed by Deutschmann  (1988) in rat and mouse liver after 1,3-
 28      butadiene exposure, and Johanson and Filser claimed that no model parameters were fitted to
 29      these data. Finally, simulated blood concentrations of EB approximate those observed by Bond
 30      et al.  (1986) in the mouse but are slightly higher than those observed in the rat.
 31            No sensitivity analysis for the model parameters was reported.
 32            The results of Johanson and Filser's (1993) model simulations suggest that the internal
 33      dose of EB, expressed as the concentration of EB or the area under the concentration-time curve
         1/28/98                                    8-9        DRAFT-DO NOT CITE OR QUOTE
-------

Is
pt






W
o
s












o
1
ss
CS
P-i



5->
0 S ^
in S1 o
cs m f-
m
r— ^


s
JD
0
t
*A IO 2
r\] p^ .__( C
cs o
'•S
5

p
s.








"« "«
G G
"Is la
•S | -H |
(3 S CB «
T3 "3 "O "3
1 1 1 1
w c)o 55 55



1
»d
Physiological data
Body weight (g)
Alveolar ventilation (rr

CO VO /-N IO
oo \o ^ cs




a
,5
o
^_^
f^. OI \Q ^^ ly-j
<— t Q VO CN
o
a.
5^
0*








1 |
'§ 1 "S
11 J
11 1^1
co co ,2; pr, 1-4




IS"
Cardiac output (mL/mi
Blood flows
(% of cardiac output)

~°c^








,_, ID O V~>
tn












3 o
«3 S
S-a
^ c3
Sji
g1 8 S
§ a « .5
_J S PL, nJ





Compartment volumes
(% of body weight)




>o en en
0^ 1 CM O











o"
>
"§
CB
73
OT
1
Lung and arterial,
Fat
Blood





Partition coefficients'"
1,3 -Butadiene




^0
^
o









t-4
g
T3
TO

73

G
1
bp
3





a
I
o
•» — i




^^
* en


















«!
2 S






1/28/98
8-10
DRAFT-DO NOT CITE OR QUOTE
-------
I

 Si
•h^

 O
•o
 o
T3
 a
 «

 e
 o
 en

 «


I
•O
 tu
 en




 1

oo

—
S
OS
H
•8

O
BB
o
s


i,
1
«s

















B5
5
Metabolic const
r*-
o oi >n
en


CNI
o en m
en

?
iv
.O.£T-
p(.HB T/jouirl) wa
jSui.j.urai.jorau) """^
) urajojd ^Buiosojoijrtj

a
o
CO
•a
•52
1,3 -Butadiene o

0 f-
en i— i



O ON
en i— i

I
•i.
.S ^TN
Microsomal protein (
Vmax (nmol-min-'-mg-





EB hydrolysis

0 ^


10 ^
*~t f-vj

5
1
X" "N ^
uJ cu
Apparent KM (mmol/
Intrinsic KM (% of ap







irt i — i o O
ON .— 1 O
l-H


^H
vn to o o
ON T— i o

w
ff^l
Cytosolic protein (m§
Vmax/KMofEB(nL-m
KM toward EB (mmo
KM toward glutathion





EB conjugation
Sb



en
oo
"o
ncentration (m
Initial steady-state co



X
•&
Glutathione kine

*^d" ^^


•&1 r-J
in o

i!,
•a
CO
Elimination rate cons






                                                    1 1
                                                                ON
                                                                OO
                                                                ON
                    r-l 03
                          OO

                          ON
||tr
-3
                                                      43

                                                      

'•S I"
i> O
'§g &
  i
                                                                      > w
                                   s
                                  t
 1/28/98
8-11
DRAFT-DO NOT CITE OR QUOTE
-------
 1      in the venous blood, the other compartments, or the whole body, is at most about three times
 2      greater in the mouse than in the rat for a given exposure concentration. The greatest differences
 3      in internal dose of EB between the two species result from 1,3-butadiene exposure concentrations
 4      of above 1,000 ppm, when glutathione depletion occurs in the mouse but not in the rat after 6 to
 5      9 h of exposure.  Once again, the relatively small interspecies differences in body burden of EB
 6      indicated by PBPK modeling cannot explain the striking differences in cancer response between
 7      mice and rats exposed to 1,3-butadiene. Johanson and Filser suggested that differences in the
 8      kinetics of DEB or nonmetabolic factors, such as differences in immune response or in the
 9      expression of oncogenes, may be responsible for the interspecies differences in cancer response.
10      A major advancement found in the PBPK model of Johanson and Filser (1993) is the use of
11      experimentally derived partition coefficients, especially because these values differ substantially
12      from the theoretically estimated values. A further strength of their analysis is that they compared
13      the simulation results with data from several different experiments. The Johanson and Filser
14      model also incorporates hepatic glutathione turnover and depletion as well as intrahepatic first-
15      pass hydrolysis of EB, although the significance of these refinements is unknown. Some of the
1 6      limitations of the model include the exclusion of extrahepatic metabolism and of further
17      metabolism of EB to DEB. In addition, the values of the KM for 1,3-butadiene oxidation and of
18      the intrinsic KM for intrahepatic first-pass hydrolysis of EB were obtained by fitting in vivo data.
19      Finally, no sensitivity analysis was reported, although, for example, it was acknowledged that
20      wide ranges of glutathione concentrations and turnover rates have been observed.  Therefore, it is
21      unknown how sensitive the model is to changes in these and other parameters. Johanson and
22      Filser are reportedly working on a corresponding PBPK model for humans, but it has not yet
23      been published.

24      8.2.5.  Evelo et al. (1993)
25             Evelo et al. (1993) present a PBPK model for the uptake, distribution, and metabolic
26      clearance of 1,3-butadiene in mice and rats. Their stated objective was to investigate the relative
27      importance of liver and lung metabolism at different 1,3-butadiene exposure concentrations. The
28      Evelo et al. model has six physiological compartments: liver, fat, muscle, a vessel-rich group,
29      the bronchial area of the lung, and the alveolar area of the lung. A chamber compartment is also
30      included for validation against the data from closed-chamber experiments.  1,3-Butadiene
31      metabolism is assigned to both  the alveolar and bronchial areas of the lung and to the liver. Gas
32      exchange occurs in the alveolar area of the lung.
33             Values for the standard physiological parameters were allometrically scaled from the data
34     of Travis (1988). Volumes and blood flows for the two separate lung compartments were taken

        1/28/98                                   8-12       DRAFT-DO NOT CITE OR QUOTE
-------
        from Greep and Weis (1977). Tissue:blood and blood:air partition coefficients were theoretically
        estimated using the regression analysis method of Fiserova-Bergerova and Diaz (1986), as was
        done previously by Hattis and Wasson (1987).
  4            To describe the oxidation of 1,3-butadiene to EB, Evelo et al. (1993) calculated the ratios
  5     of the maximum metabolic activity between the liver and the lung from the in vitro data of
  6     Schmidt and Loesser (1985) for the mouse and for the rat.  Then, the total (whole-body)
  7     maximum metabolic activities, the K^s, and "the most probable distribution" of metabolic
  8     activity between the alveolar and bronchial areas of the lung were derived by optimizing the
  9     model against the closed-chamber data of Kreiling et al. (1986b) for the mouse and Bolt et al.
10     (1984) for the rat.  The only options considered for the distribution of the metabolic activity of
11      the lung were that all the metabolism took place in either one of the two areas, that it was equal
1 2     in each area, or that it was distributed relative to the volumes of each area; the best fit was found
1 3     using the latter distribution.  The values of the physiological and metabolic parameters used in
14     the Evelo et al. model are summarized in Table 8-4.
1 5            The only independent validation of the model was against the whole-body extraction
1 6     ratios reported by Dahl et al. (1990).  Evelo et al. (1993) calculated extraction ratios of 8.4% for
1 7     the mouse and 5.2% for the rat, whereas Dahl et al. found ratios of 12.8% for the mouse and
1 8     4.3% for the rat. Evelo et al. also noted that the whole-body Vmax value obtained for the rat by
        fitting the model to the data of Bolt et al. (1984) does not fall within the range of values allowed
20     by experimental error based on the gas-uptake studies of Laib et al, (1992).
21             Evelo et al. (1993) stated that sensitivity analyses found the model optimization to be
22     relatively insensitive to variability in the value of KM. No other sensitivity analysis results are
23     reported.
24            The model simulations of Evelo et al. (1993)  suggest that the relative importance of 1,3-
2 5     butadiene metabolism in the mouse lung is greater than the distribution of metabolic activity
26     would imply,  especially at exposure concentrations of less than 200 ppm and for KM values of
27     less than the "best fit" value. Evelo et al. concluded that there is a strong first-pass effect in the
28     mouse lung. At higher concentrations, alveolar metabolism is saturated, and liver metabolism
29     becomes relatively more important.  The relative importance of lung metabolism also increases
30     with decreasing exposure concentration for the rat and human, especially with lower values of
31      KM; however, unlike for the mouse, the lung metabolism never exceeds the liver metabolism.
32     Evelo et al. suggested that the higher rate of metabolic activation in the mouse lung could be
33     responsible for the mouse's greater sensitivity to developing lung carcinomas and heart
34     hemangiosarcomas from exposure to 1,3-butadiene.
        1/28/98                                   8-13       DRAFT-DO NOT CITE OR QUOTE
-------
       Table 8-4. Parameter values used in the Evelo et al. (1993) PBPK model
Parameter
Physiological parameters
Body mass (kg)
Cardiac output (mL/min)
Alveolar ventilation (mL/min)
Blood flows (mL/min):
Liver
Fat
Muscle
Vessel-rich tissue
Bronchial lung area
Alveolar lung area
Volumes (mL):
Liver
Fat
Muscle
Vessel-rich tissue
Bronchial lung area
Alveolar lung area
Partition coefficients'
Blood:air
Fat:blood
Livenblood
Muscle:blood
Kidney:blood
Lung:blood
Brain:blood
Vessel rich:bloodb
Metabolic parameters
V.nst.toai (umol-hr'-kg'1)
Vn-iivt, (umol-hr'-kg'1)
V^teo^M (nmol-hr-'-kg-1)
KsToiM)
Mice

0.0275
24.83
24.5

6.14
2.34
3.81
10.75
1.79
23.04

1.65
2.94
19.09
1.17
0.2
0.18
Rats

0.215
75.93
118.7

19.17
6.52
11.13
33.60
5.514
70.42

8.63
14.0
162.7
9.49
1.29
1.63

0.894
32.362
2.675
1.871
1.690
1.272
2.355
2.02

465
318
77
70
8

200
171
13
16
5
"Same for all species.
bMean value of kidneyrblood and brain:blood.
1/28/98
8-14
DRAFT-DO NOT CITE OR QUOTE
-------
  1             The Evelo et al. (1993) model suffers from a number of serious weaknesses. Several
  12      important parameters are not empirically derived. The partition coefficients are estimated
  3      theoretically, and the whole-body Vmax and KM are optimized. For the rat, this exercise generated
  4      a Ymax value that was inconsistent with other in vivo data.  Furthermore, sensitivity analyses
  5      revealed that the optimization was insensitive to variability in the value of KM, so there is
  6      considerable uncertainty in the actual value of this parameter. The results pertaining to the
  7      relative importance of lung metabolism, however, are highly sensitive to the value of K^.  The
  8      separation of the lung into alveolar and bronchial  areas and the "optimized" distribution of lung
  9      metabolism between the two areas also appear tenuous.  Other limitations of the model are that
 1 0      metabolism is limited to the lung and the liver and that further metabolism of EB is not
 1 1      incorporated.  In addition, the model was not adequately validated, and only limited sensitivity
 1 2      analyses are described.  Finally, results for humans are discussed; however, the parameters used
 1 3      for the human model are not fully reported.

 14      8.2.6. Medinsky et al. (1994)
 1 5             The most recent PBPK model published for butadiene is the model of Medinsky et al.
 1 6      (1994) for 1,3-butadiene and EB uptake and metabolism in mice and rats.  The Medinsky et al.
J 7      model is a venous equilibration, flow-limited model with six physiological compartments—liver,
 F8      lung, fat,  slowly perfused tissue group, rapidly perfused tissue group, and blood—and a •
 1 9      compartment representing the air in closed-chamber experiments.  The model describes the
 20      oxidative metabolism of 1,3-butadiene in the liver and lung, as well as hydrolysis and glutathione
 21      conjugation of EB in the liver. In the mouse, hepatic oxidation of EB is also included.  In
 22      addition to measuring actual partition coefficients, Medinsky et al. conducted closed-chamber
 23      experiments of 1,3-butadiene uptake with both mice and rats to test the predictions of their
 24      model.
 25             Medinsky et al. (1994) measured partition coefficients for 1,3-butadiene and EB
 26      experimentally in vitro for both mouse and rat tissues.  They found no significant differences
 27      between the two species, except for the muscle:air partition coefficient for 1,3-butadiene and the
 28      fatair coefficient for EB (although the ultimate fatblood coefficient was not significantly
 29      different). Organ and body weights were taken from specific experiments on 1,3-butadiene. The
 30      remaining physiological parameters were based on average literature  values, with the exception
 31      of alveolar ventilation rate.  Alveolar ventilation rates,  conventionally defined as 70% of
 32      measured total ventilation rates, yielded overestimates of 1,3-butadiene uptake at low
 33      concentrations, consistent with observations by Johanson and Filser (1992) for other volatile
 |4      organic chemicals. Therefore, "apparent" alveolar ventilation rates were obtained  by

         1/28/98                                   8-15        DRAFT-DO NOT CITE OR QUOTE
-------
  1      optimization to provide rates that yielded the best fit of the model to the EB uptake data. The
  2     optimized rates represented 63% of alveolar ventilation for both rats and mice.
  3            Oxidation of 1,3 -butadiene and EB (the latter in mouse liver only) and hydrolysis of EB
  4     were described using Michaelis-Menten kinetics. Glutathione conjugation of EB was assumed to
  5     be first order, based on the large KM value reported by Csanady et al. (1992).
  6            Rate constants for the metabolism of 1,3-butadiene and EB were taken from the in vitro
  7     data of Csanady et al. (1992).  Apparent enzyme affinities (Kjy,) measured in vitro were used
  8     directly, whereas maximum metabolic rates (V^ were scaled to the whole organs. However,
  9     when the organ microsomal concentrations reported by Csanady et al. are used to scale the
10     metabolic rates similarly reported by Csanady et al., "[1,3-butadiene] uptake from the closed
11      chamber is underestimated." Therefore,  Medinsky et al. (1994) used literature values that were
12     two to six times greater for microsomal concentrations in the liver and lung in order to
13     successfully simulate the chamber study results. The parameter values used in the Medinsky et
14     al. model are summarized in Table 8-5.
1 5            For validation of the model components pertaining to EB uptake and metabolism, model
1 6     predictions were compared with the EB uptake data from the closed-chamber experiments of
17     Filser and Bolt (1984) for rats and Kreiling et al. (1987) for mice, although these were the same
1 8     data used to optimize the alveolar ventilation rates.  The model predictions were deemed
19     "adequate," although EB uptake was overestimated at the highest exposure concentration,
20     especially for the rats (3,000 ppm). Medinsky et al. (1994) then compared model simulations of
21      1,3-butadiene uptake to their own closed-chamber data for mice and rats exposed to 1,3-
22     butadiene and to data from the closed-chamber experiments of Bolt et al. (1984) for rats and
23     Kreiling et al. (1986b) for mice and concluded that the model adequately predicted the in vivo
24     uptake results. Medinsky et al. also compared model predictions with the 1,3-butadiene retention
25     data of Bond et al. (1986) and found the results similar for exposure concentrations up to about
26     100 ppm. At higher concentrations, the model overestimated butadiene retention observed in
27     mice.  Furthermore, the blood concentrations ofEB following 1,3-butadiene exposure,  as
28     reported by Bond et al. were overestimated by the model for both mice (except at the lowest
29     exposure) and rats by about two- to fourfold, although Medinsky et al. suggested that the
30     discrepancy might be attributable to EB loss from the blood during sampling.
31             No comprehensive  sensitivity analysis for the model parameters was reported. Medinsky
32     et al. (1994) did note that use of the microsomal concentrations reported by Csanady et al. (1992)
33     resulted in underestimation of the 1,3-butadiene uptake from chamber studies. In addition, they
34     investigated whether the model was sensitive to the different values obtained for the muscle: air
        1/28/98                                   8-16       DRAFT-DO NOT CITE OR QUOTE
-------
      Table 8-5. Parameter values used in the Medinsky et al. (1993) PBPK model
Parameter
Physiological parameters:
Alveolar ventilation (L/hr/kg)a
Cardiac output (L/hr/kg)b
Body weight (kg)0
Blood flows (fraction of cardiac output):
Liver
Fat
Lung
Slowly perfused tissues
Rapidly perfused tissues
Organ volumes (fraction of body weight):
Liver
Fat
Lung
Slowly perfused tissues
Rapidly perfused tissues
Partition coefficients for 1,3-butadiene:
Blood: air
Livenblood
Lung:blood
Muscle:blood
Fatblood
Partition coefficients for EB:
Blood: air
Livenblood
Lung:blood
Musclerblood
Fatblood
Tissue concentrations
Liver microsomal concentration (mg/g liver)
Lung microsomal concentration (mg/g lung)
Liver cytosolic concentration (mg/g liver)d
Rat

17
• 17
0.215-0.475

0.25
0.09
1.0
0.15
0.51

0.05
0.09
0.0053
0.71
0.0347

1.49
0.799
0.617
0.987
14.9

50.4
1.43
1.09
0.393
2.74'

35
20
108
Mouse

41
41
0.028-0.035

0.25
0.09
1.0
0.15
0.51

0.0624
0.10
0.005
0.70
0.0226

1:34'
1.01
1.10
2.99
14.3

36.6
1.15
1.54
0.645
2.49

35
20
82.8
1/28/98
8-17
DRAFT-DO NOT CITE OR QUOTE
-------
       Table 8-5.  Parameter values used in the Medinsky et al. (1993) PBPK model
       (continued)
Parameter
Rate constants for oxidative metabolism of
l,3-butadiened
Liver V^. (nmol/kg/hr)
KM (umol/L)
Lung V^ (|imol/kg/hr)
KM (umol/L)

Rate constants for EB metabolism in the Iiverd
Oxidation Vmax (umol/kg/br)
KM ((omol/L)
Hydrolysis VmJX (umol/kg/hr)
KM (umol/L)
glutathione conjugation K (L/kg/hr)
Rat


62
3.75

1.01
7.75



260
260
5.66
Mouse


338
2.00

21.6
5.01

26
15.6
754
1590
4.36
'Obtained by optimization.
bVentilation/perfusion= 1.
"Depending on experiment simulated.
•"From Csanady et al. (1992), with Vmix values scaled to whole organ using above microsomal concentrations.

EB: l,2-epoxy-3-butene.
1/28/98
8-18
DRAFT-DO NOT CITE OR QUOTE
-------
  1      partition coefficients for the mouse and rat and determined that the species-specific coefficients
  2      provided the best fits to their 1,3-butadiene uptake results for the two species.  Medinsky et al.
 '3      also determined that the inclusion of lung metabolism improves the model fit for the mouse,
  4      especially at lower exposure concentrations, but has little affect for the rat.
  5             Based on their model simulations, Medinsky et al. (1994) suggested that lung metabolism
  6      may play an important role in 1,3-butadiene uptake and cartinogenesis.  Their model predicts
  7      locally generated concentrations of EB that are 15 times greater in the mouse lung than in the rat
  8      lung, for a 6-h exposure to 10 ppm. Medinsky et al. recommended that more research be done to
  9      characterize 1,3-butadiene metabolism and target cells in the mouse lung and to understand the
 1 0      pharmacokinetics of DEB in different species. They further claimed that "quantitation of the
 11      concentrations of [1,3-butadiene], [EB], and [DEB] in target and non-target tissues of rats and
 1.2      mice after exposure to [ 1,3 -butadiene] is essential for validation of existing models before these
 1 3      models can be applied to predict behavior in humans."
 14             One of the major strengths of the Medinsky et al. (1994) model is that they
 1 5      experimentally measured partition coefficients and confirmed the results of Johanson and Filser
 1 6      (1993), suggesting that the empirical values for the partition coefficients for 1,3-butadiene differ
 1 7      significantly from the theoretical values used in previous models. Medinsky et al. also
         conducted  closed-chamber experiments to obtain validation data for their model and investigated
         the role of lung metabolism in 1,3-butadiene uptake.  Some limitations of the model include the
         fact that metabolism was restricted to the liver and lung, although other tissues are known to
21       metabolize 1,3-butadiene as well (Chapter 3).  In addition, the alveolar ventilation rates were
22      determined by fitting experimental closed-chamber data, and there are uncertainties about the
23      actual values for organ microsomal contents.  Finally, only 1,3-butadiene oxidation was
24      described in the lung, although rate constants for further metabolism of EB are also available
25      from Csanadyetal. (1992).

26      8.3. SUMMARY
27            Pharmacokinetic modeling of 1,3-butadiene has not yet elucidated the reasons for the
28      interspecies differences in carcinogenic response between mice and rats.  It appears that either
29      the PBPK models are not sufficiently sophisticated to adequately model the relevant
30      pharmacokinetics (e.g., the models may need to incorporate the production and  disposition of
31       DEB) or a pharmacodynamic component(s) (e.g., DNA susceptibility or repair)  is required to
32      accurately correlate dose to response.
3 3            Furthermore,  uncertainties in the existing PBPK models and data make them unreliable
        for use in risk assessment.  Serious uncertainties exist pertaining to the model structures,

         1/28/98                                   8-19       DRAFT-DO NOT CITE OR QUOTE
-------
  1      parameter values, and validation.  For example, there are discrepancies among the models and
  2     data as to the importance of extrahepatic and extrapulmonary metabolism, competitive
  3     interaction between 1,3-butadiene and EB for oxidative metabolism, and glutathione depletion,
  4     and none of the models fully describe the kinetics for DEB.
  5            With respect to the parameter values, there are disagreements about the ventilation rate,
  6     which is a key parameter for determining 1,3-butadiene delivery, and about metabolic
  7     parameters.  For example, measurements of Vmax and KM for the oxidation of 1,3-butadiene to EB
  8     in mouse, rat, and human liver microsomes by Csanady et al. (1992) and by Duescher and Elfarra
  9     (1994) differ by up to 80-fold, and Seaton et al. (1995) measured reaction rates for the oxidation
10     of EB to DEB by rat and human liver microsomes that Csanady et al. were unable to detect
11      (Chapter 3). Use of the in vitro metabolic data of Csanady et al. (1992) in the 1,3-butadiene
12     PBPK models appears to result in an underprediction of total metabolism. Such
13     underprediction could result from (1) an inability of the in vitro data to reflect the in vivo
14     metabolic potency, (2) inaccuracies in the measurement of metabolic reaction rates or
1 5     microsomal protein content in the tissues, or (3) a deficiency hi the models such that they do not
1 6     fully characterize 1,3-butadiene metabolism (e.g., by not including metabolism in other tissues).
17     This is a critical issue for any PBPK-based  extrapolation of carcinogenic risk from rodents to
18     humans because there are no appropriate human in vivo PBPK data for 1,3-butadiene and thus
19     interspecies extrapolation must rely on in vitro data or allometric  scaling. There is also a paucity
20     of human in vitro data for extension of the PBPK models to humans. The few measurements that
21      have been made on a few metabolic parameters show a high amount of variability.
22            Another area of uncertainty is that of model validation.  The existing models have been
23     subjected to a very limited validation, mostly by comparison of simulation results with chamber
24     uptake data. Virtually all of the model reports claim that the existing models adequately fit the
25     validation data, despite important differences among the models.  In some cases, this is not
26     surprising because  some of the model parameters have been determined by optimization against
27     data similar to those being used for validation.  In other cases, it suggests that the chamber data
28     are relatively insensitive to various features of the models and might be of limited use for model
29     validation. For the PBPK models to be more reliable, they should be validated against tissue
30     concentration data for various metabolites in various tissues. More recently, these data have
31     become available (Chapter 3), although they must be interpreted with caution because it appears
32     that metabolites hi  some of the tissues are subject to further metabolism during the lag time
33     between the termination of exposure and the measurement of tissue concentrations.  The results
34     of simulations using the Medinsky et al. (1994) model suggest that the model does not conform
        1/28/98                                   8-20       DRAFT-DO NOT CITE OR QUOTE
-------
         adequately to the tissue concentration data. Any PBPK model for 1,3 -butadiene would require
         more rigorous validation before it could be considered reliable for use in risk assessment.

  3      8.4. CONCLUSIONS
  4            As discussed above, the existing PBPK models and data cannot explain the interspecies
  5      differences in 1,3-butadiene carcinogenicity.  Uncertainties in the model structures and parameter
  6      values also prohibit their use in refining risk assessment dosimetry at this time. Some areas in
  7      which more research is needed include (1) evaluation of the kinetics of DEB in rodents as well as
  8      in humans, (2) investigation of the validity of the in vitro metabolic data for extrapolating to in
  9      vivo exposure, (3) clarification of the values of various physiological parameters such as the
10      ventilation rate, (4) better characterization of the distribution of values for the human metabolic
11       rates, and (5) more measurement of tissue concentrations of metabolites for model validation. It
1 2      is possible that more information on the specific mechanisms of action is required to explain
1 3      interspecies differences in the various target tissues.
14            In any event, the existing PBPK models and data are inadequate for developing a reliable
1 5      alternative to the default methodology of using exposure to the parent compound as a dose
1 6      surrogate for extrapolation of the carcinogenic risk from animals to humans.  Any attempt to
17      extrapolate the risk in rodents to humans,  given the dramatic and unresolved interspecies
         differences between the mouse and rat, would involve far greater uncertainties than basing  a risk
1 9      assessment on the occupational data of Delzell et al. (Chapter 7). Ideally, a reliable, well-
20      validated PBPK model with parameter values for humans could also be applied to analyzing
21       different human exposure scenarios (e.g., extrapolating from occupational to environmental
22      exposures).  However, there are too many uncertainties in the PBPK modeling for that to be
23      practicable at this time.
        1/28/98                                   8-21        DRAFT-DO NOT CITE OR QUOTE
-------
-------
                   9. QUANTITATIVE RISK ASSESSMENT FOR 1,3-BUTADIENE

        9.1. EPIDEMIOLOGICALLY BASED CANCER RISK ASSESSMENT
 2      9.1.1. Exposure-Response Modeling
 3            In general, it is preferable to use high-quality epidemiologic data when they are available
 4      over toxicologic data for quantitative risk assessment purposes. In the past, available
 5      epidemiologic data on 1,3-butadiene have been inadequate for quantitative risk assessment, and
 6      previous risk assessments relied primarily on models based on the NTP mouse bioassay studies
 7      (reviewed in Chapter 1).
 8            The recently reported findings by Delzell et al. (1995) from a retrospective cohort
 9      mortality study of synthetic production workers exposed to 1,3-butadiene (reviewed in Chapter
10      7) present an opportunity to perform a quantitative risk assessment based on human data. The
1 1      investigators developed a job exposure matrix (JEM) for 1,3-butadiene, styrene, and benzene
1 2      based on industrial hygiene data, which contained estimates of the average daily exposure (in
1 3      ppm based on the 8-h TWA) and the number of annual peaks (defined as > 100 ppm for 1,3-
14      butadiene and 50 ppm for styrene) for each area and job code for each study year.  The
1 5      investigators were then able to estimate cumulative exposures (ppm*years and peak*years) by
1 6      linking the JEM with the study subject's work histories.
              Delzell et al. (1995) investigated the relationship between cumulative exposure to 1,3-
1 8      butadiene and leukemia mortality using Poisson regression analysis (Frome and Checkoway,
1 9      1 985). The models controlled for the potentially confounding effects of age (40-49,  50-59, 60-
20      69, 70-79, 80+), years since hire (10-19, 20-29, 30+),. calendar period (1950-59, 1960-69, 1970-
21      79, 1980-89, 1990-91), and race (black, other). Plant was considered as a possible confounder
22      but was dropped from the final models because it did not affect the estimated parameters for 1 ,3-
23      butadiene or styrene. Few subjects were exposed to benzene, and benzene did not appear to
24      confound the relationship between 1,3-butadiene or styrene exposure and leukemia mortality.
25      Hence, the model results presented in the report did not control for benzene exposure.
26            Different functional forms of the relationship between the relative rate (RR) and measures
27      of exposure were evaluated by Delzell et al. (1995) including the following:
28            (1) Multiplicative:
29            (2) Power:  RR = e
30            (3) Linear Excess: RR = 1 "+ pX
31            (4) Polynomial Excess:  RR=l + p1Xp + p2Xq+....
        1/28/98                                   9-1        DRAFT-DO NOT CITE OR QUOTE
-------
 1      where X represents the 1,3-butadiene or styrene exposure categories using the midpoints of the
 2      intervals, p represents the estimated model parameters, and the powers "p" and "q" are fixed real
 3      numbers. Although many polynomial functions (model 4) were considered, only the results from
 4      a square root model were presented because this was considered to provide the best fit. This
 5      model may be represented as:

 6             (5)  Square Root: RR=1+ pjX*

 7      The Poisson regression analyses revealed a positive exposure-response relationship between
 8      cumulative exposure to 1,3-Butadiene or styrene and leukemia mortality. This relationship was
 9      evident both in models that represented these exposures as categorical variables (see Table 59 in
10      Delzell et al.,  1995) and in models where exposure was represented using continuous variables as
11      described above. 1,3-Butadiene and styrene exposures among exposed study subjects were
12      found to be moderately correlated (Spearman's rank correlation, r=0.53). The relationship
13      between 1,3-butadiene cumulative exposure and leukemia mortality appeared to be independent
14      of the styrene exposure and was not appreciably altered by inclusion of styrene cumulative
15      exposure in the model. On the other hand, the relationship between styrene cumulative exposure
16      and leukemia mortality was weakened and irregular when 1,3-butadiene cumulative exposure
17      was controlled for.  These findings suggest that 1,3-butadiene cumulative exposure is a more
18      likely explanation for the leukemia excess observed hi this cohort than styrene cumulative
19      exposure.
20             Analyses of peak years indicated an association between this variable and leukemia
21      mortality even after controlling for cumulative  exposure, but this relationship was irregular in the
22      categorical regression analyses. Excluding exposures that occurred within 5 or 10 years of death
23      (i.e., lagging exposures) only slightly increased the exposure-response relationship for 1,3-
24      butadiene cumulative exposure; whereas excluding exposures within 20 years of death weakened
25      and almost eliminated the relationship (i.e., see Table 63 in Delzell et al., 1995).
26             The results that were obtained by the investigators from fitting the alternative relative rate
27      models described above are summarized in Table 9-1. These results are from models that
28      simultaneously evaluated the effects of 1,3-butadiene and styrene exposure.  The regression
29      parameter for 1,3-butadiene cumulative exposure was found to be statistically significantly
30      greater than 0 (p<0.05) in all of the models evaluated, whereas a nonsignificant and weaker
31      relationship was observed for styrene.
32             The power and square root models were found to provide the best fit to the data based on
33      comparison of the model deviances. However, the differences in deviances between the various

        1/28/98                                    9-2        DRAFT-DO NOT CITE OR QUOTE
-------
                Table 9-1. Results from exposure-response models of continuous cumulative
                exposure to 1,3-butadiene and styrene using alternative structural forms
                reported by Delzell et al.a
Structural
model form
Multiplicative:
RR = ePx
Linear:
RR=l+pX
Power:
RR = eP[ln(l+X)]
Square root:
RR=1 + P:X1/2
1,3-Bu
Model
deviance
486.0
486.0
485.6
485.6
tadiene (ppm-years)
P estimate
(S.E.)b
0.0041
(0.0019)
0.0068
(0.0050)
0.2028
(0.0972)
0.1293
(0.1024)
LRT
/?-va!uec
0.04
0.04
0.03
0.03
Styrene (ppm-years)
Model
deviance
485.9
485.7
485.2
485.4
P Estimate
(S.E.)b
0.0052
(0.0053)
0.0079
(0.0088)
0.1494
(0.1183)
0.0968
(0.1090)
LRT
/>-valuec
0.34
0.30
0.21
0.23
        a Adapted from Table 67 in Delzell et al. (1995).  Results presented are adjusted for age, calendar year, years since
         hire, race, and exposure to 1,3-butadiene or styrene.
        b S.E. is the standard error for the exposure parameter estimates.
        0 LRT, likelihood ratio test for the exposure effect (1,3 -butadiene or styrene).
 2
 3
 4
 5
 6
 7
 8
models are slight. The authors expressed a preference for the square root model as the best
model based on its goodness of fit and its simplicity. This model was refined into a "final
model" by omitting styrene and race because the effect of these variables on the estimated
parameter for 1,3-butadiene exposure was considered to have been minimal. In addition, certain
age, calendar year, and years since hire categories were collapsed for the final model for similar
reasons. The final model is summarized in Table 9-2. The relationship between cumulative 1,3-
butadiene exposure and leukemia mortality was highly statistically significant in this model
 9
10
11
12
13
9.1.2. Prediction of Lifetime Excess Risk of Leukemia
       The relative rate models presented in the report by Delzell et al., which are summarized in
Tables 9-1 and 9-2, were used as a basis for predicting the lifetime excess risk of leukemia
mortality for varying levels of continuous environmental exposures to 1,3-butadiene. These
lifetime risk estimates were made using the relative rate estimates and an actuarial program that
        1/28/98
                                           9-3
                                                              DRAFT-DO NOT CITE OR QUOTE
-------
             Table 9-2. Results from "final" square root exposure-response model of
             continuous cumulative exposure to 1,3-butadiene reported by Delzell et al.a
Variable
peta
Estimate | S.E.b
Likelihood ratio test
X2 (d.f.)c 1 p-value
Loglinear terms
Constant
Age:
40-69
70-79
80+
Calendar year:
1950-89
1990-91
Years since hire:
10-19
20+
-10.02

0
0.89
1.71

0
0.72

0
1.09
0.47


0.33
0.48


0.34


0.44

13.2 (2)



3.85(1)


7.64 (1)



0.001



0.050


0.006


Linear term
(1,3-butadiene ppm-
years)0-5
0.17
0.10
9.41 (1)
0.002
       'This table is an adaptation of Table 68 in Delzell et al. (1995).
       b S.E. is the standard error of the parameter estimate.
       eChi-square (%2) and degrees of freedom (d.f.) based on the likelihood ratio statistic.

1      takes into account the effects of competing causes of death.1  U.S. age-specific mortality rates for
2      all race and gender groups combined (NCHS, 1993) were used to specify the leukemia and all-
3      cause background rates in the actuarial program. Exposures to 1,3-butadiene were assumed to be
4      continuous for the entire lifetime, and the risks were computed up to age 85. The occupational
5      1,3-butadiene exposures hi the epidemiologic study were converted to continuous environmental
6      exposures by multiplying the occupational exposure estimates by a factor to account for
              'This program is an adaptation of the approach that was previously used in BEIRIV.
       Health Risks of Radon and Other Internally Deposited Alpha Emitters. National Academy Press,
       Washington, DC, 1988, pp. 131-134.
       1/28/98
9-4
DRAFT-DO NOT CITE OR QUOTE
-------
         differences in the number of days exposed per year (365/240 days) and another factor to account
         for differences in the amount of air inhaled per day (20/10 m3). The reported standard errors for
         the 1,3-butadiene regression coefficients were used to compute the upper 95% confidence limits
  4      for the relative rates based on a normal approximation.
  5            Point estimates and one-sided upper 95% confidence limits for lifetime risk of leukemia
  6      associated with varying levels of environmental exposure to 1,3-butadiene based on the
  7      alternative model forms are illustrated in Figures 9-1 to 9-5. Estimates of risks and exposure
  8      levels corresponding to levels of risk of potential regulatory interest are presented in Tables 9-3
  9      and 9-4. These estimates appear to vary by several orders of magnitude depending on the model
 10      used.  For example, at the 1 in a million risk level, the 95% upper confidence intervals for 1,3-
 11      butadiene exposure range from 0.1 ppb (parts per billion) (based on the multiplicative model) to
 12      1 e-6 ppb (based on the final square root model).
 13            Consistent with the proposed EPA cancer guidelines, these results were also used to
 14      estimate the exposure level (ECp; "effective concentration") and 95% lower confidence intervals
 15      (LECP) associated with varying levels of risk (p) ranging from 0.1 to 10%, which are summarized
 1 6      in Table 9-5. Although the new EPA guidelines emphasize the derivation of exposure levels
 17      associated with a 10% risk level, this does not seem reasonable in this instance. The 10% level
 18      of risk is associated with exposure levels that are higher than most of the exposures experienced
         by the workers in this epidemiologic study. Furthermore, based on the actuarial program
20      described above, a relative rate of 19 would be required for adults over the age of 20 to increase
21      the lifetime risk of leukemia death by 10%, but the leukemia standardized mortality ratios
22      (SMRs) reported by Delzell et al. (1995) were considerably lower.2 Hence, these considerations
23      suggest that using a 10% risk level would be an upward extrapolation in this case. A 1% or even
24      a lower (e.g., 0.1%) risk level would seem to be a more reasonable choice in this circumstance.
25      The analogous relative rates for increased risks of 1% or 0.1% are 2.7 and 1.17, respectively,
26      which better correspond with the set of SMRs reported by Delzell et al. (1995).  The exposure
27      levels corresponding to a 1% risk level are illustrated in Figures 9-1 to 9-5. When a 1% risk
28      level is used, the LEQ from these analyses ranges from 0.07 to 0.6 ppm based on the different
29     relative rate models. Using the final model presented by Delzell etal. (1995) would yield an
30     LEQ of 0.12 ppm.
               2The maximum reported SMR was 13.33. This SMR was based on two leukemia deaths
        among black men from plant #2 with at least 10 years of work (not all of which was salaried) and
        at least 20 years of elapsed time since hired.  (See Table 29 of Delzell et al., 1995.)
        1/28/98                                   9-5        DRAFT-DO NOT CITE OR QUOTE
-------
     § _
                                    One-sided 95% UCL
                                                                   EC, = 1.12
                                                                   LEG, = O.64
           O.O
                            O.5
                                             1.0

                                     Exposure Concentration (ppm)
                                                              1.5
                                                                               2.0
Figure 9-1. Excess risk and 95% upper confidence limit excess risk estimates based on the
multiplicative model reported by Delzell et al., 1995.*
* Multiplicative model: RR= e^x
                                                 One-sided 95% UCL  	
                                               Maximum likelihood estimate
                                                                  EC, = O.45
                                                                  LEG, = 0.12
          O.O
                           0.5
                                            1.O

                                    Exposure Concentration (ppm)
                                                             1.5
                                                                              2.0
Figure 9-2. Excess risk and 95% upper confidence limit excess risk estimates based on the
power model reported by Delzell et al., 1995.*
* Power model:  RR= epllv(1+x)1
1/28/98
9-6
DRAFT-DO NOT CITE OR QUOTE
-------
     & -
     G>
     3 -
«
                                                  One-sided 95% UCL _.-•"
                                                Maximum likelihood estimate
                                                                  EC, = 1.16
                                                                 LEC, = O.525
          O.O
                           0.5
        T

        1.0

exposure Concentration (ppm)
 r
1.5
                                                                              2.O
Figure 9-3. Excess risk and 95% upper confidence limit excess risk estimates based on the
linear excess relative rate model reported by Delzell et al., 1995.*
* Linear excess model:  RR=1 +
                                                 One-sided 95% UCL
     CO
     CD
     <=>
                                              Maximum likelihood Estimate
                                                                   EC, = O.57
                                                                  LEG, = O.O66
          0.0
                           O.5               1.0

                                     Exposure Concentration (ppm)
                         1.5
                                           2.O
Figure 9-4. Excess risk and 95% upper confidence limit excess risk estimates based on the
final square root model reported by Delzell et al., 1995.*
* Final square root model: RR=l + px'/2
1/28/98
       9-7       DRAFT-DO NOT CITE OR QUOTE
-------
    s  _
                                                    One-sided 95% UCL
                                                 Maximum likelihood estimate
                                                                 EC, = 0.77
                                                                LEG, =0.145
                          0.5
        1.0

Exposure Concentration (ppm)
                                                            1.5
                                                                             2.0
Figure 9-5.  Excess risk and 95% upper confidence limit excess risk estimates based on the
square root model reported by Delzell et al., 1995.*
* Square root model:  RR=1 + px'x'
 1/28/98
       9-8
DRAFT-DO NOT CITE OR QUOTE
-------
              Table 9-3. Maximum likelihood estimates (MLEs) of excess risk with one-sided
              95% upper confidence limits (95% UCL) from several models reported by
              Delzell et al. (1995) for continuous lifetime exposures to varying concentrations
              of 1,3-butadiene
Model
Multiplicative:
RR = ePx
Power:
RR = ePDn(l+X)]
Linear:
RR=1 + PX
Initial square root:
RR=1+ pjX1/2

Final square root:
RR=1+ p,X1/2

Concentration 1 MLE I 95% UCL
(ppm) 1 excess risk || excess risk
l.OE-04
l.OE-03
l.OE-02
l.OE-04
l.OE-03
l.OE-02
l.OE-04
l.OE-03
l.OE-02
l.OE-04
l.OE-03
l.OE-02
l.OE-04
l.OE-03
l.OE-02
5.2E-07
5.2E-06
5.3E-05
2.6E-05
2.4E-04
1.6E-03
8.7E-07
8.7E-06
8.7E-05
1.1E-04
3.6E-04
1.1E-03
1.5E-04
4.8E-04
1.5E-03
9.2E-07
9.2E-06
9.3E-05
4.6E-05
4.4E-04
3.1E-03
1.9E-06
1.9E-05
1.9E-04
2.6E-04
8.4E-04
2.6E-03
3.0E-04
9.4E-04
3.0E-03
1
2
3
4
5
6
7
       Ratios are also presented in Table 9-5 that were calculated by dividing the excess risk (p)
by the corresponding LECp for each model. Each ratio is the slope of the line segment
connecting the point (LECp, p) with the origin.  Based on the LECl5 these ratios vary by
approximately one order of magnitude from 0.016 to 0.15. If these LECrbased ratios were used
to calculate the concentration corresponding to a 1 in a million excess lifetime risk by linear
interpolation3, the values would range from 7 to 64 parts per trillion. The final model presented
by Delzell et al. (1995) would yield a corresponding exposure level of 12 parts per trillion.
             3 Linear interpolation between the origin and the point (LECp, p) is also referred to as
       ;'linear extrapolation."
       1/28/98
                                          9-9
DRAFT-DO NOT CITE OR QUOTE
-------
             Table 9-4. MLEs of parts per million continuous exposure concentrations
             associated with varying excess risk levels with one-sided 95% lower confidence
             limits (95% LCL) based on relative rate results of several models reported by
             Delzell et al. (1995) and U.S. population rates
Model
Multiplicative:
RR = eP*

Power:
RRsePIKI+X)]

Linear:
RR=l + pX

Initial square root:
RR=1 + P,X1/2

Final square root:
RR=1+ PiX1/2

[Excess
risk
1E-6
1E-5
1E-4
1E-6
1E-5
1E-4
1E-6
1E-5
1E-4
1E-6
1E-5
1E-4
1E-6
1E-5
1E-4
MLE
(ppm)
1.9E-4
1.9E-3
1.9E-2
3.9E-6
3.9E-5
4.0E-4
1.1E-4
1.1E-3
1.1E-2
7.6E-9
7.6E-7
7.6E-5
4.4E-9
4.4E-7
4.4E-5
95% LCL
(ppm)
1.1E-4
1.1E-3
1.1E-2
2.2E-6
2.2E-5
2.2E-4
0.52E-4
0.52E-3
0.52E-2
1.4E-9
1.4E-7
1.4E-5
1.1E-9
1.1E-7
1.1E-5
1      9.1.3. Sources of Uncertainty
2            It is apparent from the results presented in Table 9-5 that one major source of uncertainty
3      is the choice of the model for the prediction of risk. The range of values of the LEG at either of
4      the 1% and 10% excess risk levels spanned approximately one order of magnitude, whereas the
5      range for the 0.1% level spanned nearly two orders. In this instance, it seems more reasonable to
6      utilize the results at the 1 % risk level because this corresponds to exposures that are within the
7      range of this epidemiologic study. However, it is not possible to clearly choose one of the
8      relative rate models as the best for risk assessment purposes because none of the models has a
9      biologic basis.  Furthermore, all the models summarized in Table 9-1 fit the observed data nearly
       1/28/98
9-10
DRAFT-DO NOT CITE OR QUOTE
-------
              Table 9-5. Maximum likelihood (ECp) and 95% lower-bound (LECp)
              estimates of the continuous exposure concentrations associated with varying
              levels of excess risk (p)
Structural
model form
Multiplicative model:
RR = ePx
Power:
RR = eP[ln(l+X)]
Linear model:
RR=l+pX
Initial square root:
Final square root:
RR=1+ p,XI/2
Percentage
excess risk
(P)
10
1
0.1
10
1
0.1
10
1
0.1
10
1
0.1
10
1
0.1
1,3-Butadiene exposure
levels (ppm)
Maximum
likelihood
(ECD)
3.3
1.12
0.18
1000
0.57
0.0054
12.5
1.16
0.116
88
0.77
0.0076
51
0.45
0.0044
Lower 95%
bound
(LECP)
1.87
0.64
0.10
15
0.066
0.0025
5.65
0.525
0.0525
16.8
0.145
0.00144
13.5
0.12
0.0012
Ratio3
5.3 E-2
1.6 E-2
1.0 E-2
6.7 E-3
1.5E-1
4.0 E-l
1.8 E-2
1.9 E-2
1.9 E-2
5.9 E-3
6.9 E-2
6.9 E-l
7.4 E-3
8.3 E-2
8.3 E-l
1
2
3
4
5
6
3 The ratio is the excess risk (p/100%) divided by the one-sided lower 95% confidence limit on the exposure
  estimate (LECp).

as well. Moreover, for a given linear extrapolation, the ratios in Table 9-5 show that the
sensitivity of the result to the choice of excess risk level varies considerably for these models.,
with the linear model being least sensitive and the two square root models being most sensitive.
Of the two square root models, however, the final relative rate model could be advantageous to
the other model if the omitted parameters for the effects of race and styrene exposure are
unnecessary.
       1/28/98
                                          9-11
DRAFT-DO NOT CITE OR QUOTE
-------
 1             A major source of uncertainty in this analysis is the potential for misclassification of
 2      exposures in the study by Delzell et al. (1995). This is a frequent limitation of nearly all
 3      epidemiologic studies of this type for quantitative risk assessment purposes. The exposures of
 4      this study were based on modeling a relatively extensive set of data. However, questions have
 5      been raised concerning the accuracy of exposure estimates, particularly for some ill-defined tasks
 6      (letter from Elizabeth Moran, CMA, March 25,1996). For example, the work histories of
 7      maintenance laborers do not indicate whether they were vessel cleaners (a high-exposure
 8      category) or building cleaners (a low-exposure category).  The full impact of this potential for
 9      exposure misclassification is unknown, but preliminary analyses suggest that it may have
10      dampened and possibly distorted the observed dose-response relationship (letter from Delzell and
11      Macaluso to Aparna Koppikar, April 2,1996).
12             Another concern about the study has been expressed regarding the assignment of peak
13      exposures in the analysis, which was defined as  average exposures equal to or greater than 100
14      ppm over 15 min. It has been suggested that there were tasks with extremely high peak
15      exposures (thousands of ppm) over very short time periods (seconds to a few minutes) (letter
16      from Delzell and Macaluso to Aparna Koppikar, April 2,1996).  The models used in this risk
17      assessment assume a constant dose-rate effect and do not consider the potential for the effects of
18      peak exposures. The potential impact of work area assignments and butadiene peaks on leukemia
19      mortality in this study population is an active area of research among the investigators at the
20      University of Alabama who conducted the study by Delzell et al. (1995).

21      9.1.4. Summary and Conclusions
22             Risk estimates for environmental exposures are derived from an analysis by Delzell et al.
23      (1995) of an occupational retrospective cohort mortality study of approximately  16,000 workers
24      in six North American styrene-butadiene rubber manufacturing plants. The analysis of this study
25      is based on follow-up during 1943-1991, with an average follow-up of 25 years and about 25%
26      of the cohort deceased. While overall mortality and all cancer mortality were below expected
27      values based on general population regional rates, the increase in leukemias was statistically
28      significant (SMR = 1.43, 95% C.I. = 1.04-1.91)  for all ever-hourly men (Delzell et al., 1996).
29      The consistency of this leukemia result with other findings from previous epidemiology studies
30      with 1,3-butadiene plus other data led to the conclusion that this  increase was due to 1,3-
31      butadiene and to the decision to perform a quantitative risk assessment with this database.
32             While this cohort had been previously studied (Matanoski et al., 1987, 1988, 1989, 1990,
33      1994), the Delzell et al. update and analyses are especially noteworthy for their extensive work
34      on exposure estimation based on detailed reviews of individual j ob histories and a j ob exposure

        1/28/98                                   9-12       DRAFT-DO NOT CITE OR QUOTE
-------
         matrix (Delzell et al., 1995; Macaluso et al., 1996). The careful work on exposure allowed better
         estimates of risk and dose response. Exposure metrics included cumulative ppm-years and
         number of years with peak exposures of at least 100 ppm for at least 15 min.  Additional
  4      individual worker exposure information on both styrene and benzene allowed analyses to adjust
  5      for these potential confounding exposures. The Delzell et al. (1995) report includes these
  6      analyses.
  7            The Delzell et al. analysis used Poisson regression analysis with nine categories for
  8      cumulative exposure of 1,3-butadiene  and nine categories of exposure for styrene.  The analysis
  9      also included, as covariates, adjustments for age, race, calendar year, and years since hire.
 10      Relative rate models run within the Poisson analysis included the (1) multiplicative, (2) power,
 11      (3) linear, and (4) square root models.  The parameter representing cumulative 1,3-butadiene
 12      exposure was found to be statistically significant in all the models evaluated, and all models fit
 13      the data adequately in the observable range. The cumulative styrene exposure parameter was
 14      positive for all the models, but not statistically significant. While Delzell et al. selected the
 15      square root model as their final choice because of a slightly better likelihood fit, none of the
 1 6      models fit the  data significantly better  or worse than the others.
 1 7            The quantitative risk analysis presented here uses the results of the Delzell et al. analyses,
         which include the styrene exposure variable as a covariate, to extrapolate risk from occupational
         work-time exposure to lifetime environmental continuous exposure.  This is done by adjusting
 20      the 1,3-butadiene parameter estimates calculated by Delzell et al. to reflect continuous rather than
 21      work-time exposures and by using life table modeling techniques to convert the relative rate
 22      exposure-response relationship to a lifetime additional risk dose-response relationship. These
 23      techniques have been used before by EPA as well as other governmental agencies.
 24            After calculation of the exposure-response relationship, the low-exposure extrapolation is
 25      done in two ways reflecting the different approaches used in EPA's 1986 Guidelines for
 26      Carcinogen Risk Assessment (U.S. EPA, 1986) and those currently proposed for revision (U.S.
 27      EPA, 1996). For the 1986 Guidelines, the risk estimates are calculated as a potency or slope
 28      factor derived from  fitting a linear model (default case) to the observed data and applying the
 29      same model to lower exposure concentrations. For the proposed guidelines revisions, the risk
 30      estimates are obtained by first calculating a "point of departure" within the range of observation
 31      using any of the appropriate models and then extrapolating to 0 by means of a straight line. The
 32     LED10 (i.e., lower confidence limit on  a dose associated with 10% extra risk) is proposed in the
33     guidelines revisions as the standard point of departure; however, the LEC01 and EC01 are used
34     here because 1% is within the observable range of increased leukemia deaths for the different
         1,3-butadiene exposure groups in the Delzell et al. study, because exposure levels are expressed

         1/28/98                                   9-13        DRAFT-DO NOT CITE OR QUOTE
-------
 1      as exposure concentrations rather than doses, and because the issue of whether to use LEDs or
 2      EDs in the final guidelines has not yet been resolved.
 3            The results of the extrapolations using the four relative rate models are shown in the
 4      quantitative risk analysis and presented in Figures 9-1 to 9-4. They show that although in the
 5      observable risk range of 1%, the MLEs of required continuous exposure (EC01) are close, varying
 6      2.6-fold from 0.45 to 1.16 ppm, the LEC01 estimates range from 0.066 to 0.64, or about 10-fold.
 7      Furthermore, as the risk extrapolation decreases 10-fold to a 0.001 risk level, the ML exposure
 8      estimates for the various models diverge much more rapidly, a 45-fold range from 0.004 ppm to
 9      0.18 ppm. At the 10'5 risk level, the exposure estimates diverge by nearly four orders of
10      magnitude. Clearly, the final risk estimates based on the 1986 guidelines extrapolation
11      procedures are highly dependent on the choice of model, but those of the proposed guidelines
12      revisions, which extrapolate from the LEC01, are less affected.
13            For the 1986 guidelines approach, the model of choice is the linear default.  This choice is
14      based more on historical precedence and biological plausibility arguments than on statistical fit
15      or conservatism. In fact, for a 10"6 risk level the linear model is much less protective of public
16      health, by nearly five orders of magnitude, than is the Delzell et al. square root model choice.
17      For this approach, the maximum likelihood potency (slope) estimate is:

18                                       B = 8.7 x 10-3 (ppm)-1.

19            For the suggested default approach under the proposed guidelines revisions, the EC01
20      level is chosen because that is within the observable response range of leukemia deaths. At the
21      EC0i level, the different models provide dose estimates ranging from 0.45 ppm to 1.16 ppm and
22      the 95% LCLs on dose ranging from 0.066 to 0.64.  Without specific directions for choice from
23      the proposed guidelines, potency estimates based on each of the models examined by Delzell et
24      al. are presented in Table 9-6.
25            The cancer potency estimates using EC01s  as the point of departure range from 8.7 x
26      10"3/ppm (linear model) to 0.022/ppm (final square root model). The square root model was the
27      model preferred by Delzell et al. based on goodness of fit and simplicity; thus they chose that
28      model for various refinements, resulting in the final  square root model.  The cancer potency
29      estimates based on LEC01s range from 0.016/ppm to 0.15/ppm, with the final square root model
30      yielding 0.083/ppm while the linear model yields  0.019/ppm. Although the proposed Guidelines
31      do not offer explicit guidance on choice of model, it may be appropriate in this particular case to
32      use the final square root model to obtain the point of departure because this  model benefits from
33      the refinements performed by Delzell et al.

        1/28/98                                   9-14       DRAFT-DO NOT CITE OR QUOTE
-------
               Table 9-6. Cancer potency (unit risk) estimates based on linear extrapolation
               from the LEC01 or EC01 calculated from the models presented by Delzell et al.
Model
Multiplicative
Power
Linear
Initial square
root
Final square root
EC01 (ppm)
1.12
0.57
1.16
0.77
0.45
Potency estimate
(ppm-1)
(Le., 0.01/EC01)
8.9 x 10-3
0.018
8.7 x lO'3
0.013
0.022
LEC01 (ppm)
0.64
0.066
0.525
0.145
0.12
Potency estimate
(ppm-1)
(Le., 0.01/LEC01)
0.016
0.15
0.019
0.069
0.083
  1            As the estimates of choice, the MLEs of both the potency and EC01 are chosen.  The main
  2      reason for this choice is that these estimates are based on human data from a large, well-
  3      conducted study. Although EPA has historically used upper-limit potency estimates for animal-
        to-human extrapolations, these upper limits derive their use more from computational
        instabilities of the MLEs in the quantitative risk models used.  Human-to-hurnan extrapolations
  6      typically use a simpler linear model form that does not have these instabilities. Furthermore, the
  7      human data inherently engender far less uncertainty in the risk estimates, so one may have more
  8      confidence in the use of MLEs from human data than from animal data.

  9      9.2. CANCER RISK ESTIMATES BASED ON RODENT BIO ASSAYS
10      9.2.1. Rat-Based Estimates
11            Cancer risk estimates based on the 1981 Hazelton rat inhalation study of 1,3-butadiene
1 2      were presented in EPA's 1985 1,3-butadiene risk assessment (U.S. EPA, 1985). 95% upper-limit
13      incremental lifetime unit cancer risk estimates for humans were calculated using the linearized
14      multistage (LMS) model, after estimating the equivalent human dose assuming 1,3-butadiene
1 5      retention based on results of a 1985 NTP absorption study (NTP, 1985; see EPA's 1985 report
16      for further details). The upper limit based on the male rat tumor incidence data for Ley dig cell
17      tumors, pancreatic exocrine tumors, and/or Zymbal gland carcinomas was 4.2 x 10"3 per ppm
18      1,3-butadiene exposure. The upper limit based on the female rat tumor incidence data for
        1/28/98
9-15
DRAFT-DO NOT CITE OR QUOTE
-------
  1      mammary gland carcinomas, thyroid follicular tumors, and/or Zymbal gland carcinomas was 5.6
  2      * 10~2 per ppm 1,3-butadiene exposure.
  3             These rat-based estimates are not considered the most appropriate estimates of human
  4      risk; they are merely presented for comparison purposes.  EPA believes that the mouse is likely
  5      to represent a better rodent model for human cancer risk from 1,3-butadiene (see below) and that
  6      the cancer risk estimates derived from the epidemiologic data are the best available estimates for
  7      human risk.

  8      9.2.2. Mouse-Based Estimates
  9             Cancer risk estimates based on the 1984 NTP mouse inhalation study were presented in
10      EPA's  1985  1,3-butadiene risk assessment; however, revisions to these estimates are warranted
11      because of the new data provided by the 1993 NTP mouse inhalation bioassay, which examined
12      cancer response from exposure to lower 1,3-butadiene concentrations than those used in the 1984
13      study (NTP 1984,1993; see Chapter 6). Groups of male and female B6C3Fj mice were exposed
14      to 1,3-butadiene concentrations of 0, 6.25,20, 62.5,200, or 625 ppm 1,3-butadiene for 6
15      hours/day, 5  days/week, for up to 104 weeks. Significant increases in tumor incidence were
16      observed at multiple sites: the hematopoietic system (lymphomas; histiocytic sarcomas [males]),
17      heart (hemangiosarcomas), lung, forestomach, Harderian gland, liver, preputial gland (males),
18      ovary (females), and mammary gland (females), when adjusted for intercurrent mortality
19      (Melnick and Huff, 1993).  Significant increases in lung cancer incidence were observed in
20      female mice  at 1,3-butadiene exposure levels down to 6.25 ppm, the lowest level tested.

21      9.2.2.1. Oiiantal
22             When EPA estimates cancer risks for humans from rodent bioassay data, the risk
23      estimates are generally calculated from the incidence of rodents of the most sensitive species,
24      strain, and sex bearing tumors at any of the sites displaying treatment-attributable increases. In
25      the case of 1,3-butadiene, so many sites demonstrated significant tumor increases attributable to
26      1,3-butadiene that background levels of tumor-bearing animals obfuscate the effects of 1,3-
27      butadiene when all these tumor sites are combined. Therefore, risk estimates were derived from
28      the incidence of female (most sensitive sex) mice with malignant lymphomas, heart
29      hemangiosarcomas, lung tumors (alveolar/bronchiolar adenomas or carcinomas), mammary
30      gland tumors (carcinomas, adenocanthomas, or malignant mixed tumors), or benign or malignant
31      ovary granulosa cell tumors (Table 9-7). These sites were considered to be the most relevant
32      sites with low background tumor incidence.  Most of the impact on the low-dose linear
33      extrapolation is from the lung tumor response, because the lung tumor incidences show the

        1/28/98                                  9-16       DRAFT-DO NOT CITE OR QUOTE
-------
               Table 9-7. Dose-response data for linearized multistage model
Administered exposure
(ppm)
Human equivalent
exposure (ppm)
Number of mice with
tumors3
Number of mice at riskb
Control
0
6/50
6.25
1.1 ,
19/49
20
3.6
26/50
62.5
11
31/50
200
36
46/49
        "Lymphocytic lymphomas, heart hemangiosarcomas, alveolar/bronchiolar adenomas or carcinomas, mammary gland
         tumors (carcinomas, adenocanthonomas, malignant mixed tumors), or benign or malignant ovary granulosa cell
         tumors.
        'Female mice surviving to the time of the first significant tumor, which was a lymphocytic lymphoma at day 203.
  1      largest increases at the lowest exposures. The 625 ppm exposure group was not included in the
  2      dose-response analysis because all of the mice were dead by week 65, and the tumor response
  3      was already virtually saturated in the 200 ppm exposure group. Note also that mice that died
  4      before the time of observation of the first tumor were considered to be not at risk and were
  5      excluded from the incidence denominators.
               Human equivalent exposures were based on ppm 1,3-butadiene exposure, adjusted for
        continuous daily exposure (e.g., 6.25 ppm * 6/24 x 5/7 =1.12 ppm). No attempt was made to
  8      adjust for internal doses of reactive 1 ,3-butadiene metabolites because the PBPK data were
  9      inadequate to develop reliable PBPK models (Chapter 8).  No adjustments were made for 1,3-
1 0      butadiene absorption because there are no adequate human data. Furthermore, there is no reason
11      to expect nonlinearities in absorption at the lowest exposures (at least < 625 ppm).
12             A 95% upper-limit incremental lifetime unit cancer risk (extra risk) for humans was
1 3      calculated from the incidence data in Table 9-7 using the LMS model. The multistage model has
14      the form:
15
1 6      where P(d) represents the lifetime risk (probability) of cancer at dose d, and parameters q; > 0,
1 7      for 1=0, 1 , ..., k. Extra risk over the background tumor rate is defined as
18
        1/28/98
9-17
DRAFT-DO NOT CITE OR QUOTE
-------
  1            Point estimates of the dose coefficients (q;s), and consequently the extra risk function, at
  2     any dose d, are calculated by maximizing the likelihood function with respect to the tumor
  3     incidence data. The incremental lifetime unit cancer risk for humans (qj*) is defined as the 95%
  4     UCL on the parameter ql9 which is the linear dose coefficient, for extra risk. This 95% UCL
  5     represents a plausible upper bound for the true risk.  The 95% UCL was calculated using the
  6     computer program GLOBAL86 (Howe and Van Landingham, 1986).  Both the model and the
  7     curve-fitting methodology used are described in detail by Anderson et al. (1983).
  8            The tumor incidence data in Table 9-7 generated the following results using the LMS
  9     model (GLOBAL86):
10            MLEs of dose coefficients:
11            q0 = 0.2629
12            q, = 0.07643
13            q2 = 0.0
14            q3 = 0.0                                                        .
15            q4 = 0.0
16           p-vaiue for chi-square goodness of fit > 0.01
17            q,* = 0.10

18     Thus, the incremental unit cancer risk estimate (95% UCL) for humans calculated from the
19     mouse  1993 NTP inhalation bioassay results is 0.10 per ppm for continuous lifetime inhalation
20     exposure to 1,3-butadiene.  The MLE of risk appears to be nearly linear between 1 ppm and 1
21      ppb and is about 0.075 per ppm 1,3-butadiene exposure.
22            Under EPA's proposed new cancer risk assessment guidelines (U.S. EPA,  1996),  unit
23     cancer risk estimates for genotoxic chemicals, such as 1,3-butadiene, would be derived by
24     straight linear extrapolation to 0 from the LEDj0 (estimated 95% UCL on the dose corresponding
25     to a 10% cancer risk). Using the LEC10 generated for the LMS model by GLOBAL86 yields a
26     unit cancer risk of 0.10/1.0 ppm = 0.10 per ppm, the  same as the q^. Using the EC10 yields
27     0.10/1.4 = 7.1 x l O'2 per ppm.
28            MLE of risk for a dose of 1 ppm = 7.4 x 1Q-2
29            MLE of risk for a dose of 1 ppb = 7.6 x l O'5
30            MLE of dose for a risk of 0.10 (EC10) = 1.4 ppm
31             95% UCL on dose for a risk of 0.10 (LEC10) = 1.0 ppm
32            The unit cancer risk estimate (95% UCL) derived above is intended to depict a plausible
33     upper limit on the risk of developing any 1,3-butadiene-attributable tumor over a full (70-year)
34     lifetime. However, using the quantal incidence data for total tumor-bearing mice in each

        1/28/98                                  9-18       DRAFT-DO NOT CITE OR  QUOTE
-------
         exposure group does not fully characterize the cancer potency reflected by the mouse bioassay
  2      results. First, the methodology does not take into account the fact that many of the mice in the
 '3      higher exposure groups had tumors at multiple significant sites. Second, the methodology
  4      ignores the fact that survival was significantly decreased in female mice exposed to 20 ppm or
  5      more 1,3-butadiene as a result of fatal 1,3-butadiene-attributable tumors.  Time-to-tumor
  6      analyses conducted for specific tumor sites are presented below and can be used to evaluate the
  7      time component of the cancer risk.

  8      9.2.2.2. Time-to-Tumor
  9            The mouse inhalation bioassay results demonstrate different dose-response relationships
 10      for different tumor sites. To assess the characteristics of the dose-response relationships for
 11      different tumor sites, time-to-tumor analyses were performed to adjust for competing mortality
 12      from cancer at other sites.
 13            Time-to-tumor analyses were conducted from the individual mice data, including the 9-
 14      month and 15-month interim sacrifice data, for sites demonstrating an increased cancer
 15      incidence. Benign and malignant tumors were combined for sites where appropriate. Thus time-
 16      to-tumor analyses were performed for lung alveolar/bronchiolar adenomas or carcinomas;
 1 7      lymphocytic lymphomas; heart hemangiosarcomas; hepatocellular adenomas or carcinomas;
 J8      Harderian gland adenomas or carcinomas; forestomach squamous cell papillomas or carcinomas;
 1 9     malignant or benign ovary granulosa cell tumors (female); and mammary gland adenocanthomas,
 20      carcinomas, or malignant mixed tumors (female). Preputial gland carcinomas in male mice were
 21     not analyzed because not all the tissues were examined microscopically.
 22            Data from the 625 ppm exposure groups were excluded from analysis because of
 23     excessive early mortality, as in the quanta! analysis discussed above. In addition, data from
 24     interim sacrifices for specific sites were excluded for dose groups for which it appeared that
 25     complete histopathological examination for that site was not performed on the entire interim
 26     sacrifice group.
 27           Human equivalent exposures were based on ppm 1,3-butadiene exposure, adjusted for
 28     continuous daily exposure, as described above.
 29           The general model used for the time-to-tumor (or time-to-response) analyses was the
30     multistage Weibull model, which has the form

31
        1/28/98                                   9-19       DRAFT-DO NOT CITE OR QUOTE
-------
  1      where P(d,t) represents the probability of a tumor (or other response) by age t (in bioassay
  2      weeks) for dose d (human equivalent exposure), and parameters z^ 1, t0^0, and q^O for i=0,1,...,
  3      k, where k = the number of dose groups - 1. The parameter t0 represents the time between when
  4      a potentially fatal tumor becomes observable and when it causes death (see below). The analyses
  5      were conducted using the computer software TOX_RISK version 3.5 (Crump et al., ICF Kaiser
  6      International, Ruston, LA), which is based on Weibull models taken from Krewski et al. (1983).
  7      Parameters are estimated using the method of maximum likelihood.
  8             Specific n-stage Weibull models were selected for the individual tumor types for each sex
  9      based on the values of the log likelihoods according to the strategy used by NIOSH (1991 a). If
10      twice the difference in log likelihoods was less than a chi-square with degrees of freedom equal
11      to the difference in the number of stages included in the models being compared, then the models
12      were considered comparable and the most parsimonious model (i.e., the lowest-stage model) was
13      selected.
14             Tumor types were categorized by tumor context as either fatal or incidental tumors.
15      Incidental tumors are those tumors thought not to have caused the death of an animal, while fatal
16      tumors are thought to have resulted in animal death. Lymphocytic lymphomas, histiocytic
17      sarcomas, and heart hemangiosarcomas were treated as fatal tumors, unless observed at an
18      interim or terminal sacrifice, in which case they were considered incidental. Furthermore, these
19      fatal tumors were deemed rapidly fatal, and to was set equal to 0 (it was felt that there were
20      insufficient data to reliably estimate t0 in any event).  Tumors at all other sites were treated as
21      incidental. This is basically the same determination as that made by NIOSH (1991a), except the
22      NIOSH report dealt with preliminary data that did not distinguish histiocytic sarcomas from
23      lymphomas. NIOSH further cited the work of Portier et al. (1986) analyzing tumor types in NTP
24      historical controls to lend support to these tumor context assumptions.
25             Parameter estimates for the time-to-tumor analyses for each tumor type are presented in
26      Tables 9-8 (based on female mouse data) and 9-9 (male mice). For all tumor types except the
27      heart hemangiosarcomas (both sexes) and the forestomach (male mice), the one-stage Weibull
28      was the preferred model. For male mice, the heart hemangiosarcomas and forestomach tumors
29      were best described by the two-stage model, while for female mouse heart hemangiosarcomas, a
30      three-stage model was preferred.
31             Human unit cancer risk (or potency) estimate results (extra risk) are presented in Tables
32      9-10 (based on female mouse data) and 9-11  (male mice). Mouse lung rumors convey the
3 3      greatest amount of extrapolated risk to humans from both the female mouse data (q l * = 0.14/ppm
34      1,3-butadiene exposure) and the male mouse data (qt* = 0.10/ppm). Note that the unit risk
35      estimate of 0.14/ppm generated from the female mouse lung tumor data using a time-to-tumor

        1/28/98                                 9-20       DRAFT-DO NOT CITE OR QUOTE
-------
      Table 9-8. Parameter estimates for multistage Weibull
      model based on female mouse tumor incidence, w/o 625
                time-to-tumor
                 ppm group
Tissue
Lymphocytic
lymphoma
Heart
hemangiosarcoma
Lung
Mammary
Liver
Forestomach
Harderian gland
Ovary
Histiocytic sarcoma
QO
6.23 x 10-10
0
5.83 x 1Q-9
2.47 x lO'6
2.12 x lO'8
0
1.50 x 10-5
7.83 x IQ-9
3.68 x lO-14
Ql
1.67 x 1Q-10
0
3.40 x IQ-9
5.42 x 1Q-5
2.11 xlQ-9
1.29 x 10'9
2.06 x lO-6
1.48'x lO'8
1.23 x lO'14
Q2
-
0
-
_
-
_
_
_
-
Q3
-
2.88 x 10-
17
_
_
-
_
_
_
-
'• z
3.92
6.10
3.69
1.27
3.58
3.43
2.03
3.05
6.03
      Table 9-9. Parameter estimates for multistage Weibull time-to-tumor
      model based on male mouse tumor incidence, w/o 625 ppm group
Tissue
Lymphocytic
lymphoma
Heart
hemangiosarcoma
Lung
Liver
Forestomach
Harderian gland
Histiocytic sarcoma
QO
1.84 x lO'8
0.0
1.38 x lO'7
1.40 x 1Q-4
9.68 x lO'10
1.65 x lO'7
0.0
Ql
1.28 x IQ-9
0.0
9.53 x lO'8
5.57 x IQ-6
0.0
7.45 x 1Q-8
1.04 x IQ-13
Q2
.
1.14 x lO-23
_
_
3.83 x 10-11
_
-
Z
3.08
10.0
3.27
1.83
3.39
2.90
5.50
1/28/98
9-21
DRAFT-DO NOT CITE OR QUOTE
-------
      Table 9-10.  Human unit cancer risk estimates (extra risk, computed for risks of
      10"6) based on female mouse tumor incidences, w/o 625 ppm group using multistage
      Weibull time-to-tumor model
Tissue
Lymphocytic lymphoma
Heart hemangiosarcoma
Lung
Mammary
Liver
Forestomach
Harderian gland
Ovary
Histiocytic sarcoma
Ql*
(ppm-1)
0.0239
4.27 x lO'3
0.1404
0.0321
0.0631
0.0215
0.0443
0.0358
0.1283
MLE
(ppm-1)
0.0128
3.99 x lO'6
0.0980
0.0203
0.0366
0.0112
0.0258
0.0218
3.36 x lO'3
EC10
(ppm)
8.08
11.6
1.06
5.09
2.82
9.22
4.00
4.74
30.8
LECjo
(ppm)
4.33
9.24
0.737
3.23
1.64
4.80
2.33
2.89
0.806
0.1/LECjo
(ppm-1)
0.0231
0.0108
0.1357
0.0310
0.0610
0.0208
0.0429
0.0346
0.1241
      Table 9-11. Human unit cancer risk estimates (extra risk, computed for risks of
      10"6) based on male mouse tumor incidences, w/o 625 ppm group using multistage
      Weibull time-to-tumor model
Tissue
Lymphocytic lymphoma
Heart hemangiosarcoma
Lung
Liver
Forestomach
Harderian gland
Histiocytic sarcoma
Ql*
(pprn'1)
6.437 x 10-3
0.01266
0.1023
0.04447
4.258 x lO'3
0.07402
0.02162
MLE
(ppm-1)
2.220 x 10-3
4.040 x 1Q-3
0.06998
0.02720
1.660 x 10-5
0.05398
0.01394
EC10
(ppm)
46.6
12.0
1.48
3.80
19.2
1.92
7.42
LEG™
(ppm)
16.1
7.59
1.01
2.33
13.3
1.40
4.78
0.1/LEC10
(ppm-1)
6.224 x 1Q-3
0.01318
0.09890
0.04300
7.517 x 10-3
0.07157
0.02090
1/28/98
9-22
DRAFT-DO NOT CITE OR QUOTE
-------
        model is greater than the unit risk estimate of 0.10/ppm generated above from multiple female
        mouse tumor sites when only the quantal data were used and decreased survival time was not
        taken into account.                               '
 4           Although the time-to-tumor modeling does help account for decreased survival times in
 5     the mice, considering the tumor sites individually does not convey the total amount of risk
 6     potentially arising from the sensitivity of multiple sites. To get some indication of the total unit
 7     risk from multiple rumor sites, assuming the multiple sites are mechanistically independent, the
 8     MLEs of the unit potency from the Weibull time-to-tumor models were summed across tumor
 9     sites and estimates of the 95% upper bound on the summed unit potency were calculated. The
10     TOXJRISK software provides MLEs and 95% UCL's for human risk at various exposure levels,
11      allowing for the calculation of unit potency estimates at those exposure levels.
12           When the MLEs of unit potency calculated at 1 ppb from the female mouse data were
1 3     summed across the female mouse tumor sites, the MLE of the total unit risk was 0.23/ppm
14     continuous lifetime  1,3-butadiene exposure. A 95% upper bound for the total potency was
1 5     calculated by assuming a normal distribution for the risk estimates, deriving the variance of the
16     risk estimate for each tumor site from its 95% UCL according to the formula

17                                   95% UCL = MLE+1.645o,

1 8     where the standard deviation a is the square root of the variance, summing the variances across
19     tumor sites to obtain the variance of the sum of the MLEs, and calculating the 95% UCL on the
20     sum from the variance of the sum using the same formula. The resulting 95% UCL on the unit
21      potency for the total unit risk was 0.38/ppm. In comparison, summing the q^s across the female
22     mouse tumor sites yielded 0.50/ppm.
23           The unit potencies were also summed using a Monte Carlo analysis  and the software
24     Crystal Ball version 4.0 (Decisioneering, Denver, CO).  Normal distributions were assumed for
25     the unit potency for each tumor site, with the mean equal to the MLE and a as calculated from
26     the above formula.  A distribution around the sum of the MLEs was then generated by simulating
27     the sum of unit potencies picked from the distributions for each tumor site (according to
28     probabilities determined by those distributions) 10,000 times. The mean for the sum and the
29     95th percentile on the distribution were the same as the sum of MLEs and 95% UCL calculated
30     above, as they should be.  However, a sensitivity analysis based on the contribution to variance
31      revealed that variability associated with the unit potency estimate for the histiocytic sarcomas
32     was contributing more than 83% of the variance on the sum, and some of the simulated sums
        were negative (the distributions for the unit potency estimates were not constrained for the

        1/28/98                                  9-23       DRAFT-DO NOT CITE OR QUOTE
-------
  1      summation analyses). Excluding the histiocytic sarcomas yielded the same MLE of total risk of
  2     0.23/ppm; however, the 95% UCL decreased to 0.29/ppm. The lung, which then contributes the
  3     most to the sum, contributed about 55% of the variance, followed by the liver with 20%, and no
  4     simulated sums were negative.
  5            The same analyses were performed summing the estimates of unit potency derived from
  6     the male mouse data for the different tumor sites. The resulting MLE for the total unit risk was
  7     0.18/ppm lifetime 1,3-butadiene exposure with a 95% UCL of 0.22/ppm. The lung contributed
  8     about 56% to the variance, followed by the Harderian gland with about 20%. Histiocytic
  9     sarcomas contributed only 3% in this case, and all simulated sums were positive.
10            Finally, the summation analyses were repeated for unit potency estimates calculated at 1
11      ppm exposure for comparison with the estimates calculated at 1 ppb. For the female mouse-
12     based risks (excluding histiocytic sarcomas), the sum of the MLEs was 0.22/ppm (2% lower than
13     at 1 ppb)  and the 95% UCL on the sum was 0.28/ppm (4% lower than at 1 ppb). Thus, the total
14     unit potency estimates are reasonably linear up to 1 ppm continuous lifetime exposure. Recall
15     from Table 9-8 that the selected model for the heart hemangiosarcomas in the female mouse was
16     nonlinear; however, the unit risk estimates based on the heart hemangiosarcomas at these
17     extrapolated doses are lower than for the other sites and do not affect the total risk summed
18     across tumor sites.  Similarly, the male mouse based- results (both the sum of the MLEs and the
1 9     95% UCL on the sum) calculated at 1 ppm were 2% lower than those calculated at 1 ppb. For
20     the male mice, the selected models for both the heart hemangiosarcomas and the forestomach
21      tumors were nonlinear (Table 9-9); however, as with the female heart hemangiosarcomas, the
22     risks from these sites have little impact on the total risk.
23            The results of these summation analyses are summarized in Table 9-12.

24     9.2.3. Discussion
25             Based on the analyses discussed above, the best estimate for an upper bound on human
26     extra cancer risk from continuous lifetime exposure to 1,3-butadiene derived from animal data is
27     about 0.3/ppm. This estimate reflects the time-to-tumor response as well as the exposure-
28      response relationships for the multiple tumor sites (excluding histiocytic sarcomas) in the most
29      sensitive species and sex (the female mouse).  Histiocytic sarcomas were excluded because they
30     introduced excessive variance into the upper bound while contributing only negligibly to the
31      MLE of total unit risk.
32            The greatest source of uncertainty in this estimate is from the interspecies extrapolation
33      of risk from the mouse to humans. The two rodent species for which bioassay data were
        1/28/98                                   9-24       DRAFT-DO NOT CITE OR QUOTE
-------
               Table 9-12. Unit potency estimates (extra risk) summed across tumor sites

Female mouse tumor sites
(calculated at 1 ppb)
Female sites excluding
histiocytic sarcomas (at 1 ppb)
Female sites excluding
histiocytic sarcomas (at 1 ppm)
Male mouse tumor sites
(at 1 ppb)
Male mouse tumor sites
(at 1 ppm)
Sum of MLEs
(ppm'1)
0.23
0.23
0.22
0.18
0.17
95% UCL on
sum (ppm""1)
0.38
0.29
0.28
0.22
0.21
Sum of qj*s
(ppm-1)
0.50
0.37
0.36
0.27
0.26
        available—the mouse and the rat—varied significantly in their carcinogenic responses to 1,3-
        butadiene, in terms of both site specificity and degree of response (Chapter 6). The mouse and
        rat also exhibit substantial quantitative differences in their metabolism of 1,3-butadiene to
        potentially reactive metabolites (Chapter 3).  Unfortunately, existing pharmacokinetic models
        have been unable to explain the species differences in carcinogenic response (Chapter 8), and it
 6      is likely that there are pharmacodynamic as well as pharmacokinetic differences between the
 7      mouse and rat with respect to their sensitivities to 1,3-butadiene.
 8            The mouse was the more sensitive species to the carcinogenic effects of 1,3-butadiene
 9      exposure and, hence, the more conservative (public health protective) for the extrapolation of risk
10      to humans. In addition, the mouse appears to be the more relevant species for extrapolation to
11      humans  in terms of site specificity, as 1,3-butadiene induces tumors of the lymphohematopoietic
12      system in both mice and humans. Melnick and Kohn (1995) further suggest that the genetic
13      mutations observed in 1,3-butadiene-induced mouse tumors are analogous to genetic alterations
14      frequently observed in human tumors.                                  •
15            In addition to uncertainties pertaining to the relevance of the rodent models to human
1 6      risk, there is uncertainty in quantitatively scaling the animal risks to humans.  Ideally, a PBPK
17      model for the internal dose of the reactive metabolite(s) would decrease some of the quantitative
18      uncertainty in interspecies extrapolation; however, current PBPK models are inadequate for this
1 9      purpose  (Chapter 8).  In vitro metabolism data for humans suggest that interhuman variability in
        1/28/98
9-25
DRAFT-DO NOT CITE OR QUOTE
-------
  1      the capacity to metabolically activate 1,3 -butadiene nearly spans the range between rats and mice
  2     (Chapter 3).
  3            Another major source of uncertainty in the unit potency estimate of 0.3/ppm is the
  4     extrapolation of high-dose risks observed in the mouse bioassay to lower doses that would be of
  5     concern from human environmental exposures. A multistage Weibull time-to-tumor model was
  6     the preferred model because it can take into account the differences in mortality between the
  7     exposure groups in the mouse bioassay; however, it is unknown how well this model is
  8     predicting the low-dose extrapolated risks for 1,3-butadiene.
  9            There are also uncertainties pertaining to the specific assumptions used in conducting
10     these multistage Weibull time-to-tumor analyses.  Some alternative analyses were performed to
11      consider the sensitivity of the results to some of these assumptions. For example, for each of the
12     tumor types assumed to be fatal, alternative analyses were conducted in which the modeling
13     software estimated t0. In each case, the resulting q^s, ECi0s, and LEC10s were identical to those
14     generated when to was set equal to 0 a priori.
15            In addition, analyses were performed on the lymphocytic lymphoma data including the
16     625 ppm group, as this was the exposure group most affected by lymphocytic lymphomas and
17     relatively few animals in this group survived to develop tumors at other sites. From the female
18     mouse data, the resulting q,* was 0.515/ppm,  or roughly twice that calculated when the 625 ppm
19     group was excluded. From male mice, the q,* was 0.0215/ppm, or roughly 3 times higher than
20     that obtained when the 625 ppm group was excluded.
21             NIOSH (1991a) examined the sensitivity of its results for each tumor type to (1) model
22     selection (i.e., stage of Weibull model) from among models deemed to be comparable,  (2) tumor
23     context assumptions, and (3) exclusion/inclusion of the 625 ppm exposure group, and generally
24     found only small discrepancies in the results.  Moreover, uncertainties in some of the model
25     assumptions are trivial compared with the major uncertainties introduced by the interspecies and
26     high-to-low dose extrapolations.
27            In conclusion, because of the high uncertainty in extrapolating 1,3-butadiene cancer risks
28     from rodents to humans and the existence of good-quality occupational epidemiology data with
29     exposure measures, the epidemiology-based risk estimates presented at the beginning of this
30     chapter are the preferred human risk estimates. The rodent-based estimates are presented
31      primarily for comparison purposes. Realizing that different quantitative methodologies and
32     assumptions were used to calculate the various risk estimates, recall that the estimated upper
33     bound (95% UCL) on human incremental lifetime unit cancer risk from continuous 1,3-butadiene
34     exposure was 6 x 10'2/ppm based on the female rat tumors, 3 x 10'Vppm based on the female
35     mouse tumors, and 2 x 10'2/ppm and 6 x 10'3/ppm based on lymphocytic lymphomas in female

        1/28/98                                  9-26      DRAFT-DO NOT CITE OR QUOTE
-------
         and male mice, respectively (lymphocytic lymphomas being the tumor site that most closely
         resembles the lymphohematopoietic cancers observed in male workers exposed to 1,3-
         butadiene).  The best estimate (MLE) of human incremental lifetime unit cancer risk extrapolated
  4      from the leukemias observed hi occupational epidemiology studies was 9 x 10"3/ppm.

  5      9.3. REPRODUCTIVE AND DEVELOPMENTAL TOXICITY
  6      9.3.1.  Introduction
  7            The reproductive and developmental effects of 1,3-butadiene are among the effects (both
  8      cancer and noncancer) observed at the lowest exposure levels following short-term or chronic
  9      inhalation exposure.  Data on reproductive and developmental effects were available from three
 10      types of studies for modeling and calculation of a benchmark concentration (BMC).  In the first
 11      type of study, developmental toxiciry of 1,3-butadiene was evaluated in studies in mice and rats
 12      that included 10-day exposures via inhalation at 0, 40, 200, and 1,000 ppm on gestation days (gd)
 13      6-15 for 6 h/day (Hackett et al., 1987a, b).  In rats, no effects were detected at any exposure level
 14      for developmental toxicity (200 ppm was the NOAEL for maternal toxicity), while reduced fetal
 15      weights were seen in mice at all exposure levels (Table 9-13).  Thus, 40 ppm was considered a
 16      LOAEL for mice.
J 7            In the second type of study, male-mediated effects of 1,3-butadiene were evaluated in a
         dominant lethal study hi which CD-I mice were exposed to 0,12.5, or-1,250 ppm for 6 h/day, 5
 1 9      days/week, 10 weeks (Anderson et al., 1993,  1995). One group of females at each exposure level
 20      was killed on gd 17, while another was allowed to litter. At 12.5 ppm, the frequency of late
 21      deaths and congenital abnormalities on gd 17 were increased, while hi litters allowed to deliver
 22      their pups, changes in implantation numbers,  postimplantation loss, litter size, and weight at birth
 23      and at weaning were significantly different only at 1,250 ppm. In addition, body weights of Fj
 24      males at all tune points and of F, females at several time points between 8 and 71 weeks of age
 25      were significantly increased above controls at both 12.5 ppm and 1,250 ppm. Based on the data
 26      from animals killed on gd 17, there was no NOAEL for dominant lethal effects in the study
 27      (Table 9-14).
 28            In the third type of study, reproductive effects of 1,3-butadiene were seen in lifetime
 29      studies in mice after chronic inhalation exposure to 6.25,20, 62.5,200, and 625 ppm for 6 h/day,
 30      5 days/week (NTP, 1993). The lowest exposure level studied in mice (6.25 ppm) showed
 31      increased ovarian  atrophy and was considered a LOAEL (Table 9-15).  Minimal data from
 32      studies on rats suggested their lesser sensitivity to chronic exposure than for mice in that effects
 33      on fertility were noted only at high exposure levels (600 ppm and above).
        1/28/98                                  9-27       DRAFT-DO NOT CITE OR QUOTE
-------
       Table 9-13. Prenatal (developmental) toxicity study (Hackett et al., 1987b)
Species/strain:      Pregnant CD-I mice
Exposure time:     Gestational day (GD) 6-15
Exposure regimen:  6 h/day
Exposure levels:    0,40,200, or 1,000 ppm
Fetal Weight Data
Exposure level
0
40 ppm
200 ppm
1000 ppm
No. litters
18
19
21
20
Mean fetal
weight/litter
1.341
1.282
1.126
1.038
      Table 9-14. Male-mediated developmental toxicity (Anderson et al., 1993,1995)

Species/strain:      CD-I mice, adult males
Exposure time:     10 weeks
Exposure regimen:  6 h/day, 5 days/week
Exposure levels:    0,12.5 ppm, 1250 ppm
Exposure level
0
12.5 ppm
1250 ppm
Exposure level
0
12.5 ppm
1250 ppm
Number exposed
25
25
50
Mean litter size at
birth
(no. litters)
12.22(18)
11.14(21)
9.06 (33)
No. implants
(no. preg. females)
12.09 (23)
12.75 (24)
10.68 (38)
Mean no. implants
(no. litters)
12.81 (16)
12.35 (17)
10.47 (32)
% Early and late
deaths
4.68
7.52
22.91
% Post-
implantation loss
4.88
9.05
23.88
% Live implants
94.6
92.2
76.8
Mean litter size at
weaning
(no. litters)
12.17(18)
10.95 (20)
9.03 (33)
1/28/98
9-28
DRAFT-DO NOT CITE OR QUOTE
-------
     w  s
     .0  ca
       CN
     m


     ffl
O\
S   I
S    T7-^
CK
o
O
o
VO

o"
>
"5.
o
<*-»
03
.2
'C
C3
6



1
O
e2«
***
"*
°
•s
e

1
03
X
CU
6
Z
•a
-4-»
0

s»
°





•a
a>
s
S
03
X

6




•o

ju
£
s
CA
O
&,.
X
U

vo
^^
oo





Os





o









o








0










o




o



00

CO
CO




OS





o









o








!










1




a.
a.
CN



j^
VO
VO
VO




00





o









o








!










1
t




1
o
o
0
CN


o
p
^
oo




o
>n





o
Os









o








o










o




1
o
""3
CN
^o


o
C5
vo
oo




o
•o





t~~









o








o
O\










o




1
o
-o
0
o
CN


^.
en
£-^
oo




Os





o
o









CN








o
o
^^









00




a.
o.
o
tz>
CN















s
_o
•B
-i
s
12
J 4)
fl
&•
^
h'arian f
O .
















2
o
s

o

^_^
22

en


oo
OS

. o


^^^
§,
•VO

g


o


o1
Os
OS




o

o

1
o
o
o
o
CN

5
en
_
VI

r-


c?f
Os

O


^\
O
**~^s
CN

0


o

/_v
o
r— (
OO




o

o

o,
o
p
CN
 1/28/98
                               9-29
DRAFT-DO NOT CITE OR QUOTE
-------
a

'•£5
C
O
w
G\
V)

u
• «*
H



I
u
Cv














J-?
o
•»-•
t.
—
"3
.w
11
•o
e
CS
o
_c
c>
1
CS
Is
Q
o
u
s
s
"3


















£
CS
o
£
A
ex
2
**••
C5

k<
_cs
"3
en

H


o
£
^
c.
o
IM
•4^
cs

—
"a
en
-------
  1             In conclusion, each of these three types of studies indicated the potential for 1,3-
 .2      butadiene to affect reproduction and development in mice at low levels of exposure.
 W                                '    •       ••',-          i
  3      9.3.2. Fetal Weight Modeling
  4             Fetal weight data (Table 9-13) were fit using a log-logistic model for developmental
  5      toxicity, as described by Allen et'al. (1994a). The TERALOG model software (ICF Kaiser
  6      International, KS Crump Group) was used for this purpose. This model allows for nesting of
  7      fetal data within litters and takes into account intralitter correlations and litter size. To apply this
  8      model, the individual fetal weights were converted to dichotomous data using two different
  9      values as the cutoff for defining an adverse level of response:
10             (1) a decrease below the 5th percentile of the control distribution, or
11             (2) a decrease below the 10th percentile.
1 2      The model was used to estimate: (a) the EC05* and the LEC05** associated with a 5% additional
13      risk of obtaining a fetal weight below the 5th percentile of the controls, or (b) the EC10 and LEC10
14      associated with a 10%  additional risk of obtaining a fetal weight below the 1 Oth percentile of
15      controls, based on Kavlocketal. (1995). The model can be expressed as:

16                    P(d, s) = a + 0,s + [1 - a - 6ls]l{ 1 + exp[/? + 62s - y log (d-d0)]}

1 7      where P(d, s) is the probability of a low-weight fetus at dose d and litter size s, and the
1 8      parameters a, /?, y, 8l3  and 02 are estimated by methods of maximum likelihood.  In order to get
19      an acceptable fit, an intercept parameter (d0~) was included in the model (sometimes referred to as
20      a threshold parameter,  i.e., the point at which the model can no longer distinguish from
21      background). The parameter constraints were: d0 > 0; y > 1; 0 < « - 0^ <, 1.
22             Fetal weight also was modeled as the average of mean fetal weights per litter using the
23      continuous power model (Allen et al, 1994b). The THWC model software (ICF Kaiser
24      International, KS Crump Group) was used for this purpose. Several cutoff values were used,
25      based on Kavlocketal. (1995):
26             (1) a 5% reduction in mean fetus weight/litter from the control mean,
               The EC is the effective (exposure) concentration associated with a given level of risk,
        5% in this case.
               "The LEG is the lower confidence limit on the effective concentration associated with a
        given level of risk. The LEG is also known as the benchmark concentration.
        1/28/98                                   9-31       DRAFT-DO NOT CITE OR QUOTE
-------
  1             (2) a reduction in mean fetus weight/litter to the 25th percentile of the control
  2              distribution, and
  3            (3) a reduction in mean fetus weight/litter to 0.5 SD below the control mean.

  4            The continuous power model can be expressed as:

  5                                      md
  6     where m(d) is the mean of the mean fetus weight/litter for dose d, and a, /?, and y are parameters
  7     estimated by maximum likelihood methods. The parameter constraints were: « ^ 0; y ^ 1.
  8            Goodness of fit was determined by a %*• test for the log-logistic model, and by an F test
  9     for the continuous power model (U.S. EPA, 1 995, Appendix A).  The model was considered to
1 0     provide an acceptable fit if the p value was greater than 0.05 and a graphical display of the data
1 1      showed a good fit of the model.
12            A third approach used to model fetal weight data was the hybrid approach proposed by
1 3     Gaylor and Slikker (1990) and further developed by Crump (1995). The BENCH_C model
1 4     software  (ICF Kaiser International, KS Crump Group) was used for this purpose. This approach
1 5     uses all of the information contained in the original observations by modeling changes in mean
1 6     response  as a function of exposure concentration, but defines ECs and LECs in terms of
1 7     probability  of response. The continuous data are fit using a model that incorporates parameters
1 8     from the  quantal model. Several models are possible within the software for both continuous
1 9     data and quantal risk estimates.  For this study, the log-logistic model (not including litter size)
20     was used for the quantal risk estimates and the following model for the continuous portion of the
21      hybrid model:

22                        m(d) = m(0) + o[N-'(l-P0) - N-H(l-P0)[-l/[l+(^)]]}]

23     where N is the standard normal distribution function, m(d) is the mean response at dose d, a is
24     the standard deviation of the response fixed for all dose groups, and /? and k are the log-logistic
25     model parameters estimated by the maximum likelihood method. The parameter constraints
26     were: kz !;/?;> 0.
27            Crump (1995) indicated that a background rate (P0) of 5% and an EC corresponding to
28     1 0% additional risk corresponds to a change from the control mean of 0.61 SD. Since a change
29     in mean fetal weight/litter of 0.5 SD corresponded on average to a NOAEL in studies by Kavlock
30     et al. (1995), a P0 of 0.05 and an EC10 (10% additional risk) were used here.
        1/28/98                                  9-32       DRAFT-DO NOT CITE OR QUOTE
-------
  1             Results of the modeling approaches for fetal weight are shown in Table 9-16 and Figures
  2      9-6 to 9-8.  The log-logistic model resulted in an adequate fit of the data. Since the log-logistic
 *3      model requires converting continuous data to quantal responses, the continuous power model
  4      was also applied, but did not give an adequate fit with all four exposure levels. When fit to the
  5      first three exposure levels, an adequate fit was obtained. The continuous power model gave
  6      similar ECs and LECs but these were somewhat larger than those obtained with the log-logistic
  7      model except for the one based on a cutoff using the 25th percentile. The hybrid approach
  8      resulted in a quantal estimate of dose at the LEC10 that was lower than that for either the log-
  9      logistic or continuous power model.
10             All three models have strengths and limitations that must be considered. The log-logistic
11      model accounts for intralitter correlation and litter size, but requires conversion of continuous
1 2      data to quantal responses. Neither the continuous power nor the hybrid model are currently
13      structured to account for intralitter correlation or litter size. The version of the hybrid model
14      used here does not allow use of the standard deviation (a) for individual dose groups, so the a at
1 5      dose d0 was used for all dose groups. The continuous power and hybrid models take advantage
1 6      of the power of modeling the continuous data, but the hybrid model expresses the EC and LEC
17      as a quantal estimate of risk, allowing direct comparison with ECs and LECs for quantal
18      endpoints.  Given the various advantages and limitations of these models, the hybrid model is:
        considered the preferred approach for modeling continuous data.
               Table 9-16. Fetal weight modeling (LOAEL = 40 ppm)
Model
Log-logistic (l-4)a
Continuous power
(l-3)a
Hybrid model3
(1-4)
Response
Individual fetal weight
Mean fetal
weight/litter
Mean fetal
weight/litter
Cutoff
5th percentile
10th percentile
5% relative
reduction
25th percentile
0.5 SD absolute
reduction
P0 = 0.05
EC
EC05 = 46.85
EC10 = 49.69
65.08
45.10
50.99
EC10 = 28.19
LEC
LEC05 =
27.02
LEC10 =
38.89
53.51
36.66
41.44
LECjQ =
13.67
/>-Value
0.079
0.067
0.77
0.08
        "Exposure levels modeled in each case are shown in parentheses.
        1/28/98
9-33
DRAFT-DO NOT CITE OR QUOTE
-------
                                                                  • Observed P(d)
                                                                   Predicted P(d)
                200       400       600
                        EXPOSURE (ppm)
        800
        1000
Figure 9-6. Observed versus predicted dose (exposure) probability P(d) of fetal weight
reduction below the 10th percentile of controls using log-logistic model.
                                                                      - Observed
                                                                      - Predicted
                     50           100          150
                            EXPOSURE (ppm)
                      200
Figure 9-7.  Observed versus predicted mean fetal weight per litter using continuous model.
1/28/98
9-34
DRAFT-DO NOT CITE OR QUOTE
-------
        o
        o
        "=
        Q.
        (O
        CO
        HI
        UJ
        o
        OS
        UJ
        Q.
100
                                                                        Observed
                                                                        Predicted
            20--
             0
                          200        400        600
                                    EXPOSURE (ppm)
                                                800
           1000
        Figure 9-8. Observed versus predicted percent of mean fetal weights per litter less than the
        5th percentile of controls (P0 = 0.05) using hybrid model.
  1      9.3.3. Male-Mediated Developmental Toxicity Modeling
  2             Several endpoints from animals killed at gd 17 and after birth were modeled using a log-
  3      linear model:
  4                                      y(d) = a+/3x [ln(l + d)J

  5      This model was used because of the wide spacing of doses and the lack of linearity in the dose-
  6      response relationship. The data were limited in that only two exposure levels in addition to
  7      controls were used, and the exposure levels differed by two orders of magnitude.
  8             Although a statistically significant effect was noted at 12.5 ppm and 1,250 ppm for the
  9      incidence of late deaths in the original paper (Anderson et al., 1993), the response in late deaths
10      at the higher exposure level was lower than at 12.5 ppm, probably because there were so many
11      early deaths at the higher level. For the same reason, the incidence of congenital abnormalities
12      was higher at 12.5 ppm than at 1,250 ppm. When early and late deaths were combined, a
1 3      consistently increasing response with increasing exposure level was seen. When combined, the
14      incidence in the controls was increased from 0 to 13 (4.68% of total implants), while in the 12.5
        ppm group the incidence increased from 7 (2.29%) to 23 (7.52%).
        1/28/98
                                     9-35
DRAFT-DO NOT CITE OR QUOTE
-------
  1             Unfortunately, fetal weights were not reported in the prenatal portion of the dominant
  2      lethal study, and only total litter weights (which are confounded by the number of live pups)
  3      were reported in the postnatal portion of the study. When mean pup weight per litter was
  4      calculated, there was no difference among F, controls and treated offspring, and in some cases, a
  5      slight increase was seen (data not shown). This is interesting in light of the fact that treated Fj
  6      male and female weights were increased above controls at 8 through 71 weeks of age. No
  7      modeling of these data was conducted.
  8             Table 9-17 shows the results of modeling the dominant lethal data. The ECs and LECs
  9      for both 5% and 10% responses are shown.  The log-linear model gave a good fit for all the data
10      except for the number of implants in the prenatal study (see Figures 9-9 to 9-15; note that "dose"
11      refers to 24-h adjusted exposure). This apparently was due to the fact that the number of
12      implants was somewhat higher in the 12.5 ppm group than in controls or the 1,250 ppm group.
13      Given that these data are from fetuses or pups within litters, it is likely that an EC05 and LEC05
14      can be estimated from the data with some degree of reliability.  Also, based on the studies of
15      Allen et al.  (1994a and b), the  LEC05 (BMC05) for such endpoints was similar to the NOAEL on
16      average.  Although certain endpoints not modeled here (late fetal deaths and congenital
17      malformations) were statistically increased in both the 12.5 ppm and 1,250 ppm exposure
               Table 9-17. ECs and LECs for male-mediated developmental toxicity3
Prenatal data
Estimate
EC05
LEC05
ECIO
LEC10
j?-Value
NOAEL
No.
implants
0.21
0.12
0.47
0.26
0.12
220
ppm
Early and
late deaths
3.4
2.4
18
10
0.66
2.2
ppm
Live
implants
3.5
2.4
19
11
0.99
2.2
ppm
Postnatal data
No.
implants
0.12
0.08
0.26
0.17
0.95
220
ppm
Post-
implantation
loss
3.2
2.2
16
9.0
0.99
2.2
ppm
Litter
size at
birth
0.1
0.07
0.20
0.15
0.54
2.2
ppm
Litter size
at weaning
0.1
0.07
0.20
0.15
0.45
2.2
ppm
        'Exposures were adjusted to 24-h daily exposures (e.g., 12.51  6) 15 ] = 2.2 ppm).
                                                     \ 24/ \ 11
        1/28/98
9-36
DRAFT-DO NOT CITE OR QUOTE
-------
                                                                   - Observed Mean
                                                                   - Predicted Mean
                         2
         3       4
    LN(1+ DOSE)
Figure 9-9. Observed versus predicted mean number of implants (prenatal) using log-
linear model.
                                                              - Observed Proportion
                                                              • Predicted Proportion
           0
2      3      4
   LN(1+ DOSE)
5
Figure 9-10. Observed versus predicted proportion of early and late deaths per
implantation (prenatal) using log-linear model.
1/28/98
              9-37
   DRAFT-DO NOT CITE OR QUOTE
-------
                                                              - Observed Proportion
                                                              - Predicted Proportion
        01       23456
                          LN(1+ DOSE)
Figure 9-11. Observed versus predicted proportion of live implants (prenatal) using log-
linear model.
to
u.
O
t£

    14
12--

10--
 8--
 6--

 4-

 2-
                  - Observed Mean
                  - Predicted Mean
                            LN(1+ DOSE)

Figure 9-12. Observed versus predicted mean number of implants (postnatal) using log-
linear model.
 1/28/98
                                   9-38
DRAFT-DO NOT CITE OR QUOTE
-------
          0
2             4
 LN(1+ DOSE)
                                                             Observed Proportion
                                                             Predicted Proportion
Figure 9-13. Observed versus predicted proportion of post-implantation losses (postnatal)
using log-linear model.
                                                                    -Observed Mean
                                                                    - Predicted Mean
                        23456
                             LN(1+ DOSE)

Figure 9-14. Observed versus predicted mean litter size at birth using log-linear model.
1/28/98
               9-39
DRAFT-DO NOT CITE OR QUOTE
-------
                                                                              Observed Mean
                                                                             - Predicted Mean
               0
234
   LN(1+ DOSE)
        Figure 9-15. Observed versus predicted mean litter size at weaning using log-linear model.

 1      groups, no other endpoints showed a statistically significant increase at 12.5 ppm by pairwise
 2      comparison.  However, there was a trend toward an increase in the incidence of early and late
 3      fetal deaths and percent postimplantation loss, and a decrease in percent live implants and litter
 4      size at birth and at weaning in the 12.5 ppm exposure group.  Given the overall effect seen on
 5      development in this study, the NOAEL for most endpoints was considered to be much closer to
 6      12.5 ppm than to 1,250 ppm. Since litter size at birth and at weaning showed the lowest ECs and
 7      LECs, these endpoints will be used for calculation of an RfC.

 8      9.3.4. Ovarian, Uterine, and Testicular Atrophy Modeling
 9            The quantal Wiebull model was used initially to model all data. In cases where this
10      model did not provide a good fit of the data, a log-logistic model was used.  The  15-month and
11      chronic ovarian atrophy data could not be fit adequately using the quantal Weibull model. A log-
12      logistic model similar to that used for fetal weight (setting 02S and 02S to zero) was found to fit
13      the data well. The model was run to determine the probability of additional  risk and extra risk.
14      Goodness of fit was determined by a £ test.  The model was considered to give a good fit if the p
15      value was greater than 0.05 and a graphical display of the data showed a good fit of the model.
16            An attempt was made to model various levels of severity in the lesions seen, based on the
17      data shown in Table 9-15. The data for moderate lesions were fit using the quantal Weibull
18      model (Allen et al., 1994b) for dichotomous data. This model can be expressed as:
        1/28/98
                9-40
DRAFT-DO NOT CITE OR QUOTE
-------
        where P(d) is the probability of response at exposure level d and a., j3, and y are parameters that
 3      are estimated from the observed dose-response data. Parameter constraints were a z 0; p ^ 0; y
 4      > 0.  The model was run to determine the probability of additional risk. Goodness of fit was
 5      determined by a %* test.  The model was considered to provide an acceptable fit if the/? value was
 6      greater than 0.05 and a graphical display of the data showed a good fit of the model.
 7             Table 9-18 gives the results of fitting the log-logistic model to the 2-year ovarian atrophy
 8      data for exposure groups 1-5 and 1-4. The model gave a poor fit for all six exposure groups,
 9      because of leveling off of the response at exposures above 62.5 ppm (36 ppm adjusted for
10      continuous exposure). The best fit of the model was for exposure groups  1-4, although the model
11      also fit exposure groups  1-5 well (Figure 9-16; exposures  adjusted for continuous exposure), and
12      the ECi0s and LEC10s  obtained for groups 1-4 and  1-5 were similar. As expected, LECi0s were
1 3      lowest for ovarian atrophy at 2 years. Moderate ovarian atrophy at 2 years also was modeled
14      using the quantal Weibull model with exposure groups 1-5 or 1-4. The EC10 and LEC10 were
15      higher than those for all lesions. Ovarian atrophy data for all six exposure groups at 9 and 15
16      months were fit using the quantal Weibull or log-logistic model.
               Table 9-18.  ECs and LECs for ovarian, uterine, and testicular atrophy using the
               quantal Weibull and log-logistic models3
Endpoint
Ovarian atrophy - 2 years
Ovarian atrophy - 2 year
Moderate lesions only
Ovarian atrophy - 15 mos
Ovarian atrophy - 9 mos
Uterine atrophy
Testicular atrophy
Model
Log-logistic (l-5)b
Log-logistic (1-4)
Quantal Weibull (1-5)
Quantal Weibull (1-4)
Log-logistic (1-6)
Quantal Weibull (1-6)
Quantal Weibull (1-6)
Quantal Weibull (1-6)
NOAEL/LOAEL
l.lppm(LOAEL)
1.1 ppm
1.1 ppm
11 ppm
1 1 ppm
36 ppm
EC10
0.32
0.29C
0.27
0.24C
3.02
2.31
2.10
20.04
29.37
40.59
LEC10
0.22
0.21°
0.18
0.1 T
2.35
1.67
0.72
9.95
18.43
25.64
/7-Value
0.11
0.96
0.55
0.96
0.66
0.83
0.66
0.55
        "Exposures were adjusted for continuous exposure (e.g., 6.25 / 6 \UL\= 1-1
        bExposure levels included in the model.                 124/ \7
        °Extra risk. All other values are estimates of additional risk.
                ppm)
        1/28/98
9-41
DRAFT-DO NOT CITE OR QUOTE
-------
Q
Ul
&
m
u.
%
H-
        o
        o;
        in
        a.
                                                                                    - Observed
                                                                                    - Predicted
                            50
                               100        150
                              EXPOSURE (ppm)
 200
250
        Figure 9-16.  Ovarian atrophy (groups 1-5) using log-logistic model.
  1             Uterine and testicular atrophy data also were modeled using the quantal Weibull model.
  2      The quantal Weibull model resulted hi an acceptable fit of the 2-year uterine atrophy and
  3      testicular atrophy data (Table 9-18 and Figures 9-17 and 9-18; exposures adjusted for continuous
  4      exposure). However, the ECi0s and LEC]0s were much higher for these endpoints than for 9-
  5      month, 15-month or 2-year ovarian atrophy data. LEC10s were estimated because it has been
  6      shown (Allen et al., 1994b) that, for quantal responses, the LEC10 is near or below the range of
  7      detectable responses.  Also, the Proposed Guidelines for Carcinogen Risk Assessment (EPA,
  8      1996) propose use of an LED10 as the default point of departure for low-dose extrapolation, and
  9      use of an LECj0 as a default for noncancer estimation of an RfC would be consistent with this
10      approach.
11            Although some 9- and 12-month interim sacrifice data were available for ovarian, uterine,
12      and testicular atrophy (Table 9-15), these were less than ideal for modeling because smaller
13      numbers of animals were killed and not all dose groups were represented.  In addition, some
14      animals died or became moribund and were killed before the 2-year death time point.  To account
15      for the variability in time of death, time-to-response analyses were done using the multistage
16      Weibull model as discussed in Section 9.2.2.2. Exposures were adjusted to the equivalent
        1/28/98
                                         9-42
DRAFT-DO NOT CITE OR QUOTE
-------
 Q
 UJ
 UJ
 u.
 u.
 111
 o
 tt
 UJ
 0.
                  20
 40       60       80


     EXPOSURE (ppm)
                                                                         - Observed

                                                                         - Predicted
                                                     100
 120
 Figure 9-17.  Uterine atrophy (groups 1-6) using quantal Weibull model.
Q
UJ


o
Ul
LL
U.
UJ
o
(£
UJ
Q.
                                                                         - Observed

                                                                         - Predicted
                 20
40       60       80

   EXPOSURE (ppm)
                                                     100
120
Figure 9-18. Testicular atrophy (groups 1-6) using quantal Weibull model.
1/28/98
             9-43
                                                  DRAFT-DO NOT CITE OR QUOTE
-------
 1      continuous lifetime exposures. An EC10 and an LEC10 were calculated in each case. All the
 2      reproductive responses were treated as incidental, not fatal. Parameter estimates for each
 3      reproductive endpoint are presented in Table 9-19.
 4            Results of the Weibull time-to-response model are shown in Table 9-20.  The ECs and
 5      LECs were similar to those from other models used for ovarian atrophy, uterine atrophy, and
 6      testicular atrophy, with the exception of those from the'modeling of testicular atrophy including
 7      the highest exposure group, for which the Weibull time-to-response model yields results roughly
 8      five times lower than the quanta! Weibull model. The quantal Weibull model results for uterine
 9      and testicular atrophy were for additional risk, while the Weibull time-to-response results were
10      for extra risk; however, because of the low background rates of both uterine and testicular
11      atrophy, additional risk and extra risk should be nearly the same. The results of the time-to-
12      response model are used in the calculation of an RfC.
13            The time-to-response model also allows for the calculation of risks at ages less than full
14      lifetime.  Thus, if one is concerned about ovarian or uterine atrophy primarily during a woman's
15      reproductive years, one can calculate corresponding EC10s and LEC10s. Assuming reproductive
1 6      capabilities until 45 years of age yields EC10 = 1.3 ppm and LEQo =  1.1 ppm for ovarian atrophy
17      (625 ppm dose group excluded) and EC10 = 31 ppm and LEC10 = 22 ppm for uterine atrophy (625
18      ppm group included).
               Table 9-19.  Parameters for Weibull time-to-response model used to model
               reproductive effects observed in mice based on ppm butadiene exposure1
Response
Ovarian
atrophy
Uterine
atrophy
Testicular
atrophy
625 ppm
group
included
no
yes
no
yes
no
yes
QO
4.86 x lO'6
9.01 x IQ-7
6.73 x IQ-5
9.08 x 10-5
4.28 x 10-4
1.60 x 1Q-4
Ql
7.06 x lO'6
1.32x lO'6
5.28 x lO'5
9.74 x lO'6
2.24 x lO'5
1.52 x 10-4
Q2
-
-
-
1.31 x ID'6
-
-
Z
2.21
2.58
1.0
1.0
1.0
1.0
        'Each response was considered to be incidental with induction time, T0=0. See Section 9.2.2.2 on time-to-tumor
         modeling of the mouse carcinogenicity data for a discussion of the Weibull model structure and selection.
         1/28/98
9-44
DRAFT-DO NOT CITE OR QUOTE
-------
               Table 9-20.  Human benchmark 1,3-butadiene exposure concentrations
               calculated for reproductive effects observed in mice using a Weibull
               time-to-response model (extra risk)
Response
Ovarian
atrophy
Uterine
atrophy
Testicular
atrophy
625 ppm
group
included
no
yes
no
yes
no
yes
Based on ppm butadiene exposure
ECio
0.497
0.473
18.8
24.0
44.3
6.54
LEC10
0.382
0.369
12.0
15.6
15.9
5.39
  1
  2
  3
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
9.3.5.  Summary and Conclusions
       ECs and LECs were estimated for three types of exposure scenarios to 1,3-butadiene
based on different endpoints:
       1.  Short-term exposure (10 days)—fetal weight reduction
       2.  Subchronic exposure (10 weeks)—male-mediated developmental toxicity
       3.  Chronic exposure—ovarian, uterine and testicular atrophy
These analyses demonstrate approaches for estimation of ECs and LECs based on continuous
and quantal data.
       Results of the fetal weight analysis illustrate how both continuous and quantal modeling
approaches can be used for continuous data. All of the LECs calculated were below the LOAEL
of 40 ppm, except for two LECs calculated using the continuous power model, which were near
this value.  Since the hybrid modeling approach is considered the preferred method for modeling
continuous data, the LECIO  of 13.7 ppm from this model will be used for calculating the
reference concentration for developmental toxicity for short-term exposure (RfCDT).
       Results of the analysis for male-mediated developmental toxicity following 10 weeks of
exposure gave ECs and LECs much lower than those from the 10-day exposures based on fetal
weight. Therefore, the LEC10 for the dominant lethal study will be used to calculate an RfC for a
subchronic exposure scenario.
       Modeling of the 2-year ovarian atrophy data, the effect occurring at the lowest chronic
exposure level, gave a good fit with the log-logistic model, but only when the highest exposure
        1/28/98
                                        9-45
DRAFT-DO NOT CITE OR QUOTE
-------
 1      level was dropped. This approach was justified because the responses leveled off for the top
 2      three exposure groups.  The LECs derived for a 10% increase in additional risk or extra risk were
 3      5- to 6-fold below the LOAEL of 6.25 ppm. When the time-to-response model was applied to
 4      account for interim sacrifice data and early mortality, an LECIO of 0.38 ppm (extra risk) was
 5      calculated, a value similar to that using the log-logistic model.
 6             Ovarian atrophy has been shown to be related to the amount of the diepoxide metabolite
 7      in the tissue (Doerr et al., 1996). Uterine  atrophy may be secondary to ovarian atrophy, and thus
 8      may also be related to the amount of diepoxide metabolite formation. Modeling of the ovarian
 9      atrophy and uterine atrophy data was considered based on internal dose of the diepoxide
10      metabolite. However, an adequate pharmacokinetic model was not available to estimate levels of
11      the diepoxide metabolite (Chapter 8).
12             RfC calculations will be made for both ovarian atrophy, the reproductive effect occurring
13      at the lowest chronic exposure level, and testicular atrophy, the reproductive effect observed in
14      male mice following chronic exposure.

15      9.4. REFERENCE CONCENTRATIONS FOR REPRODUCTIVE AND
16          DEVELOPMENTAL EFFECTS
17      9.4.1.  Introduction
18             As discussed in Chapter 5 and Section 9.3, a variety of reproductive and developmental
19      effects have been observed in mice and rats exposed to 1,3-butadiene by inhalation. (There are
20      no human reproductive or developmental data available for 1,3-butadiene.) In this section,
21      sample reference concentrations (RfCs) are calculated for the most sensitive reproductive and
22      developmental endpoints, i.e., those effects exhibiting responses at the lowest exposure
23      concentrations from various exposure scenarios, using both the traditional NOAEL/LOAEL
24      approach and the "benchmark dose" approach (Crump, 1984).  A reference concentration (or
25      dose) is an estimate of a daily exposure to humans that is "likely to be without an appreciable
26      risk of deleterious [noncancer] effects during a lifetime" (Barnes et al.,  1988). The final reported
27      RfC will be based on the endpoint resulting in the lowest calculated RfC level. This RfC will be
28      solely an RfC for reproductive and developmental effects (R/D RfC) and not a true RfC because
29      other noncancer endpoints were not considered.

30      9.4.2.  Calculation of RfCs
31             The most sensitive developmental effect was decreased fetal weight in the mouse.  The
32      most sensitive reproductive effects observed in subchronic exposure studies were decreased litter
33      size at birth and at weaning in dominant lethal studies of mice (i.e., male mice are  exposed to

        1/28/98                                  9-46       DRAFT-DO NOT CITE OR QUOTE
-------
         1,3-butadiene and effects on litters are measured after mating to unexposed females). These
         effects are highly correlated and both yielded the same modeled effective dose results (Table 9-
         17).  Litter size at birth reflects both decreased implants and increased fetal deaths, while litter
  4      size at weaning also reflects neonatal deaths. Dominant lethal effects in humans would likely be
  5      manifested as spontaneous abortions, miscarriages, stillbirths, or very early deaths. From chronic
  6      exposure studies (2-year bioassays), the most sensitive reproductive endpoints were ovarian
  7      atrophy in female mice and testicular atrophy in male mice.
  8            Table 9-21 summarizes the EC)0 (i.e., the exposure concentration resulting in a 10%
  9      increase in risk based on modeling the exposure-response data in the observable range), the
 10      LEG]0 (i.e., the 95% lower confidence limit on the exposure concentration estimated to result in
 11      a 10% increase in risk), and the NOAEL (i.e., no observed adverse effect level) or LOAEL (i.e.,
 12      lowest observed adverse effect level; reported when no NOAEL was observed) for these 1,3-
 13      butadiene-induced effects. Table 9-21 also provides sample calculations of RfCs using the
 14      NOAEL (or LOAEL) as well as the LECj0 as "points of departure." Uncertainty factors are then
 15      applied to the "point of departure" to calculate the RfC.
 16            Typically, a factor of 10 is used for interspecies uncertainty when the "point of departure"
 17      is based on nonhuman data; however, when ppm equivalence across species is assumed as was
 18      done here,  a factor of 3 is used instead. Thus, in Table 9-21, an interspecies uncertainty factor of
         3 was used for all endpoints except ovarian atrophy. For ovarian atrophy, there is convincing
 20      evidence that the diepoxide metabolite (1,2:3,4-diepoxybutane, DEB) is required to elicit the
 21      effect and,  while the differences cannot be quantified without an adequate physiologically based
 22      pharmacokinetic (PBPK) model, it is expected that humans produce lower concentrations of
 23      DEB than mice, based on differences in metabolic rates. Thus, an uncertainty factor of 1.5 was
 24      used for ovarian atrophy to account for differences between mice and humans in the amount of
 25      DEB produced, yet allow that humans may be more sensitive to DEB.
 26            A large degree of human variability has been observed in metabolic activities that could
 27      affect 1,3-butadiene toxicity. For example, Seaton et al. (1995) measured a 60-fold variation in
 28      the initial rate of oxidation of 1,2-epoxy-3-butene (EB) to DEB in microsomes from 10 different
 29      human livers. However, overall variability in total metabolism and susceptibility is unknown,
 30      thus the conventional intraspecies uncertainty factor of 10 for human variability was used for
31       each endpoint in Table 9-21.
32            With respect to the acute/subchronic-to-chronic uncertainty factor, none was needed for
33      ovarian or testicular atrophy because these effects were based on chronic studies.  No acute-to-
34      chronic uncertainty factor was used for fetal weight either, because only exposures during
        1/28/98                 .                  9-47       DRAFT-DO NOT CITE OR QUOTE
-------
O
a
'•&
a
53
A
*!»
,|
O
42
o
^«*
42
a
S
a
•2
^
•o
a
CJ
CJ
.£;

"3
3
•a
o
2
&
£*
s*
a
CM
0
42
"3
i-5
o»
o
2
&
H
1 I
ws £;
A 2
S S
s
o
•3
•" S "^
y 6 g
PS 2 S

'h
££ a o
,2 13 cj
K 2 iS

o 3^
3 *j '1
W U3 t: u
sill

,
» o -^*
13 "S •—
2 o ^
•* "S .S t* O
3| v gti
JJ 3 O CS


3 >.
? -H
eu j§
S g S
^ e S
*^ ^ *^
c/l ^
'i.1
y» ~n **
o ° £

>S ~ »2

§1


^^
. r H
w§

|}l


5
o








en











*° ,






O




en




2



00
ts




Decreased
fetal weight




en
d








o
•"•





*^_i






o




en











40 (LOAEL)
(10 d, 6h/d,
GD6-15)



0
o
o








m











en






o




en




o



cb
o



x->
lilili
liilli




§
o








o
^^





CO






0




CO











2.2 (LOAEL)
(10 week,
adjusted to 24
h/d)



CO
o
o
0








CO











t— 1






o




^
^



CO
o


•a
O
o




.1^
§!
O K




O
0
0








o
"~"





~*






o




Jo
*^










W o
si W ^



o








CO











»-*






o




CO




^



VI
vd




S3
~3 >>
•2-g.
S S
f- to




ts









_






^H






0




m











S
2
_ ^..i.ss'
en G.^ S



                                                                          •-
                                                                        ix a
                                                                        < *
                                                                        O "3
                                                                               8

                                                                               'i
                                                                               o.

                                                                               o
                                                                         2 2

                                                                         trt T3
                                                                         1 a
                                                                         co «
                                                                         8.5
                                                                         f ^>a
                                                                         KJ Q



                                                                         1-^N
                                                                         S 5 " c

                                                                         1 S|"S
                                                                         o 3 -a •-

                                                                         £ 8 •? §

                                                                         « C "!f r3
                                    o *-» 3
                                                                         8-S
                                                                         O P
                                                                          O
                                                                         *o jr;



                                                                         ?l
                                                                            •sr J3 so i
                                                                            (N o c 'J
                                                                            2 Q e '
                                    < H 13
1/28/98
9-48
DRAFT-DO NOT CITE OR QUOTE
-------
         gestation are relevant. Although dominant lethal effects appear to occur with exposure during a
         specific time period of spermatogenesis (i.e., only certain stages of developing sperm appear
         susceptible), chronic exposure might result in continuous induction of these effects, so a factor of
  4      3 was used.
  5             Under the NOAEL/LOAEL approach, the NOAEL is defined as the exposure level for
  6      which there is no observed adverse effect, although it is circumscribed by the detection limit of
  7      the study.  For endpoints for which there is no NOAEL, an uncertainty factor of 10 is typically
  8      used to attempt to extrapolate from the LOAEL to a level at which there are presumed to be no
  9      detectable effects. In the benchmark dose approach, the typical "point of departure" corresponds
 10      to a 10% increased response level, which is explicitly not a no-effect level.  In this risk
 11      assessment, a risk reduction factor of 3 was used to extrapolate to a level below which no
 12      detectable effects would be expected, analogous to the LOAEL-to-NOAEL uncertainty factor.
 13      Final guidance on this methodology is still being developed by EPA.
 14            In addition to the sample RfCs presented in Table 9-21 for lifetime 1,3-butadiene
 15      exposure, two RfCs were calculated for subchronic exposure. An RfCDT of 0.14 ppm for
 1 6      developmental toxicity from short-term exposures was calculated for decreased fetal weight,
 17      using the same factors depicted in Table 9-21. This RfCDT is identical to the sample RfC
 18      calculated for decreased fetal weight because no subchronic-to-chronic uncertainty factor was
         used in that calculation. Finally, an RfC for subchronic exposure was  calculated for the
 20      decreased litter size endpoints from the subchronic dominant lethal study. Using the LEC10 of
 21       0.15 ppm and uncertainty factors of 3 for interspecies extrapolation, 10 for intraspecies
 22      variability, and 3 for risk reduction (analogous to the LOAEL-to-NOAEL uncertainty factor), as
 23      described above, yields an R/D RfC for subchronic exposure of 0.0015 ppm.

 24      9.4.3. Discussion
 25            Tne EC,0s in Table 9-21 suggest that the dominant lethal (male-mediated) effect is the
 26      most sensitive reproductive/developmental endpoint (i.e., the "critical" endpoint), and thus   •
 27      should be the basis for the final R/D RfC.  The dominant lethal effect also yields the lowest
 28      sample RfC of 0.5 ppb. To arrive at the final R/D RfC, a further uncertainty factor of 3 is used to
 29      account for the lack of comprehensive reproductive testing, especially the absence of a
 30     multigenerational study. This final calculation yields an R/D RfC of 0.15 ppb.
31            There are substantial uncertainties in estimating low-exposure human risks for
32      reproductive and developmental effects observed in animals exposed to high concentrations of an
33      agent. It is generally believed that there is a nonlinear low-dose exposure-response relationship
        for noncancer effects, and perhaps a threshold, although this is difficult to demonstrate

         1/28/98                                    9-49        DRAFT-DO NOT CITE OR QUOTE
-------
 1      empirically.  The shape of this low-dose exposure-response relationship is unclear, however, so
 2      RfCs are calculated for noncancer effects rather than exposure-based risk estimates. The major
 3      uncertainties considered in deriving an RfC include the extrapolation of effects observed in
 4      animals to humans (interspecies extrapolation), the potential existence of sensitive human
 5      subpopulations resulting from human (intraspecies) variability, and various deficiencies in the
 6      database. These areas of uncertainty are addressed to some extent by the uncertainty factors.
 7      Other methodological uncertainties arise in the determination of the "point of departure" and in
 8      the selection of the relevant exposure metric for equating animal exposure-response relationships
 9      to humans.
10             There are a number of limitations in using the NOAEL/LOAEL approach for obtaining a
11      "point of departure"; these have inspired the development of an alternative "benchmark dose" (or
12      concentration) methodology.  Fkst, the NOAEL/LOAEL approach relies on one exposure level
13      and ignores the rest of the exposure-response data. Second, the NOAEL/LOAELs depend
14      explicitly on the specific exposure levels selected for the study. They are also a function of study
15      power because a LOAEL is the lowest exposure level with a statistically significant increase in
16      an adverse effect, whereas a NOAEL could represent an increase that failed to attain statistical
17      significance. Finally, NOAEL/LOAELs are not readily comparable across endpoints or studies
18      because they can refer to different response levels.
1 9             The alternative benchmark concentration approach involves modeling the full exposure-
20      response curve in the observable range and calculating an effective concentration (EC)
21      corresponding to some level of response (e.g., 10%) that can be used as a point of comparison
22      across endpoints and  studies (the 10% effect level is typically at the low end of the observable
23      range, although sometimes a lower level of response can be estimated). The LEC10 is being
24      considered as the default "point of departure" to take into account statistical variability around
25      the ECIO estimate. While the benchmark concentration approach alleviates some of the
26      limitations of the NOAEL/LOAEL approach, there are  still uncertainties regarding the
27      appropriate exposure-response model to use.  It is generally expected that models that provide a
28      good fit to the data in the observable range should yield reasonably similar EC10s, as shown for
29      quantal models by Allen et al. (1994b).
30             As shown in Table 9-21, these two approaches yielded nearly identical RfCs for
31      decreased fetal weight and for ovarian atrophy. For the dominant lethal effect of decreased litter
32      size, the RfCs were similar, with that from the NOAEL/LOAEL approach four-fold higher than
33      that from the benchmark concentration approach. For testicular atrophy, on the other hand, the
34      NOAEL-based RfC is over 20 tunes higher than the LEC10-based RfC. At least part of this
35      discrepancy is likely attributable to the fact that the time-to-response modeling conducted to

        1/28/98                                   9-50        DRAFT-DO NOT  CITE OR QUOTE
-------
        derive the LEC!0 took into account the decreased survival times in the higher exposure groups in
        the chronic study. This had the effect of increasing the effective percent affected in the midrange
        of the exposure-response curve, which otherwise is fairly flat.  This assessment advances the use
 4     of the benchmark dose/concentration approach.
 5            Uncertainties also exist in the choice of exposure metric. Ideally, NOAEL or LOAELs
 6     and LEC10s (or EC10s) should be converted to appropriate human equivalent exposures before
 7     using these exposure levels as "points of departure." Theoretically, this is best accomplished by
 8     using a PBPK model to convert animal exposures to biologically effective doses to the target
 9     organ and then to convert these tissue concentrations back to human exposures to the parent
10     compound. Unfortunately, the current PBPK data and models are inadequate for use in risk
11      assessment; therefore, exposure concentrations of 1,3-butadiene are used as the default exposure
12     metric (this risk assessment assumes equivalence of effects from equivalent ppm exposures
13     across species). For the lifetime chronic exposure study, demonstrating ovarian and testicular
14     atrophy, mouse exposure concentrations were adjusted to human equivalent continuous chronic
15     exposures.
16            For the subchronic and acute studies, however, the appropriate time frame for exposure
17     averaging is unclear. Typically,  daily exposures resulting in nondevelopmental effects have been
        adjusted to an equivalent 24-h exposure, while exposures resulting in developmental effects have
        not been adjusted (U.S. EPA IRIS online database, 1997). Consistent with this approach, 1,3-
20     butadiene exposures resulting in dominant lethal effects have been adjusted to a 24-h exposure,
21      whereas exposure levels from fetal weight studies have not been adjusted. The exposure
22     concentrations for these subchronic and acute effects have not been adjusted to reflect total
23     duration of exposure because the critical time frames are unknown. Thus, for example, a
24     1-day exposure is treated equivalently to a 10-week exposure to the same daily level. Also, for
25     developmental effects, a 4-h exposure to 50 ppm would be treated equivalently to an 8-h
26     exposure to 50 ppm.
27            Finally, there are uncertainties in the uncertainty factors used to derive the RfC from the
28     "point of departure." These factors are largely arbitrary. In particular, the shape of the
29     exposure-response curve below the observable range is unknown,  and it is uncertain that the
30     NOAEL or the LOAEL/10 or the LEC10/3 actually represent no-effect levels, independent of the
31      application of the interspecies and intraspecies uncertainty factors.

32     9.4.4. Conclusions
33            In conclusion, an R/D RfC of 0.15 ppb was calculated for the critical endpoint of the
        dominant lethal effect of decreased litter size at birth (or at weaning), based on mouse data.  This

        1/28/98                                   9-51       DRAFT-DO NOT CITE OR QUOTE
-------
 1      reference concentration, the uncertainties discussed above notwithstanding, is presumed to
 2      represent a daily exposure to humans that is likely to be without an appreciable risk of
 3      reproductive or developmental effects during a lifetime.
 4             In addition, an RfCDT of 0.14 ppm for developmental toxicity from short-term exposures
 5      was calculated based on fetal weight data for mice, and an R/D RfC for subchronic exposure of
 6      0.0015 ppm was obtained based on the dominant lethal results in mice. Each of these RfCs was
 7      calculated using benchmark concentration methodology.

 8      9.5. CONCLUSIONS ON QUANTITATIVE RISK ESTIMATES
 9             In this chapter, a lifetime extra unit cancer risk (MLE) of 9 x 10"3 per ppm of continuous
10      1,3-butadiene exposure was calculated based on linear modeling and extrapolation of the excess
11      leukemia mortality reported in a high-quality occupational epidemiology study. Using this
12      cancer potency estimate, the chronic exposure level resulting in an increased cancer risk of 10"6
13      can be estimated as follows:  (10-6)/(9 x lO'Vppm) = 1 x 1Q-4 ppm = 0.1 ppb.  The 95% UCL on
14      the unit cancer risk was 2 x l O'2 per ppm.
15             A range of human cancer potency estimates from 4x10'3/ppm to 0.29/ppm was also
16      calculated based on a variety of tumors observed in mice and rats exposed to 1,3-butadiene.
17      These risk estimates are considered inferior to those based on the epidemiological data, primarily
18      because of the large uncertainties in extrapolating 1,3-butadiene cancer risks across species in
19      light of the large unexplained differences in responses of rats and mice.
20             In addition, benchmark doses and reference concentrations were calculated for an
21      assortment of reproductive and developmental effects observed in mice exposed to 1,3-butadiene.
22      An R/D RfC of 0.15 ppb was obtained for the critical effect of decreased litter size at birth (or at
23      weaning) observed in dominant lethal studies of mice, using a benchmark concentration
24      approach to obtain the "point of departure."  This R/D RfC is presumed to be a chronic exposure
25      level without "appreciable risk" of reproductive or developmental effects.  Although other
26      noncancer effects were not examined, the reproductive endpoints were quite  sensitive,  and it is
27      likely that the R/D RfC is protective against  other noncancer effects as well.
28             Finally, an RiCDT of 0.1 ppm for developmental toxicity from short-term exposures was
29      calculated from mouse fetal weight data, and an R/D RfC for subchronic exposures of 0.0015
30      ppm was derived from the dominant lethal results hi mice, each using benchmark concentration
31      methodology.
        1/28/98                                  9-52       DRAFT-DO NOT CITE OR QUOTE
-------
                                     10. WEIGHT OF EVIDENCE

        10.1.  EVALUATION
  2            1,3-Butadiene is a colorless, odorless chemical that exists in ambient air in gaseous form.
  3     This extremely volatile chemical is very slightly soluble in water and is not found in soil and
  4     food.  Thus, exposure to 1,3-butadiene is mainly via inhalation. Increased mortality from
  5     leukemias and lymphomas was observed among male workers occupationally exposed to 1,3-
  6     butadiene in polymer and monomer production, respectively.  No information is available in
  7     females.  The data from one Canadian and seven U.S. polymer production plants show that
  8     exposure to 1,3-butadiene is causally associated with occurrence of leukemias (cell type is not
  9     known at this time).
10            Two lifetime inhalation studies in mice and one lifetime inhalation study in rats found
11      occurrence of malignant tumors in multiple sites in both mice and rats. Increased occurrence of
12     lymphomas in a 1-year inhalation study in Swiss mice indicated that the presence of retro virus
13     was not an essential factor for the development of 1,3-butadiene-induced lymphomas.
14            Once inhaled, 1,3-butadiene is distributed throughout the body. The relative distribution
15     of 1,3-butadiene in different organs is unknown at this time.  1,3-butadiene is metabolized by
16     oxidation to a monoepoxide, diepoxide, and epoxy diol.  Which metabolite(s) is responsible for
        the causation of cancer is still uncertain. Differences in measured concentration levels of these
        metabolites in mice and rats do not provide an explanation for the differences observed in
19      malignancies in these two species. All three of these metabolites have been shown to be
20      mutagenic in vivo and in vitro.

21      10.2. CONCLUSION
22             Based on the overall evidence from human, animal, and mutagenicity studies, 1,3-
23      butadiene is concluded to be a known human carcinogen.
        1/28/98                                   10-1       DRAFT-DO NOT CITE OR QUOTE
-------
-------
                                   11. RISK CHARACTERIZATION

        11.1. INTRODUCTION
 2             The U.S. Environmental Protection Agency (EPA) first published a health assessment of
 3      1,3-butadiene in 1985. The 1985 assessment concluded that 1,3-butadiene was a possible human
 4      carcinogen and calculated an upper bound cancer potency estimate of 0.25/ppm based on mouse
 5      data. Since then, a number of new studies on 1,3-butadiene have been completed in various
 6      disciplines such as epidemiology, toxicology, and pharmacokinetics. The purpose of this effort
 7      was to review the new information and determine if any changes were needed to the earlier
 8      conclusions.
 9            This reassessment is intended to serve as a source document for risk assessors inside and
10      outside the Agency.  Its development, however, was prompted primarily by a request from EPA's
11      Office of Mobile Sources (OMS) to support decision making regarding the Air Toxic Rule's
12      Section 202 (1) (2) of the Clean Air Act Amendment. The scope of the document has been
13      limited to address only the health effects specifically requested by OMS: carcinogenicity,
14      mutagenicity, and reproductive/developmental toxicity. Similarly, a detailed exposure
15      assessment was not requested and not conducted. For background purposes, however, some
16      exposure information has been included.
              The major findings of this report are as follows. First, sufficient evidence exists to
        consider 1,3-butadiene a known human carcinogen.  The evidence for this includes findings in
19      epidemiologic studies as well as clear evidence that  1,3-butadiene is an animal carcinogen and is
20      metabolized into  genotoxic metabolites by experimental animals and humans.
21            Second, based on linear modeling of human data, the best estimate of lifetime extra
22      cancer risk from continuous 1,3-butadiene exposure  is about 9 * 10"3/ppm, or 9 x 10"6/ppb. In
23      other words, it is  estimated that 9 persons in 1 million exposed to 1 ppb 1,3-butadiene
24      continuously for their lifetimes would develop cancer as a result of their exposure. Lower
25      cumulative exposures are expected to result in risks that are proportionately lower.
26            Third, although there are no human data on reproductive or developmental effects, a
27      variety of such effects have been observed in mice and rats exposed to 1,3-butadiene. A
28      reproductive/developmental reference concentration (RfC) of 0.05 ppb was calculated based on
29      the critical reproductive effect of reduced litter sizes, reflecting increased prenatal mortality,
30      observed among the offspring of male mice exposed to 1,3-butadiene.
31            Fourth, there are insufficient data to determine if children or other special subpopulations
32      are differentially affected by exposure to 1,3-butadiene. Heavy smokers are likely to be more
        heavily exposed than the general population.


        1/28/98                                   11-1       DRAFT-DO NOT CITE OR QUOTE
-------
  1             This chapter will briefly summarize and integrate the critical data and analyses on which
  2      these findings are based and discuss the strengths and weaknesses of those data and the resulting
  3      confidence hi the findings. With the exception of the section on special subpopulations, all of
  4      the sections in this chapter discuss material presented in the earlier chapters of this assessment.
  5
  6      11.2. EXPOSURE OVERVIEW
  7             Approximately 3 billion pounds of 1,3-butadiene are produced annually hi the United
  8      States.  1,3-Butadiene is used primarily in the manufacture of styrene-butadiene rubber, plastics,
  9      and thermoplastic resins. Environmental releases occur from process vents during these
10      operations. 1,3-Butadiene is not a component of gasoline or diesel fuel, but is formed as a by-
11      product of incomplete combustion. Mobile sources, including both on-road and nonroad
12      engines, are estimated to account for 79% of all 1,3-butadiene emissions (EPA, 1992). 1,3-
13      Butadiene emissions from vehicles are reduced by catalytic converters; total emissions may
14      decline as older cars without converters are removed from service.
15             The compound is highly volatile and slightly soluble in water. Thus, environmental
16      releases result primarily hi emissions to the atmosphere. In the atmosphere, 1,3-butadiene
17      undergoes rapid destruction by photoinitiated reactions, and 50% of it is removed in
18      approximately 6 hours (U.S. DHHS, 1992).  Although it is degraded rapidly in the atmosphere,
19      1,3-butadiene is almost always present at low concentrations hi urban and suburban areas.
20      Because of this, the general population is exposed to some levels via inhalation.  1,3-Butadiene is
21      not found hi significant amounts in food, soil, water, plants, fish, or sediment. Therefore, the
22      predominant pathway of exposure is via inhalation.
23             Monitoring done from 1987 to 1994 by Aerometric Information Retrieval System at more
24      than 20 different urban and suburban locations detected ambient air levels of 1,3-butadiene
25      ranging from 0.22 to 1.02 jig/m3" (0.10 to 0.46 ppb).  Indoor air levels are likely to be higher than
26      ambient levels when smoking occurs. 1,3-Butadiene emissions from cigarettes have been
27      measured to be 200 to 400 [j.g/cigarette, and levels in smoke-filled bars have been found to range
28      from 2.7 to 19 ug/m3 (1.2 to 8.4 ppb) (Lofroth et al., 1989; Brunnemann et al., 1990).
29
30      11.3. CANCER HAZARD ASSESSMENT
31      11.3.1.  Human Evidence
32            Sufficient evidence exists to consider 1,3-butadiene a known human carcinogen.
33             In most situations, epidemiologic data are used to delineate the causality of certain health
34      effects. Several cancers have been causally associated with exposure to agents for which there is
35      no direct biological evidence. Insufficient knowledge about the biological bases  for diseases in
36      humans makes it difficult to identify exposure to an agent as causal, particularly for malignant

        1/28/98                                   11 -2       DRAFT-DO NOT CITE OR QUOTE
-------
  1      diseases when the exposure was in the distant past.  Consequently, epidemiologists and biologists
  >2      have provided a set of criteria supportive of a causal relationship between an exposure and a
  3      health outcome. A causal interpretation is enhanced for studies that meet these criteria. None of
  4      these criteria actually proves causality; actual proof is rarely attainable when dealing with
  5      environmental carcinogens. None of these criteria should be considered either necessary (except
  6      temporality of exposure) or sufficient in itself. The  absence of any one or even several of these
  7      criteria does not prevent a causal interpretation. However, if more criteria apply, it provides
  8      credible evidence for causality. The following discussion addresses the strengths and limitations
  9      of the epidemiologic studies of workers occupationally exposed to 1,3-butadiene, from which the
 10      human evidence is derived, and then summarizes how adequately the causality criteria apply.
 11             The conclusion of "sufficient evidence" of human carcinogenicity is based on more than
 12      10 epidemiologic studies examining five different, groups of workers.  These studies are
 13      summarized in Table 11-1.
 14             The strongest evidence comes from the follow-up study of a cohort of 15,000 synthetic
 15      rubber workers (UAB cohort) conducted by Delzell  et al. (1996) and Macaluso et al. (1996) and
 16      reported in two components. The cohort was derived from seven U.S. plants and one Canadian
 17      plant. The follow-up was from 1943 to 1994. Investigators estimated the exposures to 1,3-
 18      butadiene, styrene, and benzene for each worker (Macaluso et al., 1996).  Quantitative  exposures
         were calculated and limited validation of exposure estimates were attempted by various means.
         Cumulative and peak exposures were calculated for  each worker. Comparison with the U.S.
 21      population resulted in significant excesses for leukemia in ever-hourly workers (43% higher than
 22      general population) and its subcohbrt of blacks (127%)  (Delzell et al., 1996).  Significant
 23      excesses were also found in the ever-hourly subcohort for year of death (87% for 1985+), year of
 24      hire (100% for 1950-59), age at death (79% for <55  years), and for more than 10 years
 25      employment and more than 20 years since hire (92% for whites and 336% for blacks).
 26      Laboratory workers, maintenance workers, and polymerization workers also showed higher risks
 27    '  of 331%, 165%, and 151%, respectively. All these analyses were conducted adjusting  for styrene
 28      and benzene. When internal comparison was carried out using the estimated ppm-years
 29      exposure data, risk ratios increased with increasing exposures. These findings demonstrate
 30      specificity and strength of association. A fairly consistent association between exposure to
 31      butadiene and occurrence of leukemia across the plants  was also  found. Furthermore, the trend
 32      test for increasing risk of leukemia with increasing exposure to 1,3-butadiene was statistically
 33      significant (dose response).
 34            The major strengths of this study are as follows. First, the study had detailed and
J35      comprehensive quantitative exposure estimations for 1,3-butadiene, styrene, and benzene for


         1/28/98                                   11-3       DRAFT-DO NOT CITE OR QUOTE
-------
       Table 11-1. Summary of epidemiologic studies
Plants
7 U.S. and 1
Canadian
polymer
production plants
(UAB cohort)*
7 U.S. and 1
Canadian
polymer
production plants
(JHU cohort)1
1 U.S. monomer
production plant
(Texaco cohort)
3 U.S. monomer
production plants
(Union Carbide
cohort)
I U.S. monomer
production plant
(Shell Oil Deer
Park cohort)
Number of
workers,
dates studied
15,000,
1943-1994
13,500,
1943 - 1985
2,800,
1943-1994
364,
1940-1990
614,
1948-1989
Authors
Delzell et al.,
1996
Macaluso et al.,
1996
Matanoski and
Schwartz, 1987
Matanoski et al.,
1989, 1990, and
1993
Santos-Burgoa et
al., 1992
Downs etal., 1987
Divine, 1990
Divine etal., 1993
Divine and
Hartman, 1996
Ward etal., 1995
and 1996a
Cowles et al.,
. 1994
Approach
Cohort study
using
quantitative
exposure
estimates for
each worker
Cohort studies
using qualitative
exposures;
case-control
study using
estimated
quantitative
exposures for
each case and
control
Cohort studies
using qualitative
exposures, last
study made
quantitative
exposure
estimates
Cohort study
using qualitative
exposures
Cohort study
using qualitative
exposures
Significant findings
Excess mortality due to
leukemia;
leukemia risk increased
with increasing
exposure level
Excess mortality due to
lumpho- hematopoietic
cancers;
leukemia risk increased
with increasing
exposure level in case-
control study
Excess mortality due to
lymphosarcoma
in prewar workers
Excess mortality due to
lymphosarcoma in
World War II workers
No increase in mortality
or morbidity
'Six U.S. plants and one Canadian plant were common in Johns Hopkins University (JHU) and University of
Alabama, Birmingham (UAB) studies.
1/28/98
11-4
DRAFT-DO NOT CITE OR QUOTE
-------
  1     each individual. Second, the cohort was large, with a long follow-up period of 49 years. Third,
  >2     both external and internal comparison showed similar results. Fourth, adjustments for potential
  3     confounding factors were carried out. Fifth, analyses by duration of employment and for latency
  4     were conducted.
  5            The study had some limitations. First, some misclassification of exposure may have
  6     occurred with respect to certain jobs, but it is unlikely to have occurred only in leukemia cases,
  7     because the exposures were calculated a priori to health effects evaluation. Second, the excess
  8     mortality observed for leukemia was based on death certificates and was not verified by medical
  9     records.  This may have resulted in some misclassification of leukemias.  Third, histologic typing
 10     of leukemia was also not available. Thus, currently it is not known whether a single cell type or
 11     more than one cell type is associated with the exposure to 1,3-butadiene.
 12            A large cohort of synthetic rubber workers (JHU cohort)1, assembled from one  Canadian
 13     and seven U.S. plants, was also studied by Matanoski and Schwartz (1987) and then followed up
 14     by Matanoski et al. (1989,1990). The follow-up included a nested case-control  study (Santos-
 15     Burgoa et al., 1992).  Approximately 13,500 individuals were followed from 1943 to 1985. A
 16     significant excess of lymphohematopoietic cancer was observed in the cohort study.  The nested
 17     case-control study from this cohort, comprising 59 cases of lymphohematopoietic cancers and
 18     193 matched controls, found significantly increased relative odds for leukemia.  Increases of 7
        times in the high-exposure group and of 4 times in the low-exposure group were observed in the
        ever/never exposed analysis, of 9 times in the matched analysis, and of 8 times in the conditional
21     analysis (specificity and strength of association). Exposures to 1,3-butadiene and styrene were
22     estimated for each case and control using job records and levels of exposures to 1,3-butadiene
23     and styrene associated with those jobs, independently of the case or control status. A significant
24     trend of increasing risk of leukemia with increasing exposure level of 1,3 -butadiene was also
25     observed (dose response).
26            The findings of excess leukemia risk in the nested case-control study were questioned by
27     Acquavella (1989) and Cole et al. (1993), as these  findings were inconsistent with the absence of
28     excess leukemia risk in the base cohort study. Thus, Matanoski et al. (1993) reevaluated the
29     original nested case-control study by choosing a new set of three controls per case. The
30     investigators also verified the cause of death by obtaining the hospital records (25 out of 26 were
31      correctly recorded on the death certificates).  The findings of the new analysis were similar to
32     those of the earlier analysis. Although the  controversy about the cohort and case-control study is
33     still not resolved, the nested case-control study demonstrates a strong association between
34     exposure to 1,3-butadiene and occurrence of leukemias.
               1 One Canadian plant and six U.S. plants were common in the JHU and UAB studies.
        1/28/98                                   11-5        DRAFT-DO NOT CITE OR QUOTE
-------
  1            The main strengths of the JHU cohort study are as follows. First, this was the first large
  2      cohort study of polymer production workers. Second, adjustments for confounding exposures
  3      were conducted. Third, analyses by duration of employment and for latency were carried out.
  4      Fourth, the nested case-control study was well conducted and well analyzed, with quantitative
  5      estimation of exposures for each case and control as well as verification of leukemia.
  6            Limitations of the JHU cohort study included the exclusion of more than 50% of the
  7      population because of the lack of work histories, work start date, and exposure data.  In addition,
  8      the follow-up for four plants, where the starting date was 1957 to 1970, may not have been long
  9      enough for malignancies to develop. As far as the nested case-control study is concerned, the
10      estimated exposures were crude and not substantiated by air monitoring data.  Exposure
11      misclassification may have occurred based on the estimated exposures by job if the jobs were
12      incorrectly identified for higher or lower exposure.  However, the panel members were blind
13      toward the status of cases and controls; thus, the distribution of misclassification should be the
14      same in cases and controls.
15            Three different cohorts of monomer production workers were studied.  The largest cohort
16      of approximately 2,800 workers in a Texaco plant followed from 1943 to 1994 by several
17      investigators (Downs et al., 1987; Divine, 1990; Divine et al., 1993; Divine and Hartman, 1996).
18      All the investigations essentially found lower than expected mortality from all causes and total
19      cancers as compared to the general population.  The only significant excess mortality observed
20      was for lymphosarcoma in the prewar subcohort of workers who had worked for less than 10
21      years and had a latency of 0-9 years; 154% to 169% higher than the general population. Even
22      though exposures were estimated hi the last follow-up, no information about exposure levels was
23      available for the prewar period; however, it is believed that exposures were high.
24            The major strengths of this study are, first, it is the largest cohort of monomer workers.
25      Second, it had a long follow-up period of 52 years.  Third, analyses by duration of employment,
26      and for latency, as well as adjustment for potential confounding factors were conducted. Fourth,
27      the exposures in each individual were estimated in the last follow-up.
28            The main limitation was lack of exposure information in the earlier follow-ups.
29      Furthermore, although the investigators estimated the exposures for each individual in their last
30      follow-up, no information was available on work histories or levels of 1,3-butadiene exposure
31      during the prewar period, which made exposure estimation in the prewar workers impossible.
32            A small cohort of 364 individuals who had potential exposure to 1,3-butadiene at three
33      Union Carbide plants during World War II was studied by Ward et al. (1995, 1996a). This
34      investigation also found a statistically significant excess for lymphosarcoma by 477%, which was
35      based on four cases (specificity and strength of association).  The observation of excess
36      lymphosarcoma was consistent with the finding in the Texaco cohort study. The main limitations

        1/28/98                                  11-6       DRAFT-DO NOT CITE OR QUOTE
-------
         of this study are that the cohort was small and that exposures were assumed based on job
         categories. In addition, there was no analysis for latency or adjustments for potential
         confounding by exposure to other chemicals.
  •4            Cowles et al. (1994) studied the third cohort of 614 workers. This study failed to show
  5      any increased mortality or morbidity.  Due to several methodologic limitations such as lack of
  6      exposure information, short follow-up, and lack of information on confounders, this study failed
  7      to provide any negative evidence toward the causal association between exposure to 1,3-
  8      butadiene monomer and occurrence of lymphosarcoma that was observed in the other two
  9      studies.
 10            All the epidemiologic studies, cohort and nested case-control, evaluated for this
 11      assessment are observational studies in occupationally exposed populations. As such, they have
 12      various methodologic strengths and limitations as discussed above.  A common limitation to all
 13      the studies is the use of death certificates, which could lead to misclassification bias. Validation
 14      of diagnosis of lymphohematopoietic cancer was not done in any of the studies except in
 15      Matanoski et al. (1993).  This is a methodologic concern, given the fact that
 16      lymphohematopoietic cancer recording on death certificates is unreliable (Percy et al., 1981).
 17            Based on these monomer and polymer production workers' cohorts, it is obvious that an
 18      increased number of lymphohematopoietic cancers is observed in these populations. A clear
         difference is becoming apparent, though.  Increased lymphosarcomas develop in monomer
        workers, whereas excess leukemias occur in polymer workers. Furthermore, the
 21     lymphosarcomas observed in the monomer workers were among wartime workers, who were
 22     probably exposed to higher levels of 1,3-butadiene for shorter periods of time and not in long-
 23     term workers with low levels of exposure. A similar observation comes from the stop-exposure
 24     studies conducted by Melnick et al. (1990c). They observed that for a given total exposure, the
 25     incidence of lymphoma was greater among mice exposed to higher concentrations of butadiene
 26     for a shorter period of time (625 ppm for 26 weeks) than among mice exposed to a lower
 27     concentration for a longer period of time (312 ppm for 52 weeks). Consequently, this suggests
 28     that it may be the concentration of 1,3-butadiene rather than the duration of exposure that is
 29     important in the occurrence of lymphomas. There is a null relationship between exposure to 1,3-
 30     butadiene monomer and occurrence of leukemias, which are observed in polymer workers. This
 31      may be due to the exposure patterns for 1,3-butadiene in monomer production workers or to the
 32     absence of exposure to a necessary co/modifying factor or a confounding factor that occurs in
33     polymer production workers. Data are currently lacking to confirm or refute any of these
34     possibilities. The findings of the UAB study, which investigated styrene and benzene exposures
35     as well, suggest that the observed associations of leukemia with 1,3 -butadiene exposure are not
        due to confounding by exposure to other chemicals. The findings of excess leukemias in
        1/28/98
11 -7       DRAFT-DO NOT CITE OR QUOTE
-------
 1      polymer production workers are consistent with a causal association with exposure to 1,3-
 2      butadiene.
 3             Table 11-2 shows the application of the causality criteria to the studies discussed above.
 4             As these criteria are well satisfied, it is concluded that there is sufficient evidence to
 5      consider 1,3-butadiene a known human carcinogen.
 6
 7      11.3.2.  Animal Data
 8             1,3-Butadiene is an animal carcinogen.
 9             Chronic bioassay studies provide unequivocal evidence that 1,3 -butadiene is a multisite
10      carcinogen in both rats and mice.  These studies also demonstrate that the mouse is more
11      sensitive than the rat to 1,3-butadiene-induced carcinogenicity and develops tumors at different
12      sites, although the reasons for these interspecies differences are not understood at this time. The
13      most sensitive site was the female mouse lung, which exhibited significantly increased tumor
14      incidence at the lowest exposure concentration tested (6.25 ppm).
               Table 11-2. Epidemiologic causality criteria
Criteria
Temporality: exposure
occurred prior to effect
Specificity of cancer
Strength of association
Consistency
Dose-response relationship
Biological plausibility
Monomer plant workers
Yes
Lymphosarcoma
154% to 477% higher mortality
from lymphosarcoma than general
population
2 of 3 studies agree
Cannot be demonstrated due to
lack of quantitative exposure data
Yes
Polymer plant workers
Yes
Leukemia (specific cell type[s] not
known at this time)
7 to 9 times higher relative odds
for leukemia (nested case-control
study)3;
151% to 331% higher mortality
from leukemia than general
population
Fairly consistent across the plants
Yes
Yes
        * Relative odds is the ratio of the frequency of exposure to 1,3-butadiene in cases to the frequency of exposure to
        1,3-butadiene in controls, where both the cases and controls are from the same occupational cohort.
        1/28/98
11-8
DRAFT-DO NOT CITE OR QUOTE
-------
         11.3.3. Other Supportive Data
               1,3-Butadiene is metabolized into genotoxic metabolites by experimental animals and
         humans.
  4            Metabolic activation is required for 1,3 -butadiene carcinogenicity, and there is evidence
  5      that 1,3-butadiene is metabolized to at least three genotoxic metabolites:  a monoepoxide (1,2-
  6      epoxy-3-butene, EB), a diepoxide (1,2:3,4-diepoxybutane, DEB), and an epoxydiol (3,4-epoxy-
  7      1,2-butanediol). The enzymes responsible for the metabolic activation of 1,3-butadiene to these
  8      epoxide metabolites exist in humans as well as mice and rats.  EB and DEB have been measured
  9      in the blood of rats, mice, and monkeys after 1,3-butadiene exposure, and their production by
 10      human tissues has been observed in vitro. Formation of 3,4-epoxy-l,2-butanediol has been
 11      observed in vitro using tissues from mice, rats, and humans. Activation rates for 1,3-butadiene
 12      are typically higher in the mouse than in the rat, reflected by higher tissue concentrations of EB
 13      and DEB in the mouse versus the rat. Activation rates in humans exhibit a high degree of
 14      variability and appear to span the range between mice and rats.
 15            Among the genotoxic effects of 1,3-butadiene is an N7-alkylguanine adduct that has been
 16      observed in the liver DNA of exposed mice and in the urine of an exposed worker.  Similarly,
 17      increased frequencies of hprt mutations have been observed in the lymphocytes of mice and rats
 18      exposed to  1,3 -butadiene and in lymphocytes of occupationally exposed workers. Even though
        these mutations may not be directly related to tumor development, they provide in vivo evidence
        of similarities in the disposition and genotoxic action of 1,3-butadiene between mice and
21     humans.
22
23     11.3.4. Cancer Characterization
24            1,3-Butadiene is a known human carcinogen.
25            This characterization is supported by the three findings discussed above:  (1)
26     epidemiologic studies showing increased leukemias in workers occupationally exposed to
27     1,3-butadiene (by inhalation), (2) laboratory studies showing that 1,3-butadiene causes a variety
28     of tumors in mice and rats by inhalation, and (3) studies demonstrating that 1,3-butadiene is
29 •     metabolized into genotoxic metabolites by experimental animals and humans. The specific
30     mechanisms of 1,3-butadiene-induced carcinogenesis are unknown; however, it is virtually
31      certain that the carcinogenic effects are mediated by genotoxic metabolites of 1,3-butadiene.
32     Under EPA's 1986 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 1986), 1,3-butadiene
33     would be classified as a "Group A"—Human Carcinogen. It is characterized as a "Known
34     Human Carcinogen" according to EPA's  1996 Proposed Guidelines for Carcinogen Risk
        Assessment  (U.S. EPA, 1996).


        1/28/98                                  11-9        DRAFT-DO NOT CITE OR QUOTE
-------
 1     11.4. QUANTITATIVE RISK ESTIMATION FOR CANCER
 2            Lifetime extra cancer risk is estimated to be about 9 x 1 Or3 per ppm continuous
 3     1,3-butadiene exposure, based on human data.
 4            The Delzell et al. (1995) retrospective cohort study of more than 15,000 male
 5     styrene-butadiene rubber production workers provides high-quality epidemiologic data for
 6     estimating the human cancer risk from 1,3-butadiene exposure.  In the Delzell et al. study,
 7     1,3-butadiene exposure was estimated for each job and work area for each study year, and these
 8     estimates were linked to workers' work histories to derive cumulative exposure estimates for
 9     each individual worker. Consistent with EPA's 1986 Guidelines for Carcinogen Risk
10     Assessment (U.S. EPA, 1986) and evidence of the genotoxicity of 1,3-butadiene, the linear
11     relative rate exposure-response model reported by Delzell et al. was used to calculate a maximum
12     likelihood estimate (MLE) of 8.7 x 10'3/ppm (or 9 x 10'3/ppm, rounded to one significant figure)
13     for lifetime extra risk of leukemia mortality from continuous environmental 1,3-butadiene
14     exposure. The corresponding 95% upper limit on unit risk is 0.02/ppm. There were insufficient
15     exposure-response data to calculate a lymphoma risk estimate from the monomer cohorts.
16            Alternatively, interpreting the proposed new carcinogen risk assessment guidelines (U.S.
17     EPA, 1996), linear extrapolation from the LEC01 or the EC01 (i.e., the 95% lower confidence limit
18     or MLE, respectively, of the exposure concentration associated with a 1% increased risk) is
19     warranted given the clear genotoxicity of 1,3-butadiene and the fact that a 1 % increase in risk is
20     within the range of the epidemiology data.  The models presented by Delzell et al. yield LEC01
21     and EC01 values  ranging from 0.066 to 0.64 ppm and from 0.45 to 1.16 ppm, respectively. The
22     corresponding cancer potency estimates range from 0.016/ppm to 0.15/ppm (based on the LEC01)
23     and from 8.7 x lO'Vppm to 0.022/ppm (based on the EC01). The square root model provided the
24     best fit to the data and was chosen by Delzell et al. for further refinements. Thus, their final
25     square root model might be the appropriate model to select for determination of the ultimate
26     "point of departure" for linear extrapolation. Based on this model, a cancer potency estimate of
27     0.08/ppm is obtained from the LEC01 of 0.12 ppm, and a potency estimate of 0.02/ppm is
28     obtained from the EC0, of 0.45 ppm. These unit risk estimates are roughly two- and fourfold
29     higher, respectively, than the MLE and upper bound estimates calculated using the linear model
30      described above.
31             For comparison purposes, human unit cancer risk estimates based on extrapolation from
32     the results of lifetime animal inhalation studies are summarized in Table 11-3. These potency
33      estimates are 95% upper confidence limits on unit cancer risk calculated from incidence data on
34      all significantly elevated tumor sites using a linearized low-dose extrapolation model.  Such
35      estimates are generally considered by EPA to represent plausible upper bounds on the extra unit
36      cancer risk to humans. Table 11-3 also includes unit risk estimates based only on the

        1/28/98                                 11-10       DRAFT-DO NOT CITE OR QUOTE
-------
               Table 11-3.  Estimates of upper bounds on human extra unit cancer risk
               (potency) from continuous lifetime exposure to 1,3-butadiene based on animal
               inhalation bioassays
Species
Rata
Mouseb
Sex
M
F
M •
F
M
F
Tumor sites/types
Leydig cell, pancreatic exocrine cell, Zymbal
gland
Mammary gland, thyroid follicular cell, Zymbal
gland
Lymphocytic lymphomas, histiocytic sarcomas,
heart hemangiosarcomas, lung, forestomach,
Harderian gland, liver, preputial gland
Lymphocytic lymphomas, heart
hemangiosarcomas, lung, forestomach, Harderian
gland, liver, ovary, mammary gland
Lymphocytic lymphomas
Lymphocytic lymphomas
Upper bound on
potency (ppm'1)
4.2 x lO'3
5.6 x lO'2
0.22
0.29
6.4 x 10'3
2.4 x lO'2
        a From U.S. EPA's 1985 assessment; linearized multistage model.
        b Based on 1993 NTP study; Weibull multistage time-to-tumor model.
  1
  2
  3
  4
  5
  6
  7
  8
  Q
10
11
12
13
14
lymphocytic lymphomas in mice, because this was the tumor type in rodents most analogous to
the lymphohematopoietic cancers observed in workers exposed to 1,3-butadiene.
       In both rodent species, females are apparently more sensitive than males, as evidenced by
the higher risk estimates. The "best estimate" (i.e., MLE from the linear model) of 8.7 x
ICT/ppm for extra cancer risk from the human (male) leukemia data exceeds the upper bound
estimates based on the male rat data and on the male mouse data for lymphocytic lymphomas,
and is 25 times lower than the upper bound estimate based on all male mouse tumors.
       Human health risk estimates based on extrapolation from high-quality epidemiologic
results are preferable to those based on rodent data because they avoid the uncertainties inherent
in extrapolating across species and, typically, the human exposures in epidemiologic studies are
closer to anticipated environmental exposures than the high exposures used in animal studies,
thus reducing the extent of low-dose extrapolation. In the case of 1,3-butadiene, while the rat
exposures were far in excess of human exposures, the lowest EXPOSURE in the 1993 NTP
mouse study (4.7 ppm, 8 h TWA) is within the range of occupational exposures (0.7-1.7 ppm
median and 39-64 ppm max 8 h TWAs for work-area groups).  However, interspecies differences
in tumor sites and susceptibilities between rats and mice are especially pronounced, and the
        1/28/98
                                        11-11       DRAFT-DO NOT CITE OR QUOTE
-------
 1      biological bases for these differences are unresolved. A review of available pharmacokinetic
 2      data and models revealed that the state of the science is currently inadequate for either explaining
 3      interspecies differences or improving on default dosimetry assumptions. Therefore, the
 4      quantitative extrapolation of rodent risks to humans is highly uncertain for 1,3-butadiene.
 5             Even though high-quality human data were used for the quantitative cancer risk
 6      estimation for 1,3-butadiene, there are inevitable uncertainties in the calculated risk estimate.
 7      First, there are uncertainties inherent in the epidemiologic study itself. In particular, there are
 8      uncertainties hi the retrospective estimation of 1,3-butadiene exposures, which could have
 9      resulted in exposure misclassification.  Nondifferential exposure misclassification would tend to
10      bias estimates of effect toward the null, resulting in an underestimate of risk.  Differential
11      misclassification could bias results in either direction.
12             Second, there are uncertainties regarding the appropriate dose metric for dose-response
13      analysis. Although the dose surrogate of cumulative exposure (i.e., ppm x years) yielded highly
14      statistically significant exposure-response relationships, cumulative exposure is strongly
15      correlated with other possible exposure measures, and there may be a dose-rate effect (e.g., risk
16      at high exposures may be more than proportionately greater than at lower exposures) obscured in
17      the analysis, or operative at exposures below the observable range but relevant to low-dose
18      extrapolation.
19             Third, there are uncertainties pertaining to the model for low-dose extrapolation.             WU
20      Although Delzell et al. expressed preference  for the square root model based on its goodness of
21      fit, the four exposure-response models that they investigated were virtually indistinguishable on
22      statistical grounds, and because the specific mechanisms of 1,3-butadiene carcinogenesis are
23      unknown, there is no biological basis for choosing one model over another. Even though the
24      models give similar results in the observable range, they deviate substantially  at lower exposures.
25      For example, at a lifetime continuous exposure of 1 ppb, the preferred model of Delzell et al.
26      yields  a cancer potency estimate almost two orders of magnitude higher than that obtained by the
27      linear model. However, there was no apparent biological reason to depart from a default
28      assumption of linearity, so the linear model was used in this risk assessment.
29             Fourth, it is uncertain which potential modifying or confounding factors should be
30      included in the model. The linear model of Delzell et al., which was used in this risk assessment,
31      adjusted for age, calendar year, years since hire, race, and exposure to styrene. However, these
32      investigators dropped styrene and race from their preferred square root model to obtain their final
33      model. Furthermore, there may be other relevant factors that weren't included in the models at
34      all.
35             Fifth, there are uncertainties in the parameter estimates used in the models.  The study of
36      Delzell et al. is large, providing some degree of reliability in the parameter estimates; however,
         1/28/98                                    11-12      DRAFT-DO NOT CITE OR QUOTE
-------
        especially given the large human variability that has been observed in metabolic activities that
        could affect cancer risk from 1,3-butadiene exposure, the generalizability of the occupational
        results is unclear.
  4            In addition, there are important concerns raised by comparison with the rodent data.
  5     First, the rodent studies suggest that 1,3-butadiene is a multi-site carcinogen. It is possible that
  6     humans may also be at risk of 1,3-butadiene-induced carcinogenicity at other sites and that the
  7     epidemiologic study had insufficient power to detect the other excess risks.  In the mouse, for
  8     example, the lung is the most sensitive tumor site. Significant excesses of lung cancer may not
  9     have been detectable in the epidemiologic study because of the high background rates of lung
 10     cancer in humans. Delzell et al. did observe a slight increase in lung cancer among maintenance
 11     workers. The reported excess cancer risk estimate, which is based only on leukemias, may be an
 12     underestimate if other sites are also at risk.
 13            Second, both the rat and mouse studies suggest that females are more sensitive to
 14     1,3-butadiene-induced carcinogenicity than males, and the mammary gland in females was the
 15     only tumor site common to both species.  If female humans are also more sensitive than males,
 16     then the male-based risk estimates calculated from the epidemiology study would underestimate
 17     risks to females.
 18            Despite these uncertainties, confidence in the excess cancer risk estimate of 9 x
        10-3/ppm is relatively high. First, the estimate is based on human data. Furthermore, these data
 ?0     are from a large, high-quality epidemiologic study in which 1,3-butadiene exposures were
21      estimated for each individual a priori to conducting the exposure-response analysis. Although
22     there are uncertainties  in the exposure estimation, a serious attempt was made to reconstruct
23     historical exposures for specific tasks and work areas. It is virtually unprecedented to have such
24     a comprehensive exposure  assessment for individual workers in such a large occupational
25     epidemiologic study.  In addition, the assumption of linearity for low-dose extrapolation is
26     reasonable given the clear evidence of genotoxicity by 1,3-butadiene metabolites.
27            Using the cancer potency estimate of 9 * lO'Vppm, the chronic (70 year) exposure level
28     resulting in an increased cancer risk of 1O'6 (i.e., one in a million) can be estimated as follows:
29     (10-6)/(9 x 10-3/ppm) = 1 x  lO^ppm = 0.1 ppb.
30
31      11.5. SUMMARY OF REPRODUCTIVE/DEVELOPMENTAL EFFECTS
32            A variety of reproductive and developmental effects have been observed in mice and
33      rats exposed to 1,3-butadiene by inhalation.  There are no human data on reproductive or
34      developmental effects.
               The most sensitive developmental endpoint was decreased fetal weight in the mouse.
        Decreases were observed at the lowest exposure concentration (40 ppm, 6 h/day, gestation days

        •1/28/98                                   11-13       DRAFT-DO NOT CITE OR QUOTE
-------
 1      6-15); thus there was no NOAEL for this effect. Generally, however, it is thought that there is an
 2      exposure threshold, and while effects on fetal growth in humans cannot be ruled out, they are not
 3      expected to occur from low environmental exposures to 1,3-butadiene. No developmental
 4      toxiclty was observed in rats.
 5             The most sensitive reproductive endpoints observed in subchronic exposure studies were
 6      litter size at birth and at weaning hi dominant lethal studies of mice (i.e., male mice are exposed
 7      to 1,3-butadiene and effects on litters are measured after mating to unexposed females). Litter
 8      size at birth reflects both decreased implants and increased fetal deaths, while litter size at
 9      weaning also reflects neonatal deaths. Dominant lethal effects in humans would likely be
10      manifested as spontaneous abortions, miscarriages, stillbirths, or very early deaths. The
11      dominant lethal responses are believed to represent a genotoxic effect; however, a large number
12      of sperm would have to be affected to result in any meaningful increase in risk, because the
13      chances of any single sperm both having a critical mutation and fertilizing an egg are minuscule.
14      Thus, dominant lethal effects are not expected in humans exposed to low environmental
15      exposures, although the possibility of such effects or of transmissible genetic mutations cannot
16      be ruled out.
17             From chronic exposure studies (2-year bioassays), the most sensitive reproductive effects
18      were ovarian atrophy in female mice and testicular atrophy in male mice. Testicular atrophy was
19      primarily a high-exposure effect and likely has an exposure threshold.  Ovarian atrophy, on the
20      other hand, was observed at the lowest exposure level (6.25 ppm, 6 h/day, 5 days/week, for 2
21      years), although an exposure threshold is assumed for this endpoint as well.  Uterine atrophy was
22      also observed in the highest exposure groups: however, this is thought to be a secondary effect
23      of the ovarian atrophy. The mechanisms of ovarian atrophy are unknown, although there is
24      strong evidence that the effect is mediated by the diepoxide metabolite. It is further expected,
25      based on metabolic data, that humans would produce lower concentrations of this metabolite than
26      do mice. Thus, it is likely that humans are less sensitive to 1,3-butadiene-induced ovarian
27      atrophy than are mice.  No reproductive effects were reported in the 2-year rat study. In
28      conclusion, ovarian atrophy is not expected in humans from environmental exposures to 1,3-
29      butadiene; although, the effect cannot be ruled out.
30
31      11.6.  QUANTITATIVE ESTIMATION (RfC) FOR REPRODUCTIVE/
32            DEVELOPMENTAL EFFECTS
33             An RfC for reproductive and developmental effects ofO.lSppb was obtained for the
34      critical effect of decreased litter size at birth (or at weaning), based on subchronic dominant
35      lethal studies in the mouse.

        1/28/98                                  11-14       DRAFT-DO  NOT CITE OR QUOTE
-------
                A reference concentration (RfC) is an estimate of the daily exposure to humans that is
         "likely to be without appreciable risk of deleterious [noncancer] effects during a lifetime." The
         RfC is calculated for the "critical  [noncancer] effect," i.e., the effect for which an increased
  4     response is observed at the lowest concentration used in the study, or for which benchmark
  5     concentration modeling yields the lowest EC10. In this assessment, the RfC is only for
  6     reproductive and developmental effects (R/D RfC), because other noncancer effects were not
  7     considered. Of the 1,3-butadiene  reproductive/developmental effects, the critical effect was
  8     decreased litter size at birth or at weaning (both of these effects yielded the same EC i Q), as
  9     observed in dominant lethal studies of male mice.  An R/D RfC was calculated based on the
 10     LEC10, which was calculated using benchmark concentration methodology, and uncertainty
 11      factors for interspecies extrapolation (3), intraspecies variability (10), extrapolation from
 12     subchronic study to chronic exposure (3), the absence of multigenerational studies (3), and "risk
 13     reduction" to extrapolate to a level at which no detectable effects are expected (analogous to the
 14     LOAEL-to-NOAEL uncertainty factor) (3).  The resulting R/D RfC is 0.15 ppb [0.15
 15     ppm/(3xlOx3x3x3)]. The actual risks at low exposure levels are unknown; the R/D RfC merely
 16     provides a bound on chronic exposure below which no "appreciable risk" of reproductive or
 17     developmental effects is expected.
 18            Although other noncancer  effects were not examined, the reproductive endpoints were
         quite sensitive, and it is likely that the R/D RfC is protective against other noncancer effects as
 R)     well.
 21             In addition, a RfCDT of 0.1 ppm for developmental toxicity from short-term exposures
 22      was calculated from the mouse fetal weight data, and a R/D RfC for subchronic exposures of
 23      0.0015 ppm was derived from the  dominant lethal results in mice, each using benchmark
 24      concentration methodology to obtain the "point of departure" for applying uncertainty factors.
 25
 26      11.7. SPECIAL SUBPOPULATIONS
 27      11.7.1. Sensitive Subpopulations
 28            It is uncertain whether children or other subpopulations have greater susceptibility to
 29      exposure to 1,3-butadiene than the general population.
 30   •          There is no information available on health effects in children from exposure to 1,3-
 31      butadiene at this time.  Occurrence of leukemia is causally associated with exposure to 1,3-
 32      butadiene in adults, and leukemia is one of the most common cancers in children. Furthermore,
 33      leukemia risk in children has been shown to increase with simultaneous exposure to multiple risk
 34      factors (Gibson et al., 1968). Thus, exposure to 1,3-butadiene may be an additional risk factor
_35      increasing the leukemia risk further in children.


         1/28/98                                   11-15      DRAFT-DO NOT CITE OR QUOTE
-------
 1             Tobacco smoke contains 1,3-butadiene as well as other carcinogens, and there are a few
 2      studies suggesting that parental smoking increases the risk of leukemia or lymphoma in children
 3      (John et al., 1991; Stjernfeldt et al., 1986).  The overall evidence, however, is inconclusive
 4      because other studies observed no increased risk. Furthermore, if there is an effect in children
 5      from parental smoking, it is unclear whether it is attributable to preconception effects on fathers'
 6      sperm, in utero exposure of the fetus, and/or postnatal exposure to environmental tobacco smoke.
 7             Because metabolic activation of 1,3-butadiene to epoxide metabolites is believed to be
 8      necessary for carcinogenicity, it is possible that genetic differences in metabolic or detoxification
 9      enzymes could result hi different risks to different human subpopulations. For example,
10      investigators have observed that polymorphism in glutathione-S-transferase genes confers
11      differential susceptibility to the induction of sister chromatid exchanges by butadiene metabolites
12      in cultured human lymphocytes. However, the critical/rate-limiting mechanistic steps are
13      unknown at present; thus, it is unknown whether or not there are actual human subpopulations
14      that may have notably different susceptibility to 1,3-butadiene.
15
16      11.7.2. Highly Exposed Subpopulations
17             Some subpopulations may be at greater risk than the general population as a  result of
18      higher exposure to 1,3-butadiene.
19             Heavy smokers may be highly exposed to 1,3-butadiene due to its formation in tobacco
20      smoke.  Cigarette smoke has been shown to be a risk factor for various types of leukemias. It
21      should be noted, however, that known and suspected leukemogenic constituents of tobacco
22      smoke include benzene, polonium-210, nitrosamines, and hydrocarbons in addition to 1,3-
23      butadiene (Schottenfeld and Fraumeni, 1996).
24
25      11.8. FUTURE RESEARCH NEEDS
26             Although 1,3-butadiene is classified as a known human carcinogen in this assessment,
27      there are some data gaps in various areas which, if filled, will refine the assessment. The specific
28      research needs are as follows:
29
30      Epidemiology
31             •  The medical records for the leukemia cases in the studies by Delzell et al. and
32                Macaluso et al. should be reviewed to verify the cell types of leukemias.
33             •  Further follow-up of these studies is recommended because it will give an opportunity
34                to observe whether any noncancer effects, such as cardiovascular, or any cancers with
35                a longer latency period are associated with exposure to 1,3-butadiene.
36
        1/28/98                                  11-16      DRAFT-DO NOT CITE OR QUOTE
-------
  1             •  Studies in other polymer facilities around the world could also add to the human
  2                evidence of carcinogenicity.
 |3
  4             •  All epidemiologic studies to date have examined male cohorts. Some butadiene
  5                production facilities around the world (e.g., China) employ women in their
  6                laboratories. If the number of women in these facilities is large enough, a
  7                reproductive/developmental study would help determine if female workers are at risk
  8                of reproductive effects or if exposed fetuses are at risk of developmental effects
  9
10             •  A reproductive study of exposed males is also needed to examine potential dominant
11                lethal effects in humans.
12
13      Toxicology
14             •  Elucidation of the mechanisms responsible for the interspecies differences in
15                sensitivity to 1,3-butadiene could assist in resolving questions about the human risk
16                for reproductive effects and for cancer at sites for which the Delzell et al. study may
17                have had insufficient power to detect an effect.
18
19      Molecular biology
20             •  Once the mechanisms of 1,3-butadiene-induced health effects are better understood,
21                information on polymorphisms in human metabolic enzymes (or DNA repair
22                enzymes, etc.) could help define sensitive subpopulations.
23
        11.9.  SUMMARY AND CONCLUSIONS
               The purpose of this effort was to review the new information that has become available
26      since EPA's  1985 health assessment of 1,3-butadiene and to determine if any changes were
27      needed to the earlier conclusions.
28             1,3-Butadiene is a gas used commercially in the production of styrene-butadiene rubber,
29      plastics, and  thermoplastic resins.  The major environmental source of 1,3-butadiene is the
30      incomplete combustion of fuels from mobile sources (e.g., automobile exhaust).  Tobacco smoke
31      can be a significant source of 1,3-butadiene in indoor air.
32             This assessment concludes that 1,3-butadiene is a known human carcinogen, based on
33      three types of evidence: (1) epidemiologic studies showing increased leukemias in workers
34      occupationally exposed to 1,3-butadiene (by inhalation), (2) studies showing that 1,3-butadiene
35      causes a variety of tumors in mice and rats by inhalation, and (3) studies demonstrating that 1,3-
36      butadiene is metabolized into genotoxic metabolites by experimental animals and humans.
37      The specific  mechanisms of 1,3-butadiene-induced carcinogenesis are unknown; however, it is
38      virtually certain that the carcinogenic effects are mediated by genotoxic metabolites of
39      1,3-butadiene.
               The best estimate of human lifetime extra cancer risk from chronic exposure to
        1,3-butadiene is 9 x 10~3perppm based on linear modeling and extrapolation of the increased

        1/28/98                                 11-17       DRAFT-DO NOT CITE OR QUOTE
-------
 1      leukemia risks observed in occupationally exposed workers.  Although there is uncertainty in
 2      extrapolating from occupational exposures to lower environmental exposures, this risk estimate
 3      has the advantage of being based on a large, high-quality human study, and linear extrapolation is
 4      warranted by the known genotoxicity of 1,3-butadiene metabolites. The corresponding estimate
 5      of the chronic exposure level of 1,3-butadiene resulting in an extra cancer risk of 10"6 (i.e., one in
 6      a million) is 0.1 ppb. The 95% upper bound on unit risk from the linear model is 0.02/ppm.
 7            1,3-Butadiene also causes a variety of reproductive and developmental effects in mice
 8      and rats; no human data on these effects are available. The most sensitive effect was reduced
 9      litter size at birth and at weaning observed in studies in which exposed male mice were mated
10      with unexposed females. In humans, such an effect might be manifested as an increased risk of
11      spontaneous abortions, miscarriages, stillbirths, or very early deaths. Based on this critical effect
12      of reduced litter size, a reference concentration (i.e., a chronic exposure level presumed to be
13      "without appreciable risk") of 0.15 ppb for reproductive and developmental effects was
14      calculated from the modeled benchmark concentration (LED ,0) of 0.15 ppm. The actual risks at
15      low exposure levels are unknown; this RfC merely provides a bound on chronic exposure below
16      which no "appreciable risk" of reproductive or developmental effects is expected.
17            There are insufficient data from which to draw any conclusions on potentially sensitive
18      subpopulations.
19            In summary, the primary changes in EPA's conclusions about the health effects of 1,3-
20      butadiene from the 1985 document to this one are:
21            •   The cancer classification has been changed from probable to known human
22                carcinogen.
23
24            •   The unit cancer risk estimate has been changed from 0.25/ppm (upper bound based on
25                mouse data) to 0.009/ppm (best estimate based on linear modeling and extrapolation
26                of human data).
27
28            •   For the first time, an RFC (0.15 ppb) is calculated for reproductive/developmental
29                effects.
30
        1/28/98                                  11-18       DRAFT-DO NOT CITE OR QUOTE
-------
                                                12.  REFERENCES
  1      Acquavella, JF. (1989) The paradox of butadiene epidemiology. Exp Pathol 37-114
  2
  3      Adler, ID; Anderson, D. (1994) Dominant lethal effects after inhalation exposure to 1,3-butadiene Mutat Res
  4      309:295-297.
  5
  6      Adler, ID; Cao, J; Filser, JG; et al. (1994) Mutagenicity of 1,3-butadiene inhalation in somatic and germinal cells of
  7      mice. Mutat Res 309:307-314.
  8
  9      Adler, ID; Filser, JG; Gassner, P; et al. (1995) Heritable translocations induced by inhalation exposure of male mice
 10      to 1,3-butadiene. Mutat Res 347:121-127.
 11
 12      Agency for Toxic Substances and Disease Registry (ATSDR) (1992) Toxicological profile for 1,3-butadiene. TP-
 13      91/07, PB93-11690. Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.
 14
 15      Albertini, RJ; Castle, KL; Borcherding, WR. (1982) T-cell cloning to detect the mutant 6-thioguanine-resistant
 16      lymphocytes present in human peripheral blood. Proc Natl Acad Sci USA 79:6617-6621
 17
 18      Allen, BC; Kavlock, RJ; Kimmel, CA; et al. (1994a) Dose-response assessment for development toxicity: III.
 19      Statistical models. Fundam Appl Toxicol 23:496-509.
 20
 21      Allen, BC; Kavlock, RJ; Kimmel, CA; et al. (1994b) Dose-response assessment for developmental toxicity: II.
 22      Comparison of generic benchmark dose estimates with NOAELs. Fundam Appl Toxicol 23 '487-495
 23
 2A      American Congress of Government Industrial Hygienists (ACGIH). (1994) World-wide limits for toxic and  •
         hazardous chemicals in air, water, and soil.  ACGIH Pub. No. 9565.

 27      Anderson, EL; U.S. EPA Carcinogen Assessment Group. (1983) Quantitative approaches in use to assess cancer
 28      risk. Risk Anal 3:277-295.
 29
 30      Anderson, D; Edwards, AJ; Brinkworth, MH. (1993) Male-mediated F, effects in mice exposed to 1,3-butadiene. In:
 31      Butadiene and styrene: assessment of health hazards. IARC Scientific Publications. Vol. 127. Sorsa, M; Peltonen, K;
 32      Vainio, H; et al., eds. Lyon, France: International Agency for Research on Cancer, pp. 171-181.
 33
 34      Anderson, D; Edwards, AJ; Brinkworth, MH. (1995) The detection of dominant lethal mutations and foetal
 35      malformations, tumours and chromosome damage in the offspring of male mice treated sub-chronically with
 36      butadiene. BIBRA Report No.  1060/5/95.
 37
 38      Andjelkovich, D; Taulbee J; Symons, M.  (1976) Mortality experience of a cohort of rubber workers, 1964-1973  J
 39      Occup Med 18:387-394.
 40
 41       Andjelkovich, D; Taulbee, J; Symons, M. (1977) Mortality of rubber workers with reference to work experience. J
42      Occup Med 19:397-405.
43
44      Anon. (1991) Chemical profile: butadiene. Chem Mark Rep 239:50.
45
46      Arce, GT; Vincent, DR; Cunningham, MJ; et al. (1990) In vitro and in vivo genotoxicity of 1,3-butadiene.  Environ
47      Health Perspect 86:75-78.
48
49      Armitage, P. (1971) Statistical methods in medical research. New York, NY: John Wiley and Sons, pp. 362-365.
         1/28/98                                        12-1         DRAFT-DO NOT CITE OR QUOTE
-------
 1       Arms, A; Travis, C. (1988) Reference physiological parameters in pharmacokinetic modelling. U.S. Environmental
 2       Protection Agency. EPA/600/6-88/004.
 3
 4       Au, WW; Bechtold, WE; Wharton, EB, Jr.; et al. (1995) Chromosome aberrations and response to gamma-ray
 5       challenge in lymphocytes of workers exposed to 1,3-butadiene. Mutat Res  334:125-130.
 6
 7       Bailer, AJ; Portier, CJ. (1988) Effects of treatment-induced mortality and tumor-induced mortality on tests for
 8       carcinogenicity in small samples. Biometrics 44:417-431.
 9
10       Barnes, DG; Dourson, M; U.S. EPA Reference Dose Work Group. (1988)  Reference dose (RfD): Description and
11       use in health risk assessments.  Reg Toxicol Pharmacol 8:471-86.
12
13       Beaumont, JJ; Breslow, NE. (1981) Power considerations in epidemiologic studies of vinyl chloride workers. Am J
14       Epidemiol 114:725-734.
15
16       Bechtold, WE; Strunk, MR; Chang, I-Y; et al. (1994) Species differences in urinary metabolite ratios between mice,
17       rats, and humans. Toxicol Appl Pharmacol  127:44-49.
18
19       Bechtold, WE; Strunk, MR; Thornton-Manning, JR;  et al. (1995) Analysis  of butadiene, butadiene monoxide and
20       butadiene diepoxide in blood by gas chromatography/gas chromatography/mass spectroscopy. Chem Res Toxicol
21       8:182-187.
22
23       Bemadini, S; Pelin, K; Peltonen, K; et al. (1996) Induction of sister chromatid exchange by 3,4-epoxybutane-l,2-
24       diol in cultured human lymphocytes of different GSTT1 and GSTM1 genotypes. Mutat Res 361:121-127.
25
26       Bolt, HM; Schmeidel, G; Filser, JG; et al. (1983) Biological activation of 1,3-butadiene to vinyl oxirane by rat liver
27       microsomes and expiration of the reactive metabolite by exposed rats. J Cancer Res Clin Oncol 106:112-116.
28
29       Bolt, HM; Filser, JG; StSrmer,  F. (1984) Inhalation pharmacokinetics based on gas uptake studies. Arch Toxicol
30       55:213-218.
31
32       Bond, JA; Csanady, GA. (1991) Species and organ differences in the in vitro metabolism of butadiene to butadiene
33       monoepoxide. Proc Am Assoc Cancer Res 32:119.
34
35       Bond, JA; Dahl, AR; Henderson, RF; et al. (1986) Species differences in the disposition of inhaled butadiene.
36       Toxicol Appl Pharmacol 84:617-627.
37
38       Bond, JA; Dahl, AR; Henderson, RF; et al. (1987) Species differences in the distribution of butadiene in tissues. Am
39       IndHyg Assoc J 48:867-872.
40
41       Boogaard, PJ; Bond, JA. (1996) The role of hydrolysis in the detoxication of l,2:3,4-diepoxybutane by human, rat,
42       and mouse liver and lung in vitro. Toxicol Appl Pharmacol 141:617.
43
44       Boogaard, PJ; Sumner, SCJ; Bond, JA. (1996) Glutathione conjugation of l,2:3,4-diepoxybutane in human liver and
45       rat and mouse liver and lung in vitro. Toxicol Appl Pharmacol 136:307-316.
46
47       Brunnemann, KD; Kagan, MR; Cox, JE; et al. (1990) Analysis of 1,3-butadiene and other selected gas-phase
48       components in cigarette mainstream and sidestream smoke by gas chromatography-mass selective detection,
49       Carcinogenesis 11:1863.
50
51       Budavari, S, ed. (1989) The Merck index, 11th ed. Rahway, NJ:Merck & Co., pp. 230-231.
52
53       California Air Resources Board. (1991) Proposed identification of 1,3-butadiene as a toxic air contaminant.
54       Prepared by the Office of Environmental Health Hazard Assessment, Sacramento, CA.


         1/28/98                                        12-2        DRAFT-DO NOT CITE OR QUOTE
-------
         California Air Resources Board. (1992) Proposed identification of 1,3-butadiene as a toxic air contaminant. Part A.
         Exposure assessment. Stationary Source Division, Sacramento, GA.

         Campbell, DL; Mangino, J. (1994) Evaluation and improvement of the Puget Sound toxic air contaminants
  5      emissions inventory. Prepared by Radian Corporation for the Office of Research and Development, U.S.
  6      Environmental Protection Agency, Washington, DC, under EPA contract no. 68-D1-0031
  7
  8      Carpenter, CP; Shaffer, CB; Weil, CS; et al. (1944) Studies on the .inhalation of 1,3-butadiene; with a comparison of
  9      its narcotic effect with benzol, toluol and styrene, and a note on the elimination of sryrene by the human J Ind Hyg
 10      Toxicol 26:69-78.
 11
 12      Checkoway, H; Williams, TM. (1982) A hematology survey of workers at a styrene-butadiene synthetic rubber
 13      manufacturing plant. Am Ind Hyg Assoc J 43:164-169.
 14
 15      Cheng, X;  Ruth, JA. (1993) A simplified methodology for quantitation of butadiene metabolites. Application of the
 16      study of butadiene metabolism by rat liver microsomes. Drug Metab Dispos 21:121-124
 17
 18      Citti, L; Gervasi, PG; Turchi, G; et al. (1984) The reaction of 3,4-epoxy-l-butene with deoxyguanosine and DNA in
 19      vitro: synthesis and characterization of the main adducts. Carcinogenesis 5:47-52
 20
 21      Cochrane, JE; Skopek, TR. (1993) Mutagenicity of 1,3-butadiene and its epoxide metabolites in human TK6 cells
 22      and splenic T  cells isolated from exposed B6C3F1 mice. In: Butadiene and styrene: assessment of health hazards.
 23      IARC Scientific Publications.  Vol. 127. Sorsa, M; Peltonen, K; Vainio, H; et al., eds. Lyon, France: International
 24      Agency for Research on Cancer, pp. 195-204.
 25
 26      Cochrane, JE; Skopek, TR. (1994a) Mutagenicity of butadiene and its epoxide metabolites: I. Mutagenic potential
 27      of 1,2-epoxybutene, 1,2,3,4-diepoxybutane and 3,4-epoxy-1,2-butanediol in cultured human lymphoblasts
 28      Carcinogenesis 15:713-717.

 r   •          '
 30      Cochrane, JE; Skopek, TR. (1994b) Mutagenicity of butadiene and its epoxide metabolites: II. Mutational spectra of
 31      butadiene,  1,2-epoxybutene, and diepoxybutane at the hprt locus in splenic T cells from exposed B6C3F, mice.
 32      Carcinogenesis 15:719-723.
 33
 34      Cole, P; Delzell, E; Acquavella, J. (1993) Exposure to butadiene and lymphatic and hematopoietic cancer.
 35      Epidemiology 4:86.
 36
 37      Concawe. (1987) A survey of exposures to gasoline vapour. Report No. 4/87. de Hague- Concawe
 38                                                         -
 39      Cowles, SR; Tsai, SP; Snyder, PJ; et al. (1994) Mortality, morbidity, and haematologic results from a cohort of
40      long-term workers involved hi 1,3-butadiene monomer production. Occup Environ Med 51 -323-329
41
42      Cox, DR. (1972) Regression models and life tables. J R Stat Soc B34:187-220.
43
44      Crisp, TM. (1992) Organization of the ovarian follicle and events in its biology: oogenesis, ovulation or atresia
45      Mutat Res 296:89-106.
46
47      Crump, KS. (1995) Calculation of benchmark doses from continuous data. Risk Anal 15:79-89
48
49      Crump, KS. (1984) A new method for determining allowable daily intakes. Fundam Appl Toxicol 4-854-871
50
51       Csanady, GA;  Bond, JA. (1991) Species differences in the biotransformation of 1,3-butadiene to DNA-reactive
52      epoxides: role  in cancer risk assessment. Chemical Industry Institute of Toxicology, CUT Activities 1 l-i-8
53
         1/28/98                                         12-3        DRAFT-DO NOT CITE OR QUOTE
-------
 1       Csan&dy, GA; Guengerich, FP; Bond, JA. (1992) Comparison of the biotransformation of 1,3-butadiene and its
 2       metabolite, butadiene monoepoxide, by hepatic and pulmonary tissue from humans, rats, and mice. Carcinogenesis
 3       13:1143-1153.
 4
 5       Dahl, AR; Birnbaum, LS; Bond, JA; et al. (1987) The fate of isoprene inhaled by rats: comparison to butadiene.
 6       Toxicol Appl Pharmacol 89:237-242.
 7
 8       Dahl, AR; Bechtold, WE; Bond, JA; et al. (1990) Species differences in the metabolism and disposition of inhaled
 9       1,3-butadiene and isoprene. Environ Health Perspect 86:65-69.
10
11       Dahl, AR; Sun, DJ; Birnbaum, LS; et al. (1991) Toxicokinetics of inhaled 1,3-butadiene in monkeys: comparison to
12       toxicokinetics in rats and mice. Toxicol Appl Pharmacol 110:9-19.
13
14       de Meester, C. (1988) Genotoxic properties of 1,3-butadiene. Mutat Res 195:273-281.
15
16       de Meester, C; Poncelet, F; Roberfroid, M; et al. (1980) The mutagenicity of butadiene towards Salmonella
17       typhimurium. Toxicol Lett 6:125-130.
18
19       Del Monte, M; Citti, L; Gervasi, PG. (1985) Isoprene metabolism by liver microsomal monooxygenases.
20       Xenobiotica 15:591-597.
21
22       Delzell, E; Sathiakumar, N; Macaluso, M. (1995) A follow-up study of synthetic rubber workers. Final report
23       prepared under contract to International Institute of Synthetic Rubber Producers.
24
25       Delzell, E; Sathiakumar, N; Hovinga, M. (1996) A follow-up study of synthetic rubber workers. Toxicology
26       113:182-189.
27
28       Deutschmann, S. (1988) Vergleichende Untersuchungen zur Depletion von zellularem Glutathion durch 1,3-
29       Butadien bei Maus und Ratte.  Thesis. Ruhr-Universita't, Bochum, Germany.
30
31       Deutschmann, S; Laib, RJ. (1989) Concentration-dependent depletion of non-protein sulfhydryl (NPSH) content in
32       lung, heart and liver tissue of rats and mice after acute inhalation exposure to butadiene. Toxicol Lett 45:175-183.
33
34       Divine, BJ. (1990) An update  on mortality among workers at a 1,3-butadiene facility— preliminary results. Environ
35       Health Perspect 86:119-128.
36
37       Divine, BJ; Hartman, CM. (1996) Mortality update of butadiene production workers. Toxicology 113:169-181.
38
39       Divine, BJ; Wendt, JK; Hartman, CM. (1993) Cancer mortality among workers at a butadiene facility. In: Butadiene
40       and styrene: assessment of health hazards. IARC Scientific Publications. Vol.  127. Sorsa, M; Peltonen, K; Vainio,
41       H; et al., eds. Lyon, France: International Agency for Research on Cancer, pp. 345-362.
42
43       Doerr, JK; Hooser, SB; Smith, BJ; et al. (1995) Ovarian toxicity of 4-vinylcyclohexene and related olefins in
44       B6C3F, mice: role of diepoxides. Chem Res Toxicol 8:963-969.
45
46       Doerr, JK; Hollis, EA; Sipes, IG. (1996) Species difference in the ovarian toxicity of 1,3-butadiene epoxides in
47       B6C3F, mice and Sprague-Dawley rats. Toxicology 113:128-136.
48
49       Downs, TD; Crane, MM; Kim, KW. (1987) Mortality among workers at a butadiene facility. Am J Ind Med
50       12:311-329.
51
52       Downs, T; Pier, S; Crane, M. (1992) Cause-specific mortality in a cohort of 1,000 ABS workers. Abstr Symp.
53       Butadiene and styrene: assessment of health hazards. Espoo, Finland.
54
         1/28/98                                        12-4        DRAFT-DO NOT CITE OR QUOTE
-------
   1       Duescher, RJ; Elfarra, AA. (1992) 1,3-Butadiene oxidation by human myeloperoxidase: role of chloride ion in
   2       catalysis of divergent pathways. J Biol Chem 267:19859
   13
   4       Duescher, RJ; Elfarra, AA. (1993) Chloroperoxidase-mediated Oxidation of 1,3-butadiene to 3-butenal, a
   5       crotonaldehyde precursor. Chem Res Toxicol 6:669-673
   6                                                                                                       •
   7       Duescher, RJ; Elfarra, AA. (1994) Human liver microsomes are efficient catalysts for 1,3-butadiene oxidation:
   8       evidence for major roles by cytochrome P450 2A6 and 2E1. Arch Biochem Biophys 311:342-349.
   9
 10       Dunnick, JK; Eustis, SL; Piegorsch, WW; et al. (1988) Respiratory tract lesions in F344/N rats and B6C3F, mice
 11       after inhalation exposure to 1,2-epoxybutane. Toxicology 50:69-82
 12
 13       Elfarra, AA; Duescher, RJ; Pasch, CM. (1991)  Mechanisms of 1,3-butadiene oxidations to butadiene monoxide and
 14       crotonaldehyde by mouse liver microsomes and chloroperoxidase. Arch Biochem Biophys 286:244-251.
 J O
 16       Elfarra, AA; Sharer, JE; Duescher, RJ. (1995) Synthesis and characterization of N-acetyl-L-cysteine S-conjugates of
 17       butadiene monoxide and then- detection and quantitation in urine of rats and mice given butadiene monoxide Chem
 18       Res Toxicol 8:68-76.
 19
 20       Evelo, CTA; Oostendorp, JGM; ten Berge, WF; et al. (1993) Physiologically based toxicokinetic modeling of 1,3-
 21       butadiene lung metabolism in mice becomes more important at low doses. Environ Health Perspect 101 -496-502
 22                                                                    •      .
 23       Filser, JG. (1992) The closed chamber technique—uptake, endogenous production, excretion, steady-state kinetics "
 24       and rates of metabolism of gases and vapors. Arch Toxicol 66:1-10.
 25
 26       Filser, JG; Bolt, HM. (1984) Inhalation pharmacokinetics based on gas uptake studies. Arch Toxicol 55:219-223.

 28       Filser, JG; Althaler, B; Welter, HF; et al. (1992) Metabolism of 1,3-butadiene hi microsomes from livers of mouse,
"""       rat, and man. Naunyn-Schmiedeberg's Arch Toxicol Suppl 345 :R31.

 31       Fiserova-Bergerova, V; Diaz, ML. (1986) Determination and prediction of tissue gas partition coefficients Int Arch
 32       Occup Environ Health 58:75-87.
 33
 34       Frank, PM. (1978) Introduction to system sensitivity theory. New York: Academic Press
 35
 36       Frome, EL; Checkoway, H. (1985) Use of Poisson regression models in estimating incidence rates and ratios Am J
 37       Epidemiol 121(2):309-22.
 38
 39       Gart, JJ;  Chu, KC; Tarone, RE. (1979) Statistical issues in interpretation of chronic bioassay for carcinogenicity J
 40      Natl Cancer Inst 62:957-974.
 41
 42      Gaylor, DW; Slikker, W. (1990) Risk assessment for neurotoxic effects. Neurotoxicology 11-211-218
 43
 44      Gervasi,  PG; Longo, V. (1990) Metabolism and mutagenicity of isoprene. Environ Health Perspect 86:85-87.

 46      Gervasi,  PG; Citti, L; Del Monte, M; et al. (1985) Mutagenicity and chemical reactivity of epoxide intermediates of
 47      the isoprene metabolism and other structurally related compounds. Mutat Res 156-77-82
 48
 49      Gibson, RW; Bross, IDJ; Lilienfeld, AM; et al. (1968) Leukemia in children exposed to multiple risk factors New
 50      Engl J Med 279:906-909.
 51
 52      Goodrow, T; Reynolds,  S; Maronpot, R; et al. (1990) Activation of K-ras by codon 13 mutations in C57BL/6 x
 53      C3HF, mouse tumors induced by exposure to 1,3-butadiene. Cancer Res 50:4818-4823
 54
         Greep, RO; Weis,  L. (1977) Histology. New York: McGraw-Hill.


         1/28/98                                         12-5         DRAFT-DO NOT CITE OR QUOTE
-------
 1       Hackett, PL; Sikov, MR; Mast, TJ; et al. (1987a) Inhalation developmental toxicology studies of 1,3-butadiene in
 2       the rat (final report). Richland, WA: Pacific Northwest Laboratory; PNL Report No. PNL-6414 UC-48; NIH Report
 3       No. NIH-401-ES-41031; 104 p. Prepared for NIEHS, NTP, under a Related Services Agreement with the U.S.
 4       Department of Energy under contract DE-AC06-76RLO-1830.
 5
 6       Hackett, PL; Sikov, MR; Mast, TJ; et al. (1987b) Inhalation developmental toxicology studies: teratology study of
 7       1,3-butadiene in mice (final report). Richland, WA: Pacific Northwest Laboratory; PNL Report No. PNL-6412 UC-
 8       48; NIH Report No. NIH-401-ES-40131; 92 p. Prepared for NIEHS, NTP, under a Related Services Agreement with
 9       the U.S. Department of Energy under contract DE-AC06-76RLO-1830.
10
11       Hackett, PL; McClanahan, BJ; Brown, MG; et al. (1988a). Sperm-head morphology study in B6C3F, mice
12       following inhalation exposure to 1,3-butadiene (final technical report). Richland, WA: Pacific Northwest
13       Laboratory; PNL Report No. PNL-6459 UC-48; NIH Report No. NIH-Y01-ES-70153; 51 p. Prepared for NIEHS,
14       NTP, under a Related Services Agreement with the U.S. Department of Energy under contract DE-AC06-76RLO-
15       1830.
16
17       Hackett, PL; McClanahan, BJ; Mast, TJ; et al. (1988b). Dominant lethal study in CD-I mice following inhalation
18       exposure to 1,3-butadiene (final technical report). Richland, WA: Pacific Northwest Laboratory;  PNL no. PNL-6545
19       UC-408; NIH no. NIH-Y01-ES-70153; 85 p. Prepared for NIEHS, NTP, under a Related Services Agreement with
20       the U.S. Department of Energy under contract DE-AC06-76RLO-1830.
21
22       Haley, TJ, Jr. (1978) Chloroprene (2-chloro- 1,3-butadiene)—what is the evidence for its carcinogenicity?  Clin
23       Toxicol 13:153-170.
24
25       Hallenbeck, W. (1992) Cancer risk assessment for the inhalation of 1,3-butadiene using physiologically based
26       pharmacokinetic modelling. Bull Environ Contain Toxicol 49:66-70.
27
28       Hattis, D; Wasson, J. (1987) Pharmacokinetic/mechanism-based analysis of the carcinogenic risk of butadiene.
29       PBS8-20817. Rockville, MD: National Institute for Occupational Safety and Health, pp. 1-122.
30
31       Hayes, RB; Xi, L; Bechtel, W; et al. (1996) hprt Mutation frequency among workers exposed to  1,3-butadiene in
32       China. Toxicology 113:100-105.
33
34       Hazleton Laboratories Europe, Ltd. (HLE). (1981) 1,3-Butadiene: inhalation teratogenicity study in the rat (final
35       report). Harrowgate, England: Hazleton Labs; addendum no. 2788-522/3.
36
37       Himmelstein, MW; Turner, MJ; Asgharian, B; et al.  (1994) Comparison of blood concentrations  of 1,3-butadiene
38       and butadiene epoxides in mice and rats exposed to 1,3-butadiene by inhalation. Carcinogenesis 15:1479-1486.
39
40       Himmelstein, MW; Asgharian, B; Bond, JA. (1995)  High concentrations of butadiene epoxides in livers and lungs
41       of mice compared to rats exposed to  1,3-butadiene. Toxicol Appl Pharmacol 132:281-288.
42
43       Himmelstein, MW; Acquavella, JF; Recio, L; et al. (1997) Toxicology and epidemiology of 1,3-butadiene. Crit Rev
44       Toxicol 27(1):1-108.
45
46       Howe, RB; Van Landingham, C. (1986) GLOBAL86: a computer program to extrapolate quantal animal toxicity
47       data to low doses. EPA Contract No. 68-01-6826. KS Crump and Company, Ruston, LA. Unpublished.
48
49       Huff, JE; Melnick, RL; Solleveld, HA; et al. (1985) Multiple organ carcinogenicity of 1,3-butadiene in B6C3F,
50       mice after 60 weeks of inhalation exposure. Science 227:548-549.
51
52       Hunt, WF, Jr.; Faoro, R; Hudischewskyj, AB; et al.  (1988) The EPA interim database for  air toxics volatile
53       organic chemicals. Systems Applications, Inc., and U.S. Environmental Protection Agency, Office of Air  Quality
54       Planning and Standards, Research Triangle Park, NC. EPA/450/4-88-014.


         1/28/98                                         12-6        DRAFT-DO NOT CITE OR QUOTE
-------
  1       International Agency for Research on Cancer (IARC). (1982) IARC monographs on the evaluation of carcinogenic
  2       risk of chemicals to humans. Vol. 28. The rubber industry. Lyon, France: International Agency for Research on
  3       Cancer.
 M
  5       International Agency for Research on Cancer (IARC). (1986) IARC monographs on the evaluation of carcinogenic
  6       risk of chemicals to humans. Vol. 39. Some chemicals used in plastics and elastomers butadiene. Lyon, France:
  7       International Agency for Research on Cancer, pp. 155-179.
  8
  9       International Agency for Research on Cancer (IARC). (1992) IARC monographs on the evaluation of carcinogenic
 10       risks to humans. Vol. 54. 1,3-Butadiene. Occupational exposures to mists and vapours from strong inorganic acids
 11       and other industrial chemicals. Lyon, France: International Agency for Research on Cancer, pp. 237-287.
 12
 13       International Institute of Synthetic Rubber Producers (IISRP). (1982) 1,3-Butadiene: inhalation teratogenicity in the
 14       rat (final report with cover letter dated 08/11/82). Report no. 2788-522/3; submission 8EHQ-0382-0441.
 15       Harrowgate, England: Hazleton Laboratories Europe, Ltd.
 16
 17       Irons, RD. (1990) Studies on the mechanism of 1,3-butadiene-induced leukemogenesis: the potential role of
 18       endogenous murine leukemia virus. Environ Health Perspect 86:49-55.
 19
 20       Irons, RD; Smith, CN; Stillman, WS; et al. (1986a) Macrocytic-megaloblastic anemia in male B6C3F, Swiss mice
 21       following chronic exposure to 1,3-butadiene. Toxicol Appl Pharmacol 83:95-100.                     '
 22
 23       Irons, RD; Smith, CN; Stillman, WS; et al. (1986b) Macrocytic-megaloblastic anemia in male NIH Swiss mice
 24       following repeated exposure to 1,3-butadiene. Toxicol Appl Pharmacol 85:450-455.
 25
 26       Irons, RD; Stillman, WS; Cloyd, MW. (1987) Selective activation of endogenous ecotropic retrovirus in
 27       hematopoietic tissues of B6C3F! mice during the preleukemic phase of 1,3-butadiene exposure. Virology 161-457-
.28       462.
 19
 30       Irons, RD; Cathro, HP; Stillman, WS; et al. (1989) Susceptibility to 1,3-butadiene-induced leukemogenesis
 31       correlates with endogenous ecotropic retroviral background in the mouse. Toxicol Appl Pharmacol 101:170-176.
 32
 33       Jacobson-Kram, D; Rosenthal, S. (1995) Molecular and genetic toxicology of 1,3-butadiene Mutat Res 339-121-
 34       130.
 35
 36       Jelitto, B; Vangala, RR; Laib, RJ. (1989) Species differences in DNA damage by butadiene: role of diepoxybutane
 37       Arch Toxicol 13 (suppl.):246-249.
 38
 39       Johanson, G; Filser, JG. (1992) Experimental data from closed chamber gas uptake studies in rodents suggest lower
 40       uptake rate of chemical than calculated from literature values on alveolar ventilation. Arch Toxicol 66-291-295
 41
 42       Johanson, G; Filser, JG. (1993) A physiologically based pharmacokinetic model for butadiene and its metabolite
 43       butadiene monoxide in rat and mouse and its significance for risk extrapolation. Arch Toxicol 67:151 -163
 44
 45       John, EM; Savitz, DA; Sandier, DP. (1991) Prenatal exposure to parents' smoking and childhood cancer. Am J
 46       Epidemiol 133:123-132.
 47
 48       Kaplan, EL; Meyer, P. (1958) Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-
 49       481.
 5.0
 51       Kavlock, R; Allen, BC; Kimmel, CA; et al. (1995) Dose-response assessment for developmental toxicity:  IV.
 52       Benchmark doses for fetal weight changes. Fundam Appl Toxicol 26:211-222.
 53
         1/28/98                                         12-7        DRAFT-DO NOT CITE OR QUOTE
-------
 1       Kelsey, KT; Wiencke, JK; Ward, J; et al. (1995) Sister-chromatid exchanges, glutathione S-transferase theta
 2       deletion and cytogenetic sensitivity to diepoxybutane in lymphocytes from butadiene monomer production workers.
 3       Mutat Res 335:267-273.
 4
 5       Kirshenbaum, I. (1978) Butadiene. In: Kirk-Othmer encyclopedia of chemical technology, 3rd ed., vol. 4. Mark,
 6       HF; Othmer, DF; Overberger, CG; et al., eds. New York, NY: John Wiley and Sons, pp. 313-337.
 7
 8       Kohn, MC; Melnick, RL. (1993) Species differences in the production and clearance of 1,3-butadiene metabolites: a
 9       mechanistic model indicates predominantly physiological, not biochemical, control. Carcinogenesis 14:619-628.
10
11       Kosaric, N; Duvnjak, Z; Farkas, A; et al. (1987) Ethanol. In: Ullmann's encyclopedia of industrial chemistry, 5th
12       Rev. Ed., Vol. A9. Gerhartz, W, ed. New York, NY:VCH Publishers, p. 590.
13
14       Kraybill, HF. (1980) Evaluation of public health aspects of carcinogenic biorefractories in drinking water. Prev Med
15       9:212-218.
16
17       Kreiling, R. (1987) Formation of 7-(l-hydroxy-3-butene-2-yl)guanine in liver DNA of mice exposed to MC-
18       butadiene. Naunyn-Schmiedebergs Arch Pharmacol 335 (suppl):31.
19
20       Kreiling, R; Laib, RJ; Bolt, HM. (1986a) Alkylation of nuclear proteins and DNA after exposure of rats and mice to
21       [1,4-HC] 1,3-butadiene. Toxicol Lett 30:131-136.
22
23       Kreiling, R; Laib, RJ; Filser, JG; et al. (1986b) Species differences in butadiene metabolism between mice and rats
24       evaluated by inhalation pharmacokinetics. Arch Toxicol 58:235-238.
25
26       Kreiling, R; Laib, RJ; Filser, JG; et al. (1987) Inhalation pharmacokinetics of l,2-epoxybutene-3 reveal species
27       differences between rats and mice sensitive to butadiene induced carcinogenesis. Arch Toxicol 61:7-11.
28
29       Kreiling, R; Laib, RJ; Bolt, HM. (1988) Depletion of hepatic non-protein sulfhydryl content during exposure of rats
30       and mice to butadiene. Toxicol Lett 41:209-214.
31
32       Kreuzer, P; Kessler, W; Welter, HF; et al. (1991) Enzyme specific kinetics of l,2-epoxybutene-3 in microsomes and
33       cytosol from livers of mouse,  rat, and man. Arch Toxicol  65:59-67.
34
35       Krewski, D; Crump, KS; Farmer, J; et al. (1983) A comparison of statistical methods for low dose extrapolation
36       utilizing time-to-tumor data. Fundam Appl Toxicol 3:140-160.
37
38       Laib, RJ; Filser, JG; Kreiling, R; et al. (1990) Inhalation pharmacokinetics of 1,3-butadiene and l,2-epoxybutene-3
39       in rats and mice. Environ Health Perspect 86:57-63.
40
41       Laib, RJ; Tucholski, M; Filser, JG; et al. (1992) Pharmacokinetic interaction between 1,3-butadiene and styrene in
42       Sprague-Dawley rats. Arch Toxicol 66:310-314.
43
44       Legator, MS; Au, WW; Ammenheuser, M; et al. (1993) Elevated somatic cell mutant frequencies and altered DNA
45       repair responses in nonsmoking workers exposed to 1,3-butadiene. In: Butadiene and styrene: assessment of health
46       hazards. IARC Scientific Publications. Vol. 127. Sorsa, M; Peltonen, K; Vainio, H; et al., eds. Lyon, France:
47       International Agency for Research on Cancer, pp. 253-263.
48
49       Lemen, RA; Meinhardt, TJ; Crandall, MS; et al. (1990) Environmental epidemiologic investigations in the
50       styrene-butadiene rubber production industry. Environ Health Perspect 86:103-106.
51
52       Lemieux, PM; DeMarini, DM. (1992). Mutagenicity of emissions from the simulated open burning and scrap rubber
53       tires. EPA-600/R-92-127. Washington, DC: U.S. Environmental Protection Agency.
54


         1/28/98                                         12-8         DRAFT-DO  NOT CITE OR QUOTE
-------
         Lieser (1983) Tierexperimentelle Pharmakokinetik von 1,3-Butadiene. Thesis. Johannes-Gutenberg-Universitat,
         Mainz, Germany.

         Ligocki, MP. (1993) Review of sources and emission estimates for 1,3-butadiene. SYSAPP-93/074. Prepared by
  o      Systems Applications International, Inc., San Rafael, CA.
  6
  7      Linet, MS. (1985) The Leukemias: epidemiologic aspects. Monographs in epidemiology and biostatistics. Vol. 6.
  8      New York: Oxford University Press, p. 16.
  9
 10      LSfgren, L; Berglund, P; Nordlinder, R; et al. (1991) Selective assessment of C2-C0 alkenes in air by adsorption
 11      sampling and gas chromatography. Int J Environ Anal Chem 45:39-44.
 12
 13      Lofroth, G; Burton, RM, Forehand, L; et al. (1989) Characterization of environmental tobacco smoke  Environ Sci
 14      Technol23:610.
 15
 16      Lyman, WJ; Reehl, WF; Rosenblatt, DH. (1990) Handbook of chemical property estimation methods. Washington,
 17      DC: American Chemical Society.
 18
 19      Macaluso, M; Larson, R; Delzell, E. (1996) Leukemia and cumulative exposure to butadiene, styrene and benzene
 20      among workers in the synthetic rubber industry. Toxicology 113:190-202.
 21
 22      Maldotti, A; Chiorboli, C; Bignozzi, CA; et al. (1980) Photooxidation of 1,3-butadiene containing systems: rate
 23      constant determination for the reaction of acrolein with -OH radicals. Int J Chem Kinet 12-905-913
 24
 25      Malvoisin, E; Roberfroid, M. (1982) Hepatic microsomal metabolism of 1,3-butadiene. Xenobiotica 12-137-144
 26
 27      Malvoisin, E; Lhoest, G; Poncelet, F; et al. (1979a) Identification and quantitation of l,2-epoxybutene-3 as the
         primary metabolite of 1,3-butadiene. JChromatogr 178:419-425.

         Malvoisin, E; Evrard, E; Roberfroid, M; et al. (1979b) Determination of Kovats retention indices with a capillary
 31      column and electron capture detection: application to the assay of the enzymatic conversion of 3,4-epoxy-l-butene
 32      into diepoxybutane. J Chromatogr 179:81.
 33
 34      Malvoisin, E; Mercier, M; Roberfroid, M. (1981) Enzymatic hydration of butadiene monoxide and its importance in
 35      the metabolism of butadiene. Biol Reactive Intermed II, 136A:437-444.
 36
 37      Maniglier-Poulet, C; Cheng, X. (1996) Kinetic and stereochemical differences in butadiene monoepoxide hydrolysis
 38      by rat and mouse liver microsomes. (Abstract.) Toxicologist 15:265.
 39
40      Maniglier-Poulet, C; Cheng, X; Ruth, JA; et al. (1995) Metabolism of 1,3-butadiene to butadiene monoxide in
41       mouse and human bone marrow cells. Chem-Biol Interact 97:119-129.
42
43      Mannervik, B. (1985) The isoenzymes of glutathione transferase. Adv Enzymol 57:357-417.
44
45      Maronpot, RR. (1987) Ovarian toxicity and carcinogenicity in eight recent National Toxicology Program studies.
46      Environ Health Perspect 73:125-130.
47
48      Matanoski, GM; Schwartz, L. (1987) Mortality of workers in styrene-butadiene polymer production. J Occup Med
49      29:675-680.
 50
51       Matanoski, GM; Santos-Burgoa, C. (1994) Butadiene and lymphatic and hematopoietic cancer (letter to the editor).
52      Epidemiology 5:261.
53
         1/28/98                                        12-9        DRAFT-DO NOT CITE OR QUOTE
-------
 1      Matanoski, GM; Schwartz, L; Sperrazza, J; et al. (1982) Mortality of workers in the styrene-butadiene rubber
 2      polymer manufacturing industry. A report prepared under contract to International Institute of Synthetic Rubber
 3      Producers.
 4
 5      Matanoski, GM; Santos-Burgoa, C; Zeger, SL; et al. (1989) Epidemiologic data related to health effects of
 6      1,3-butadiene. In: Assessment of inhalation hazards. Mohr, U; Bates, DV; Dungworth, DL; et al., eds. New York,
 7      NY: Springer-Verlag, pp. 201-214.
 8
 Q      Matanoski, GM; Santos-Burgoa, C; Schwartz, L. (1990) Mortality of a cohort of workers in the styrene-butadiene
10      polymer manufacturing industry (1943-1982). Environ Health Perspect 86:107-117.
11
12      Matanoski, GM; Francis, M; Correa-Villasenor, A. (1993) Cancer epidemiology among styrene-butadiene rubber
13      workers. In: Butadiene and styrene: assessment of health hazards. Scientific Publications. Vol. 127. Sorsa, M;
14      Peltonen, K; Vainio, H; et al., eds. Lyon, France: International Agency for Research on Cancer, pp. 363-374.
15
16      Mattison, DR; Plowchalk, DR; Meadows, MJ; et al. (1990) Reproductive toxicity: male and female reproductive
17      systems as targets for chemical injury. Med Clin N Am 74:391-411.
18
19      McAlister, RA; Moore, WH; Rice, J; et al. (1989) Nonmethane organic compound monitoring program final report
20      volume II: urban air toxics monitoring program, 1988. EPA-450/4-89-005. Office of Air Quality Planning and
21      Standards, U.S. Environmental Protection Agency.
22
23      McAlister, RA; Moore, WH; Rice, J; et al. (1990) Urban air toxics monitoring program, 1989. EPA-450/4-91-001.
24      Research Triangle Park, NC: Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency.
25
26      McAlister, RA; Epperson, DL; Jongleux, RF. (1991) Urban air toxics monitoring program aldehyde results, 1989.
27      EPA-450/4-91-006. Research Triangle Park, NC: Radian Corp. and Office of Air Quality Planning and Standards,
28      U.S. Environmental Protection Agency.
29
30      McCabe, RW; Siegl, O; Chun, W; et al. (1992) Speciated hydrocarbon emissions from the combustion of single
31      component fuels. II. Catalyst effects. J Air Waste Manage Assoc 42:1071-1077.
32
33      McMichael, AJ; Spirtas, R; Kupper, LL. (1974) An epidemiologic study of mortality within a cohort of rubber
34      workers, 1964-1972. J Occup Med 16:458-464.
35
36      McMichael, AJ; Spirtas, R; Gamble, JF; et al. (1976) Mortality among rubber workers: relationship to specific jobs.
37      J Occup Med 18:178-185.
38
39      Medinsky, MA; Leavens, TL; Csanady, GA; et al. (1994) In vivo metabolism of butadiene by mice and rats: a
40      comparison of physiological model predictions and experimental data. Carcinogenesis 15:1329-1340.
41
42      Meinhardt, TJ; Lemen, RA; Crandall, MS; et al. (1982) Environmental epidemiologic investigations  of the
43      styrene-butadiene rubber industry: mortality patterns with discussion of the hematopoietic and lymphatic
44      malignancies. Scand J Work Environ Health 8:250-259.
45
46      Melnick, RL; Huff, JE. (1992) 1,3-Butadiene: toxicity and carcinogeniciry in laboratory animals and in humans.
47      Rev Environ Contain Toxicol 124:111-144.
48
49      Melnick, RL; Huff, JE. (1993) 1,3-Butadiene induces cancer in experimental animals at all concentrations from 6.25
50      to 8,000 parts per million. In: Butadiene and styrene: assessment of health hazards. IARC Scientific Publications.
51      Vol. 127. Sorsa, M; Peltonen, K; Vainio, H; et al., eds. Lyon, France: International Agency for Research on Cancer,
52      pp. 309-322.
53
         1/28/98                                        12-10        DRAFT-DO NOT CITE OR QUOTE
-------
         Melnick, RL; Kohn, MC. (1995) Mechanistic data indicate that 1,3-butadiene is a human carcinogen.
         Carcinogenesis 16:157-163.

 F4      Melnick, RL; Huff, JE; Chou, BJ; et al. (1990a) Carcinogenicity of 1,3-butadiene in C57BL/6 x C3HFl mice at low
  5      exposure concentrations. Cancer Res 50:6592-6599.
  6
  7      Melnick, RL; Huff, JE; Bird, MG; et al. (1990b) Symposium overview: toxicology, carcinogenesis, and human
  8      health aspects of 1,3-butadiene. Environ Health Perspect 86:3-5.
  9
 10      Melnick, RL; Huff, JE; Roycroft, JH; et al. (1990c) Inhalation toxicology and carcinogeniciry of 1,3-butadiene in
 11      B6C3F, mice following 65 weeks of exposure.  Environ Health Perspect 86:27-36
 12
 13      Melnick, RL; Sills, RC; Roycroft, JH; et al. (1994) Isoprene, an endogenous hydrocarbon and industrial chemical,
 14      induces multiple organ neoplasia in rodents after 26 weeks of inhalation exposure. Cancer Res 54:5333-5339
 15
 16      Meng, Q; Recio, L; Reilly, AA; et al. (1996) The in vivo mutagenic potential of 1,3-butadiene, evaluated at the hprt
 17      locus of T-cells, is greater in mice than in rats exposed concurrently by inhalation. Abstract presented at the Twelfth
 18      Health Effects Institute Annual Conference.
 19                                                      .   ,
 20      Miller, LM. (1978) Investigations of selected potential environmental contaminants: butadiene and its oligomers.
 21      U.S. Environmental Protection Agency, Washington, DC. EPA-560/2-78-008
 22
 23      Miller, RA; Boorman, GA. (1990) Morphology of neoplastic lesions by 1,3-butadiene in B6C3F, mice  Environ
 24      Health Perspect 86:37-48.
 25
 26      Monson, RR. (1974) Analysis of relative survival and proportional mortality. Comput Biomed Res 7:325-332
 27
 28      Monson, RR; Nakano, KK. (1976) Mortality among rubber workers. I. White male union employees in Akron,
         Ohio. Am J Epidemiol 103:284-296.

 31      Morrow, NL. (1990) The industrial production and use of 1,3-butadiene. Environ Health Perspect 86-7-8
 32
 33      National Institute for Occupational Safety and Health (NIOSH). (1984) 1,3-Butadiene. Current Intelligence Bulletin
 34      41, Publication No. 84-105.U.S. Department of Health and Human Services, Cincinnati, OH
 35
 36      National Institute for Occupational Safety and Health (NIOSH). (1991a) A quantitative assessment of the risk of
 37      cancer associated with exposure to 1,3-butadiene, based on a low dose inhalation study in B6C3F, mice. Submitted
 38      to the Occupational Safety and Health Administration (OSHA) for Butadiene Docket (#H-041). U.S. Department of
 39      Health and Human Services, Washington, DC.
40
41       National Institute for Occupational Safety and Health (NIOSH). (1991b) Testimony from the National Institute for
42      Occupational Safety and Health on the Occupational Safety and Health Administration's proposed rule on
43      occupational exposure to 1,3-butadiene. U.S. Department of Health and Human Services, Washington, DC.
44
45      National Toxicology Program (NTP), U.S. Public Health Service, U.S. Department of Health and Human Services.
46      (1984) Toxicology and carcinogenesis studies of 1,3-butadiene (CAS No.  106-99-0) in B6C3F, mice (inhalation
47      studies). NTP TR 288, NIH Pub. No. 84-2544. Research Triangle Park, NC.
48
49      National Toxicology Program (NTP), U.S. Public Health Service, U.S. Department of Health and Human Services.
50      (1985) Quarterly report from Lovelace Research Institute, January 1 through March 31, 1985. Interagency
51       agreement 22-Y01-ES-0092. (L. Birnbaum, NTP Project Officer).
52
         1/28/98                                         12-11        DRAFT-DO NOT CITE OR QUOTE
-------
 1      National Toxicology Program (NTP), U.S. Public Health Service, U.S. Department of Health and Human Services.
 2      (1986) Toxicology and carcinogenesis studies of 4-vinylcyclohexene (CAS No. 100-40-3) in F344/N rats and
 3      B6C3F, mice (gavage studies). NTP TR 303. Research Triangle Park, NC.
 4
 5      National Toxicology Program (NTP), U.S. Public Health Service, U.S. Department of Health and Human Services.
 Q      (1989) Toxicology and carcinogenesis studies of 4-vinyl-l-cyclohexene diepoxide (CAS No. 106-87-6) in F344/N
 7      rats and B6C3F, mice (dermal studies). NTP TR 362, NIH Pub. No. 90-2817. Research Triangle Park, NC.
 8
 9      National Toxicology Program (NTP), U.S. Public Health Service, U.S. Department of Health and Human Services.
10      (1993) Toxicology and carcinogenesis studies of 1,3-butadiene (CAS No. 106-99-0) in B6C3F, mice (inhalation
11      studies). NTP TR 434. Research Triangle Park, NC.
12
13      Nauhaus, SK; Fennell, TR; Asgharian, B; et al. (1996) Characterization of urinary metabolites from Sprague-
14      Dawley rats and B6C3F, mice exposed to [l,2,3,4-C-13]butadiene. Chem Res Technol 9:764-773.
15
16      Nordlinder, RG; Nilsson, RI; Buskhe, AB. (1996) 1,3-butadiene in gasoline: an analytical conclusion. J Occup
17      Environ Med 38:463.
18
19      Norppa, H; Sorsa, M. (1993) Genetic toxicology of 1,3-butadiene and styrene. In: Butadiene and styrene:
20      assessment of health hazards. IARC Scientific Publications. Vol. 127. Sorsa, M; Peltonen, K; Vainio, H; et al., eds.
21      Lyon, France: International Agency for Research on Cancer, pp. 185-193.
22
23      Norppa, H; Hirvonen, A; Jarventaus, H; et al. (1995) Role of GSTT1 and GSTM1 genotypes in determining
24      individual sensitivity to sister chromatid exchange induction by diepoxybutane in cultured human lymphocytes.
25      Carcinogenesis 16:1261-1264.
26
27      Osterman-Golkar, SM; Kautianen, A; Bergmark, E; et al. (1991) Hemoglobin adducts and urinary mercapturic acids
28      in rats as biological indicators of butadiene exposure. Chem Biol Interactions 80:291-302.
29
30      Osterman-Golkar, SM; Bond, JA; Ward, JB, Jr; et al. (1993) Use of hemoglobin adducts for biomonitoring exposure
31      of 1,3-butadiene. In: Butadiene and styrene: assessment of health hazards. IARC Scientific Publications. Vol. 127.
32      Sorsa, M; Peltonen, K; Vainio, H; et al., eds. Lyon, France: International Agency for Research on Cancer, pp. 127-
33       134.
34
35      Osterman-Golkar, SM; Peltonen, K; Anttinen-Klementi, T; et al. (1996) Haemoglobin adducts as biomarkers of
36      occupational exposure to 1,3-butadiene. Mutagenesis 11:145-149.
37
38       Owen, PE; Glaister, JR. (1990) Inhalation toxicity and carcinogenicity of 1,3-butadiene in Sprague-Dawley rats.
39      Environ Health Perspect 86:19-25.
40
41       Owen, PE; Pullinger, DH; Glaister, JR; et al. (1985) 1,3 Butadiene: two-year inhalation toxicity/carcinogenicity
42       study in the rat (abstract no. P34). In: 26th Congress of the European Society of Toxicology, 16-19 June, Kuopio,
43       University of Kuopio. Hanhijarvi, H., ed. p. 69.  .
44
45       Owen, PE; Glaister, JR; Gaunt, IF; et al. (1987) Inhalation toxicity studies with 1,3-butadiene. 3. Two-year
46       toxicity/carcinogenicity study in rats. Am Ind Hyg Assoc J 48(5):407-413.
47
48       Percy, C; Stanek, E; Gloeckler L. (1981) Accuracy of cancer death certificates and its effects on cancer mortality
49      statistics. Am J Public Health 71:242-250.
50
51       Peter, H; Wiegand, HJ; Bolt, HM; et al. (1987) Pharmacokinetics of isoprene in mice and rats. Toxicol Lett 36:9-14.
52
          1/28/98                                        12-12        DRAFT-DO NOT CITE OR QUOTE
-------
         Peterson, J; Ward, D. (1989) An inventory of particulate matter and air toxic emissions from prescribed fires in the
         United States for 1989. Final Report LAG #DW12934736-01-0-1989. U.S. Department of Agriculture, Forest
         Service.

         Portier CJ; Bailer AJ. (1989) Testing for increased carcinogenicity using a survival-adjusted quantal response test.
 6       Fundam Appl Toxicol 12:731-737.
 7
 8       Portier, CJ; Hedges, JC; Hoel, DG. (1986) Age-specific models of mortality and tumor onset for historical control
 9       animals in the National Toxicology Program's carcinogenicity experiments. Cancer Res 46:4372-4378
10
11       Ramnas, O; Ostermark, U; Pettersson, G. (1994) Characterization of sixty alkenes in a cat-cracked gasoline naphtha
12       by gas chromatography. Chromatographia 3 8:222-223.
13
14       Recio, L; Goldsworthy, TL. (1995) The use of transgenic mice for studying mutagenicity induced by 1,3-butadiene.
15       Toxicol Lett 82:607-612.
16
17       Recio, L; Cochrane, J; Simpson, D; et al. (1990) DNA sequence analysis of in vivo hprt mutation in human T-
18       lymphocytes. Mutagenesis 5:505-510.
19
20       Reynolds, SH; Stowers, SJ; Patterson, RM; et al. (1987) Activated oncogenes in B6C3F, mouse liver tumors:
21       implications for risk assessment. Science 237:1309-1316.
22
23       Robinson, JP; Wiley, JA; Piazza, T; et al. (1989) Activity patterns of California residents and their implications for
24       potential exposure to pollution. Draft final report. California Air Resources Board, Sacramento, CA.
25       CARB-46-177-33.
26
27       Rolzhauser, HP. (1985) Pharmakokinetik von Butadienmonoxid. Thesis, Johannes-Gutenberg-Uniyersitat, Mainz,
28       Germany.

\P
30       Rosenthal, SL. (1985) The reproductive effects assessment group's report on the mutagenicity of 1,3-butadiene and
31       its reactive metabolites. Environ Mutagen 7:933-945.
32
33       Rothman, K; Boice, J. (1982) Epidemiologic analysis with a programmable calculator. Boston, MA: Epidemiology
34       Resources, Inc., pp. 11-17.
35
36       Sabourin, PJ; Burka, LT; Bechtold, WE; et al. (1992) Species differences in urinary butadiene metabolites;
37       identification of l,2-dihydroxy-4-(JV-acetylcysteinyl) butane, a novel metabolite of butadiene. Carcinogenesis
38       13:1633-1638.
39
40       Sandberg, DV; Pickford, SG; Darley, EF. (1975) Emissions from slash burning and the influence of flame retardant
41       chemicals. J Air Pollut Control Assoc 25:278-281.
42
43       Santodonato, J. (1985) Monograph on human exposure to chemicals in the workplace: 1,3-butadiene (PB86-
44       147261). Washington, DC: U. S. National Technical Information Service.
45
46       Santos-Burgoa, C. (1988) Case-control study of lymphohematopoietic malignant neoplasms within a cohort of
47       styrene-butadiene polymerization workers. Thesis. Baltimore, MD, Johns Hopkins School of Hygiene and Public
48       Health.
49
50       Santos-Burgoa, C; Matanoski, GM; Zeger, S; et al. (1992) Lymphohematopoietic cancer in styrene-butadiene
51       polymerization workers. Am J Epidemiol 136:843-854.
52
53       Sax, NI; Lewis, RJ, Sr. (1987) Hawley's condensed chemical dictionary, 11th ed. New York, NY: Van Nostrand
         Reinhold, p. 177.


         1/28/98                                        12-13         DRAFT-DO NOT CITE OR QUOTE
-------
  1       Schmidt, U; Loeser, E. (1985) Species differences in the formation of butadiene monoxide from 1,3-butadiene.
  2       Arch Toxicol 57:222-225.
  3
  4       Schottenfeld, D; Fraumeni, JF, Jr. (1996) Cancer epidemiology and prevention. 2nd ed.  New York: Oxford Univ
  5       Press, p. 856.
  6
  7       Seaton, MJ; Follansbee, MH; Bond, JA. (1995) Oxidation of l,2-epoxy-3-butene to l,2:3,4-diepoxybutane by
  8       cDNA expressed human cytochromes P450 2E1 and 3A4 and human, mouse and rat liver microsomes.
  9       Carcinogenesis 16:2287-2293.
10
11       Shah, JJ; Heyerdahl, EK. (1988) National ambient volatile organic compounds database update (project report).
12       U.S. Environmental Protection Agency, Atmospheric Sciences Research Laboratory, Research Triangle Park, NC.
13       EPA/600/3-88/010(a).
14
15       Sharer, JE; Elfarra, AA. (1992) S-(2-Hydroxy-3-buten-l-yl)glutathione and S-(l-hydroxy-3-buten-yl-)glutathione
16       are in vivo metabolites of 1,3-butadiene monoxide: detection and quantitation in bile. Chem Res Toxicol 5:787-790.
17
18       Sharer, JE; Duescher, RJ; Elfarra, AA. (1991) Formation, stability, and rearrangements of glutathione conjugates of
19       butadiene monoxide: evidence for the formation of stable sulfrane intermediates. Chem Res Toxicol 4:430-436.
20
21       Sharer, JE; Duescher, RJ; Elfarra, AA. (1992) Species and tissue differences in the microsomal oxidation of 1,3-
22       butadiene and the glutathione conjugation of butadiene monoxide in mice and rats. Drug Metab Dispos 20:658-664.
23
24       Shields, PG; Xu, GX; Blot, WJ; et al. (1995) Mutagens from heated Chinese and U.S. cooking oils. J Natl Cancer
25       Inst 87:836-841  [abstract from Toxline 95:258350].
26
27       Siematycki, J; ed. (1991) Risk factors for cancer in the workplace. Boca Raton, FL: CRC Press.
28
29       SjSblom, T; Lahdetie, J. (1996) Micronuclei are induced in rat spermatids in vitro by 1,2,3,4-diepoxybutane but not
30       by l,2-epoxy-3-butene and l,2-dihydroxy-3,4-epoxybutane. Mutagenesis 11:525-528.
31
32       Soderkvist, P; Goodrow, T; Cochran, C; et al. (1992) Characterization of p53 mutations  and allelic losses on
33       chromosome 4, 11, and 14 in butadiene-induced tumors of B6C3F, mice. J Cell Biochem 16B: 138.
34
35       Sorsa, M; Autio, K; Demopoulos, NA; et al. (1994) Human cytogenetic biomonitoring of occupational exposure to
36       1,3-butadiene. Mutat Res 309:321-326.
37
38       Sorsa, M; Osterman-Golkar, S; Peltonen, K; et al. (1996) Assessment of exposure to butadiene in the process
39       industry. Toxicology 113:77-83.
40
41       SRI International, 1993.1993 directory of chemical producers-USA. Menlo Park,CA: SRI International.
42
43       Startin, JR; Gilbert, J. (1984) Single ion monitoring of butadiene in plastics and foods by coupled mass
44       spectrometry-automatic headspace gas chromatography. J Chromatogr  294:427-430.
45
46       Stjemfeldt, M; Berglund, K; Lindsten, J; Ludrigsson, J. (1986) Maternal smoking during pregnancy and risk of
47       childhood cancer. Lancet 1(8494):1350-1352.
48
49       Sun, JD; Dahl, AR; Bond, JA; et al.  (1989a) Characterization of hemoglobin adduct formation in mice and rats after
50       administration of [MC]butadiene or [14C]isoprene. Toxicol Appl Pharmacol 100:86-95.
51
52       Sun, JD; Dahl, AR; Bond, JA; et al.  (1989b) Metabolism of inhaled butadiene to monkeys: comparison to rodents.
53       ExpPathol 37:133-135.



         1/28/98                                       12-14       DRAFT-DO NOT CITE OR QUOTE
-------
         Tarone, RE. (1975) Tests for trend in life table analysis. Biometrika 62:679-682.

         Tates, AD; van Dam, FJ; de Zwart, FA; et al. (1996) Biological effect monitoring in industrial workers from the
         Czech Republic exposed to low levels of butadiene. Toxicology 113-91-99
  5
  6      Thier, R; Mueller, M; Taylor, JB; et al. (1995) Enhancement of bacterial mutagenicity of bifunctional alkylating
  7      agents by expression of mammalian glutathione S-transferase. Chem Res Toxicol 8'465
  8
  9      Thornton-Manning, JR; Dahl, AR; Bechtold, WE; et al. (1995a) Disposition of butadiene monoepoxide and
 10      butadiene diepoxide in various tissues of rats and mice following a low-level inhalation exposure to 1,3-butadiene.
 11      Carcinogenesis 16:1723-1731.
 12
 13      Thornton-Manning, JR; Dahl, AR; Bechtold, WE; et al. (1995b) Gender differences in the metabolism of 1,3-
 14      butadiene in Sprague-Dawley rats following low-level inhalation exposure.  Carcinogenesis 16:2875-2878
 15
 16      Thornton-Manning, JR; Bechtold, WE; Griffith, RF; et al. (1996) Tissue concentrations of butadiene epoxides in
 17      rodents following multiple and single exposures to 1,3-butadiene (BD) by inhalation. (Abstract) Toxicolo^ist
 18      30:215.
 19
 20      Thurmond, LM; Lauer, LD; House, RV; et al. (1986) Effect of short-term inhalation exposure to 1,3-butadiene on
 21      murine immune functions. Toxicol Appl Pharmacol 86:170-179.
 22
 23      Travis, CC. (1988) Carcinogen risk assessment: contemporary issues in risk analysis. Vol. 3. New York, NY:
 24      Plenum Press.
 25
 26      U.S. Department of Health and Human Services (U.S. DHHS), Public Health Service. (1992) Toxicological profile
 27      for 1,3-butadiene.
 18                                                                                                       ' '
         U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). (1990, August 10)
 30      Occupational exposure to 1,3-butadiene; proposed rule and notice of hearing. Federal Register 55:32736-32776
 31
 32      U.S. Department of Labor, Occupational Safety and Health Administration (OSHA). (1996, November 4)
 33      Occupational exposure to 1,3-butadiene; final rale. Federal Register 61(214):56745-56795
 34
 35      U.S. Environmental Protection Agency. (1978) Interim primary drinking water regulations. Federal Register
 36      43:20135-29150.
 37
 38      U.S. Environmental Protection Agency. (1985) Mutagenieity and carcinogenicity assessment of 1,3-butadiene.
 39      Office of Health and Environmental Assessment, Office of Research and Development, Washington DC
40      EPA/600/8-85/004F.
41                                                                                   ,
42      U.S. Environmental Protection Agency. (1986, Sept. 24) Guidelines for carcinogen risk assessment Federal
43      Register 51(185):33992-34003.
44
45      U.S. Environmental Protection Agency. (1989a) AIRS user's guide, volumes I-VII. Office of Air Quality Planning
46      and Standards, Research Triangle Park, NC.
47
48      U.S. Environmental Protection Agency. (1989b) 1989 Urban Air Toxics Monitoring Program. Office of Air Quality
49      Planning and Standards, Research Triangle Park, NC. EPA/450/4-91 -001
50
51       U.S. Environmental Protection Agency. (1989c) 1989 Urban Air Toxics Monitoring Program; aldehyde results.
52      Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA/450/4-91-006
53
         1/28/98                                        12-15        DRAFT-DO NOT CITE OR QUOTE
-------
  1       U.S. Environmental Protection Agency. (1990a) 1989 Urban Air Toxics Monitoring Program.  Office of Air Quality
  2       Planning and Standards, Research Triangle Park, NC. EPA/450/4-91 -024.
  3
  4       U.S. Environmental Protection Agency (1990b) 1990 Urban Air Toxics Monitoring Program carbonyl results.
  5       Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA/450/4-91-025.
  6
  7       U.S. Environmental Protection Agency. (1991) Non-road engine and vehicle emission study. Office of Mobile
  8       Sources, Ann Arbor, MI. EPA 21A-2001.
  9
10       U.S. Environmental Protection Agency. (1992) Compilation of air pollutant emission factors (AP-42). Supplement
11       E, October 1992. U.S. Environmental Protection Agency.
12
13       U.S. Environmental Protection Agency. (1993a) Motor vehicle-related air toxics study. Office of Mobile Sources,
14       Ann Arbor, MI. EPA/420-R-93-005.
15
16       U.S. Environmental Protection Agency. (1993b) The Integrated Air Cancer Project, Boise Study; research to
17       improve risk assessment of area sources, woodstoves and mobile sources. Office of Research and Development.
18       Presentation to the Office of Mobile Sources, August 1993.
19
20       U.S. Environmental Protection Agency. (1994a) Locating and estimating air emissions from sources of 1,3-
21       butadiene. (September 1994 draft.) Office of Ah- Quality Planning and Standards, Research Triangle Park, NC.
22
23       U.S. Environmental Protection Agency. (1994b) Memo from Rich Cook, Office of Mobile Sources to Dennis
24       Beauregard, Office of Air Quality Planning and Standards. November 9, 1994.
25
26       U.S. Environmental Protection Agency. (1995) The use of the benchmark-dose approach in health risk assessment.
27       Office of Research and Development, Washington, DC. EPA/630/R-94/007.
28
29       U.S. Environmental Protection Agency. (1996, April 23) Proposed guidelines for carcinogen risk assessment.
30       Federal Register 61(79):17960-18011.
31
32       U.S. International Trade Commission (USITC). (1990) Synthetic organic chemicals; United States production and
33       sales. 1989. Pub. 2338. Washington, DC: Government Printing Office.
34
35       U.S. National Library of Medicine. (1991) Toxic Chemical Release Inventory (TRI) Data Bank, Bethesda, MD.
36
37       Uuskula, M; Jarventaus, H; Hirvonen, A; et al. (1995) Influence of GSTM1 genotype on sister chromatid exchange
38       induction by styrene-7,8-oxide and l,2-epoxy-3-butene in cultured human lymphocytes. Carcinogenesis 16:947-
39       950.
40
41       Van Duuren, BL; Nelson, N; Orris, L; et al. (1963) Carcinogenicity of epoxides, lactones, and peroxy compounds. J
42       Natl Cancer Inst 31:41-55.
43
44       Van Duuren, BL; Langseth, L; Orris, L; et al. (1966) Carcinogenicity of epoxides, lactones, and peroxy compounds.
45       IV. Tumor response in epithelial and connective tissue in mice and rats. J Natl Cancer Inst 37:825-838.
46
47       Verschueren, K. (1983) Handbook of environmental data on organic chemicals, 2nd ed. New York, NY: Van
48       Nostrand Reinhold, pp. 295-297.
49
50       Ward, DE; Hao, WM. (1992) Air toxic emissions from burning of biomass globally—preliminary estimates. Paper
51       92-98.03 Presented at the 85th Annual Meeting of the Air and Waste Management Association, Kansas City, MO.
52
53       Ward, JB, Jr;  Au, WW; Ammenheuser, MM; et al. (1994) hprt Mutant lymphocyte frequencies in workers at a 1,3-
54       butadiene production plant. Environ Health Perspect 102:79-85.


         1/28/98                                        12-16       DRAFT-DO NOT CITE OR QUOTE
-------
         Ward, EM; Fajen, JM; Ruder, A. (1995) Mortality study of workers in 1,3-butadiene production units identified
         from a chemical workers cohort. Environ Health Perspect 103:598-603.

         Ward, JB, Jr; Ammenheuser, MM; Bechtold, WE. (1996a) Assessment of butadiene exposure in a rubber plant
  a      using urine metabolite and HPRT mutant frequency as biological markers of exposure and effect. (Abstract.)
  6      Environ MolecMutag27(suppl.27):73.
  7
  8      Ward, JB, Jr.; Ammenheuser, MM; Wharton, EB, Jr.; et al. (1996b) Biological monitoring for mutagenic effects of
  9      occupational exposure to butadiene. Toxicology 113:84-90
10
11       Ward, EM; Fajen, JM; Ruder, A. (1996c) Mortality study of workers employed in 1,3-butadiene production units
12      identified from a large chemical workers cohort. Toxicology 113:157-168.
13
14      Weast, RC, ed. (1989) CRC handbook of chemistry and physics, 70th ed. Boca Raton, FL:CRC Press, p. C-160
15
16      Wijnberg, L; Faoro, R. (1989) Urban air toxics monitoring program—results of aldehyde monitoring fiscal year
17      1988. PEI Associates and the U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards.
18      Research Triangle Park, NC. September 1989.
19
20      Wistuba, D; Nowotny, H-P; Tragerf, O; et al. (1989) Cytochrome P-450-catalyzed asymmetric epoxidation of
21       simple prochiral and chiral aliphatic alkenes: species dependence and effect of enzyme induction of enantioselective
22       oxirane formation. Chirality 1:127.
23
24       Xiao, Y; Tates, AD (1995) Clastogenic effects of 1,3-butadiene and its metabolites 1,2-epoxybutene and 1,2,3,4-
25       diepoxybutane in splenocytes and germ cells of rats and mice in vivo. Environ Mol Mutagen 26:97-108.
        1/28/98.                                        12-17       DRAFT-DO NOT CITE OR QUOTE
-------
-------